Linux-2.6.12-rc2

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!

drivers/s390/Kconfig

@@ -0,0 +1,209 @@
+config CCW
+	bool
+	default y
+
+source "drivers/block/Kconfig"
+
+source "drivers/md/Kconfig"
+
+
+menu "Character device drivers"
+
+config UNIX98_PTYS
+	bool "Unix98 PTY support"
+	---help---
+	  A pseudo terminal (PTY) is a software device consisting of two
+	  halves: a master and a slave. The slave device behaves identical to
+	  a physical terminal; the master device is used by a process to
+	  read data from and write data to the slave, thereby emulating a
+	  terminal. Typical programs for the master side are telnet servers
+	  and xterms.
+
+	  Linux has traditionally used the BSD-like names /dev/ptyxx for
+	  masters and /dev/ttyxx for slaves of pseudo terminals. This scheme
+	  has a number of problems. The GNU C library glibc 2.1 and later,
+	  however, supports the Unix98 naming standard: in order to acquire a
+	  pseudo terminal, a process opens /dev/ptmx; the number of the pseudo
+	  terminal is then made available to the process and the pseudo
+	  terminal slave can be accessed as /dev/pts/<number>. What was
+	  traditionally /dev/ttyp2 will then be /dev/pts/2, for example.
+
+	  The entries in /dev/pts/ are created on the fly by a virtual
+	  file system; therefore, if you say Y here you should say Y to
+	  "/dev/pts file system for Unix98 PTYs" as well.
+
+	  If you want to say Y here, you need to have the C library glibc 2.1
+	  or later (equal to libc-6.1, check with "ls -l /lib/libc.so.*").
+	  Read the instructions in <file:Documentation/Changes> pertaining to
+	  pseudo terminals. It's safe to say N.
+
+config UNIX98_PTY_COUNT
+	int "Maximum number of Unix98 PTYs in use (0-2048)"
+	depends on UNIX98_PTYS
+	default "256"
+	help
+	  The maximum number of Unix98 PTYs that can be used at any one time.
+	  The default is 256, and should be enough for desktop systems. Server
+	  machines which support incoming telnet/rlogin/ssh connections and/or
+	  serve several X terminals may want to increase this: every incoming
+	  connection and every xterm uses up one PTY.
+
+	  When not in use, each additional set of 256 PTYs occupy
+	  approximately 8 KB of kernel memory on 32-bit architectures.
+
+source "drivers/char/watchdog/Kconfig"
+
+comment "S/390 character device drivers"
+
+config TN3270
+	tristate "Support for locally attached 3270 terminals"
+	help
+	  Include support for IBM 3270 terminals.
+
+config TN3270_TTY
+	tristate "Support for tty input/output on 3270 terminals"
+	depends on TN3270
+	help
+	  Include support for using an IBM 3270 terminal as a Linux tty.
+
+config TN3270_FS
+	tristate "Support for fullscreen applications on 3270 terminals"
+	depends on TN3270
+	help
+	  Include support for fullscreen applications on an IBM 3270 terminal.
+
+config TN3270_CONSOLE
+	bool "Support for console on 3270 terminal"
+	depends on TN3270=y && TN3270_TTY=y
+	help
+	  Include support for using an IBM 3270 terminal as a Linux system
+	  console.  Available only if 3270 support is compiled in statically.
+
+config TN3215
+	bool "Support for 3215 line mode terminal"
+	help
+	  Include support for IBM 3215 line-mode terminals.
+
+config TN3215_CONSOLE
+	bool "Support for console on 3215 line mode terminal"
+	depends on TN3215
+	help
+	  Include support for using an IBM 3215 line-mode terminal as a
+	  Linux system console.
+
+config CCW_CONSOLE
+ 	bool
+ 	depends on TN3215_CONSOLE || TN3270_CONSOLE
+ 	default y
+ 
+config SCLP
+	bool "Support for SCLP"
+	help
+	  Include support for the SCLP interface to the service element.
+
+config SCLP_TTY
+	bool "Support for SCLP line mode terminal"
+	depends on SCLP
+	help
+	  Include support for IBM SCLP line-mode terminals.
+
+config SCLP_CONSOLE
+	bool "Support for console on SCLP line mode terminal"
+	depends on SCLP_TTY
+	help
+	  Include support for using an IBM HWC line-mode terminal as the Linux
+	  system console.
+
+config SCLP_VT220_TTY
+	bool "Support for SCLP VT220-compatible terminal"
+	depends on SCLP
+	help
+	  Include support for an IBM SCLP VT220-compatible terminal.
+
+config SCLP_VT220_CONSOLE
+	bool "Support for console on SCLP VT220-compatible terminal"
+	depends on SCLP_VT220_TTY
+	help
+	  Include support for using an IBM SCLP VT220-compatible terminal as a
+	  Linux system console.
+
+config SCLP_CPI
+	tristate "Control-Program Identification"
+	depends on SCLP
+	help
+	  This option enables the hardware console interface for system
+	  identification. This is commonly used for workload management and
+	  gives you a nice name for the system on the service element.
+	  Please select this option as a module since built-in operation is
+	  completely untested.
+	  You should only select this option if you know what you are doing,
+	  need this feature and intend to run your kernel in LPAR.
+
+config S390_TAPE
+	tristate "S/390 tape device support"
+	help
+	  Select this option if you want to access channel-attached tape
+	  devices on IBM S/390 or zSeries.
+	  If you select this option you will also want to select at
+	  least one of the tape interface options and one of the tape
+	  hardware options in order to access a tape device.
+	  This option is also available as a module. The module will be
+	  called tape390 and include all selected interfaces and
+	  hardware drivers.
+
+comment "S/390 tape interface support"
+	depends on S390_TAPE
+
+config S390_TAPE_BLOCK
+	bool "Support for tape block devices"
+	depends on S390_TAPE
+	help
+	  Select this option if you want to access your channel-attached tape
+	  devices using the block device interface.  This interface is similar
+	  to CD-ROM devices on other platforms.  The tapes can only be
+	  accessed read-only when using this interface.  Have a look at
+	  <file:Documentation/s390/TAPE> for further information about creating
+	  volumes for and using this interface.  It is safe to say "Y" here.
+
+comment "S/390 tape hardware support"
+	depends on S390_TAPE
+
+config S390_TAPE_34XX
+	tristate "Support for 3480/3490 tape hardware"
+	depends on S390_TAPE
+	help
+	  Select this option if you want to access IBM 3480/3490 magnetic
+	  tape subsystems and 100% compatibles.
+	  It is safe to say "Y" here.
+
+
+
+config VMLOGRDR
+	tristate "Support for the z/VM recording system services (VM only)"
+	depends on IUCV
+	help
+	  Select this option if you want to be able to receive records collected
+	  by the z/VM recording system services, eg. from *LOGREC, *ACCOUNT or
+	  *SYMPTOM.
+	  This driver depends on the IUCV support driver.
+
+config MONREADER
+	tristate "API for reading z/VM monitor service records"
+	depends on IUCV
+	help
+	  Character device driver for reading z/VM monitor service records
+
+endmenu
+
+menu "Cryptographic devices"
+
+config Z90CRYPT
+	tristate "Support for PCI-attached cryptographic adapters"
+        default "m"
+        help
+	  Select this option if you want to use a PCI-attached cryptographic
+	  adapter like the PCI Cryptographic Accelerator (PCICA) or the PCI
+	  Cryptographic Coprocessor (PCICC).  This option is also available
+	  as a module called z90crypt.ko.
+
+endmenu

drivers/s390/Makefile

@@ -0,0 +1,9 @@
+#
+# Makefile for the S/390 specific device drivers
+#
+
+obj-y += s390mach.o sysinfo.o
+obj-y += cio/ block/ char/ crypto/ net/ scsi/
+
+drivers-y += drivers/s390/built-in.o
+

drivers/s390/block/Kconfig

@@ -0,0 +1,68 @@
+if ARCH_S390
+
+comment "S/390 block device drivers"
+	depends on ARCH_S390
+
+config BLK_DEV_XPRAM
+	tristate "XPRAM disk support"
+	depends on ARCH_S390
+	help
+	  Select this option if you want to use your expanded storage on S/390
+	  or zSeries as a disk.  This is useful as a _fast_ swap device if you
+	  want to access more than 2G of memory when running in 31 bit mode.
+	  This option is also available as a module which will be called
+	  xpram.  If unsure, say "N".
+
+config DCSSBLK
+	tristate "DCSSBLK support"
+	help
+	  Support for dcss block device
+
+config DASD
+	tristate "Support for DASD devices"
+	depends on CCW
+	help
+	  Enable this option if you want to access DASDs directly utilizing
+	  S/390s channel subsystem commands. This is necessary for running
+	  natively on a single image or an LPAR.
+
+config DASD_PROFILE
+	bool "Profiling support for dasd devices"
+	depends on DASD
+	help
+	  Enable this option if you want to see profiling information
+          in /proc/dasd/statistics.
+
+config DASD_ECKD
+	tristate "Support for ECKD Disks"
+	depends on DASD
+	help
+	  ECKD devices are the most commonly used devices. You should enable
+	  this option unless you are very sure to have no ECKD device.
+
+config DASD_FBA
+	tristate "Support for FBA  Disks"
+	depends on DASD
+	help
+	  Select this option to be able to access FBA devices. It is safe to
+	  say "Y".
+
+config DASD_DIAG
+	tristate "Support for DIAG access to Disks"
+	depends on DASD && ARCH_S390X = 'n'
+	help
+	  Select this option if you want to use Diagnose250 command to access
+	  Disks under VM.  If you are not running under VM or unsure what it is,
+	  say "N".
+
+config DASD_CMB
+	tristate "Compatibility interface for DASD channel measurement blocks"
+	depends on DASD
+	help
+	  This driver provides an additional interface to the channel measurement
+	  facility, which is normally accessed though sysfs, with a set of
+	  ioctl functions specific to the dasd driver.
+	  This is only needed if you want to use applications written for
+	  linux-2.4 dasd channel measurement facility interface.
+
+endif

drivers/s390/block/Makefile

@@ -0,0 +1,17 @@
+#
+# S/390 block devices
+#
+
+dasd_eckd_mod-objs := dasd_eckd.o dasd_3990_erp.o dasd_9343_erp.o
+dasd_fba_mod-objs  := dasd_fba.o dasd_3370_erp.o dasd_9336_erp.o
+dasd_diag_mod-objs := dasd_diag.o
+dasd_mod-objs      := dasd.o dasd_ioctl.o dasd_proc.o dasd_devmap.o \
+			dasd_genhd.o dasd_erp.o
+
+obj-$(CONFIG_DASD) += dasd_mod.o
+obj-$(CONFIG_DASD_DIAG) += dasd_diag_mod.o
+obj-$(CONFIG_DASD_ECKD) += dasd_eckd_mod.o
+obj-$(CONFIG_DASD_FBA)  += dasd_fba_mod.o
+obj-$(CONFIG_DASD_CMB)  += dasd_cmb.o
+obj-$(CONFIG_BLK_DEV_XPRAM) += xpram.o
+obj-$(CONFIG_DCSSBLK) += dcssblk.o

drivers/s390/block/dasd_3370_erp.c

@@ -0,0 +1,104 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_3370_erp.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 2000
+ *
+ * $Revision: 1.9 $
+ */
+
+#define PRINTK_HEADER "dasd_erp(3370)"
+
+#include "dasd_int.h"
+
+
+/*
+ * DASD_3370_ERP_EXAMINE 
+ *
+ * DESCRIPTION
+ *   Checks only for fatal/no/recover error. 
+ *   A detailed examination of the sense data is done later outside
+ *   the interrupt handler.
+ *
+ *   The logic is based on the 'IBM 3880 Storage Control Reference' manual
+ *   'Chapter 7. 3370 Sense Data'.
+ *
+ * RETURN VALUES
+ *   dasd_era_none	no error 
+ *   dasd_era_fatal	for all fatal (unrecoverable errors)
+ *   dasd_era_recover	for all others.
+ */
+dasd_era_t
+dasd_3370_erp_examine(struct dasd_ccw_req * cqr, struct irb * irb)
+{
+	char *sense = irb->ecw;
+
+	/* check for successful execution first */
+	if (irb->scsw.cstat == 0x00 &&
+	    irb->scsw.dstat == (DEV_STAT_CHN_END | DEV_STAT_DEV_END))
+		return dasd_era_none;
+	if (sense[0] & 0x80) {	/* CMD reject */
+		return dasd_era_fatal;
+	}
+	if (sense[0] & 0x40) {	/* Drive offline */
+		return dasd_era_recover;
+	}
+	if (sense[0] & 0x20) {	/* Bus out parity */
+		return dasd_era_recover;
+	}
+	if (sense[0] & 0x10) {	/* equipment check */
+		if (sense[1] & 0x80) {
+			return dasd_era_fatal;
+		}
+		return dasd_era_recover;
+	}
+	if (sense[0] & 0x08) {	/* data check */
+		if (sense[1] & 0x80) {
+			return dasd_era_fatal;
+		}
+		return dasd_era_recover;
+	}
+	if (sense[0] & 0x04) {	/* overrun */
+		if (sense[1] & 0x80) {
+			return dasd_era_fatal;
+		}
+		return dasd_era_recover;
+	}
+	if (sense[1] & 0x40) {	/* invalid blocksize */
+		return dasd_era_fatal;
+	}
+	if (sense[1] & 0x04) {	/* file protected */
+		return dasd_era_recover;
+	}
+	if (sense[1] & 0x01) {	/* operation incomplete */
+		return dasd_era_recover;
+	}
+	if (sense[2] & 0x80) {	/* check data erroor */
+		return dasd_era_recover;
+	}
+	if (sense[2] & 0x10) {	/* Env. data present */
+		return dasd_era_recover;
+	}
+	/* examine the 24 byte sense data */
+	return dasd_era_recover;
+
+}				/* END dasd_3370_erp_examine */
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-indent-level: 4 
+ * c-brace-imaginary-offset: 0
+ * c-brace-offset: -4
+ * c-argdecl-indent: 4
+ * c-label-offset: -4
+ * c-continued-statement-offset: 4
+ * c-continued-brace-offset: 0
+ * indent-tabs-mode: 1
+ * tab-width: 8
+ * End:
+ */

drivers/s390/block/dasd_9336_erp.c

@@ -0,0 +1,61 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_9336_erp.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 2000
+ *
+ * $Revision: 1.8 $
+ */
+
+#define PRINTK_HEADER "dasd_erp(9336)"
+
+#include "dasd_int.h"
+
+
+/*
+ * DASD_9336_ERP_EXAMINE 
+ *
+ * DESCRIPTION
+ *   Checks only for fatal/no/recover error. 
+ *   A detailed examination of the sense data is done later outside
+ *   the interrupt handler.
+ *
+ *   The logic is based on the 'IBM 3880 Storage Control Reference' manual
+ *   'Chapter 7. 9336 Sense Data'.
+ *
+ * RETURN VALUES
+ *   dasd_era_none	no error 
+ *   dasd_era_fatal	for all fatal (unrecoverable errors)
+ *   dasd_era_recover	for all others.
+ */
+dasd_era_t
+dasd_9336_erp_examine(struct dasd_ccw_req * cqr, struct irb * irb)
+{
+	/* check for successful execution first */
+	if (irb->scsw.cstat == 0x00 &&
+	    irb->scsw.dstat == (DEV_STAT_CHN_END | DEV_STAT_DEV_END))
+		return dasd_era_none;
+
+	/* examine the 24 byte sense data */
+	return dasd_era_recover;
+
+}				/* END dasd_9336_erp_examine */
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-indent-level: 4 
+ * c-brace-imaginary-offset: 0
+ * c-brace-offset: -4
+ * c-argdecl-indent: 4
+ * c-label-offset: -4
+ * c-continued-statement-offset: 4
+ * c-continued-brace-offset: 0
+ * indent-tabs-mode: 1
+ * tab-width: 8
+ * End:
+ */

drivers/s390/block/dasd_9343_erp.c

@@ -0,0 +1,22 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_9345_erp.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 2000
+ *
+ * $Revision: 1.13 $
+ */
+
+#define PRINTK_HEADER "dasd_erp(9343)"
+
+#include "dasd_int.h"
+
+dasd_era_t
+dasd_9343_erp_examine(struct dasd_ccw_req * cqr, struct irb * irb)
+{
+	if (irb->scsw.cstat == 0x00 &&
+	    irb->scsw.dstat == (DEV_STAT_CHN_END | DEV_STAT_DEV_END))
+		return dasd_era_none;
+
+	return dasd_era_recover;
+}

drivers/s390/block/dasd_cmb.c

@@ -0,0 +1,145 @@
+/*
+ * linux/drivers/s390/block/dasd_cmb.c ($Revision: 1.6 $)
+ *
+ * Linux on zSeries Channel Measurement Facility support
+ *  (dasd device driver interface)
+ *
+ * Copyright 2000,2003 IBM Corporation
+ *
+ * Author: Arnd Bergmann <arndb@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/init.h>
+#include <linux/ioctl32.h>
+#include <linux/module.h>
+#include <asm/ccwdev.h>
+#include <asm/cmb.h>
+
+#include "dasd_int.h"
+
+static int
+dasd_ioctl_cmf_enable(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+
+	device = bdev->bd_disk->private_data;
+	if (!device)
+		return -EINVAL;
+
+	return enable_cmf(device->cdev);
+}
+
+static int
+dasd_ioctl_cmf_disable(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+
+	device = bdev->bd_disk->private_data;
+	if (!device)
+		return -EINVAL;
+
+	return disable_cmf(device->cdev);
+}
+
+static int
+dasd_ioctl_readall_cmb(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+	struct cmbdata __user *udata;
+	struct cmbdata data;
+	size_t size;
+	int ret;
+
+	device = bdev->bd_disk->private_data;
+	if (!device)
+		return -EINVAL;
+	udata = (void __user *) args;
+	size = _IOC_SIZE(no);
+
+	if (!access_ok(VERIFY_WRITE, udata, size))
+		return -EFAULT;
+	ret = cmf_readall(device->cdev, &data);
+	if (ret)
+		return ret;
+	if (copy_to_user(udata, &data, min(size, sizeof(*udata))))
+		return -EFAULT;
+	return 0;
+}
+
+/* module initialization below here. dasd already provides a mechanism
+ * to dynamically register ioctl functions, so we simply use this. */
+static inline int
+ioctl_reg(unsigned int no, dasd_ioctl_fn_t handler)
+{
+	int ret;
+	ret = dasd_ioctl_no_register(THIS_MODULE, no, handler);
+#ifdef CONFIG_COMPAT
+	if (ret)
+		return ret;
+
+	ret = register_ioctl32_conversion(no, NULL);
+	if (ret)
+		dasd_ioctl_no_unregister(THIS_MODULE, no, handler);
+#endif
+	return ret;
+}
+
+static inline void
+ioctl_unreg(unsigned int no, dasd_ioctl_fn_t handler)
+{
+	dasd_ioctl_no_unregister(THIS_MODULE, no, handler);
+#ifdef CONFIG_COMPAT
+	unregister_ioctl32_conversion(no);
+#endif
+
+}
+
+static void
+dasd_cmf_exit(void)
+{
+	ioctl_unreg(BIODASDCMFENABLE,  dasd_ioctl_cmf_enable);
+	ioctl_unreg(BIODASDCMFDISABLE, dasd_ioctl_cmf_disable);
+	ioctl_unreg(BIODASDREADALLCMB, dasd_ioctl_readall_cmb);
+}
+
+static int __init
+dasd_cmf_init(void)
+{
+	int ret;
+	ret = ioctl_reg (BIODASDCMFENABLE, dasd_ioctl_cmf_enable);
+	if (ret)
+		goto err;
+	ret = ioctl_reg (BIODASDCMFDISABLE, dasd_ioctl_cmf_disable);
+	if (ret)
+		goto err;
+	ret = ioctl_reg (BIODASDREADALLCMB, dasd_ioctl_readall_cmb);
+	if (ret)
+		goto err;
+
+	return 0;
+err:
+	dasd_cmf_exit();
+
+	return ret;
+}
+
+module_init(dasd_cmf_init);
+module_exit(dasd_cmf_exit);
+
+MODULE_AUTHOR("Arnd Bergmann <[email protected]>");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("channel measurement facility interface for dasd\n"
+		   "Copyright 2003 IBM Corporation\n");

drivers/s390/block/dasd_devmap.c

@@ -0,0 +1,772 @@
+/*
+ * File...........: linux/drivers/s390/block/dasd_devmap.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *		    Horst Hummel <Horst.Hummel@de.ibm.com>
+ *		    Carsten Otte <Cotte@de.ibm.com>
+ *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2001
+ *
+ * Device mapping and dasd= parameter parsing functions. All devmap
+ * functions may not be called from interrupt context. In particular
+ * dasd_get_device is a no-no from interrupt context.
+ *
+ * $Revision: 1.37 $
+ */
+
+#include <linux/config.h>
+#include <linux/ctype.h>
+#include <linux/init.h>
+
+#include <asm/debug.h>
+#include <asm/uaccess.h>
+
+/* This is ugly... */
+#define PRINTK_HEADER "dasd_devmap:"
+
+#include "dasd_int.h"
+
+kmem_cache_t *dasd_page_cache;
+EXPORT_SYMBOL(dasd_page_cache);
+
+/*
+ * dasd_devmap_t is used to store the features and the relation
+ * between device number and device index. To find a dasd_devmap_t
+ * that corresponds to a device number of a device index each
+ * dasd_devmap_t is added to two linked lists, one to search by
+ * the device number and one to search by the device index. As
+ * soon as big minor numbers are available the device index list
+ * can be removed since the device number will then be identical
+ * to the device index.
+ */
+struct dasd_devmap {
+	struct list_head list;
+	char bus_id[BUS_ID_SIZE];
+        unsigned int devindex;
+        unsigned short features;
+	struct dasd_device *device;
+};
+
+/*
+ * Parameter parsing functions for dasd= parameter. The syntax is:
+ *   <devno>		: (0x)?[0-9a-fA-F]+
+ *   <busid>		: [0-0a-f]\.[0-9a-f]\.(0x)?[0-9a-fA-F]+
+ *   <feature>		: ro
+ *   <feature_list>	: \(<feature>(:<feature>)*\)
+ *   <devno-range>	: <devno>(-<devno>)?<feature_list>?
+ *   <busid-range>	: <busid>(-<busid>)?<feature_list>?
+ *   <devices>		: <devno-range>|<busid-range>
+ *   <dasd_module>	: dasd_diag_mod|dasd_eckd_mod|dasd_fba_mod
+ *
+ *   <dasd>		: autodetect|probeonly|<devices>(,<devices>)*
+ */
+
+int dasd_probeonly =  0;	/* is true, when probeonly mode is active */
+int dasd_autodetect = 0;	/* is true, when autodetection is active */
+
+/*
+ * char *dasd[] is intended to hold the ranges supplied by the dasd= statement
+ * it is named 'dasd' to directly be filled by insmod with the comma separated
+ * strings when running as a module.
+ */
+static char *dasd[256];
+/*
+ * Single spinlock to protect devmap structures and lists.
+ */
+static DEFINE_SPINLOCK(dasd_devmap_lock);
+
+/*
+ * Hash lists for devmap structures.
+ */
+static struct list_head dasd_hashlists[256];
+int dasd_max_devindex;
+
+static struct dasd_devmap *dasd_add_busid(char *, int);
+
+static inline int
+dasd_hash_busid(char *bus_id)
+{
+	int hash, i;
+
+	hash = 0;
+	for (i = 0; (i < BUS_ID_SIZE) && *bus_id; i++, bus_id++)
+		hash += *bus_id;
+	return hash & 0xff;
+}
+
+#ifndef MODULE
+/*
+ * The parameter parsing functions for builtin-drivers are called
+ * before kmalloc works. Store the pointers to the parameters strings
+ * into dasd[] for later processing.
+ */
+static int __init
+dasd_call_setup(char *str)
+{
+	static int count = 0;
+
+	if (count < 256)
+		dasd[count++] = str;
+	return 1;
+}
+
+__setup ("dasd=", dasd_call_setup);
+#endif	/* #ifndef MODULE */
+
+/*
+ * Read a device busid/devno from a string.
+ */
+static inline int
+dasd_busid(char **str, int *id0, int *id1, int *devno)
+{
+	int val, old_style;
+ 
+	/* check for leading '0x' */
+	old_style = 0;
+	if ((*str)[0] == '0' && (*str)[1] == 'x') {
+		*str += 2;
+		old_style = 1;
+	}
+	if (!isxdigit((*str)[0]))	/* We require at least one hex digit */
+		return -EINVAL;
+	val = simple_strtoul(*str, str, 16);
+	if (old_style || (*str)[0] != '.') {
+		*id0 = *id1 = 0;
+		if (val < 0 || val > 0xffff)
+			return -EINVAL;
+		*devno = val;
+		return 0;
+	}
+	/* New style x.y.z busid */
+	if (val < 0 || val > 0xff)
+		return -EINVAL;
+	*id0 = val;
+	(*str)++;
+	if (!isxdigit((*str)[0]))	/* We require at least one hex digit */
+		return -EINVAL;
+	val = simple_strtoul(*str, str, 16);
+	if (val < 0 || val > 0xff || (*str)++[0] != '.')
+		return -EINVAL;
+	*id1 = val;
+	if (!isxdigit((*str)[0]))	/* We require at least one hex digit */
+		return -EINVAL;
+	val = simple_strtoul(*str, str, 16);
+	if (val < 0 || val > 0xffff)
+		return -EINVAL;
+	*devno = val;
+	return 0;
+}
+
+/*
+ * Read colon separated list of dasd features. Currently there is
+ * only one: "ro" for read-only devices. The default feature set
+ * is empty (value 0).
+ */
+static inline int
+dasd_feature_list(char *str, char **endp)
+{
+	int features, len, rc;
+
+	rc = 0;
+	if (*str != '(') {
+		*endp = str;
+		return DASD_FEATURE_DEFAULT;
+	}
+	str++;
+	features = 0;
+
+	while (1) {
+		for (len = 0; 
+		     str[len] && str[len] != ':' && str[len] != ')'; len++);
+		if (len == 2 && !strncmp(str, "ro", 2))
+			features |= DASD_FEATURE_READONLY;
+		else if (len == 4 && !strncmp(str, "diag", 4))
+			features |= DASD_FEATURE_USEDIAG;
+		else {
+			MESSAGE(KERN_WARNING,
+				"unsupported feature: %*s, "
+				"ignoring setting", len, str);
+			rc = -EINVAL;
+		}
+		str += len;
+		if (*str != ':')
+			break;
+		str++;
+	}
+	if (*str != ')') {
+		MESSAGE(KERN_WARNING, "%s",
+			"missing ')' in dasd parameter string\n");
+		rc = -EINVAL;
+	} else
+		str++;
+	*endp = str;
+	if (rc != 0)
+		return rc;
+	return features;
+}
+
+/*
+ * Try to match the first element on the comma separated parse string
+ * with one of the known keywords. If a keyword is found, take the approprate
+ * action and return a pointer to the residual string. If the first element
+ * could not be matched to any keyword then return an error code.
+ */
+static char *
+dasd_parse_keyword( char *parsestring ) {
+
+	char *nextcomma, *residual_str;
+	int length;
+
+	nextcomma = strchr(parsestring,',');
+	if (nextcomma) {
+		length = nextcomma - parsestring;
+		residual_str = nextcomma + 1;
+	} else {
+		length = strlen(parsestring);
+		residual_str = parsestring + length;
+        }
+	if (strncmp ("autodetect", parsestring, length) == 0) {
+		dasd_autodetect = 1;
+		MESSAGE (KERN_INFO, "%s",
+			 "turning to autodetection mode");
+                return residual_str;
+        }
+        if (strncmp ("probeonly", parsestring, length) == 0) {
+		dasd_probeonly = 1;
+		MESSAGE(KERN_INFO, "%s",
+			"turning to probeonly mode");
+                return residual_str;
+        }
+        if (strncmp ("fixedbuffers", parsestring, length) == 0) {
+		if (dasd_page_cache)
+			return residual_str;
+		dasd_page_cache =
+			kmem_cache_create("dasd_page_cache", PAGE_SIZE, 0,
+					  SLAB_CACHE_DMA, NULL, NULL );
+		if (!dasd_page_cache)
+			MESSAGE(KERN_WARNING, "%s", "Failed to create slab, "
+				"fixed buffer mode disabled.");
+		else
+			MESSAGE (KERN_INFO, "%s",
+				 "turning on fixed buffer mode");
+                return residual_str;
+        }
+	return ERR_PTR(-EINVAL);
+}
+
+/*
+ * Try to interprete the first element on the comma separated parse string
+ * as a device number or a range of devices. If the interpretation is
+ * successfull, create the matching dasd_devmap entries and return a pointer
+ * to the residual string.
+ * If interpretation fails or in case of an error, return an error code.
+ */
+static char *
+dasd_parse_range( char *parsestring ) {
+
+	struct dasd_devmap *devmap;
+	int from, from_id0, from_id1;
+	int to, to_id0, to_id1;
+	int features, rc;
+	char bus_id[BUS_ID_SIZE+1], *str;
+
+	str = parsestring;
+	rc = dasd_busid(&str, &from_id0, &from_id1, &from);
+	if (rc == 0) {
+		to = from;
+		to_id0 = from_id0;
+		to_id1 = from_id1;
+		if (*str == '-') {
+			str++;
+			rc = dasd_busid(&str, &to_id0, &to_id1, &to);
+		}
+	}
+	if (rc == 0 &&
+	    (from_id0 != to_id0 || from_id1 != to_id1 || from > to))
+		rc = -EINVAL;
+	if (rc) {
+		MESSAGE(KERN_ERR, "Invalid device range %s", parsestring);
+		return ERR_PTR(rc);
+	}
+	features = dasd_feature_list(str, &str);
+	if (features < 0)
+		return ERR_PTR(-EINVAL);
+	while (from <= to) {
+		sprintf(bus_id, "%01x.%01x.%04x",
+			from_id0, from_id1, from++);
+		devmap = dasd_add_busid(bus_id, features);
+		if (IS_ERR(devmap))
+			return (char *)devmap;
+	}
+	if (*str == ',')
+		return str + 1;
+	if (*str == '\0')
+		return str;
+	MESSAGE(KERN_WARNING,
+		"junk at end of dasd parameter string: %s\n", str);
+	return ERR_PTR(-EINVAL);
+}
+
+static inline char *
+dasd_parse_next_element( char *parsestring ) {
+	char * residual_str;
+	residual_str = dasd_parse_keyword(parsestring);
+	if (!IS_ERR(residual_str))
+		return residual_str;
+	residual_str = dasd_parse_range(parsestring);
+	return residual_str;
+}
+
+/*
+ * Parse parameters stored in dasd[]
+ * The 'dasd=...' parameter allows to specify a comma separated list of
+ * keywords and device ranges. When the dasd driver is build into the kernel,
+ * the complete list will be stored as one element of the dasd[] array.
+ * When the dasd driver is build as a module, then the list is broken into
+ * it's elements and each dasd[] entry contains one element.
+ */
+int
+dasd_parse(void)
+{
+	int rc, i;
+	char *parsestring;
+
+	rc = 0;
+	for (i = 0; i < 256; i++) {
+		if (dasd[i] == NULL)
+			break;
+		parsestring = dasd[i];
+		/* loop over the comma separated list in the parsestring */
+		while (*parsestring) {
+			parsestring = dasd_parse_next_element(parsestring);
+			if(IS_ERR(parsestring)) {
+				rc = PTR_ERR(parsestring);
+				break;
+			}
+		}
+		if (rc) {
+			DBF_EVENT(DBF_ALERT, "%s", "invalid range found");
+			break;
+		}
+	}
+	return rc;
+}
+
+/*
+ * Add a devmap for the device specified by busid. It is possible that
+ * the devmap already exists (dasd= parameter). The order of the devices
+ * added through this function will define the kdevs for the individual
+ * devices. 
+ */
+static struct dasd_devmap *
+dasd_add_busid(char *bus_id, int features)
+{
+	struct dasd_devmap *devmap, *new, *tmp;
+	int hash;
+
+	new = (struct dasd_devmap *)
+		kmalloc(sizeof(struct dasd_devmap), GFP_KERNEL);
+	if (!new)
+		return ERR_PTR(-ENOMEM);
+	spin_lock(&dasd_devmap_lock);
+	devmap = 0;
+	hash = dasd_hash_busid(bus_id);
+	list_for_each_entry(tmp, &dasd_hashlists[hash], list)
+		if (strncmp(tmp->bus_id, bus_id, BUS_ID_SIZE) == 0) {
+			devmap = tmp;
+			break;
+		}
+	if (!devmap) {
+		/* This bus_id is new. */
+		new->devindex = dasd_max_devindex++;
+		strncpy(new->bus_id, bus_id, BUS_ID_SIZE);
+		new->features = features;
+		new->device = 0;
+		list_add(&new->list, &dasd_hashlists[hash]);
+		devmap = new;
+		new = 0;
+	}
+	spin_unlock(&dasd_devmap_lock);
+	if (new)
+		kfree(new);
+	return devmap;
+}
+
+/*
+ * Find devmap for device with given bus_id.
+ */
+static struct dasd_devmap *
+dasd_find_busid(char *bus_id)
+{
+	struct dasd_devmap *devmap, *tmp;
+	int hash;
+
+	spin_lock(&dasd_devmap_lock);
+	devmap = ERR_PTR(-ENODEV);
+	hash = dasd_hash_busid(bus_id);
+	list_for_each_entry(tmp, &dasd_hashlists[hash], list) {
+		if (strncmp(tmp->bus_id, bus_id, BUS_ID_SIZE) == 0) {
+			devmap = tmp;
+			break;
+		}
+	}
+	spin_unlock(&dasd_devmap_lock);
+	return devmap;
+}
+
+/*
+ * Check if busid has been added to the list of dasd ranges.
+ */
+int
+dasd_busid_known(char *bus_id)
+{
+	return IS_ERR(dasd_find_busid(bus_id)) ? -ENOENT : 0;
+}
+
+/*
+ * Forget all about the device numbers added so far.
+ * This may only be called at module unload or system shutdown.
+ */
+static void
+dasd_forget_ranges(void)
+{
+	struct dasd_devmap *devmap, *n;
+	int i;
+
+	spin_lock(&dasd_devmap_lock);
+	for (i = 0; i < 256; i++) {
+		list_for_each_entry_safe(devmap, n, &dasd_hashlists[i], list) {
+			if (devmap->device != NULL)
+				BUG();
+			list_del(&devmap->list);
+			kfree(devmap);
+		}
+	}
+	spin_unlock(&dasd_devmap_lock);
+}
+
+/*
+ * Find the device struct by its device index.
+ */
+struct dasd_device *
+dasd_device_from_devindex(int devindex)
+{
+	struct dasd_devmap *devmap, *tmp;
+	struct dasd_device *device;
+	int i;
+
+	spin_lock(&dasd_devmap_lock);
+	devmap = 0;
+	for (i = 0; (i < 256) && !devmap; i++)
+		list_for_each_entry(tmp, &dasd_hashlists[i], list)
+			if (tmp->devindex == devindex) {
+				/* Found the devmap for the device. */
+				devmap = tmp;
+				break;
+			}
+	if (devmap && devmap->device) {
+		device = devmap->device;
+		dasd_get_device(device);
+	} else
+		device = ERR_PTR(-ENODEV);
+	spin_unlock(&dasd_devmap_lock);
+	return device;
+}
+
+/*
+ * Return devmap for cdev. If no devmap exists yet, create one and
+ * connect it to the cdev.
+ */
+static struct dasd_devmap *
+dasd_devmap_from_cdev(struct ccw_device *cdev)
+{
+	struct dasd_devmap *devmap;
+
+	devmap = dasd_find_busid(cdev->dev.bus_id);
+	if (IS_ERR(devmap))
+		devmap = dasd_add_busid(cdev->dev.bus_id,
+					DASD_FEATURE_DEFAULT);
+	return devmap;
+}
+
+/*
+ * Create a dasd device structure for cdev.
+ */
+struct dasd_device *
+dasd_create_device(struct ccw_device *cdev)
+{
+	struct dasd_devmap *devmap;
+	struct dasd_device *device;
+	int rc;
+
+	devmap = dasd_devmap_from_cdev(cdev);
+	if (IS_ERR(devmap))
+		return (void *) devmap;
+	cdev->dev.driver_data = devmap;
+
+	device = dasd_alloc_device();
+	if (IS_ERR(device))
+		return device;
+	atomic_set(&device->ref_count, 2);
+
+	spin_lock(&dasd_devmap_lock);
+	if (!devmap->device) {
+		devmap->device = device;
+		device->devindex = devmap->devindex;
+		if (devmap->features & DASD_FEATURE_READONLY)
+			set_bit(DASD_FLAG_RO, &device->flags);
+		else
+			clear_bit(DASD_FLAG_RO, &device->flags);
+		if (devmap->features & DASD_FEATURE_USEDIAG)
+			set_bit(DASD_FLAG_USE_DIAG, &device->flags);
+		else
+			clear_bit(DASD_FLAG_USE_DIAG, &device->flags);
+		get_device(&cdev->dev);
+		device->cdev = cdev;
+		rc = 0;
+	} else
+		/* Someone else was faster. */
+		rc = -EBUSY;
+	spin_unlock(&dasd_devmap_lock);
+
+	if (rc) {
+		dasd_free_device(device);
+		return ERR_PTR(rc);
+	}
+	return device;
+}
+
+/*
+ * Wait queue for dasd_delete_device waits.
+ */
+static DECLARE_WAIT_QUEUE_HEAD(dasd_delete_wq);
+
+/*
+ * Remove a dasd device structure. The passed referenced
+ * is destroyed.
+ */
+void
+dasd_delete_device(struct dasd_device *device)
+{
+	struct ccw_device *cdev;
+	struct dasd_devmap *devmap;
+
+	/* First remove device pointer from devmap. */
+	devmap = dasd_find_busid(device->cdev->dev.bus_id);
+	if (IS_ERR(devmap))
+		BUG();
+	spin_lock(&dasd_devmap_lock);
+	if (devmap->device != device) {
+		spin_unlock(&dasd_devmap_lock);
+		dasd_put_device(device);
+		return;
+	}
+	devmap->device = NULL;
+	spin_unlock(&dasd_devmap_lock);
+
+	/* Drop ref_count by 2, one for the devmap reference and
+	 * one for the passed reference. */
+	atomic_sub(2, &device->ref_count);
+
+	/* Wait for reference counter to drop to zero. */
+	wait_event(dasd_delete_wq, atomic_read(&device->ref_count) == 0);
+
+	/* Disconnect dasd_device structure from ccw_device structure. */
+	cdev = device->cdev;
+	device->cdev = NULL;
+
+	/* Disconnect dasd_devmap structure from ccw_device structure. */
+	cdev->dev.driver_data = NULL;
+
+	/* Put ccw_device structure. */
+	put_device(&cdev->dev);
+
+	/* Now the device structure can be freed. */
+	dasd_free_device(device);
+}
+
+/*
+ * Reference counter dropped to zero. Wake up waiter
+ * in dasd_delete_device.
+ */
+void
+dasd_put_device_wake(struct dasd_device *device)
+{
+	wake_up(&dasd_delete_wq);
+}
+
+/*
+ * Return dasd_device structure associated with cdev.
+ */
+struct dasd_device *
+dasd_device_from_cdev(struct ccw_device *cdev)
+{
+	struct dasd_devmap *devmap;
+	struct dasd_device *device;
+
+	device = ERR_PTR(-ENODEV);
+	spin_lock(&dasd_devmap_lock);
+	devmap = cdev->dev.driver_data;
+	if (devmap && devmap->device) {
+		device = devmap->device;
+		dasd_get_device(device);
+	}
+	spin_unlock(&dasd_devmap_lock);
+	return device;
+}
+
+/*
+ * SECTION: files in sysfs
+ */
+
+/*
+ * readonly controls the readonly status of a dasd
+ */
+static ssize_t
+dasd_ro_show(struct device *dev, char *buf)
+{
+	struct dasd_devmap *devmap;
+	int ro_flag;
+
+	devmap = dasd_find_busid(dev->bus_id);
+	if (!IS_ERR(devmap))
+		ro_flag = (devmap->features & DASD_FEATURE_READONLY) != 0;
+	else
+		ro_flag = (DASD_FEATURE_DEFAULT & DASD_FEATURE_READONLY) != 0;
+	return snprintf(buf, PAGE_SIZE, ro_flag ? "1\n" : "0\n");
+}
+
+static ssize_t
+dasd_ro_store(struct device *dev, const char *buf, size_t count)
+{
+	struct dasd_devmap *devmap;
+	int ro_flag;
+
+	devmap = dasd_devmap_from_cdev(to_ccwdev(dev));
+	if (IS_ERR(devmap))
+		return PTR_ERR(devmap);
+	ro_flag = buf[0] == '1';
+	spin_lock(&dasd_devmap_lock);
+	if (ro_flag)
+		devmap->features |= DASD_FEATURE_READONLY;
+	else
+		devmap->features &= ~DASD_FEATURE_READONLY;
+	if (devmap->device) {
+		if (devmap->device->gdp)
+			set_disk_ro(devmap->device->gdp, ro_flag);
+		if (ro_flag)
+			set_bit(DASD_FLAG_RO, &devmap->device->flags);
+		else
+			clear_bit(DASD_FLAG_RO, &devmap->device->flags);
+	}
+	spin_unlock(&dasd_devmap_lock);
+	return count;
+}
+
+static DEVICE_ATTR(readonly, 0644, dasd_ro_show, dasd_ro_store);
+
+/*
+ * use_diag controls whether the driver should use diag rather than ssch
+ * to talk to the device
+ */
+static ssize_t 
+dasd_use_diag_show(struct device *dev, char *buf)
+{
+	struct dasd_devmap *devmap;
+	int use_diag;
+
+	devmap = dasd_find_busid(dev->bus_id);
+	if (!IS_ERR(devmap))
+		use_diag = (devmap->features & DASD_FEATURE_USEDIAG) != 0;
+	else
+		use_diag = (DASD_FEATURE_DEFAULT & DASD_FEATURE_USEDIAG) != 0;
+	return sprintf(buf, use_diag ? "1\n" : "0\n");
+}
+
+static ssize_t
+dasd_use_diag_store(struct device *dev, const char *buf, size_t count)
+{
+	struct dasd_devmap *devmap;
+	ssize_t rc;
+	int use_diag;
+
+	devmap = dasd_devmap_from_cdev(to_ccwdev(dev));
+	if (IS_ERR(devmap))
+		return PTR_ERR(devmap);
+	use_diag = buf[0] == '1';
+	spin_lock(&dasd_devmap_lock);
+	/* Changing diag discipline flag is only allowed in offline state. */
+	rc = count;
+	if (!devmap->device) {
+		if (use_diag)
+			devmap->features |= DASD_FEATURE_USEDIAG;
+		else
+			devmap->features &= ~DASD_FEATURE_USEDIAG;
+	} else
+		rc = -EPERM;
+	spin_unlock(&dasd_devmap_lock);
+	return rc;
+}
+
+static
+DEVICE_ATTR(use_diag, 0644, dasd_use_diag_show, dasd_use_diag_store);
+
+static ssize_t
+dasd_discipline_show(struct device *dev, char *buf)
+{
+	struct dasd_devmap *devmap;
+	char *dname;
+
+	spin_lock(&dasd_devmap_lock);
+	dname = "none";
+	devmap = dev->driver_data;
+	if (devmap && devmap->device && devmap->device->discipline)
+		dname = devmap->device->discipline->name;
+	spin_unlock(&dasd_devmap_lock);
+	return snprintf(buf, PAGE_SIZE, "%s\n", dname);
+}
+
+static DEVICE_ATTR(discipline, 0444, dasd_discipline_show, NULL);
+
+static struct attribute * dasd_attrs[] = {
+	&dev_attr_readonly.attr,
+	&dev_attr_discipline.attr,
+	&dev_attr_use_diag.attr,
+	NULL,
+};
+
+static struct attribute_group dasd_attr_group = {
+	.attrs = dasd_attrs,
+};
+
+int
+dasd_add_sysfs_files(struct ccw_device *cdev)
+{
+	return sysfs_create_group(&cdev->dev.kobj, &dasd_attr_group);
+}
+
+void
+dasd_remove_sysfs_files(struct ccw_device *cdev)
+{
+	sysfs_remove_group(&cdev->dev.kobj, &dasd_attr_group);
+}
+
+
+int
+dasd_devmap_init(void)
+{
+	int i;
+
+	/* Initialize devmap structures. */
+	dasd_max_devindex = 0;
+	for (i = 0; i < 256; i++)
+		INIT_LIST_HEAD(&dasd_hashlists[i]);
+	return 0;
+
+}
+
+void
+dasd_devmap_exit(void)
+{
+	dasd_forget_ranges();
+}

drivers/s390/block/dasd_diag.c

@@ -0,0 +1,541 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_diag.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Based on.......: linux/drivers/s390/block/mdisk.c
+ * ...............: by Hartmunt Penner <hpenner@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ *
+ * $Revision: 1.42 $
+ */
+
+#include <linux/config.h>
+#include <linux/stddef.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/hdreg.h>	/* HDIO_GETGEO			    */
+#include <linux/bio.h>
+#include <linux/module.h>
+#include <linux/init.h>
+
+#include <asm/dasd.h>
+#include <asm/debug.h>
+#include <asm/ebcdic.h>
+#include <asm/io.h>
+#include <asm/s390_ext.h>
+#include <asm/todclk.h>
+
+#include "dasd_int.h"
+#include "dasd_diag.h"
+
+#ifdef PRINTK_HEADER
+#undef PRINTK_HEADER
+#endif				/* PRINTK_HEADER */
+#define PRINTK_HEADER "dasd(diag):"
+
+MODULE_LICENSE("GPL");
+
+struct dasd_discipline dasd_diag_discipline;
+
+struct dasd_diag_private {
+	struct dasd_diag_characteristics rdc_data;
+	struct dasd_diag_rw_io iob;
+	struct dasd_diag_init_io iib;
+	unsigned int pt_block;
+};
+
+struct dasd_diag_req {
+	int block_count;
+	struct dasd_diag_bio bio[0];
+};
+
+static __inline__ int
+dia250(void *iob, int cmd)
+{
+	int rc;
+
+	__asm__ __volatile__("    lhi   %0,3\n"
+			     "	  lr	0,%2\n"
+			     "	  diag	0,%1,0x250\n"
+			     "0:  ipm	%0\n"
+			     "	  srl	%0,28\n"
+			     "	  or	%0,1\n"
+			     "1:\n"
+#ifndef CONFIG_ARCH_S390X
+			     ".section __ex_table,\"a\"\n"
+			     "	  .align 4\n"
+			     "	  .long 0b,1b\n"
+			     ".previous\n"
+#else
+			     ".section __ex_table,\"a\"\n"
+			     "	  .align 8\n"
+			     "	  .quad  0b,1b\n"
+			     ".previous\n"
+#endif
+			     : "=&d" (rc)
+			     : "d" (cmd), "d" ((void *) __pa(iob))
+			     : "0", "1", "cc");
+	return rc;
+}
+
+static __inline__ int
+mdsk_init_io(struct dasd_device * device, int blocksize, int offset, int size)
+{
+	struct dasd_diag_private *private;
+	struct dasd_diag_init_io *iib;
+	int rc;
+
+	private = (struct dasd_diag_private *) device->private;
+	iib = &private->iib;
+	memset(iib, 0, sizeof (struct dasd_diag_init_io));
+
+	iib->dev_nr = _ccw_device_get_device_number(device->cdev);
+	iib->block_size = blocksize;
+	iib->offset = offset;
+	iib->start_block = 0;
+	iib->end_block = size;
+
+	rc = dia250(iib, INIT_BIO);
+
+	return rc & 3;
+}
+
+static __inline__ int
+mdsk_term_io(struct dasd_device * device)
+{
+	struct dasd_diag_private *private;
+	struct dasd_diag_init_io *iib;
+	int rc;
+
+	private = (struct dasd_diag_private *) device->private;
+	iib = &private->iib;
+	memset(iib, 0, sizeof (struct dasd_diag_init_io));
+	iib->dev_nr = _ccw_device_get_device_number(device->cdev);
+	rc = dia250(iib, TERM_BIO);
+	return rc & 3;
+}
+
+static int
+dasd_start_diag(struct dasd_ccw_req * cqr)
+{
+	struct dasd_device *device;
+	struct dasd_diag_private *private;
+	struct dasd_diag_req *dreq;
+	int rc;
+
+	device = cqr->device;
+	private = (struct dasd_diag_private *) device->private;
+	dreq = (struct dasd_diag_req *) cqr->data;
+
+	private->iob.dev_nr = _ccw_device_get_device_number(device->cdev);
+	private->iob.key = 0;
+	private->iob.flags = 2;	/* do asynchronous io */
+	private->iob.block_count = dreq->block_count;
+	private->iob.interrupt_params = (u32)(addr_t) cqr;
+	private->iob.bio_list = __pa(dreq->bio);
+
+	cqr->startclk = get_clock();
+
+	rc = dia250(&private->iob, RW_BIO);
+	if (rc > 8) {
+		DEV_MESSAGE(KERN_WARNING, device, "dia250 returned CC %d", rc);
+		cqr->status = DASD_CQR_ERROR;
+	} else if (rc == 0) {
+		cqr->status = DASD_CQR_DONE;
+		dasd_schedule_bh(device);
+	} else {
+		cqr->status = DASD_CQR_IN_IO;
+		rc = 0;
+	}
+	return rc;
+}
+
+static void
+dasd_ext_handler(struct pt_regs *regs, __u16 code)
+{
+	struct dasd_ccw_req *cqr, *next;
+	struct dasd_device *device;
+	unsigned long long expires;
+	unsigned long flags;
+	char status;
+	int ip;
+
+	/*
+	 * Get the external interruption subcode. VM stores
+	 * this in the 'cpu address' field associated with
+	 * the external interrupt. For diag 250 the subcode
+	 * needs to be 3.
+	 */
+	if ((S390_lowcore.cpu_addr & 0xff00) != 0x0300)
+		return;
+	status = *((char *) &S390_lowcore.ext_params + 5);
+	ip = S390_lowcore.ext_params;
+
+	if (!ip) {		/* no intparm: unsolicited interrupt */
+		MESSAGE(KERN_DEBUG, "%s", "caught unsolicited interrupt");
+		return;
+	}
+	cqr = (struct dasd_ccw_req *)(addr_t) ip;
+	device = (struct dasd_device *) cqr->device;
+	if (strncmp(device->discipline->ebcname, (char *) &cqr->magic, 4)) {
+		DEV_MESSAGE(KERN_WARNING, device,
+			    " magic number of dasd_ccw_req 0x%08X doesn't"
+			    " match discipline 0x%08X",
+			    cqr->magic, *(int *) (&device->discipline->name));
+		return;
+	}
+
+	/* get irq lock to modify request queue */
+	spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
+
+	cqr->stopclk = get_clock();
+
+	expires = 0;
+	if (status == 0) {
+		cqr->status = DASD_CQR_DONE;
+		/* Start first request on queue if possible -> fast_io. */
+		if (!list_empty(&device->ccw_queue)) {
+			next = list_entry(device->ccw_queue.next,
+					  struct dasd_ccw_req, list);
+			if (next->status == DASD_CQR_QUEUED) {
+				if (dasd_start_diag(next) == 0)
+					expires = next->expires;
+				else
+					DEV_MESSAGE(KERN_WARNING, device, "%s",
+						    "Interrupt fastpath "
+						    "failed!");
+			}
+		}
+	} else 
+		cqr->status = DASD_CQR_FAILED;
+
+	if (expires != 0)
+		dasd_set_timer(device, expires);
+	else
+		dasd_clear_timer(device);
+	dasd_schedule_bh(device);
+
+	spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
+}
+
+static int
+dasd_diag_check_device(struct dasd_device *device)
+{
+	struct dasd_diag_private *private;
+	struct dasd_diag_characteristics *rdc_data;
+	struct dasd_diag_bio bio;
+	long *label;
+	int sb, bsize;
+	int rc;
+
+	private = (struct dasd_diag_private *) device->private;
+	if (private == NULL) {
+		private = kmalloc(sizeof(struct dasd_diag_private),GFP_KERNEL);
+		if (private == NULL) {
+			DEV_MESSAGE(KERN_WARNING, device, "%s",
+				"memory allocation failed for private data");
+			return -ENOMEM;
+		}
+		device->private = (void *) private;
+	}
+	/* Read Device Characteristics */
+	rdc_data = (void *) &(private->rdc_data);
+	rdc_data->dev_nr = _ccw_device_get_device_number(device->cdev);
+	rdc_data->rdc_len = sizeof (struct dasd_diag_characteristics);
+
+	rc = diag210((struct diag210 *) rdc_data);
+	if (rc)
+		return -ENOTSUPP;
+
+	/* Figure out position of label block */
+	switch (private->rdc_data.vdev_class) {
+	case DEV_CLASS_FBA:
+		private->pt_block = 1;
+		break;
+	case DEV_CLASS_ECKD:
+		private->pt_block = 2;
+		break;
+	default:
+		return -ENOTSUPP;
+	}
+
+	DBF_DEV_EVENT(DBF_INFO, device,
+		      "%04X: %04X on real %04X/%02X",
+		      rdc_data->dev_nr,
+		      rdc_data->vdev_type,
+		      rdc_data->rdev_type, rdc_data->rdev_model);
+
+	/* terminate all outstanding operations */
+	mdsk_term_io(device);
+
+	/* figure out blocksize of device */
+	label = (long *) get_zeroed_page(GFP_KERNEL);
+	if (label == NULL)  {
+		DEV_MESSAGE(KERN_WARNING, device, "%s",
+			    "No memory to allocate initialization request");
+		return -ENOMEM;
+	}
+	/* try all sizes - needed for ECKD devices */
+	for (bsize = 512; bsize <= PAGE_SIZE; bsize <<= 1) {
+		mdsk_init_io(device, bsize, 0, 64);
+		memset(&bio, 0, sizeof (struct dasd_diag_bio));
+		bio.type = MDSK_READ_REQ;
+		bio.block_number = private->pt_block + 1;
+		bio.buffer = __pa(label);
+		memset(&private->iob, 0, sizeof (struct dasd_diag_rw_io));
+		private->iob.dev_nr = rdc_data->dev_nr;
+		private->iob.key = 0;
+		private->iob.flags = 0;	/* do synchronous io */
+		private->iob.block_count = 1;
+		private->iob.interrupt_params = 0;
+		private->iob.bio_list = __pa(&bio);
+		if (dia250(&private->iob, RW_BIO) == 0)
+			break;
+		mdsk_term_io(device);
+	}
+	if (bsize <= PAGE_SIZE && label[0] == 0xc3d4e2f1) {
+		/* get formatted blocksize from label block */
+		bsize = (int) label[3];
+		device->blocks = label[7];
+		device->bp_block = bsize;
+		device->s2b_shift = 0;	/* bits to shift 512 to get a block */
+		for (sb = 512; sb < bsize; sb = sb << 1)
+			device->s2b_shift++;
+		
+		DEV_MESSAGE(KERN_INFO, device,
+			    "capacity (%dkB blks): %ldkB",
+			    (device->bp_block >> 10),
+			    (device->blocks << device->s2b_shift) >> 1);
+		rc = 0;
+	} else {
+		if (bsize > PAGE_SIZE)
+			DEV_MESSAGE(KERN_WARNING, device, "%s",
+				    "DIAG access failed");
+		else
+			DEV_MESSAGE(KERN_WARNING, device, "%s",
+				    "volume is not CMS formatted");
+		rc = -EMEDIUMTYPE;
+	}
+	free_page((long) label);
+	return rc;
+}
+
+static int
+dasd_diag_fill_geometry(struct dasd_device *device, struct hd_geometry *geo)
+{
+	if (dasd_check_blocksize(device->bp_block) != 0)
+		return -EINVAL;
+	geo->cylinders = (device->blocks << device->s2b_shift) >> 10;
+	geo->heads = 16;
+	geo->sectors = 128 >> device->s2b_shift;
+	return 0;
+}
+
+static dasd_era_t
+dasd_diag_examine_error(struct dasd_ccw_req * cqr, struct irb * stat)
+{
+	return dasd_era_fatal;
+}
+
+static dasd_erp_fn_t
+dasd_diag_erp_action(struct dasd_ccw_req * cqr)
+{
+	return dasd_default_erp_action;
+}
+
+static dasd_erp_fn_t
+dasd_diag_erp_postaction(struct dasd_ccw_req * cqr)
+{
+	return dasd_default_erp_postaction;
+}
+
+static struct dasd_ccw_req *
+dasd_diag_build_cp(struct dasd_device * device, struct request *req)
+{
+	struct dasd_ccw_req *cqr;
+	struct dasd_diag_req *dreq;
+	struct dasd_diag_bio *dbio;
+	struct bio *bio;
+	struct bio_vec *bv;
+	char *dst;
+	int count, datasize;
+	sector_t recid, first_rec, last_rec;
+	unsigned blksize, off;
+	unsigned char rw_cmd;
+	int i;
+
+	if (rq_data_dir(req) == READ)
+		rw_cmd = MDSK_READ_REQ;
+	else if (rq_data_dir(req) == WRITE)
+		rw_cmd = MDSK_WRITE_REQ;
+	else
+		return ERR_PTR(-EINVAL);
+	blksize = device->bp_block;
+	/* Calculate record id of first and last block. */
+	first_rec = req->sector >> device->s2b_shift;
+	last_rec = (req->sector + req->nr_sectors - 1) >> device->s2b_shift;
+	/* Check struct bio and count the number of blocks for the request. */
+	count = 0;
+	rq_for_each_bio(bio, req) {
+		bio_for_each_segment(bv, bio, i) {
+			if (bv->bv_len & (blksize - 1))
+				/* Fba can only do full blocks. */
+				return ERR_PTR(-EINVAL);
+			count += bv->bv_len >> (device->s2b_shift + 9);
+		}
+	}
+	/* Paranoia. */
+	if (count != last_rec - first_rec + 1)
+		return ERR_PTR(-EINVAL);
+	/* Build the request */
+	datasize = sizeof(struct dasd_diag_req) +
+		count*sizeof(struct dasd_diag_bio);
+	cqr = dasd_smalloc_request(dasd_diag_discipline.name, 0,
+				   datasize, device);
+	if (IS_ERR(cqr))
+		return cqr;
+	
+	dreq = (struct dasd_diag_req *) cqr->data;
+	dreq->block_count = count;
+	dbio = dreq->bio;
+	recid = first_rec;
+	rq_for_each_bio(bio, req) {
+		bio_for_each_segment(bv, bio, i) {
+			dst = page_address(bv->bv_page) + bv->bv_offset;
+			for (off = 0; off < bv->bv_len; off += blksize) {
+				memset(dbio, 0, sizeof (struct dasd_diag_bio));
+				dbio->type = rw_cmd;
+				dbio->block_number = recid + 1;
+				dbio->buffer = __pa(dst);
+				dbio++;
+				dst += blksize;
+				recid++;
+			}
+		}
+	}
+	cqr->buildclk = get_clock();
+	cqr->device = device;
+	cqr->expires = 50 * HZ;	/* 50 seconds */
+	cqr->status = DASD_CQR_FILLED;
+	return cqr;
+}
+
+static int
+dasd_diag_free_cp(struct dasd_ccw_req *cqr, struct request *req)
+{
+	int status;
+
+	status = cqr->status == DASD_CQR_DONE;
+	dasd_sfree_request(cqr, cqr->device);
+	return status;
+}
+
+static int
+dasd_diag_fill_info(struct dasd_device * device,
+		    struct dasd_information2_t * info)
+{
+	struct dasd_diag_private *private;
+
+	private = (struct dasd_diag_private *) device->private;
+	info->label_block = private->pt_block;
+	info->FBA_layout = 1;
+	info->format = DASD_FORMAT_LDL;
+	info->characteristics_size = sizeof (struct dasd_diag_characteristics);
+	memcpy(info->characteristics,
+	       &((struct dasd_diag_private *) device->private)->rdc_data,
+	       sizeof (struct dasd_diag_characteristics));
+	info->confdata_size = 0;
+	return 0;
+}
+
+static void
+dasd_diag_dump_sense(struct dasd_device *device, struct dasd_ccw_req * req,
+		     struct irb *stat)
+{
+	DEV_MESSAGE(KERN_ERR, device, "%s",
+		    "dump sense not available for DIAG data");
+}
+
+/*
+ * max_blocks is dependent on the amount of storage that is available
+ * in the static io buffer for each device. Currently each device has
+ * 8192 bytes (=2 pages). dasd diag is only relevant for 31 bit.
+ * The struct dasd_ccw_req has 96 bytes, the struct dasd_diag_req has
+ * 8 bytes and the struct dasd_diag_bio for each block has 16 bytes. 
+ * That makes:
+ * (8192 - 96 - 8) / 16 = 505.5 blocks at maximum.
+ * We want to fit two into the available memory so that we can immediately
+ * start the next request if one finishes off. That makes 252.75 blocks
+ * for one request. Give a little safety and the result is 240.
+ */
+struct dasd_discipline dasd_diag_discipline = {
+	.owner = THIS_MODULE,
+	.name = "DIAG",
+	.ebcname = "DIAG",
+	.max_blocks = 240,
+	.check_device = dasd_diag_check_device,
+	.fill_geometry = dasd_diag_fill_geometry,
+	.start_IO = dasd_start_diag,
+	.examine_error = dasd_diag_examine_error,
+	.erp_action = dasd_diag_erp_action,
+	.erp_postaction = dasd_diag_erp_postaction,
+	.build_cp = dasd_diag_build_cp,
+	.free_cp = dasd_diag_free_cp,
+	.dump_sense = dasd_diag_dump_sense,
+	.fill_info = dasd_diag_fill_info,
+};
+
+static int __init
+dasd_diag_init(void)
+{
+	if (!MACHINE_IS_VM) {
+		MESSAGE_LOG(KERN_INFO,
+			    "Machine is not VM: %s "
+			    "discipline not initializing",
+			    dasd_diag_discipline.name);
+		return -EINVAL;
+	}
+	ASCEBC(dasd_diag_discipline.ebcname, 4);
+
+	ctl_set_bit(0, 9);
+	register_external_interrupt(0x2603, dasd_ext_handler);
+	dasd_diag_discipline_pointer = &dasd_diag_discipline;
+	return 0;
+}
+
+static void __exit
+dasd_diag_cleanup(void)
+{
+	if (!MACHINE_IS_VM) {
+		MESSAGE_LOG(KERN_INFO,
+			    "Machine is not VM: %s "
+			    "discipline not cleaned",
+			    dasd_diag_discipline.name);
+		return;
+	}
+	unregister_external_interrupt(0x2603, dasd_ext_handler);
+	ctl_clear_bit(0, 9);
+	dasd_diag_discipline_pointer = NULL;
+}
+
+module_init(dasd_diag_init);
+module_exit(dasd_diag_cleanup);
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-indent-level: 4 
+ * c-brace-imaginary-offset: 0
+ * c-brace-offset: -4
+ * c-argdecl-indent: 4
+ * c-label-offset: -4
+ * c-continued-statement-offset: 4
+ * c-continued-brace-offset: 0
+ * indent-tabs-mode: 1
+ * tab-width: 8
+ * End:
+ */

drivers/s390/block/dasd_diag.h

@@ -0,0 +1,66 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_diag.h
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Based on.......: linux/drivers/s390/block/mdisk.h
+ * ...............: by Hartmunt Penner <hpenner@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ *
+ * $Revision: 1.6 $
+ */
+
+#define MDSK_WRITE_REQ 0x01
+#define MDSK_READ_REQ  0x02
+
+#define INIT_BIO	0x00
+#define RW_BIO		0x01
+#define TERM_BIO	0x02
+
+#define DEV_CLASS_FBA	0x01
+#define DEV_CLASS_ECKD	0x04
+
+struct dasd_diag_characteristics {
+	u16 dev_nr;
+	u16 rdc_len;
+	u8 vdev_class;
+	u8 vdev_type;
+	u8 vdev_status;
+	u8 vdev_flags;
+	u8 rdev_class;
+	u8 rdev_type;
+	u8 rdev_model;
+	u8 rdev_features;
+} __attribute__ ((packed, aligned(4)));
+
+struct dasd_diag_bio {
+	u8 type;
+	u8 status;
+	u16 spare1;
+	u32 block_number;
+	u32 alet;
+	u32 buffer;
+} __attribute__ ((packed, aligned(8)));
+
+struct dasd_diag_init_io {
+	u16 dev_nr;
+	u16 spare1[11];
+	u32 block_size;
+	u32 offset;
+	u32 start_block;
+	u32 end_block;
+	u32 spare2[6];
+} __attribute__ ((packed, aligned(8)));
+
+struct dasd_diag_rw_io {
+	u16 dev_nr;
+	u16 spare1[11];
+	u8 key;
+	u8 flags;
+	u16 spare2;
+	u32 block_count;
+	u32 alet;
+	u32 bio_list;
+	u32 interrupt_params;
+	u32 spare3[5];
+} __attribute__ ((packed, aligned(8)));
+

drivers/s390/block/dasd_eckd.h

@@ -0,0 +1,346 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_eckd.h
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *                  Horst Hummel <Horst.Hummel@de.ibm.com> 
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ *
+ * $Revision: 1.10 $
+ */
+
+#ifndef DASD_ECKD_H
+#define DASD_ECKD_H
+
+/*****************************************************************************
+ * SECTION: CCW Definitions
+ ****************************************************************************/
+#define DASD_ECKD_CCW_WRITE		 0x05
+#define DASD_ECKD_CCW_READ		 0x06
+#define DASD_ECKD_CCW_WRITE_HOME_ADDRESS 0x09
+#define DASD_ECKD_CCW_READ_HOME_ADDRESS	 0x0a
+#define DASD_ECKD_CCW_WRITE_KD		 0x0d
+#define DASD_ECKD_CCW_READ_KD		 0x0e
+#define DASD_ECKD_CCW_ERASE		 0x11
+#define DASD_ECKD_CCW_READ_COUNT	 0x12
+#define DASD_ECKD_CCW_SLCK		 0x14
+#define DASD_ECKD_CCW_WRITE_RECORD_ZERO	 0x15
+#define DASD_ECKD_CCW_READ_RECORD_ZERO	 0x16
+#define DASD_ECKD_CCW_WRITE_CKD		 0x1d
+#define DASD_ECKD_CCW_READ_CKD		 0x1e
+#define DASD_ECKD_CCW_PSF		 0x27
+#define DASD_ECKD_CCW_RSSD		 0x3e
+#define DASD_ECKD_CCW_LOCATE_RECORD	 0x47
+#define DASD_ECKD_CCW_DEFINE_EXTENT	 0x63
+#define DASD_ECKD_CCW_WRITE_MT		 0x85
+#define DASD_ECKD_CCW_READ_MT		 0x86
+#define DASD_ECKD_CCW_WRITE_KD_MT	 0x8d
+#define DASD_ECKD_CCW_READ_KD_MT	 0x8e
+#define DASD_ECKD_CCW_RELEASE		 0x94
+#define DASD_ECKD_CCW_READ_CKD_MT	 0x9e
+#define DASD_ECKD_CCW_WRITE_CKD_MT	 0x9d
+#define DASD_ECKD_CCW_RESERVE		 0xB4
+
+/*
+ *Perform Subsystem Function / Sub-Orders
+ */
+#define PSF_ORDER_PRSSD			 0x18
+
+/*****************************************************************************
+ * SECTION: Type Definitions
+ ****************************************************************************/
+
+struct eckd_count {
+	__u16 cyl;
+	__u16 head;
+	__u8 record;
+	__u8 kl;
+	__u16 dl;
+} __attribute__ ((packed));
+
+struct ch_t {
+	__u16 cyl;
+	__u16 head;
+} __attribute__ ((packed));
+
+struct chs_t {
+	__u16 cyl;
+	__u16 head;
+	__u32 sector;
+} __attribute__ ((packed));
+
+struct chr_t {
+	__u16 cyl;
+	__u16 head;
+	__u8 record;
+} __attribute__ ((packed));
+
+struct geom_t {
+	__u16 cyl;
+	__u16 head;
+	__u32 sector;
+} __attribute__ ((packed));
+
+struct eckd_home {
+	__u8 skip_control[14];
+	__u16 cell_number;
+	__u8 physical_addr[3];
+	__u8 flag;
+	struct ch_t track_addr;
+	__u8 reserved;
+	__u8 key_length;
+	__u8 reserved2[2];
+} __attribute__ ((packed));
+
+struct DE_eckd_data {
+	struct {
+		unsigned char perm:2;	/* Permissions on this extent */
+		unsigned char reserved:1;
+		unsigned char seek:2;	/* Seek control */
+		unsigned char auth:2;	/* Access authorization */
+		unsigned char pci:1;	/* PCI Fetch mode */
+	} __attribute__ ((packed)) mask;
+	struct {
+		unsigned char mode:2;	/* Architecture mode */
+		unsigned char ckd:1;	/* CKD Conversion */
+		unsigned char operation:3;	/* Operation mode */
+		unsigned char cfw:1;	/* Cache fast write */
+		unsigned char dfw:1;	/* DASD fast write */
+	} __attribute__ ((packed)) attributes;
+	__u16 blk_size;		/* Blocksize */
+	__u16 fast_write_id;
+	__u8 ga_additional;	/* Global Attributes Additional */
+	__u8 ga_extended;	/* Global Attributes Extended	*/
+	struct ch_t beg_ext;
+	struct ch_t end_ext;
+	unsigned long long ep_sys_time; /* Ext Parameter - System Time Stamp */
+	__u8 ep_format;        /* Extended Parameter format byte       */
+	__u8 ep_prio;          /* Extended Parameter priority I/O byte */
+	__u8 ep_reserved[6];   /* Extended Parameter Reserved          */
+} __attribute__ ((packed));
+
+struct LO_eckd_data {
+	struct {
+		unsigned char orientation:2;
+		unsigned char operation:6;
+	} __attribute__ ((packed)) operation;
+	struct {
+		unsigned char last_bytes_used:1;
+		unsigned char reserved:6;
+		unsigned char read_count_suffix:1;
+	} __attribute__ ((packed)) auxiliary;
+	__u8 unused;
+	__u8 count;
+	struct ch_t seek_addr;
+	struct chr_t search_arg;
+	__u8 sector;
+	__u16 length;
+} __attribute__ ((packed));
+
+struct dasd_eckd_characteristics {
+	__u16 cu_type;
+	struct {
+		unsigned char support:2;
+		unsigned char async:1;
+		unsigned char reserved:1;
+		unsigned char cache_info:1;
+		unsigned char model:3;
+	} __attribute__ ((packed)) cu_model;
+	__u16 dev_type;
+	__u8 dev_model;
+	struct {
+		unsigned char mult_burst:1;
+		unsigned char RT_in_LR:1;
+		unsigned char reserved1:1;
+		unsigned char RD_IN_LR:1;
+		unsigned char reserved2:4;
+		unsigned char reserved3:8;
+		unsigned char defect_wr:1;
+		unsigned char XRC_supported:1; 
+		unsigned char reserved4:1;
+		unsigned char striping:1;
+		unsigned char reserved5:4;
+		unsigned char cfw:1;
+		unsigned char reserved6:2;
+		unsigned char cache:1;
+		unsigned char dual_copy:1;
+		unsigned char dfw:1;
+		unsigned char reset_alleg:1;
+		unsigned char sense_down:1;
+	} __attribute__ ((packed)) facilities;
+	__u8 dev_class;
+	__u8 unit_type;
+	__u16 no_cyl;
+	__u16 trk_per_cyl;
+	__u8 sec_per_trk;
+	__u8 byte_per_track[3];
+	__u16 home_bytes;
+	__u8 formula;
+	union {
+		struct {
+			__u8 f1;
+			__u16 f2;
+			__u16 f3;
+		} __attribute__ ((packed)) f_0x01;
+		struct {
+			__u8 f1;
+			__u8 f2;
+			__u8 f3;
+			__u8 f4;
+			__u8 f5;
+		} __attribute__ ((packed)) f_0x02;
+	} __attribute__ ((packed)) factors;
+	__u16 first_alt_trk;
+	__u16 no_alt_trk;
+	__u16 first_dia_trk;
+	__u16 no_dia_trk;
+	__u16 first_sup_trk;
+	__u16 no_sup_trk;
+	__u8 MDR_ID;
+	__u8 OBR_ID;
+	__u8 director;
+	__u8 rd_trk_set;
+	__u16 max_rec_zero;
+	__u8 reserved1;
+	__u8 RWANY_in_LR;
+	__u8 factor6;
+	__u8 factor7;
+	__u8 factor8;
+	__u8 reserved2[3];
+	__u8 reserved3[10];
+} __attribute__ ((packed));
+
+struct dasd_eckd_confdata {
+	struct {
+		struct {
+			unsigned char identifier:2;
+			unsigned char token_id:1;
+			unsigned char sno_valid:1;
+			unsigned char subst_sno:1;
+			unsigned char recNED:1;
+			unsigned char emuNED:1;
+			unsigned char reserved:1;
+		} __attribute__ ((packed)) flags;
+		__u8 descriptor;
+		__u8 dev_class;
+		__u8 reserved;
+		unsigned char dev_type[6];
+		unsigned char dev_model[3];
+		unsigned char HDA_manufacturer[3];
+		unsigned char HDA_location[2];
+		unsigned char HDA_seqno[12];
+		__u16 ID;
+	} __attribute__ ((packed)) ned1;
+	struct {
+		struct {
+			unsigned char identifier:2;
+			unsigned char token_id:1;
+			unsigned char sno_valid:1;
+			unsigned char subst_sno:1;
+			unsigned char recNED:1;
+			unsigned char emuNED:1;
+			unsigned char reserved:1;
+		} __attribute__ ((packed)) flags;
+		__u8 descriptor;
+		__u8 reserved[2];
+		unsigned char dev_type[6];
+		unsigned char dev_model[3];
+		unsigned char DASD_manufacturer[3];
+		unsigned char DASD_location[2];
+		unsigned char DASD_seqno[12];
+		__u16 ID;
+	} __attribute__ ((packed)) ned2;
+	struct {
+		struct {
+			unsigned char identifier:2;
+			unsigned char token_id:1;
+			unsigned char sno_valid:1;
+			unsigned char subst_sno:1;
+			unsigned char recNED:1;
+			unsigned char emuNED:1;
+			unsigned char reserved:1;
+		} __attribute__ ((packed)) flags;
+		__u8 descriptor;
+		__u8 reserved[2];
+		unsigned char cont_type[6];
+		unsigned char cont_model[3];
+		unsigned char cont_manufacturer[3];
+		unsigned char cont_location[2];
+		unsigned char cont_seqno[12];
+		__u16 ID;
+	} __attribute__ ((packed)) ned3;
+	struct {
+		struct {
+			unsigned char identifier:2;
+			unsigned char token_id:1;
+			unsigned char sno_valid:1;
+			unsigned char subst_sno:1;
+			unsigned char recNED:1;
+			unsigned char emuNED:1;
+			unsigned char reserved:1;
+		} __attribute__ ((packed)) flags;
+		__u8 descriptor;
+		__u8 reserved[2];
+		unsigned char cont_type[6];
+		unsigned char empty[3];
+		unsigned char cont_manufacturer[3];
+		unsigned char cont_location[2];
+		unsigned char cont_seqno[12];
+		__u16 ID;
+	} __attribute__ ((packed)) ned4;
+	unsigned char ned5[32];
+	unsigned char ned6[32];
+	unsigned char ned7[32];
+	struct {
+		struct {
+			unsigned char identifier:2;
+			unsigned char reserved:6;
+		} __attribute__ ((packed)) flags;
+		__u8 selector;
+		__u16 interfaceID;
+		__u32 reserved;
+		__u16 subsystemID;
+		struct {
+			unsigned char sp0:1;
+			unsigned char sp1:1;
+			unsigned char reserved:5;
+			unsigned char scluster:1;
+		} __attribute__ ((packed)) spathID;
+		__u8 unit_address;
+		__u8 dev_ID;
+		__u8 dev_address;
+		__u8 adapterID;
+		__u16 link_address;
+		struct {
+			unsigned char parallel:1;
+			unsigned char escon:1;
+			unsigned char reserved:1;
+			unsigned char ficon:1;
+			unsigned char reserved2:4;
+		} __attribute__ ((packed)) protocol_type;
+		struct {
+			unsigned char PID_in_236:1;
+			unsigned char reserved:7;
+		} __attribute__ ((packed)) format_flags;
+		__u8 log_dev_address;
+		unsigned char reserved2[12];
+	} __attribute__ ((packed)) neq;
+} __attribute__ ((packed));
+
+struct dasd_eckd_path {
+	__u8 opm;
+	__u8 ppm;
+	__u8 npm;
+};
+
+/*
+ * Perform Subsystem Function - Prepare for Read Subsystem Data	 
+ */
+struct dasd_psf_prssd_data {
+	unsigned char order;
+	unsigned char flags;
+	unsigned char reserved[4];
+	unsigned char suborder;
+	unsigned char varies[9];
+} __attribute__ ((packed));
+
+#endif				/* DASD_ECKD_H */

drivers/s390/block/dasd_erp.c

@@ -0,0 +1,254 @@
+/*
+ * File...........: linux/drivers/s390/block/dasd.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *		    Horst Hummel <Horst.Hummel@de.ibm.com>
+ *		    Carsten Otte <Cotte@de.ibm.com>
+ *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2001
+ *
+ * $Revision: 1.14 $
+ */
+
+#include <linux/config.h>
+#include <linux/ctype.h>
+#include <linux/init.h>
+
+#include <asm/debug.h>
+#include <asm/ebcdic.h>
+#include <asm/uaccess.h>
+
+/* This is ugly... */
+#define PRINTK_HEADER "dasd_erp:"
+
+#include "dasd_int.h"
+
+struct dasd_ccw_req *
+dasd_alloc_erp_request(char *magic, int cplength, int datasize,
+		       struct dasd_device * device)
+{
+	unsigned long flags;
+	struct dasd_ccw_req *cqr;
+	char *data;
+	int size;
+
+	/* Sanity checks */
+	if ( magic == NULL || datasize > PAGE_SIZE ||
+	     (cplength*sizeof(struct ccw1)) > PAGE_SIZE)
+		BUG();
+
+	size = (sizeof(struct dasd_ccw_req) + 7L) & -8L;
+	if (cplength > 0)
+		size += cplength * sizeof(struct ccw1);
+	if (datasize > 0)
+		size += datasize;
+	spin_lock_irqsave(&device->mem_lock, flags);
+	cqr = (struct dasd_ccw_req *)
+		dasd_alloc_chunk(&device->erp_chunks, size);
+	spin_unlock_irqrestore(&device->mem_lock, flags);
+	if (cqr == NULL)
+		return ERR_PTR(-ENOMEM);
+	memset(cqr, 0, sizeof(struct dasd_ccw_req));
+	data = (char *) cqr + ((sizeof(struct dasd_ccw_req) + 7L) & -8L);
+	cqr->cpaddr = NULL;
+	if (cplength > 0) {
+		cqr->cpaddr = (struct ccw1 *) data;
+		data += cplength*sizeof(struct ccw1);
+		memset(cqr->cpaddr, 0, cplength*sizeof(struct ccw1));
+	}
+	cqr->data = NULL;
+	if (datasize > 0) {
+		cqr->data = data;
+ 		memset(cqr->data, 0, datasize);
+	}
+	strncpy((char *) &cqr->magic, magic, 4);
+	ASCEBC((char *) &cqr->magic, 4);
+	set_bit(DASD_CQR_FLAGS_USE_ERP, &cqr->flags);
+	dasd_get_device(device);
+	return cqr;
+}
+
+void
+dasd_free_erp_request(struct dasd_ccw_req * cqr, struct dasd_device * device)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&device->mem_lock, flags);
+	dasd_free_chunk(&device->erp_chunks, cqr);
+	spin_unlock_irqrestore(&device->mem_lock, flags);
+	atomic_dec(&device->ref_count);
+}
+
+
+/*
+ * dasd_default_erp_action just retries the current cqr
+ */
+struct dasd_ccw_req *
+dasd_default_erp_action(struct dasd_ccw_req * cqr)
+{
+	struct dasd_device *device;
+
+	device = cqr->device;
+
+        /* just retry - there is nothing to save ... I got no sense data.... */
+        if (cqr->retries > 0) {
+                DEV_MESSAGE (KERN_DEBUG, device, 
+                             "default ERP called (%i retries left)",
+                             cqr->retries);
+		cqr->lpm    = LPM_ANYPATH;
+		cqr->status = DASD_CQR_QUEUED;
+        } else {
+                DEV_MESSAGE (KERN_WARNING, device, "%s",
+			     "default ERP called (NO retry left)");
+		cqr->status = DASD_CQR_FAILED;
+		cqr->stopclk = get_clock ();
+        }
+        return cqr;
+}				/* end dasd_default_erp_action */
+
+/*
+ * DESCRIPTION
+ *   Frees all ERPs of the current ERP Chain and set the status
+ *   of the original CQR either to DASD_CQR_DONE if ERP was successful
+ *   or to DASD_CQR_FAILED if ERP was NOT successful.
+ *   NOTE: This function is only called if no discipline postaction
+ *	   is available
+ *
+ * PARAMETER
+ *   erp		current erp_head
+ *
+ * RETURN VALUES
+ *   cqr		pointer to the original CQR
+ */
+struct dasd_ccw_req *
+dasd_default_erp_postaction(struct dasd_ccw_req * cqr)
+{
+	struct dasd_device *device;
+	int success;
+
+	if (cqr->refers == NULL || cqr->function == NULL)
+		BUG();
+
+	device = cqr->device;
+	success = cqr->status == DASD_CQR_DONE;
+
+	/* free all ERPs - but NOT the original cqr */
+	while (cqr->refers != NULL) {
+		struct dasd_ccw_req *refers;
+
+		refers = cqr->refers;
+		/* remove the request from the device queue */
+		list_del(&cqr->list);
+		/* free the finished erp request */
+		dasd_free_erp_request(cqr, device);
+		cqr = refers;
+	}
+
+	/* set corresponding status to original cqr */
+	if (success)
+		cqr->status = DASD_CQR_DONE;
+	else {
+		cqr->status = DASD_CQR_FAILED;
+		cqr->stopclk = get_clock();
+	}
+
+	return cqr;
+
+}				/* end default_erp_postaction */
+
+/*
+ * Print the hex dump of the memory used by a request. This includes
+ * all error recovery ccws that have been chained in from of the 
+ * real request.
+ */
+static inline void
+hex_dump_memory(struct dasd_device *device, void *data, int len)
+{
+	int *pint;
+
+	pint = (int *) data;
+	while (len > 0) {
+		DEV_MESSAGE(KERN_ERR, device, "%p: %08x %08x %08x %08x",
+			    pint, pint[0], pint[1], pint[2], pint[3]);
+		pint += 4;
+		len -= 16;
+	}
+}
+
+void
+dasd_log_sense(struct dasd_ccw_req *cqr, struct irb *irb)
+{
+	struct dasd_device *device;
+
+	device = cqr->device;
+	/* dump sense data */
+	if (device->discipline && device->discipline->dump_sense)
+		device->discipline->dump_sense(device, cqr, irb);
+}
+
+void
+dasd_log_ccw(struct dasd_ccw_req * cqr, int caller, __u32 cpa)
+{
+	struct dasd_device *device;
+	struct dasd_ccw_req *lcqr;
+	struct ccw1 *ccw;
+	int cplength;
+
+	device = cqr->device;
+	/* log the channel program */
+	for (lcqr = cqr; lcqr != NULL; lcqr = lcqr->refers) {
+		DEV_MESSAGE(KERN_ERR, device,
+			    "(%s) ERP chain report for req: %p",
+			    caller == 0 ? "EXAMINE" : "ACTION", lcqr);
+		hex_dump_memory(device, lcqr, sizeof(struct dasd_ccw_req));
+
+		cplength = 1;
+		ccw = lcqr->cpaddr;
+		while (ccw++->flags & (CCW_FLAG_DC | CCW_FLAG_CC))
+			cplength++;
+
+		if (cplength > 40) {	/* log only parts of the CP */
+			DEV_MESSAGE(KERN_ERR, device, "%s",
+				    "Start of channel program:");
+			hex_dump_memory(device, lcqr->cpaddr,
+					40*sizeof(struct ccw1));
+
+			DEV_MESSAGE(KERN_ERR, device, "%s",
+				    "End of channel program:");
+			hex_dump_memory(device, lcqr->cpaddr + cplength - 10,
+					10*sizeof(struct ccw1));
+		} else {	/* log the whole CP */
+			DEV_MESSAGE(KERN_ERR, device, "%s",
+				    "Channel program (complete):");
+			hex_dump_memory(device, lcqr->cpaddr,
+					cplength*sizeof(struct ccw1));
+		}
+
+		if (lcqr != cqr)
+			continue;
+
+		/*
+		 * Log bytes arround failed CCW but only if we did
+		 * not log the whole CP of the CCW is outside the
+		 * logged CP. 
+		 */
+		if (cplength > 40 ||
+		    ((addr_t) cpa < (addr_t) lcqr->cpaddr &&
+		     (addr_t) cpa > (addr_t) (lcqr->cpaddr + cplength + 4))) {
+			
+			DEV_MESSAGE(KERN_ERR, device,
+				    "Failed CCW (%p) (area):",
+				    (void *) (long) cpa);
+			hex_dump_memory(device, cqr->cpaddr - 10,
+					20*sizeof(struct ccw1));
+		}
+	}
+
+}				/* end log_erp_chain */
+
+EXPORT_SYMBOL(dasd_default_erp_action);
+EXPORT_SYMBOL(dasd_default_erp_postaction);
+EXPORT_SYMBOL(dasd_alloc_erp_request);
+EXPORT_SYMBOL(dasd_free_erp_request);
+EXPORT_SYMBOL(dasd_log_sense);
+EXPORT_SYMBOL(dasd_log_ccw);

drivers/s390/block/dasd_fba.c

@@ -0,0 +1,607 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_fba.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ *
+ * $Revision: 1.39 $
+ */
+
+#include <linux/config.h>
+#include <linux/stddef.h>
+#include <linux/kernel.h>
+#include <asm/debug.h>
+
+#include <linux/slab.h>
+#include <linux/hdreg.h>	/* HDIO_GETGEO			    */
+#include <linux/bio.h>
+#include <linux/module.h>
+#include <linux/init.h>
+
+#include <asm/idals.h>
+#include <asm/ebcdic.h>
+#include <asm/io.h>
+#include <asm/todclk.h>
+#include <asm/ccwdev.h>
+
+#include "dasd_int.h"
+#include "dasd_fba.h"
+
+#ifdef PRINTK_HEADER
+#undef PRINTK_HEADER
+#endif				/* PRINTK_HEADER */
+#define PRINTK_HEADER "dasd(fba):"
+
+#define DASD_FBA_CCW_WRITE 0x41
+#define DASD_FBA_CCW_READ 0x42
+#define DASD_FBA_CCW_LOCATE 0x43
+#define DASD_FBA_CCW_DEFINE_EXTENT 0x63
+
+MODULE_LICENSE("GPL");
+
+static struct dasd_discipline dasd_fba_discipline;
+
+struct dasd_fba_private {
+	struct dasd_fba_characteristics rdc_data;
+};
+
+static struct ccw_device_id dasd_fba_ids[] = {
+	{ CCW_DEVICE_DEVTYPE (0x6310, 0, 0x9336, 0), driver_info: 0x1},
+	{ CCW_DEVICE_DEVTYPE (0x3880, 0, 0x3370, 0), driver_info: 0x2},
+	{ /* end of list */ },
+};
+
+MODULE_DEVICE_TABLE(ccw, dasd_fba_ids);
+
+static struct ccw_driver dasd_fba_driver; /* see below */
+static int
+dasd_fba_probe(struct ccw_device *cdev)
+{
+	int ret;
+
+	ret = dasd_generic_probe (cdev, &dasd_fba_discipline);
+	if (ret)
+		return ret;
+	ccw_device_set_options(cdev, CCWDEV_DO_PATHGROUP);
+	return 0;
+}
+
+static int
+dasd_fba_set_online(struct ccw_device *cdev)
+{
+	return dasd_generic_set_online (cdev, &dasd_fba_discipline);
+}
+
+static struct ccw_driver dasd_fba_driver = {
+	.name        = "dasd-fba",
+	.owner       = THIS_MODULE,
+	.ids         = dasd_fba_ids,
+	.probe       = dasd_fba_probe,
+	.remove      = dasd_generic_remove,
+	.set_offline = dasd_generic_set_offline,
+	.set_online  = dasd_fba_set_online,
+	.notify      = dasd_generic_notify,
+};
+
+static inline void
+define_extent(struct ccw1 * ccw, struct DE_fba_data *data, int rw,
+	      int blksize, int beg, int nr)
+{
+	ccw->cmd_code = DASD_FBA_CCW_DEFINE_EXTENT;
+	ccw->flags = 0;
+	ccw->count = 16;
+	ccw->cda = (__u32) __pa(data);
+	memset(data, 0, sizeof (struct DE_fba_data));
+	if (rw == WRITE)
+		(data->mask).perm = 0x0;
+	else if (rw == READ)
+		(data->mask).perm = 0x1;
+	else
+		data->mask.perm = 0x2;
+	data->blk_size = blksize;
+	data->ext_loc = beg;
+	data->ext_end = nr - 1;
+}
+
+static inline void
+locate_record(struct ccw1 * ccw, struct LO_fba_data *data, int rw,
+	      int block_nr, int block_ct)
+{
+	ccw->cmd_code = DASD_FBA_CCW_LOCATE;
+	ccw->flags = 0;
+	ccw->count = 8;
+	ccw->cda = (__u32) __pa(data);
+	memset(data, 0, sizeof (struct LO_fba_data));
+	if (rw == WRITE)
+		data->operation.cmd = 0x5;
+	else if (rw == READ)
+		data->operation.cmd = 0x6;
+	else
+		data->operation.cmd = 0x8;
+	data->blk_nr = block_nr;
+	data->blk_ct = block_ct;
+}
+
+static int
+dasd_fba_check_characteristics(struct dasd_device *device)
+{
+	struct dasd_fba_private *private;
+	struct ccw_device *cdev = device->cdev;	
+	void *rdc_data;
+	int rc;
+
+	private = (struct dasd_fba_private *) device->private;
+	if (private == NULL) {
+		private = kmalloc(sizeof(struct dasd_fba_private), GFP_KERNEL);
+		if (private == NULL) {
+			DEV_MESSAGE(KERN_WARNING, device, "%s",
+				    "memory allocation failed for private "
+				    "data");
+			return -ENOMEM;
+		}
+		device->private = (void *) private;
+	}
+	/* Read Device Characteristics */
+	rdc_data = (void *) &(private->rdc_data);
+	rc = read_dev_chars(device->cdev, &rdc_data, 32);
+	if (rc) {
+		DEV_MESSAGE(KERN_WARNING, device,
+			    "Read device characteristics returned error %d",
+			    rc);
+		return rc;
+	}
+
+	DEV_MESSAGE(KERN_INFO, device,
+		    "%04X/%02X(CU:%04X/%02X) %dMB at(%d B/blk)",
+		    cdev->id.dev_type,
+		    cdev->id.dev_model,
+		    cdev->id.cu_type,
+		    cdev->id.cu_model,
+		    ((private->rdc_data.blk_bdsa *
+		      (private->rdc_data.blk_size >> 9)) >> 11),
+		    private->rdc_data.blk_size);
+	return 0;
+}
+
+static int
+dasd_fba_do_analysis(struct dasd_device *device)
+{
+	struct dasd_fba_private *private;
+	int sb, rc;
+
+	private = (struct dasd_fba_private *) device->private;
+	rc = dasd_check_blocksize(private->rdc_data.blk_size);
+	if (rc) {
+		DEV_MESSAGE(KERN_INFO, device, "unknown blocksize %d",
+			    private->rdc_data.blk_size);
+		return rc;
+	}
+	device->blocks = private->rdc_data.blk_bdsa;
+	device->bp_block = private->rdc_data.blk_size;
+	device->s2b_shift = 0;	/* bits to shift 512 to get a block */
+	for (sb = 512; sb < private->rdc_data.blk_size; sb = sb << 1)
+		device->s2b_shift++;
+	return 0;
+}
+
+static int
+dasd_fba_fill_geometry(struct dasd_device *device, struct hd_geometry *geo)
+{
+	if (dasd_check_blocksize(device->bp_block) != 0)
+		return -EINVAL;
+	geo->cylinders = (device->blocks << device->s2b_shift) >> 10;
+	geo->heads = 16;
+	geo->sectors = 128 >> device->s2b_shift;
+	return 0;
+}
+
+static dasd_era_t
+dasd_fba_examine_error(struct dasd_ccw_req * cqr, struct irb * irb)
+{
+	struct dasd_device *device;
+	struct ccw_device *cdev;
+
+	device = (struct dasd_device *) cqr->device;
+	if (irb->scsw.cstat == 0x00 &&
+	    irb->scsw.dstat == (DEV_STAT_CHN_END | DEV_STAT_DEV_END))
+		return dasd_era_none;
+	
+	cdev = device->cdev;
+	switch (cdev->id.dev_type) {
+	case 0x3370:
+		return dasd_3370_erp_examine(cqr, irb);
+	case 0x9336:
+		return dasd_9336_erp_examine(cqr, irb);
+	default:
+		return dasd_era_recover;
+	}
+}
+
+static dasd_erp_fn_t
+dasd_fba_erp_action(struct dasd_ccw_req * cqr)
+{
+	return dasd_default_erp_action;
+}
+
+static dasd_erp_fn_t
+dasd_fba_erp_postaction(struct dasd_ccw_req * cqr)
+{
+	if (cqr->function == dasd_default_erp_action)
+		return dasd_default_erp_postaction;
+
+	DEV_MESSAGE(KERN_WARNING, cqr->device, "unknown ERP action %p",
+		    cqr->function);
+	return NULL;
+}
+
+static struct dasd_ccw_req *
+dasd_fba_build_cp(struct dasd_device * device, struct request *req)
+{
+	struct dasd_fba_private *private;
+	unsigned long *idaws;
+	struct LO_fba_data *LO_data;
+	struct dasd_ccw_req *cqr;
+	struct ccw1 *ccw;
+	struct bio *bio;
+	struct bio_vec *bv;
+	char *dst;
+	int count, cidaw, cplength, datasize;
+	sector_t recid, first_rec, last_rec;
+	unsigned int blksize, off;
+	unsigned char cmd;
+	int i;
+
+	private = (struct dasd_fba_private *) device->private;
+	if (rq_data_dir(req) == READ) {
+		cmd = DASD_FBA_CCW_READ;
+	} else if (rq_data_dir(req) == WRITE) {
+		cmd = DASD_FBA_CCW_WRITE;
+	} else
+		return ERR_PTR(-EINVAL);
+	blksize = device->bp_block;
+	/* Calculate record id of first and last block. */
+	first_rec = req->sector >> device->s2b_shift;
+	last_rec = (req->sector + req->nr_sectors - 1) >> device->s2b_shift;
+	/* Check struct bio and count the number of blocks for the request. */
+	count = 0;
+	cidaw = 0;
+	rq_for_each_bio(bio, req) {
+		bio_for_each_segment(bv, bio, i) {
+			if (bv->bv_len & (blksize - 1))
+				/* Fba can only do full blocks. */
+				return ERR_PTR(-EINVAL);
+			count += bv->bv_len >> (device->s2b_shift + 9);
+#if defined(CONFIG_ARCH_S390X)
+			if (idal_is_needed (page_address(bv->bv_page),
+					    bv->bv_len))
+				cidaw += bv->bv_len / blksize;
+#endif
+		}
+	}
+	/* Paranoia. */
+	if (count != last_rec - first_rec + 1)
+		return ERR_PTR(-EINVAL);
+	/* 1x define extent + 1x locate record + number of blocks */
+	cplength = 2 + count;
+	/* 1x define extent + 1x locate record */
+	datasize = sizeof(struct DE_fba_data) + sizeof(struct LO_fba_data) +
+		cidaw * sizeof(unsigned long);
+	/*
+	 * Find out number of additional locate record ccws if the device
+	 * can't do data chaining.
+	 */
+	if (private->rdc_data.mode.bits.data_chain == 0) {
+		cplength += count - 1;
+		datasize += (count - 1)*sizeof(struct LO_fba_data);
+	}
+	/* Allocate the ccw request. */
+	cqr = dasd_smalloc_request(dasd_fba_discipline.name,
+				   cplength, datasize, device);
+	if (IS_ERR(cqr))
+		return cqr;
+	ccw = cqr->cpaddr;
+	/* First ccw is define extent. */
+	define_extent(ccw++, cqr->data, rq_data_dir(req),
+		      device->bp_block, req->sector, req->nr_sectors);
+	/* Build locate_record + read/write ccws. */
+	idaws = (unsigned long *) (cqr->data + sizeof(struct DE_fba_data));
+	LO_data = (struct LO_fba_data *) (idaws + cidaw);
+	/* Locate record for all blocks for smart devices. */
+	if (private->rdc_data.mode.bits.data_chain != 0) {
+		ccw[-1].flags |= CCW_FLAG_CC;
+		locate_record(ccw++, LO_data++, rq_data_dir(req), 0, count);
+	}
+	recid = first_rec;
+	rq_for_each_bio(bio, req) bio_for_each_segment(bv, bio, i) {
+		dst = page_address(bv->bv_page) + bv->bv_offset;
+		if (dasd_page_cache) {
+			char *copy = kmem_cache_alloc(dasd_page_cache,
+						      SLAB_DMA | __GFP_NOWARN);
+			if (copy && rq_data_dir(req) == WRITE)
+				memcpy(copy + bv->bv_offset, dst, bv->bv_len);
+			if (copy)
+				dst = copy + bv->bv_offset;
+		}
+		for (off = 0; off < bv->bv_len; off += blksize) {
+			/* Locate record for stupid devices. */
+			if (private->rdc_data.mode.bits.data_chain == 0) {
+				ccw[-1].flags |= CCW_FLAG_CC;
+				locate_record(ccw, LO_data++,
+					      rq_data_dir(req),
+					      recid - first_rec, 1);
+				ccw->flags = CCW_FLAG_CC;
+				ccw++;
+			} else {
+				if (recid > first_rec)
+					ccw[-1].flags |= CCW_FLAG_DC;
+				else
+					ccw[-1].flags |= CCW_FLAG_CC;
+			}
+			ccw->cmd_code = cmd;
+			ccw->count = device->bp_block;
+			if (idal_is_needed(dst, blksize)) {
+				ccw->cda = (__u32)(addr_t) idaws;
+				ccw->flags = CCW_FLAG_IDA;
+				idaws = idal_create_words(idaws, dst, blksize);
+			} else {
+				ccw->cda = (__u32)(addr_t) dst;
+				ccw->flags = 0;
+			}
+			ccw++;
+			dst += blksize;
+			recid++;
+		}
+	}
+	cqr->device = device;
+	cqr->expires = 5 * 60 * HZ;	/* 5 minutes */
+	cqr->status = DASD_CQR_FILLED;
+	return cqr;
+}
+
+static int
+dasd_fba_free_cp(struct dasd_ccw_req *cqr, struct request *req)
+{
+	struct dasd_fba_private *private;
+	struct ccw1 *ccw;
+	struct bio *bio;
+	struct bio_vec *bv;
+	char *dst, *cda;
+	unsigned int blksize, off;
+	int i, status;
+
+	if (!dasd_page_cache)
+		goto out;
+	private = (struct dasd_fba_private *) cqr->device->private;
+	blksize = cqr->device->bp_block;
+	ccw = cqr->cpaddr;
+	/* Skip over define extent & locate record. */
+	ccw++;
+	if (private->rdc_data.mode.bits.data_chain != 0)
+		ccw++;
+	rq_for_each_bio(bio, req) bio_for_each_segment(bv, bio, i) {
+		dst = page_address(bv->bv_page) + bv->bv_offset;
+		for (off = 0; off < bv->bv_len; off += blksize) {
+			/* Skip locate record. */
+			if (private->rdc_data.mode.bits.data_chain == 0)
+				ccw++;
+			if (dst) {
+				if (ccw->flags & CCW_FLAG_IDA)
+					cda = *((char **)((addr_t) ccw->cda));
+				else
+					cda = (char *)((addr_t) ccw->cda);
+				if (dst != cda) {
+					if (rq_data_dir(req) == READ)
+						memcpy(dst, cda, bv->bv_len);
+					kmem_cache_free(dasd_page_cache,
+					    (void *)((addr_t)cda & PAGE_MASK));
+				}
+				dst = NULL;
+			}
+			ccw++;
+		}
+	}
+out:
+	status = cqr->status == DASD_CQR_DONE;
+	dasd_sfree_request(cqr, cqr->device);
+	return status;
+}
+
+static int
+dasd_fba_fill_info(struct dasd_device * device,
+		   struct dasd_information2_t * info)
+{
+	info->label_block = 1;
+	info->FBA_layout = 1;
+	info->format = DASD_FORMAT_LDL;
+	info->characteristics_size = sizeof(struct dasd_fba_characteristics);
+	memcpy(info->characteristics,
+	       &((struct dasd_fba_private *) device->private)->rdc_data,
+	       sizeof (struct dasd_fba_characteristics));
+	info->confdata_size = 0;
+	return 0;
+}
+
+static void
+dasd_fba_dump_sense(struct dasd_device *device, struct dasd_ccw_req * req,
+		    struct irb *irb)
+{
+	char *page;
+	struct ccw1 *act, *end, *last;
+	int len, sl, sct, count;
+
+	page = (char *) get_zeroed_page(GFP_ATOMIC);
+	if (page == NULL) {
+		DEV_MESSAGE(KERN_ERR, device, " %s",
+			    "No memory to dump sense data");
+		return;
+	}
+	len = sprintf(page, KERN_ERR PRINTK_HEADER
+		      " I/O status report for device %s:\n",
+		      device->cdev->dev.bus_id);
+	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+		       " in req: %p CS: 0x%02X DS: 0x%02X\n", req,
+		       irb->scsw.cstat, irb->scsw.dstat);
+	len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+		       " device %s: Failing CCW: %p\n",
+		       device->cdev->dev.bus_id,
+		       (void *) (addr_t) irb->scsw.cpa);
+	if (irb->esw.esw0.erw.cons) {
+		for (sl = 0; sl < 4; sl++) {
+			len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+				       " Sense(hex) %2d-%2d:",
+				       (8 * sl), ((8 * sl) + 7));
+
+			for (sct = 0; sct < 8; sct++) {
+				len += sprintf(page + len, " %02x",
+					       irb->ecw[8 * sl + sct]);
+			}
+			len += sprintf(page + len, "\n");
+		}
+	} else {
+	        len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+			       " SORRY - NO VALID SENSE AVAILABLE\n");
+	}
+	MESSAGE_LOG(KERN_ERR, "%s",
+		    page + sizeof(KERN_ERR PRINTK_HEADER));
+
+	/* dump the Channel Program */
+	/* print first CCWs (maximum 8) */
+	act = req->cpaddr;
+        for (last = act; last->flags & (CCW_FLAG_CC | CCW_FLAG_DC); last++);
+	end = min(act + 8, last);
+	len = sprintf(page, KERN_ERR PRINTK_HEADER
+		      " Related CP in req: %p\n", req);
+	while (act <= end) {
+		len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+			       " CCW %p: %08X %08X DAT:",
+			       act, ((int *) act)[0], ((int *) act)[1]);
+		for (count = 0; count < 32 && count < act->count;
+		     count += sizeof(int))
+			len += sprintf(page + len, " %08X",
+				       ((int *) (addr_t) act->cda)
+				       [(count>>2)]);
+		len += sprintf(page + len, "\n");
+		act++;
+	}
+	MESSAGE_LOG(KERN_ERR, "%s",
+		    page + sizeof(KERN_ERR PRINTK_HEADER));
+
+
+	/* print failing CCW area */
+	len = 0;
+	if (act <  ((struct ccw1 *)(addr_t) irb->scsw.cpa) - 2) {
+		act = ((struct ccw1 *)(addr_t) irb->scsw.cpa) - 2;
+		len += sprintf(page + len, KERN_ERR PRINTK_HEADER "......\n");
+	}
+	end = min((struct ccw1 *)(addr_t) irb->scsw.cpa + 2, last);
+	while (act <= end) {
+		len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+			       " CCW %p: %08X %08X DAT:",
+			       act, ((int *) act)[0], ((int *) act)[1]);
+		for (count = 0; count < 32 && count < act->count;
+		     count += sizeof(int))
+			len += sprintf(page + len, " %08X",
+				       ((int *) (addr_t) act->cda)
+				       [(count>>2)]);
+		len += sprintf(page + len, "\n");
+		act++;
+	}
+
+	/* print last CCWs */
+	if (act <  last - 2) {
+		act = last - 2;
+		len += sprintf(page + len, KERN_ERR PRINTK_HEADER "......\n");
+	}
+	while (act <= last) {
+		len += sprintf(page + len, KERN_ERR PRINTK_HEADER
+			       " CCW %p: %08X %08X DAT:",
+			       act, ((int *) act)[0], ((int *) act)[1]);
+		for (count = 0; count < 32 && count < act->count;
+		     count += sizeof(int))
+			len += sprintf(page + len, " %08X",
+				       ((int *) (addr_t) act->cda)
+				       [(count>>2)]);
+		len += sprintf(page + len, "\n");
+		act++;
+	}
+	if (len > 0)
+		MESSAGE_LOG(KERN_ERR, "%s",
+			    page + sizeof(KERN_ERR PRINTK_HEADER));
+	free_page((unsigned long) page);
+}
+
+/*
+ * max_blocks is dependent on the amount of storage that is available
+ * in the static io buffer for each device. Currently each device has
+ * 8192 bytes (=2 pages). For 64 bit one dasd_mchunkt_t structure has
+ * 24 bytes, the struct dasd_ccw_req has 136 bytes and each block can use
+ * up to 16 bytes (8 for the ccw and 8 for the idal pointer). In
+ * addition we have one define extent ccw + 16 bytes of data and a 
+ * locate record ccw for each block (stupid devices!) + 16 bytes of data.
+ * That makes:
+ * (8192 - 24 - 136 - 8 - 16) / 40 = 200.2 blocks at maximum.
+ * We want to fit two into the available memory so that we can immediately
+ * start the next request if one finishes off. That makes 100.1 blocks
+ * for one request. Give a little safety and the result is 96.
+ */
+static struct dasd_discipline dasd_fba_discipline = {
+	.owner = THIS_MODULE,
+	.name = "FBA ",
+	.ebcname = "FBA ",
+	.max_blocks = 96,
+	.check_device = dasd_fba_check_characteristics,
+	.do_analysis = dasd_fba_do_analysis,
+	.fill_geometry = dasd_fba_fill_geometry,
+	.start_IO = dasd_start_IO,
+	.term_IO = dasd_term_IO,
+	.examine_error = dasd_fba_examine_error,
+	.erp_action = dasd_fba_erp_action,
+	.erp_postaction = dasd_fba_erp_postaction,
+	.build_cp = dasd_fba_build_cp,
+	.free_cp = dasd_fba_free_cp,
+	.dump_sense = dasd_fba_dump_sense,
+	.fill_info = dasd_fba_fill_info,
+};
+
+static int __init
+dasd_fba_init(void)
+{
+	int ret;
+
+	ASCEBC(dasd_fba_discipline.ebcname, 4);
+
+	ret = ccw_driver_register(&dasd_fba_driver);
+	if (ret)
+		return ret;
+
+	dasd_generic_auto_online(&dasd_fba_driver);
+	return 0;
+}
+
+static void __exit
+dasd_fba_cleanup(void)
+{
+	ccw_driver_unregister(&dasd_fba_driver);
+}
+
+module_init(dasd_fba_init);
+module_exit(dasd_fba_cleanup);
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-indent-level: 4 
+ * c-brace-imaginary-offset: 0
+ * c-brace-offset: -4
+ * c-argdecl-indent: 4
+ * c-label-offset: -4
+ * c-continued-statement-offset: 4
+ * c-continued-brace-offset: 0
+ * indent-tabs-mode: 1
+ * tab-width: 8
+ * End:
+ */

drivers/s390/block/dasd_fba.h

@@ -0,0 +1,73 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_fba.h
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ *
+ * $Revision: 1.6 $
+ */
+
+#ifndef DASD_FBA_H
+#define DASD_FBA_H
+
+struct DE_fba_data {
+	struct {
+		unsigned char perm:2;	/* Permissions on this extent */
+		unsigned char zero:2;	/* Must be zero */
+		unsigned char da:1;	/* usually zero */
+		unsigned char diag:1;	/* allow diagnose */
+		unsigned char zero2:2;	/* zero */
+	} __attribute__ ((packed)) mask;
+	__u8 zero;		/* Must be zero */
+	__u16 blk_size;		/* Blocksize */
+	__u32 ext_loc;		/* Extent locator */
+	__u32 ext_beg;		/* logical number of block 0 in extent */
+	__u32 ext_end;		/* logocal number of last block in extent */
+} __attribute__ ((packed));
+
+struct LO_fba_data {
+	struct {
+		unsigned char zero:4;
+		unsigned char cmd:4;
+	} __attribute__ ((packed)) operation;
+	__u8 auxiliary;
+	__u16 blk_ct;
+	__u32 blk_nr;
+} __attribute__ ((packed));
+
+struct dasd_fba_characteristics {
+	union {
+		__u8 c;
+		struct {
+			unsigned char reserved:1;
+			unsigned char overrunnable:1;
+			unsigned char burst_byte:1;
+			unsigned char data_chain:1;
+			unsigned char zeros:4;
+		} __attribute__ ((packed)) bits;
+	} __attribute__ ((packed)) mode;
+	union {
+		__u8 c;
+		struct {
+			unsigned char zero0:1;
+			unsigned char removable:1;
+			unsigned char shared:1;
+			unsigned char zero1:1;
+			unsigned char mam:1;
+			unsigned char zeros:3;
+		} __attribute__ ((packed)) bits;
+	} __attribute__ ((packed)) features;
+	__u8 dev_class;
+	__u8 unit_type;
+	__u16 blk_size;
+	__u32 blk_per_cycl;
+	__u32 blk_per_bound;
+	__u32 blk_bdsa;
+	__u32 reserved0;
+	__u16 reserved1;
+	__u16 blk_ce;
+	__u32 reserved2;
+	__u16 reserved3;
+} __attribute__ ((packed));
+
+#endif				/* DASD_FBA_H */

drivers/s390/block/dasd_genhd.c

@@ -0,0 +1,185 @@
+/*
+ * File...........: linux/drivers/s390/block/dasd_genhd.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *		    Horst Hummel <Horst.Hummel@de.ibm.com>
+ *		    Carsten Otte <Cotte@de.ibm.com>
+ *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2001
+ *
+ * gendisk related functions for the dasd driver.
+ *
+ * $Revision: 1.48 $
+ */
+
+#include <linux/config.h>
+#include <linux/interrupt.h>
+#include <linux/fs.h>
+#include <linux/blkpg.h>
+
+#include <asm/uaccess.h>
+
+/* This is ugly... */
+#define PRINTK_HEADER "dasd_gendisk:"
+
+#include "dasd_int.h"
+
+/*
+ * Allocate and register gendisk structure for device.
+ */
+int
+dasd_gendisk_alloc(struct dasd_device *device)
+{
+	struct gendisk *gdp;
+	int len;
+
+	/* Make sure the minor for this device exists. */
+	if (device->devindex >= DASD_PER_MAJOR)
+		return -EBUSY;
+
+	gdp = alloc_disk(1 << DASD_PARTN_BITS);
+	if (!gdp)
+		return -ENOMEM;
+
+	/* Initialize gendisk structure. */
+	gdp->major = DASD_MAJOR;
+	gdp->first_minor = device->devindex << DASD_PARTN_BITS;
+	gdp->fops = &dasd_device_operations;
+	gdp->driverfs_dev = &device->cdev->dev;
+
+	/*
+	 * Set device name.
+	 *   dasda - dasdz : 26 devices
+	 *   dasdaa - dasdzz : 676 devices, added up = 702
+	 *   dasdaaa - dasdzzz : 17576 devices, added up = 18278
+	 *   dasdaaaa - dasdzzzz : 456976 devices, added up = 475252
+	 */
+	len = sprintf(gdp->disk_name, "dasd");
+	if (device->devindex > 25) {
+	        if (device->devindex > 701) {
+		        if (device->devindex > 18277)
+			        len += sprintf(gdp->disk_name + len, "%c",
+					       'a'+(((device->devindex-18278)
+						     /17576)%26));
+			len += sprintf(gdp->disk_name + len, "%c",
+				       'a'+(((device->devindex-702)/676)%26));
+		}
+		len += sprintf(gdp->disk_name + len, "%c",
+			       'a'+(((device->devindex-26)/26)%26));
+	}
+	len += sprintf(gdp->disk_name + len, "%c", 'a'+(device->devindex%26));
+
+ 	sprintf(gdp->devfs_name, "dasd/%s", device->cdev->dev.bus_id);
+
+	if (test_bit(DASD_FLAG_RO, &device->flags))
+		set_disk_ro(gdp, 1);
+	gdp->private_data = device;
+	gdp->queue = device->request_queue;
+	device->gdp = gdp;
+	set_capacity(device->gdp, 0);
+	add_disk(device->gdp);
+	return 0;
+}
+
+/*
+ * Unregister and free gendisk structure for device.
+ */
+void
+dasd_gendisk_free(struct dasd_device *device)
+{
+	del_gendisk(device->gdp);
+	device->gdp->queue = 0;
+	put_disk(device->gdp);
+	device->gdp = 0;
+}
+
+/*
+ * Trigger a partition detection.
+ */
+int
+dasd_scan_partitions(struct dasd_device * device)
+{
+	struct block_device *bdev;
+
+	/* Make the disk known. */
+	set_capacity(device->gdp, device->blocks << device->s2b_shift);
+	bdev = bdget_disk(device->gdp, 0);
+	if (!bdev || blkdev_get(bdev, FMODE_READ, 1) < 0)
+		return -ENODEV;
+	/*
+	 * See fs/partition/check.c:register_disk,rescan_partitions
+	 * Can't call rescan_partitions directly. Use ioctl.
+	 */
+	ioctl_by_bdev(bdev, BLKRRPART, 0);
+	/*
+	 * Since the matching blkdev_put call to the blkdev_get in
+	 * this function is not called before dasd_destroy_partitions
+	 * the offline open_count limit needs to be increased from
+	 * 0 to 1. This is done by setting device->bdev (see
+	 * dasd_generic_set_offline). As long as the partition
+	 * detection is running no offline should be allowed. That
+	 * is why the assignment to device->bdev is done AFTER
+	 * the BLKRRPART ioctl.
+	 */
+	device->bdev = bdev;
+	return 0;
+}
+
+/*
+ * Remove all inodes in the system for a device, delete the
+ * partitions and make device unusable by setting its size to zero.
+ */
+void
+dasd_destroy_partitions(struct dasd_device * device)
+{
+	/* The two structs have 168/176 byte on 31/64 bit. */
+	struct blkpg_partition bpart;
+	struct blkpg_ioctl_arg barg;
+	struct block_device *bdev;
+
+	/*
+	 * Get the bdev pointer from the device structure and clear
+	 * device->bdev to lower the offline open_count limit again.
+	 */
+	bdev = device->bdev;
+	device->bdev = 0;
+
+	/*
+	 * See fs/partition/check.c:delete_partition
+	 * Can't call delete_partitions directly. Use ioctl.
+	 * The ioctl also does locking and invalidation.
+	 */
+	memset(&bpart, 0, sizeof(struct blkpg_partition));
+	memset(&barg, 0, sizeof(struct blkpg_ioctl_arg));
+	barg.data = &bpart;
+	barg.op = BLKPG_DEL_PARTITION;
+	for (bpart.pno = device->gdp->minors - 1; bpart.pno > 0; bpart.pno--)
+		ioctl_by_bdev(bdev, BLKPG, (unsigned long) &barg);
+
+	invalidate_partition(device->gdp, 0);
+	/* Matching blkdev_put to the blkdev_get in dasd_scan_partitions. */
+	blkdev_put(bdev);
+	set_capacity(device->gdp, 0);
+}
+
+int
+dasd_gendisk_init(void)
+{
+	int rc;
+
+	/* Register to static dasd major 94 */
+	rc = register_blkdev(DASD_MAJOR, "dasd");
+	if (rc != 0) {
+		MESSAGE(KERN_WARNING,
+			"Couldn't register successfully to "
+			"major no %d", DASD_MAJOR);
+		return rc;
+	}
+	return 0;
+}
+
+void
+dasd_gendisk_exit(void)
+{
+	unregister_blkdev(DASD_MAJOR, "dasd");
+}

drivers/s390/block/dasd_int.h

@@ -0,0 +1,576 @@
+/* 
+ * File...........: linux/drivers/s390/block/dasd_int.h
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *                  Horst Hummel <Horst.Hummel@de.ibm.com> 
+ *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
+ *
+ * $Revision: 1.63 $
+ */
+
+#ifndef DASD_INT_H
+#define DASD_INT_H
+
+#ifdef __KERNEL__
+
+/* erp debugging in dasd.c and dasd_3990_erp.c */
+#define ERP_DEBUG
+
+
+/* we keep old device allocation scheme; IOW, minors are still in 0..255 */
+#define DASD_PER_MAJOR (1U << (MINORBITS - DASD_PARTN_BITS))
+#define DASD_PARTN_MASK ((1 << DASD_PARTN_BITS) - 1)
+
+/*
+ * States a dasd device can have:
+ *   new: the dasd_device structure is allocated.
+ *   known: the discipline for the device is identified.
+ *   basic: the device can do basic i/o.
+ *   accept: the device is analysed (format is known).
+ *   ready: partition detection is done and the device is can do block io.
+ *   online: the device accepts requests from the block device queue.
+ *
+ * Things to do for startup state transitions:
+ *   new -> known: find discipline for the device and create devfs entries.
+ *   known -> basic: request irq line for the device.
+ *   basic -> ready: do the initial analysis, e.g. format detection,
+ *                   do block device setup and detect partitions.
+ *   ready -> online: schedule the device tasklet.
+ * Things to do for shutdown state transitions:
+ *   online -> ready: just set the new device state.
+ *   ready -> basic: flush requests from the block device layer, clear
+ *                   partition information and reset format information.
+ *   basic -> known: terminate all requests and free irq.
+ *   known -> new: remove devfs entries and forget discipline.
+ */
+
+#define DASD_STATE_NEW	  0
+#define DASD_STATE_KNOWN  1
+#define DASD_STATE_BASIC  2
+#define DASD_STATE_READY  3
+#define DASD_STATE_ONLINE 4
+
+#include <linux/module.h>
+#include <linux/wait.h>
+#include <linux/blkdev.h>
+#include <linux/devfs_fs_kernel.h>
+#include <linux/genhd.h>
+#include <linux/hdreg.h>
+#include <linux/interrupt.h>
+#include <asm/ccwdev.h>
+#include <linux/workqueue.h>
+#include <asm/debug.h>
+#include <asm/dasd.h>
+#include <asm/idals.h>
+
+/*
+ * SECTION: Type definitions
+ */
+struct dasd_device;
+
+typedef int (*dasd_ioctl_fn_t) (struct block_device *bdev, int no, long args);
+
+struct dasd_ioctl {
+	struct list_head list;
+	struct module *owner;
+	int no;
+	dasd_ioctl_fn_t handler;
+};
+
+typedef enum {
+	dasd_era_fatal = -1,	/* no chance to recover		     */
+	dasd_era_none = 0,	/* don't recover, everything alright */
+	dasd_era_msg = 1,	/* don't recover, just report...     */
+	dasd_era_recover = 2	/* recovery action recommended	     */
+} dasd_era_t;
+
+/* BIT DEFINITIONS FOR SENSE DATA */
+#define DASD_SENSE_BIT_0 0x80
+#define DASD_SENSE_BIT_1 0x40
+#define DASD_SENSE_BIT_2 0x20
+#define DASD_SENSE_BIT_3 0x10
+
+/*
+ * SECTION: MACROs for klogd and s390 debug feature (dbf)
+ */
+#define DBF_DEV_EVENT(d_level, d_device, d_str, d_data...) \
+do { \
+	debug_sprintf_event(d_device->debug_area, \
+			    d_level, \
+			    d_str "\n", \
+			    d_data); \
+} while(0)
+
+#define DBF_DEV_EXC(d_level, d_device, d_str, d_data...) \
+do { \
+	debug_sprintf_exception(d_device->debug_area, \
+				d_level, \
+				d_str "\n", \
+				d_data); \
+} while(0)
+
+#define DBF_EVENT(d_level, d_str, d_data...)\
+do { \
+	debug_sprintf_event(dasd_debug_area, \
+			    d_level,\
+			    d_str "\n", \
+			    d_data); \
+} while(0)
+
+#define DBF_EXC(d_level, d_str, d_data...)\
+do { \
+	debug_sprintf_exception(dasd_debug_area, \
+				d_level,\
+				d_str "\n", \
+				d_data); \
+} while(0)
+
+/* definition of dbf debug levels */
+#define	DBF_EMERG	0	/* system is unusable			*/
+#define	DBF_ALERT	1	/* action must be taken immediately	*/
+#define	DBF_CRIT	2	/* critical conditions			*/
+#define	DBF_ERR		3	/* error conditions			*/
+#define	DBF_WARNING	4	/* warning conditions			*/
+#define	DBF_NOTICE	5	/* normal but significant condition	*/
+#define	DBF_INFO	6	/* informational			*/
+#define	DBF_DEBUG	6	/* debug-level messages			*/
+
+/* messages to be written via klogd and dbf */
+#define DEV_MESSAGE(d_loglevel,d_device,d_string,d_args...)\
+do { \
+	printk(d_loglevel PRINTK_HEADER " %s: " d_string "\n", \
+	       d_device->cdev->dev.bus_id, d_args); \
+	DBF_DEV_EVENT(DBF_ALERT, d_device, d_string, d_args); \
+} while(0)
+
+#define MESSAGE(d_loglevel,d_string,d_args...)\
+do { \
+	printk(d_loglevel PRINTK_HEADER " " d_string "\n", d_args); \
+	DBF_EVENT(DBF_ALERT, d_string, d_args); \
+} while(0)
+
+/* messages to be written via klogd only */
+#define DEV_MESSAGE_LOG(d_loglevel,d_device,d_string,d_args...)\
+do { \
+	printk(d_loglevel PRINTK_HEADER " %s: " d_string "\n", \
+	       d_device->cdev->dev.bus_id, d_args); \
+} while(0)
+
+#define MESSAGE_LOG(d_loglevel,d_string,d_args...)\
+do { \
+	printk(d_loglevel PRINTK_HEADER " " d_string "\n", d_args); \
+} while(0)
+
+struct dasd_ccw_req {
+	unsigned int magic;		/* Eye catcher */
+        struct list_head list;		/* list_head for request queueing. */
+
+	/* Where to execute what... */
+	struct dasd_device *device;	/* device the request is for */
+	struct ccw1 *cpaddr;		/* address of channel program */
+	char status;	        	/* status of this request */
+	short retries;			/* A retry counter */
+	unsigned long flags;        	/* flags of this request */
+
+	/* ... and how */
+	unsigned long starttime;	/* jiffies time of request start */
+	int expires;			/* expiration period in jiffies */
+	char lpm;               	/* logical path mask */
+	void *data;			/* pointer to data area */
+
+	/* these are important for recovering erroneous requests          */
+	struct irb irb;			/* device status in case of an error */
+	struct dasd_ccw_req *refers;	/* ERP-chain queueing. */
+	void *function; 		/* originating ERP action */
+
+	/* these are for statistics only */
+	unsigned long long buildclk;	/* TOD-clock of request generation */
+	unsigned long long startclk;	/* TOD-clock of request start */
+	unsigned long long stopclk;	/* TOD-clock of request interrupt */
+	unsigned long long endclk;	/* TOD-clock of request termination */
+
+        /* Callback that is called after reaching final status. */
+        void (*callback)(struct dasd_ccw_req *, void *data);
+        void *callback_data;
+};
+
+/* 
+ * dasd_ccw_req -> status can be:
+ */
+#define DASD_CQR_FILLED   0x00	/* request is ready to be processed */
+#define DASD_CQR_QUEUED   0x01	/* request is queued to be processed */
+#define DASD_CQR_IN_IO    0x02	/* request is currently in IO */
+#define DASD_CQR_DONE     0x03	/* request is completed successfully */
+#define DASD_CQR_ERROR    0x04	/* request is completed with error */
+#define DASD_CQR_FAILED   0x05	/* request is finally failed */
+#define DASD_CQR_CLEAR    0x06	/* request is clear pending */
+
+/* per dasd_ccw_req flags */
+#define DASD_CQR_FLAGS_USE_ERP   0	/* use ERP for this request */
+
+/* Signature for error recovery functions. */
+typedef struct dasd_ccw_req *(*dasd_erp_fn_t) (struct dasd_ccw_req *);
+
+/*
+ * the struct dasd_discipline is
+ * sth like a table of virtual functions, if you think of dasd_eckd
+ * inheriting dasd...
+ * no, currently we are not planning to reimplement the driver in C++
+ */
+struct dasd_discipline {
+	struct module *owner;
+	char ebcname[8];	/* a name used for tagging and printks */
+	char name[8];		/* a name used for tagging and printks */
+	int max_blocks;		/* maximum number of blocks to be chained */
+
+	struct list_head list;	/* used for list of disciplines */
+
+        /*
+         * Device recognition functions. check_device is used to verify
+         * the sense data and the information returned by read device
+         * characteristics. It returns 0 if the discipline can be used
+         * for the device in question.
+         * do_analysis is used in the step from device state "basic" to
+         * state "accept". It returns 0 if the device can be made ready,
+         * it returns -EMEDIUMTYPE if the device can't be made ready or
+         * -EAGAIN if do_analysis started a ccw that needs to complete
+         * before the analysis may be repeated.
+         */
+        int (*check_device)(struct dasd_device *);
+	int (*do_analysis) (struct dasd_device *);
+
+        /*
+         * Device operation functions. build_cp creates a ccw chain for
+         * a block device request, start_io starts the request and
+         * term_IO cancels it (e.g. in case of a timeout). format_device
+         * returns a ccw chain to be used to format the device.
+         */
+	struct dasd_ccw_req *(*build_cp) (struct dasd_device *,
+					  struct request *);
+	int (*start_IO) (struct dasd_ccw_req *);
+	int (*term_IO) (struct dasd_ccw_req *);
+	struct dasd_ccw_req *(*format_device) (struct dasd_device *,
+					       struct format_data_t *);
+	int (*free_cp) (struct dasd_ccw_req *, struct request *);
+        /*
+         * Error recovery functions. examine_error() returns a value that
+         * indicates what to do for an error condition. If examine_error()
+         * returns 'dasd_era_recover' erp_action() is called to create a 
+         * special error recovery ccw. erp_postaction() is called after
+         * an error recovery ccw has finished its execution. dump_sense
+         * is called for every error condition to print the sense data
+         * to the console.
+         */
+	dasd_era_t(*examine_error) (struct dasd_ccw_req *, struct irb *);
+	dasd_erp_fn_t(*erp_action) (struct dasd_ccw_req *);
+	dasd_erp_fn_t(*erp_postaction) (struct dasd_ccw_req *);
+	void (*dump_sense) (struct dasd_device *, struct dasd_ccw_req *,
+			    struct irb *);
+
+        /* i/o control functions. */
+	int (*fill_geometry) (struct dasd_device *, struct hd_geometry *);
+	int (*fill_info) (struct dasd_device *, struct dasd_information2_t *);
+};
+
+extern struct dasd_discipline *dasd_diag_discipline_pointer;
+
+struct dasd_device {
+	/* Block device stuff. */
+	struct gendisk *gdp;
+	request_queue_t *request_queue;
+	spinlock_t request_queue_lock;
+	struct block_device *bdev;
+        unsigned int devindex;
+	unsigned long blocks;		/* size of volume in blocks */
+	unsigned int bp_block;		/* bytes per block */
+	unsigned int s2b_shift;		/* log2 (bp_block/512) */
+	unsigned long flags;		/* per device flags */
+
+	/* Device discipline stuff. */
+	struct dasd_discipline *discipline;
+	char *private;
+
+	/* Device state and target state. */
+	int state, target;
+	int stopped;		/* device (ccw_device_start) was stopped */
+
+	/* Open and reference count. */
+        atomic_t ref_count;
+	atomic_t open_count;
+
+	/* ccw queue and memory for static ccw/erp buffers. */
+	struct list_head ccw_queue;
+	spinlock_t mem_lock;
+	void *ccw_mem;
+	void *erp_mem;
+	struct list_head ccw_chunks;
+	struct list_head erp_chunks;
+
+	atomic_t tasklet_scheduled;
+        struct tasklet_struct tasklet;
+	struct work_struct kick_work;
+	struct timer_list timer;
+
+	debug_info_t *debug_area;
+
+	struct ccw_device *cdev;
+
+#ifdef CONFIG_DASD_PROFILE
+	struct dasd_profile_info_t profile;
+#endif
+};
+
+/* reasons why device (ccw_device_start) was stopped */
+#define DASD_STOPPED_NOT_ACC 1         /* not accessible */
+#define DASD_STOPPED_QUIESCE 2         /* Quiesced */
+#define DASD_STOPPED_PENDING 4         /* long busy */
+#define DASD_STOPPED_DC_WAIT 8         /* disconnected, wait */
+#define DASD_STOPPED_DC_EIO  16        /* disconnected, return -EIO */
+
+/* per device flags */
+#define DASD_FLAG_RO		0	/* device is read-only */
+#define DASD_FLAG_USE_DIAG	1	/* use diag disciplnie */
+#define DASD_FLAG_DSC_ERROR	2	/* return -EIO when disconnected */
+#define DASD_FLAG_OFFLINE	3	/* device is in offline processing */
+
+void dasd_put_device_wake(struct dasd_device *);
+
+/*
+ * Reference count inliners
+ */
+static inline void
+dasd_get_device(struct dasd_device *device)
+{
+	atomic_inc(&device->ref_count);
+}
+
+static inline void
+dasd_put_device(struct dasd_device *device)
+{
+	if (atomic_dec_return(&device->ref_count) == 0)
+		dasd_put_device_wake(device);
+}
+
+/*
+ * The static memory in ccw_mem and erp_mem is managed by a sorted
+ * list of free memory chunks.
+ */
+struct dasd_mchunk
+{
+	struct list_head list;
+	unsigned long size;
+} __attribute__ ((aligned(8)));
+
+static inline void
+dasd_init_chunklist(struct list_head *chunk_list, void *mem,
+		    unsigned long size)
+{
+	struct dasd_mchunk *chunk;
+
+	INIT_LIST_HEAD(chunk_list);
+	chunk = (struct dasd_mchunk *) mem;
+	chunk->size = size - sizeof(struct dasd_mchunk);
+	list_add(&chunk->list, chunk_list);
+}
+
+static inline void *
+dasd_alloc_chunk(struct list_head *chunk_list, unsigned long size)
+{
+	struct dasd_mchunk *chunk, *tmp;
+
+	size = (size + 7L) & -8L;
+	list_for_each_entry(chunk, chunk_list, list) {
+		if (chunk->size < size)
+			continue;
+		if (chunk->size > size + sizeof(struct dasd_mchunk)) {
+			char *endaddr = (char *) (chunk + 1) + chunk->size;
+			tmp = (struct dasd_mchunk *) (endaddr - size) - 1;
+			tmp->size = size;
+			chunk->size -= size + sizeof(struct dasd_mchunk);
+			chunk = tmp;
+		} else
+			list_del(&chunk->list);
+		return (void *) (chunk + 1);
+	}
+	return NULL;
+}
+
+static inline void
+dasd_free_chunk(struct list_head *chunk_list, void *mem)
+{
+	struct dasd_mchunk *chunk, *tmp;
+	struct list_head *p, *left;
+
+	chunk = (struct dasd_mchunk *)
+		((char *) mem - sizeof(struct dasd_mchunk));
+	/* Find out the left neighbour in chunk_list. */
+	left = chunk_list;
+	list_for_each(p, chunk_list) {
+		if (list_entry(p, struct dasd_mchunk, list) > chunk)
+			break;
+		left = p;
+	}
+	/* Try to merge with right neighbour = next element from left. */
+	if (left->next != chunk_list) {
+		tmp = list_entry(left->next, struct dasd_mchunk, list);
+		if ((char *) (chunk + 1) + chunk->size == (char *) tmp) {
+			list_del(&tmp->list);
+			chunk->size += tmp->size + sizeof(struct dasd_mchunk);
+		}
+	}
+	/* Try to merge with left neighbour. */
+	if (left != chunk_list) {
+		tmp = list_entry(left, struct dasd_mchunk, list);
+		if ((char *) (tmp + 1) + tmp->size == (char *) chunk) {
+			tmp->size += chunk->size + sizeof(struct dasd_mchunk);
+			return;
+		}
+	}
+	__list_add(&chunk->list, left, left->next);
+}
+
+/*
+ * Check if bsize is in { 512, 1024, 2048, 4096 }
+ */
+static inline int
+dasd_check_blocksize(int bsize)
+{
+	if (bsize < 512 || bsize > 4096 || (bsize & (bsize - 1)) != 0)
+		return -EMEDIUMTYPE;
+	return 0;
+}
+
+/* externals in dasd.c */
+#define DASD_PROFILE_ON	 1
+#define DASD_PROFILE_OFF 0
+
+extern debug_info_t *dasd_debug_area;
+extern struct dasd_profile_info_t dasd_global_profile;
+extern unsigned int dasd_profile_level;
+extern struct block_device_operations dasd_device_operations;
+
+extern kmem_cache_t *dasd_page_cache;
+
+struct dasd_ccw_req *
+dasd_kmalloc_request(char *, int, int, struct dasd_device *);
+struct dasd_ccw_req *
+dasd_smalloc_request(char *, int, int, struct dasd_device *);
+void dasd_kfree_request(struct dasd_ccw_req *, struct dasd_device *);
+void dasd_sfree_request(struct dasd_ccw_req *, struct dasd_device *);
+
+static inline int
+dasd_kmalloc_set_cda(struct ccw1 *ccw, void *cda, struct dasd_device *device)
+{
+	return set_normalized_cda(ccw, cda);
+}
+
+struct dasd_device *dasd_alloc_device(void);
+void dasd_free_device(struct dasd_device *);
+
+void dasd_enable_device(struct dasd_device *);
+void dasd_set_target_state(struct dasd_device *, int);
+void dasd_kick_device(struct dasd_device *);
+
+void dasd_add_request_head(struct dasd_ccw_req *);
+void dasd_add_request_tail(struct dasd_ccw_req *);
+int  dasd_start_IO(struct dasd_ccw_req *);
+int  dasd_term_IO(struct dasd_ccw_req *);
+void dasd_schedule_bh(struct dasd_device *);
+int  dasd_sleep_on(struct dasd_ccw_req *);
+int  dasd_sleep_on_immediatly(struct dasd_ccw_req *);
+int  dasd_sleep_on_interruptible(struct dasd_ccw_req *);
+void dasd_set_timer(struct dasd_device *, int);
+void dasd_clear_timer(struct dasd_device *);
+int  dasd_cancel_req(struct dasd_ccw_req *);
+int dasd_generic_probe (struct ccw_device *, struct dasd_discipline *);
+void dasd_generic_remove (struct ccw_device *cdev);
+int dasd_generic_set_online(struct ccw_device *, struct dasd_discipline *);
+int dasd_generic_set_offline (struct ccw_device *cdev);
+int dasd_generic_notify(struct ccw_device *, int);
+void dasd_generic_auto_online (struct ccw_driver *);
+
+/* externals in dasd_devmap.c */
+extern int dasd_max_devindex;
+extern int dasd_probeonly;
+extern int dasd_autodetect;
+
+int dasd_devmap_init(void);
+void dasd_devmap_exit(void);
+
+struct dasd_device *dasd_create_device(struct ccw_device *);
+void dasd_delete_device(struct dasd_device *);
+
+int dasd_add_sysfs_files(struct ccw_device *);
+void dasd_remove_sysfs_files(struct ccw_device *);
+
+struct dasd_device *dasd_device_from_cdev(struct ccw_device *);
+struct dasd_device *dasd_device_from_devindex(int);
+
+int dasd_parse(void);
+int dasd_busid_known(char *);
+
+/* externals in dasd_gendisk.c */
+int  dasd_gendisk_init(void);
+void dasd_gendisk_exit(void);
+int dasd_gendisk_alloc(struct dasd_device *);
+void dasd_gendisk_free(struct dasd_device *);
+int dasd_scan_partitions(struct dasd_device *);
+void dasd_destroy_partitions(struct dasd_device *);
+
+/* externals in dasd_ioctl.c */
+int  dasd_ioctl_init(void);
+void dasd_ioctl_exit(void);
+int  dasd_ioctl_no_register(struct module *, int, dasd_ioctl_fn_t);
+int  dasd_ioctl_no_unregister(struct module *, int, dasd_ioctl_fn_t);
+int  dasd_ioctl(struct inode *, struct file *, unsigned int, unsigned long);
+
+/* externals in dasd_proc.c */
+int dasd_proc_init(void);
+void dasd_proc_exit(void);
+
+/* externals in dasd_erp.c */
+struct dasd_ccw_req *dasd_default_erp_action(struct dasd_ccw_req *);
+struct dasd_ccw_req *dasd_default_erp_postaction(struct dasd_ccw_req *);
+struct dasd_ccw_req *dasd_alloc_erp_request(char *, int, int,
+					    struct dasd_device *);
+void dasd_free_erp_request(struct dasd_ccw_req *, struct dasd_device *);
+void dasd_log_sense(struct dasd_ccw_req *, struct irb *);
+void dasd_log_ccw(struct dasd_ccw_req *, int, __u32);
+
+/* externals in dasd_3370_erp.c */
+dasd_era_t dasd_3370_erp_examine(struct dasd_ccw_req *, struct irb *);
+
+/* externals in dasd_3990_erp.c */
+dasd_era_t dasd_3990_erp_examine(struct dasd_ccw_req *, struct irb *);
+struct dasd_ccw_req *dasd_3990_erp_action(struct dasd_ccw_req *);
+
+/* externals in dasd_9336_erp.c */
+dasd_era_t dasd_9336_erp_examine(struct dasd_ccw_req *, struct irb *);
+
+/* externals in dasd_9336_erp.c */
+dasd_era_t dasd_9343_erp_examine(struct dasd_ccw_req *, struct irb *);
+struct dasd_ccw_req *dasd_9343_erp_action(struct dasd_ccw_req *);
+
+#endif				/* __KERNEL__ */
+
+#endif				/* DASD_H */
+
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-indent-level: 4 
+ * c-brace-imaginary-offset: 0
+ * c-brace-offset: -4
+ * c-argdecl-indent: 4
+ * c-label-offset: -4
+ * c-continued-statement-offset: 4
+ * c-continued-brace-offset: 0
+ * indent-tabs-mode: 1
+ * tab-width: 8
+ * End:
+ */

drivers/s390/block/dasd_ioctl.c

@@ -0,0 +1,554 @@
+/*
+ * File...........: linux/drivers/s390/block/dasd_ioctl.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *		    Horst Hummel <Horst.Hummel@de.ibm.com>
+ *		    Carsten Otte <Cotte@de.ibm.com>
+ *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2001
+ *
+ * i/o controls for the dasd driver.
+ */
+#include <linux/config.h>
+#include <linux/interrupt.h>
+#include <linux/major.h>
+#include <linux/fs.h>
+#include <linux/blkpg.h>
+
+#include <asm/ccwdev.h>
+#include <asm/uaccess.h>
+
+/* This is ugly... */
+#define PRINTK_HEADER "dasd_ioctl:"
+
+#include "dasd_int.h"
+
+/*
+ * SECTION: ioctl functions.
+ */
+static struct list_head dasd_ioctl_list = LIST_HEAD_INIT(dasd_ioctl_list);
+
+/*
+ * Find the ioctl with number no.
+ */
+static struct dasd_ioctl *
+dasd_find_ioctl(int no)
+{
+	struct dasd_ioctl *ioctl;
+
+	list_for_each_entry (ioctl, &dasd_ioctl_list, list)
+		if (ioctl->no == no)
+			return ioctl;
+	return NULL;
+}
+
+/*
+ * Register ioctl with number no.
+ */
+int
+dasd_ioctl_no_register(struct module *owner, int no, dasd_ioctl_fn_t handler)
+{
+	struct dasd_ioctl *new;
+	if (dasd_find_ioctl(no))
+		return -EBUSY;
+	new = kmalloc(sizeof (struct dasd_ioctl), GFP_KERNEL);
+	if (new == NULL)
+		return -ENOMEM;
+	new->owner = owner;
+	new->no = no;
+	new->handler = handler;
+	list_add(&new->list, &dasd_ioctl_list);
+	return 0;
+}
+
+/*
+ * Deregister ioctl with number no.
+ */
+int
+dasd_ioctl_no_unregister(struct module *owner, int no, dasd_ioctl_fn_t handler)
+{
+	struct dasd_ioctl *old = dasd_find_ioctl(no);
+	if (old == NULL)
+		return -ENOENT;
+	if (old->no != no || old->handler != handler || owner != old->owner)
+		return -EINVAL;
+	list_del(&old->list);
+	kfree(old);
+	return 0;
+}
+
+int
+dasd_ioctl(struct inode *inp, struct file *filp,
+	   unsigned int no, unsigned long data)
+{
+	struct block_device *bdev = inp->i_bdev;
+	struct dasd_device *device = bdev->bd_disk->private_data;
+	struct dasd_ioctl *ioctl;
+	const char *dir;
+	int rc;
+
+	if ((_IOC_DIR(no) != _IOC_NONE) && (data == 0)) {
+		PRINT_DEBUG("empty data ptr");
+		return -EINVAL;
+	}
+	dir = _IOC_DIR (no) == _IOC_NONE ? "0" :
+		_IOC_DIR (no) == _IOC_READ ? "r" :
+		_IOC_DIR (no) == _IOC_WRITE ? "w" : 
+		_IOC_DIR (no) == (_IOC_READ | _IOC_WRITE) ? "rw" : "u";
+	DBF_DEV_EVENT(DBF_DEBUG, device,
+		      "ioctl 0x%08x %s'0x%x'%d(%d) with data %8lx", no,
+		      dir, _IOC_TYPE(no), _IOC_NR(no), _IOC_SIZE(no), data);
+	/* Search for ioctl no in the ioctl list. */
+	list_for_each_entry(ioctl, &dasd_ioctl_list, list) {
+		if (ioctl->no == no) {
+			/* Found a matching ioctl. Call it. */
+			if (!try_module_get(ioctl->owner))
+				continue;
+			rc = ioctl->handler(bdev, no, data);
+			module_put(ioctl->owner);
+			return rc;
+		}
+	}
+	/* No ioctl with number no. */
+	DBF_DEV_EVENT(DBF_INFO, device,
+		      "unknown ioctl 0x%08x=%s'0x%x'%d(%d) data %8lx", no,
+		      dir, _IOC_TYPE(no), _IOC_NR(no), _IOC_SIZE(no), data);
+	return -EINVAL;
+}
+
+static int
+dasd_ioctl_api_version(struct block_device *bdev, int no, long args)
+{
+	int ver = DASD_API_VERSION;
+	return put_user(ver, (int __user *) args);
+}
+
+/*
+ * Enable device.
+ * used by dasdfmt after BIODASDDISABLE to retrigger blocksize detection
+ */
+static int
+dasd_ioctl_enable(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+	dasd_enable_device(device);
+	/* Formatting the dasd device can change the capacity. */
+	down(&bdev->bd_sem);
+	i_size_write(bdev->bd_inode, (loff_t)get_capacity(device->gdp) << 9);
+	up(&bdev->bd_sem);
+	return 0;
+}
+
+/*
+ * Disable device.
+ * Used by dasdfmt. Disable I/O operations but allow ioctls.
+ */
+static int
+dasd_ioctl_disable(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+	/*
+	 * Man this is sick. We don't do a real disable but only downgrade
+	 * the device to DASD_STATE_BASIC. The reason is that dasdfmt uses
+	 * BIODASDDISABLE to disable accesses to the device via the block
+	 * device layer but it still wants to do i/o on the device by
+	 * using the BIODASDFMT ioctl. Therefore the correct state for the
+	 * device is DASD_STATE_BASIC that allows to do basic i/o.
+	 */
+	dasd_set_target_state(device, DASD_STATE_BASIC);
+	/*
+	 * Set i_size to zero, since read, write, etc. check against this
+	 * value.
+	 */
+	down(&bdev->bd_sem);
+	i_size_write(bdev->bd_inode, 0);
+	up(&bdev->bd_sem);
+	return 0;
+}
+
+/*
+ * Quiesce device.
+ */
+static int
+dasd_ioctl_quiesce(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+	unsigned long flags;
+	
+	if (!capable (CAP_SYS_ADMIN))
+		return -EACCES;
+	
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+	
+	DEV_MESSAGE (KERN_DEBUG, device, "%s",
+		     "Quiesce IO on device");
+	spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
+	device->stopped |= DASD_STOPPED_QUIESCE;
+	spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
+	return 0;
+}
+
+
+/*
+ * Quiesce device.
+ */
+static int
+dasd_ioctl_resume(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+	unsigned long flags;
+	
+	if (!capable (CAP_SYS_ADMIN)) 
+		return -EACCES;
+
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+
+	DEV_MESSAGE (KERN_DEBUG, device, "%s",
+		     "resume IO on device");
+	
+	spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
+	device->stopped &= ~DASD_STOPPED_QUIESCE;
+	spin_unlock_irqrestore(get_ccwdev_lock(device->cdev), flags);
+
+	dasd_schedule_bh (device);
+	return 0;
+}
+
+/*
+ * performs formatting of _device_ according to _fdata_
+ * Note: The discipline's format_function is assumed to deliver formatting
+ * commands to format a single unit of the device. In terms of the ECKD
+ * devices this means CCWs are generated to format a single track.
+ */
+static int
+dasd_format(struct dasd_device * device, struct format_data_t * fdata)
+{
+	struct dasd_ccw_req *cqr;
+	int rc;
+
+	if (device->discipline->format_device == NULL)
+		return -EPERM;
+
+	if (device->state != DASD_STATE_BASIC) {
+		DEV_MESSAGE(KERN_WARNING, device, "%s",
+			    "dasd_format: device is not disabled! ");
+		return -EBUSY;
+	}
+
+	DBF_DEV_EVENT(DBF_NOTICE, device,
+		      "formatting units %d to %d (%d B blocks) flags %d",
+		      fdata->start_unit,
+		      fdata->stop_unit, fdata->blksize, fdata->intensity);
+
+	/* Since dasdfmt keeps the device open after it was disabled,
+	 * there still exists an inode for this device.
+	 * We must update i_blkbits, otherwise we might get errors when
+	 * enabling the device later.
+	 */
+	if (fdata->start_unit == 0) {
+		struct block_device *bdev = bdget_disk(device->gdp, 0);
+		bdev->bd_inode->i_blkbits = blksize_bits(fdata->blksize);
+		bdput(bdev);
+	}
+
+	while (fdata->start_unit <= fdata->stop_unit) {
+		cqr = device->discipline->format_device(device, fdata);
+		if (IS_ERR(cqr))
+			return PTR_ERR(cqr);
+		rc = dasd_sleep_on_interruptible(cqr);
+		dasd_sfree_request(cqr, cqr->device);
+		if (rc) {
+			if (rc != -ERESTARTSYS)
+				DEV_MESSAGE(KERN_ERR, device,
+					    " Formatting of unit %d failed "
+					    "with rc = %d",
+					    fdata->start_unit, rc);
+			return rc;
+		}
+		fdata->start_unit++;
+	}
+	return 0;
+}
+
+/*
+ * Format device.
+ */
+static int
+dasd_ioctl_format(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+	struct format_data_t fdata;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+	if (!args)
+		return -EINVAL;
+	/* fdata == NULL is no longer a valid arg to dasd_format ! */
+	device = bdev->bd_disk->private_data;
+
+	if (device == NULL)
+		return -ENODEV;
+	if (test_bit(DASD_FLAG_RO, &device->flags))
+		return -EROFS;
+	if (copy_from_user(&fdata, (void __user *) args,
+			   sizeof (struct format_data_t)))
+		return -EFAULT;
+	if (bdev != bdev->bd_contains) {
+		DEV_MESSAGE(KERN_WARNING, device, "%s",
+			    "Cannot low-level format a partition");
+		return -EINVAL;
+	}
+	return dasd_format(device, &fdata);
+}
+
+#ifdef CONFIG_DASD_PROFILE
+/*
+ * Reset device profile information
+ */
+static int
+dasd_ioctl_reset_profile(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+
+	memset(&device->profile, 0, sizeof (struct dasd_profile_info_t));
+	return 0;
+}
+
+/*
+ * Return device profile information
+ */
+static int
+dasd_ioctl_read_profile(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+
+	if (copy_to_user((long __user *) args, (long *) &device->profile,
+			 sizeof (struct dasd_profile_info_t)))
+		return -EFAULT;
+	return 0;
+}
+#else
+static int
+dasd_ioctl_reset_profile(struct block_device *bdev, int no, long args)
+{
+	return -ENOSYS;
+}
+
+static int
+dasd_ioctl_read_profile(struct block_device *bdev, int no, long args)
+{
+	return -ENOSYS;
+}
+#endif
+
+/*
+ * Return dasd information. Used for BIODASDINFO and BIODASDINFO2.
+ */
+static int
+dasd_ioctl_information(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+	struct dasd_information2_t *dasd_info;
+	unsigned long flags;
+	int rc;
+	struct ccw_device *cdev;
+
+	device = bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+
+	if (!device->discipline->fill_info)
+		return -EINVAL;
+
+	dasd_info = kmalloc(sizeof(struct dasd_information2_t), GFP_KERNEL);
+	if (dasd_info == NULL)
+		return -ENOMEM;
+
+	rc = device->discipline->fill_info(device, dasd_info);
+	if (rc) {
+		kfree(dasd_info);
+		return rc;
+	}
+
+	cdev = device->cdev;
+
+	dasd_info->devno = _ccw_device_get_device_number(device->cdev);
+	dasd_info->schid = _ccw_device_get_subchannel_number(device->cdev);
+	dasd_info->cu_type = cdev->id.cu_type;
+	dasd_info->cu_model = cdev->id.cu_model;
+	dasd_info->dev_type = cdev->id.dev_type;
+	dasd_info->dev_model = cdev->id.dev_model;
+	dasd_info->open_count = atomic_read(&device->open_count);
+	dasd_info->status = device->state;
+	
+	/*
+	 * check if device is really formatted
+	 * LDL / CDL was returned by 'fill_info'
+	 */
+	if ((device->state < DASD_STATE_READY) ||
+	    (dasd_check_blocksize(device->bp_block)))
+		dasd_info->format = DASD_FORMAT_NONE;
+	
+	dasd_info->features |= test_bit(DASD_FLAG_RO, &device->flags) ?
+		DASD_FEATURE_READONLY : DASD_FEATURE_DEFAULT;
+
+	if (device->discipline)
+		memcpy(dasd_info->type, device->discipline->name, 4);
+	else
+		memcpy(dasd_info->type, "none", 4);
+	dasd_info->req_queue_len = 0;
+	dasd_info->chanq_len = 0;
+	if (device->request_queue->request_fn) {
+		struct list_head *l;
+#ifdef DASD_EXTENDED_PROFILING
+		{
+			struct list_head *l;
+			spin_lock_irqsave(&device->lock, flags);
+			list_for_each(l, &device->request_queue->queue_head)
+				dasd_info->req_queue_len++;
+			spin_unlock_irqrestore(&device->lock, flags);
+		}
+#endif				/* DASD_EXTENDED_PROFILING */
+		spin_lock_irqsave(get_ccwdev_lock(device->cdev), flags);
+		list_for_each(l, &device->ccw_queue)
+			dasd_info->chanq_len++;
+		spin_unlock_irqrestore(get_ccwdev_lock(device->cdev),
+				       flags);
+	}
+
+	rc = 0;
+	if (copy_to_user((long __user *) args, (long *) dasd_info,
+			 ((no == (unsigned int) BIODASDINFO2) ?
+			  sizeof (struct dasd_information2_t) :
+			  sizeof (struct dasd_information_t))))
+		rc = -EFAULT;
+	kfree(dasd_info);
+	return rc;
+}
+
+/*
+ * Set read only
+ */
+static int
+dasd_ioctl_set_ro(struct block_device *bdev, int no, long args)
+{
+	struct dasd_device *device;
+	int intval;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+	if (bdev != bdev->bd_contains)
+		// ro setting is not allowed for partitions
+		return -EINVAL;
+	if (get_user(intval, (int __user *) args))
+		return -EFAULT;
+	device =  bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+	set_disk_ro(bdev->bd_disk, intval);
+	if (intval)
+		set_bit(DASD_FLAG_RO, &device->flags);
+	else
+		clear_bit(DASD_FLAG_RO, &device->flags);
+	return 0;
+}
+
+/*
+ * Return disk geometry.
+ */
+static int
+dasd_ioctl_getgeo(struct block_device *bdev, int no, long args)
+{
+	struct hd_geometry geo = { 0, };
+	struct dasd_device *device;
+
+	device =  bdev->bd_disk->private_data;
+	if (device == NULL)
+		return -ENODEV;
+
+	if (device == NULL || device->discipline == NULL ||
+	    device->discipline->fill_geometry == NULL)
+		return -EINVAL;
+
+	geo = (struct hd_geometry) {};
+	device->discipline->fill_geometry(device, &geo);
+	geo.start = get_start_sect(bdev) >> device->s2b_shift;
+	if (copy_to_user((struct hd_geometry __user *) args, &geo,
+			 sizeof (struct hd_geometry)))
+		return -EFAULT;
+
+	return 0;
+}
+
+/*
+ * List of static ioctls.
+ */
+static struct { int no; dasd_ioctl_fn_t fn; } dasd_ioctls[] =
+{
+	{ BIODASDDISABLE, dasd_ioctl_disable },
+	{ BIODASDENABLE, dasd_ioctl_enable },
+	{ BIODASDQUIESCE, dasd_ioctl_quiesce },
+	{ BIODASDRESUME, dasd_ioctl_resume },
+	{ BIODASDFMT, dasd_ioctl_format },
+	{ BIODASDINFO, dasd_ioctl_information },
+	{ BIODASDINFO2, dasd_ioctl_information },
+	{ BIODASDPRRD, dasd_ioctl_read_profile },
+	{ BIODASDPRRST, dasd_ioctl_reset_profile },
+	{ BLKROSET, dasd_ioctl_set_ro },
+	{ DASDAPIVER, dasd_ioctl_api_version },
+	{ HDIO_GETGEO, dasd_ioctl_getgeo },
+	{ -1, NULL }
+};
+
+int
+dasd_ioctl_init(void)
+{
+	int i;
+
+	for (i = 0; dasd_ioctls[i].no != -1; i++)
+		dasd_ioctl_no_register(NULL, dasd_ioctls[i].no,
+				       dasd_ioctls[i].fn);
+	return 0;
+
+}
+
+void
+dasd_ioctl_exit(void)
+{
+	int i;
+
+	for (i = 0; dasd_ioctls[i].no != -1; i++)
+		dasd_ioctl_no_unregister(NULL, dasd_ioctls[i].no,
+					 dasd_ioctls[i].fn);
+
+}
+
+EXPORT_SYMBOL(dasd_ioctl_no_register);
+EXPORT_SYMBOL(dasd_ioctl_no_unregister);

drivers/s390/block/dasd_proc.c

@@ -0,0 +1,319 @@
+/*
+ * File...........: linux/drivers/s390/block/dasd_proc.c
+ * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
+ *		    Horst Hummel <Horst.Hummel@de.ibm.com>
+ *		    Carsten Otte <Cotte@de.ibm.com>
+ *		    Martin Schwidefsky <schwidefsky@de.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999-2002
+ *
+ * /proc interface for the dasd driver.
+ *
+ * $Revision: 1.30 $
+ */
+
+#include <linux/config.h>
+#include <linux/ctype.h>
+#include <linux/seq_file.h>
+#include <linux/vmalloc.h>
+
+#include <asm/debug.h>
+#include <asm/uaccess.h>
+
+/* This is ugly... */
+#define PRINTK_HEADER "dasd_proc:"
+
+#include "dasd_int.h"
+
+static struct proc_dir_entry *dasd_proc_root_entry = NULL;
+static struct proc_dir_entry *dasd_devices_entry = NULL;
+static struct proc_dir_entry *dasd_statistics_entry = NULL;
+
+static inline char *
+dasd_get_user_string(const char __user *user_buf, size_t user_len)
+{
+	char *buffer;
+
+	buffer = kmalloc(user_len + 1, GFP_KERNEL);
+	if (buffer == NULL)
+		return ERR_PTR(-ENOMEM);
+	if (copy_from_user(buffer, user_buf, user_len) != 0) {
+		kfree(buffer);
+		return ERR_PTR(-EFAULT);
+	}
+	/* got the string, now strip linefeed. */
+	if (buffer[user_len - 1] == '\n')
+		buffer[user_len - 1] = 0;
+	else
+		buffer[user_len] = 0;
+	return buffer;
+}
+
+static int
+dasd_devices_show(struct seq_file *m, void *v)
+{
+	struct dasd_device *device;
+	char *substr;
+
+	device = dasd_device_from_devindex((unsigned long) v - 1);
+	if (IS_ERR(device))
+		return 0;
+	/* Print device number. */
+	seq_printf(m, "%s", device->cdev->dev.bus_id);
+	/* Print discipline string. */
+	if (device != NULL && device->discipline != NULL)
+		seq_printf(m, "(%s)", device->discipline->name);
+	else
+		seq_printf(m, "(none)");
+	/* Print kdev. */
+	if (device->gdp)
+		seq_printf(m, " at (%3d:%6d)",
+			   device->gdp->major, device->gdp->first_minor);
+	else
+		seq_printf(m, "  at (???:??????)");
+	/* Print device name. */
+	if (device->gdp)
+		seq_printf(m, " is %-8s", device->gdp->disk_name);
+	else
+		seq_printf(m, " is ????????");
+	/* Print devices features. */
+	substr = test_bit(DASD_FLAG_RO, &device->flags) ? "(ro)" : " ";
+	seq_printf(m, "%4s: ", substr);
+	/* Print device status information. */
+	switch ((device != NULL) ? device->state : -1) {
+	case -1:
+		seq_printf(m, "unknown");
+		break;
+	case DASD_STATE_NEW:
+		seq_printf(m, "new");
+		break;
+	case DASD_STATE_KNOWN:
+		seq_printf(m, "detected");
+		break;
+	case DASD_STATE_BASIC:
+		seq_printf(m, "basic");
+		break;
+	case DASD_STATE_READY:
+	case DASD_STATE_ONLINE:
+		seq_printf(m, "active ");
+		if (dasd_check_blocksize(device->bp_block))
+			seq_printf(m, "n/f	 ");
+		else
+			seq_printf(m,
+				   "at blocksize: %d, %ld blocks, %ld MB",
+				   device->bp_block, device->blocks,
+				   ((device->bp_block >> 9) *
+				    device->blocks) >> 11);
+		break;
+	default:
+		seq_printf(m, "no stat");
+		break;
+	}
+	dasd_put_device(device);
+	if (dasd_probeonly)
+		seq_printf(m, "(probeonly)");
+	seq_printf(m, "\n");
+	return 0;
+}
+
+static void *dasd_devices_start(struct seq_file *m, loff_t *pos)
+{
+	if (*pos >= dasd_max_devindex)
+		return NULL;
+	return (void *)((unsigned long) *pos + 1);
+}
+
+static void *dasd_devices_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	++*pos;
+	return dasd_devices_start(m, pos);
+}
+
+static void dasd_devices_stop(struct seq_file *m, void *v)
+{
+}
+
+static struct seq_operations dasd_devices_seq_ops = {
+	.start		= dasd_devices_start,
+	.next		= dasd_devices_next,
+	.stop		= dasd_devices_stop,
+	.show		= dasd_devices_show,
+};
+
+static int dasd_devices_open(struct inode *inode, struct file *file)
+{
+	return seq_open(file, &dasd_devices_seq_ops);
+}
+
+static struct file_operations dasd_devices_file_ops = {
+	.open		= dasd_devices_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
+
+static inline int
+dasd_calc_metrics(char *page, char **start, off_t off,
+		  int count, int *eof, int len)
+{
+	len = (len > off) ? len - off : 0;
+	if (len > count)
+		len = count;
+	if (len < count)
+		*eof = 1;
+	*start = page + off;
+	return len;
+}
+
+static inline char *
+dasd_statistics_array(char *str, int *array, int shift)
+{
+	int i;
+
+	for (i = 0; i < 32; i++) {
+		str += sprintf(str, "%7d ", array[i] >> shift);
+		if (i == 15)
+			str += sprintf(str, "\n");
+	}
+	str += sprintf(str,"\n");
+	return str;
+}
+
+static int
+dasd_statistics_read(char *page, char **start, off_t off,
+		     int count, int *eof, void *data)
+{
+	unsigned long len;
+#ifdef CONFIG_DASD_PROFILE
+	struct dasd_profile_info_t *prof;
+	char *str;
+	int shift;
+
+	/* check for active profiling */
+	if (dasd_profile_level == DASD_PROFILE_OFF) {
+		len = sprintf(page, "Statistics are off - they might be "
+				    "switched on using 'echo set on > "
+				    "/proc/dasd/statistics'\n");
+		return dasd_calc_metrics(page, start, off, count, eof, len);
+	}
+
+	prof = &dasd_global_profile;
+	/* prevent couter 'overflow' on output */
+	for (shift = 0; (prof->dasd_io_reqs >> shift) > 9999999; shift++);
+
+	str = page;
+	str += sprintf(str, "%d dasd I/O requests\n", prof->dasd_io_reqs);
+	str += sprintf(str, "with %d sectors(512B each)\n",
+		       prof->dasd_io_sects);
+	str += sprintf(str,
+		       "   __<4	   ___8	   __16	   __32	   __64	   _128	"
+		       "   _256	   _512	   __1k	   __2k	   __4k	   __8k	"
+		       "   _16k	   _32k	   _64k	   128k\n");
+	str += sprintf(str,
+		       "   _256	   _512	   __1M	   __2M	   __4M	   __8M	"
+		       "   _16M	   _32M	   _64M	   128M	   256M	   512M	"
+		       "   __1G	   __2G	   __4G " "   _>4G\n");
+
+	str += sprintf(str, "Histogram of sizes (512B secs)\n");
+	str = dasd_statistics_array(str, prof->dasd_io_secs, shift);
+	str += sprintf(str, "Histogram of I/O times (microseconds)\n");
+	str = dasd_statistics_array(str, prof->dasd_io_times, shift);
+	str += sprintf(str, "Histogram of I/O times per sector\n");
+	str = dasd_statistics_array(str, prof->dasd_io_timps, shift);
+	str += sprintf(str, "Histogram of I/O time till ssch\n");
+	str = dasd_statistics_array(str, prof->dasd_io_time1, shift);
+	str += sprintf(str, "Histogram of I/O time between ssch and irq\n");
+	str = dasd_statistics_array(str, prof->dasd_io_time2, shift);
+	str += sprintf(str, "Histogram of I/O time between ssch "
+			    "and irq per sector\n");
+	str = dasd_statistics_array(str, prof->dasd_io_time2ps, shift);
+	str += sprintf(str, "Histogram of I/O time between irq and end\n");
+	str = dasd_statistics_array(str, prof->dasd_io_time3, shift);
+	str += sprintf(str, "# of req in chanq at enqueuing (1..32) \n");
+	str = dasd_statistics_array(str, prof->dasd_io_nr_req, shift);
+	len = str - page;
+#else
+	len = sprintf(page, "Statistics are not activated in this kernel\n");
+#endif
+	return dasd_calc_metrics(page, start, off, count, eof, len);
+}
+
+static int
+dasd_statistics_write(struct file *file, const char __user *user_buf,
+		      unsigned long user_len, void *data)
+{
+#ifdef CONFIG_DASD_PROFILE
+	char *buffer, *str;
+
+	if (user_len > 65536)
+		user_len = 65536;
+	buffer = dasd_get_user_string(user_buf, user_len);
+	if (IS_ERR(buffer))
+		return PTR_ERR(buffer);
+	MESSAGE_LOG(KERN_INFO, "/proc/dasd/statictics: '%s'", buffer);
+
+	/* check for valid verbs */
+	for (str = buffer; isspace(*str); str++);
+	if (strncmp(str, "set", 3) == 0 && isspace(str[3])) {
+		/* 'set xxx' was given */
+		for (str = str + 4; isspace(*str); str++);
+		if (strcmp(str, "on") == 0) {
+			/* switch on statistics profiling */
+			dasd_profile_level = DASD_PROFILE_ON;
+			MESSAGE(KERN_INFO, "%s", "Statistics switched on");
+		} else if (strcmp(str, "off") == 0) {
+			/* switch off and reset statistics profiling */
+			memset(&dasd_global_profile,
+			       0, sizeof (struct dasd_profile_info_t));
+			dasd_profile_level = DASD_PROFILE_OFF;
+			MESSAGE(KERN_INFO, "%s", "Statistics switched off");
+		} else
+			goto out_error;
+	} else if (strncmp(str, "reset", 5) == 0) {
+		/* reset the statistics */
+		memset(&dasd_global_profile, 0,
+		       sizeof (struct dasd_profile_info_t));
+		MESSAGE(KERN_INFO, "%s", "Statistics reset");
+	} else
+		goto out_error;
+	kfree(buffer);
+	return user_len;
+out_error:
+	MESSAGE(KERN_WARNING, "%s",
+		"/proc/dasd/statistics: only 'set on', 'set off' "
+		"and 'reset' are supported verbs");
+	kfree(buffer);
+	return -EINVAL;
+#else
+	MESSAGE(KERN_WARNING, "%s",
+		"/proc/dasd/statistics: is not activated in this kernel");
+	return user_len;
+#endif				/* CONFIG_DASD_PROFILE */
+}
+
+int
+dasd_proc_init(void)
+{
+	dasd_proc_root_entry = proc_mkdir("dasd", &proc_root);
+	dasd_proc_root_entry->owner = THIS_MODULE;
+	dasd_devices_entry = create_proc_entry("devices",
+					       S_IFREG | S_IRUGO | S_IWUSR,
+					       dasd_proc_root_entry);
+	dasd_devices_entry->proc_fops = &dasd_devices_file_ops;
+	dasd_devices_entry->owner = THIS_MODULE;
+	dasd_statistics_entry = create_proc_entry("statistics",
+						  S_IFREG | S_IRUGO | S_IWUSR,
+						  dasd_proc_root_entry);
+	dasd_statistics_entry->read_proc = dasd_statistics_read;
+	dasd_statistics_entry->write_proc = dasd_statistics_write;
+	dasd_statistics_entry->owner = THIS_MODULE;
+	return 0;
+}
+
+void
+dasd_proc_exit(void)
+{
+	remove_proc_entry("devices", dasd_proc_root_entry);
+	remove_proc_entry("statistics", dasd_proc_root_entry);
+	remove_proc_entry("dasd", &proc_root);
+}

drivers/s390/block/dcssblk.c

@@ -0,0 +1,775 @@
+/*
+ * dcssblk.c -- the S/390 block driver for dcss memory
+ *
+ * Authors: Carsten Otte, Stefan Weinhuber, Gerald Schaefer
+ */
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/ctype.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <asm/extmem.h>
+#include <asm/io.h>
+#include <linux/completion.h>
+#include <linux/interrupt.h>
+#include <asm/ccwdev.h> 	// for s390_root_dev_(un)register()
+
+//#define DCSSBLK_DEBUG		/* Debug messages on/off */
+#define DCSSBLK_NAME "dcssblk"
+#define DCSSBLK_MINORS_PER_DISK 1
+#define DCSSBLK_PARM_LEN 400
+
+#ifdef DCSSBLK_DEBUG
+#define PRINT_DEBUG(x...) printk(KERN_DEBUG DCSSBLK_NAME " debug: " x)
+#else
+#define PRINT_DEBUG(x...) do {} while (0)
+#endif
+#define PRINT_INFO(x...)  printk(KERN_INFO DCSSBLK_NAME " info: " x)
+#define PRINT_WARN(x...)  printk(KERN_WARNING DCSSBLK_NAME " warning: " x)
+#define PRINT_ERR(x...)	  printk(KERN_ERR DCSSBLK_NAME " error: " x)
+
+
+static int dcssblk_open(struct inode *inode, struct file *filp);
+static int dcssblk_release(struct inode *inode, struct file *filp);
+static int dcssblk_make_request(struct request_queue *q, struct bio *bio);
+
+static char dcssblk_segments[DCSSBLK_PARM_LEN] = "\0";
+
+static int dcssblk_major;
+static struct block_device_operations dcssblk_devops = {
+	.owner   = THIS_MODULE,
+	.open    = dcssblk_open,
+	.release = dcssblk_release,
+};
+
+static ssize_t dcssblk_add_store(struct device * dev, const char * buf,
+				  size_t count);
+static ssize_t dcssblk_remove_store(struct device * dev, const char * buf,
+				  size_t count);
+static ssize_t dcssblk_save_store(struct device * dev, const char * buf,
+				  size_t count);
+static ssize_t dcssblk_save_show(struct device *dev, char *buf);
+static ssize_t dcssblk_shared_store(struct device * dev, const char * buf,
+				  size_t count);
+static ssize_t dcssblk_shared_show(struct device *dev, char *buf);
+
+static DEVICE_ATTR(add, S_IWUSR, NULL, dcssblk_add_store);
+static DEVICE_ATTR(remove, S_IWUSR, NULL, dcssblk_remove_store);
+static DEVICE_ATTR(save, S_IWUSR | S_IRUGO, dcssblk_save_show,
+		   dcssblk_save_store);
+static DEVICE_ATTR(shared, S_IWUSR | S_IRUGO, dcssblk_shared_show,
+		   dcssblk_shared_store);
+
+static struct device *dcssblk_root_dev;
+
+struct dcssblk_dev_info {
+	struct list_head lh;
+	struct device dev;
+	char segment_name[BUS_ID_SIZE];
+	atomic_t use_count;
+	struct gendisk *gd;
+	unsigned long start;
+	unsigned long end;
+	int segment_type;
+	unsigned char save_pending;
+	unsigned char is_shared;
+	struct request_queue *dcssblk_queue;
+};
+
+static struct list_head dcssblk_devices = LIST_HEAD_INIT(dcssblk_devices);
+static struct rw_semaphore dcssblk_devices_sem;
+
+/*
+ * release function for segment device.
+ */
+static void
+dcssblk_release_segment(struct device *dev)
+{
+	PRINT_DEBUG("segment release fn called for %s\n", dev->bus_id);
+	kfree(container_of(dev, struct dcssblk_dev_info, dev));
+	module_put(THIS_MODULE);
+}
+
+/*
+ * get a minor number. needs to be called with
+ * down_write(&dcssblk_devices_sem) and the
+ * device needs to be enqueued before the semaphore is
+ * freed.
+ */
+static inline int
+dcssblk_assign_free_minor(struct dcssblk_dev_info *dev_info)
+{
+	int minor, found;
+	struct dcssblk_dev_info *entry;
+
+	if (dev_info == NULL)
+		return -EINVAL;
+	for (minor = 0; minor < (1<<MINORBITS); minor++) {
+		found = 0;
+		// test if minor available
+		list_for_each_entry(entry, &dcssblk_devices, lh)
+			if (minor == entry->gd->first_minor)
+				found++;
+		if (!found) break; // got unused minor
+	}
+	if (found)
+		return -EBUSY;
+	dev_info->gd->first_minor = minor;
+	return 0;
+}
+
+/*
+ * get the struct dcssblk_dev_info from dcssblk_devices
+ * for the given name.
+ * down_read(&dcssblk_devices_sem) must be held.
+ */
+static struct dcssblk_dev_info *
+dcssblk_get_device_by_name(char *name)
+{
+	struct dcssblk_dev_info *entry;
+
+	list_for_each_entry(entry, &dcssblk_devices, lh) {
+		if (!strcmp(name, entry->segment_name)) {
+			return entry;
+		}
+	}
+	return NULL;
+}
+
+/*
+ * print appropriate error message for segment_load()/segment_type()
+ * return code
+ */
+static void
+dcssblk_segment_warn(int rc, char* seg_name)
+{
+	switch (rc) {
+	case -ENOENT:
+		PRINT_WARN("cannot load/query segment %s, does not exist\n",
+			   seg_name);
+		break;
+	case -ENOSYS:
+		PRINT_WARN("cannot load/query segment %s, not running on VM\n",
+			   seg_name);
+		break;
+	case -EIO:
+		PRINT_WARN("cannot load/query segment %s, hardware error\n",
+			   seg_name);
+		break;
+	case -ENOTSUPP:
+		PRINT_WARN("cannot load/query segment %s, is a multi-part "
+			   "segment\n", seg_name);
+		break;
+	case -ENOSPC:
+		PRINT_WARN("cannot load/query segment %s, overlaps with "
+			   "storage\n", seg_name);
+		break;
+	case -EBUSY:
+		PRINT_WARN("cannot load/query segment %s, overlaps with "
+			   "already loaded dcss\n", seg_name);
+		break;
+	case -EPERM:
+		PRINT_WARN("cannot load/query segment %s, already loaded in "
+			   "incompatible mode\n", seg_name);
+		break;
+	case -ENOMEM:
+		PRINT_WARN("cannot load/query segment %s, out of memory\n",
+			   seg_name);
+		break;
+	case -ERANGE:
+		PRINT_WARN("cannot load/query segment %s, exceeds kernel "
+			   "mapping range\n", seg_name);
+		break;
+	default:
+		PRINT_WARN("cannot load/query segment %s, return value %i\n",
+			   seg_name, rc);
+		break;
+	}
+}
+
+/*
+ * device attribute for switching shared/nonshared (exclusive)
+ * operation (show + store)
+ */
+static ssize_t
+dcssblk_shared_show(struct device *dev, char *buf)
+{
+	struct dcssblk_dev_info *dev_info;
+
+	dev_info = container_of(dev, struct dcssblk_dev_info, dev);
+	return sprintf(buf, dev_info->is_shared ? "1\n" : "0\n");
+}
+
+static ssize_t
+dcssblk_shared_store(struct device *dev, const char *inbuf, size_t count)
+{
+	struct dcssblk_dev_info *dev_info;
+	int rc;
+
+	if ((count > 1) && (inbuf[1] != '\n') && (inbuf[1] != '\0')) {
+		PRINT_WARN("Invalid value, must be 0 or 1\n");
+		return -EINVAL;
+	}
+	down_write(&dcssblk_devices_sem);
+	dev_info = container_of(dev, struct dcssblk_dev_info, dev);
+	if (atomic_read(&dev_info->use_count)) {
+		PRINT_ERR("share: segment %s is busy!\n",
+			  dev_info->segment_name);
+		rc = -EBUSY;
+		goto out;
+	}
+	if (inbuf[0] == '1') {
+		// reload segment in shared mode
+		rc = segment_modify_shared(dev_info->segment_name,
+					   SEGMENT_SHARED);
+		if (rc < 0) {
+			BUG_ON(rc == -EINVAL);
+			if (rc == -EIO || rc == -ENOENT)
+				goto removeseg;
+		} else {
+			dev_info->is_shared = 1;
+			switch (dev_info->segment_type) {
+				case SEG_TYPE_SR:
+				case SEG_TYPE_ER:
+				case SEG_TYPE_SC:
+					set_disk_ro(dev_info->gd,1);
+			}
+		}
+	} else if (inbuf[0] == '0') {
+		// reload segment in exclusive mode
+		if (dev_info->segment_type == SEG_TYPE_SC) {
+			PRINT_ERR("Segment type SC (%s) cannot be loaded in "
+				  "non-shared mode\n", dev_info->segment_name);
+			rc = -EINVAL;
+			goto out;
+		}
+		rc = segment_modify_shared(dev_info->segment_name,
+					   SEGMENT_EXCLUSIVE);
+		if (rc < 0) {
+			BUG_ON(rc == -EINVAL);
+			if (rc == -EIO || rc == -ENOENT)
+				goto removeseg;
+		} else {
+			dev_info->is_shared = 0;
+			set_disk_ro(dev_info->gd, 0);
+		}
+	} else {
+		PRINT_WARN("Invalid value, must be 0 or 1\n");
+		rc = -EINVAL;
+		goto out;
+	}
+	rc = count;
+	goto out;
+
+removeseg:
+	PRINT_ERR("Could not reload segment %s, removing it now!\n",
+			dev_info->segment_name);
+	list_del(&dev_info->lh);
+
+	del_gendisk(dev_info->gd);
+	blk_put_queue(dev_info->dcssblk_queue);
+	dev_info->gd->queue = NULL;
+	put_disk(dev_info->gd);
+	device_unregister(dev);
+	put_device(dev);
+out:
+	up_write(&dcssblk_devices_sem);
+	return rc;
+}
+
+/*
+ * device attribute for save operation on current copy
+ * of the segment. If the segment is busy, saving will
+ * become pending until it gets released, which can be
+ * undone by storing a non-true value to this entry.
+ * (show + store)
+ */
+static ssize_t
+dcssblk_save_show(struct device *dev, char *buf)
+{
+	struct dcssblk_dev_info *dev_info;
+
+	dev_info = container_of(dev, struct dcssblk_dev_info, dev);
+	return sprintf(buf, dev_info->save_pending ? "1\n" : "0\n");
+}
+
+static ssize_t
+dcssblk_save_store(struct device *dev, const char *inbuf, size_t count)
+{
+	struct dcssblk_dev_info *dev_info;
+
+	if ((count > 1) && (inbuf[1] != '\n') && (inbuf[1] != '\0')) {
+		PRINT_WARN("Invalid value, must be 0 or 1\n");
+		return -EINVAL;
+	}
+	dev_info = container_of(dev, struct dcssblk_dev_info, dev);
+
+	down_write(&dcssblk_devices_sem);
+	if (inbuf[0] == '1') {
+		if (atomic_read(&dev_info->use_count) == 0) {
+			// device is idle => we save immediately
+			PRINT_INFO("Saving segment %s\n",
+				   dev_info->segment_name);
+			segment_save(dev_info->segment_name);
+		}  else {
+			// device is busy => we save it when it becomes
+			// idle in dcssblk_release
+			PRINT_INFO("Segment %s is currently busy, it will "
+				   "be saved when it becomes idle...\n",
+				   dev_info->segment_name);
+			dev_info->save_pending = 1;
+		}
+	} else if (inbuf[0] == '0') {
+		if (dev_info->save_pending) {
+			// device is busy & the user wants to undo his save
+			// request
+			dev_info->save_pending = 0;
+			PRINT_INFO("Pending save for segment %s deactivated\n",
+					dev_info->segment_name);
+		}
+	} else {
+		up_write(&dcssblk_devices_sem);
+		PRINT_WARN("Invalid value, must be 0 or 1\n");
+		return -EINVAL;
+	}
+	up_write(&dcssblk_devices_sem);
+	return count;
+}
+
+/*
+ * device attribute for adding devices
+ */
+static ssize_t
+dcssblk_add_store(struct device *dev, const char *buf, size_t count)
+{
+	int rc, i;
+	struct dcssblk_dev_info *dev_info;
+	char *local_buf;
+	unsigned long seg_byte_size;
+
+	dev_info = NULL;
+	if (dev != dcssblk_root_dev) {
+		rc = -EINVAL;
+		goto out_nobuf;
+	}
+	local_buf = kmalloc(count + 1, GFP_KERNEL);
+	if (local_buf == NULL) {
+		rc = -ENOMEM;
+		goto out_nobuf;
+	}
+	/*
+	 * parse input
+	 */
+	for (i = 0; ((buf[i] != '\0') && (buf[i] != '\n') && i < count); i++) {
+		local_buf[i] = toupper(buf[i]);
+	}
+	local_buf[i] = '\0';
+	if ((i == 0) || (i > 8)) {
+		rc = -ENAMETOOLONG;
+		goto out;
+	}
+	/*
+	 * already loaded?
+	 */
+	down_read(&dcssblk_devices_sem);
+	dev_info = dcssblk_get_device_by_name(local_buf);
+	up_read(&dcssblk_devices_sem);
+	if (dev_info != NULL) {
+		PRINT_WARN("Segment %s already loaded!\n", local_buf);
+		rc = -EEXIST;
+		goto out;
+	}
+	/*
+	 * get a struct dcssblk_dev_info
+	 */
+	dev_info = kmalloc(sizeof(struct dcssblk_dev_info), GFP_KERNEL);
+	if (dev_info == NULL) {
+		rc = -ENOMEM;
+		goto out;
+	}
+	memset(dev_info, 0, sizeof(struct dcssblk_dev_info));
+
+	strcpy(dev_info->segment_name, local_buf);
+	strlcpy(dev_info->dev.bus_id, local_buf, BUS_ID_SIZE);
+	dev_info->dev.release = dcssblk_release_segment;
+	INIT_LIST_HEAD(&dev_info->lh);
+
+	dev_info->gd = alloc_disk(DCSSBLK_MINORS_PER_DISK);
+	if (dev_info->gd == NULL) {
+		rc = -ENOMEM;
+		goto free_dev_info;
+	}
+	dev_info->gd->major = dcssblk_major;
+	dev_info->gd->fops = &dcssblk_devops;
+	dev_info->dcssblk_queue = blk_alloc_queue(GFP_KERNEL);
+	dev_info->gd->queue = dev_info->dcssblk_queue;
+	dev_info->gd->private_data = dev_info;
+	dev_info->gd->driverfs_dev = &dev_info->dev;
+	/*
+	 * load the segment
+	 */
+	rc = segment_load(local_buf, SEGMENT_SHARED,
+				&dev_info->start, &dev_info->end);
+	if (rc < 0) {
+		dcssblk_segment_warn(rc, dev_info->segment_name);
+		goto dealloc_gendisk;
+	}
+	seg_byte_size = (dev_info->end - dev_info->start + 1);
+	set_capacity(dev_info->gd, seg_byte_size >> 9); // size in sectors
+	PRINT_INFO("Loaded segment %s, size = %lu Byte, "
+		   "capacity = %lu (512 Byte) sectors\n", local_buf,
+		   seg_byte_size, seg_byte_size >> 9);
+
+	dev_info->segment_type = rc;
+	dev_info->save_pending = 0;
+	dev_info->is_shared = 1;
+	dev_info->dev.parent = dcssblk_root_dev;
+
+	/*
+	 * get minor, add to list
+	 */
+	down_write(&dcssblk_devices_sem);
+	rc = dcssblk_assign_free_minor(dev_info);
+	if (rc) {
+		up_write(&dcssblk_devices_sem);
+		PRINT_ERR("No free minor number available! "
+			  "Unloading segment...\n");
+		goto unload_seg;
+	}
+	sprintf(dev_info->gd->disk_name, "dcssblk%d",
+		dev_info->gd->first_minor);
+	list_add_tail(&dev_info->lh, &dcssblk_devices);
+
+	if (!try_module_get(THIS_MODULE)) {
+		rc = -ENODEV;
+		goto list_del;
+	}
+	/*
+	 * register the device
+	 */
+	rc = device_register(&dev_info->dev);
+	if (rc) {
+		PRINT_ERR("Segment %s could not be registered RC=%d\n",
+				local_buf, rc);
+		module_put(THIS_MODULE);
+		goto list_del;
+	}
+	get_device(&dev_info->dev);
+	rc = device_create_file(&dev_info->dev, &dev_attr_shared);
+	if (rc)
+		goto unregister_dev;
+	rc = device_create_file(&dev_info->dev, &dev_attr_save);
+	if (rc)
+		goto unregister_dev;
+
+	add_disk(dev_info->gd);
+
+	blk_queue_make_request(dev_info->dcssblk_queue, dcssblk_make_request);
+	blk_queue_hardsect_size(dev_info->dcssblk_queue, 4096);
+
+	switch (dev_info->segment_type) {
+		case SEG_TYPE_SR:
+		case SEG_TYPE_ER:
+		case SEG_TYPE_SC:
+			set_disk_ro(dev_info->gd,1);
+			break;
+		default:
+			set_disk_ro(dev_info->gd,0);
+			break;
+	}
+	PRINT_DEBUG("Segment %s loaded successfully\n", local_buf);
+	up_write(&dcssblk_devices_sem);
+	rc = count;
+	goto out;
+
+unregister_dev:
+	PRINT_ERR("device_create_file() failed!\n");
+	list_del(&dev_info->lh);
+	blk_put_queue(dev_info->dcssblk_queue);
+	dev_info->gd->queue = NULL;
+	put_disk(dev_info->gd);
+	device_unregister(&dev_info->dev);
+	segment_unload(dev_info->segment_name);
+	put_device(&dev_info->dev);
+	up_write(&dcssblk_devices_sem);
+	goto out;
+list_del:
+	list_del(&dev_info->lh);
+	up_write(&dcssblk_devices_sem);
+unload_seg:
+	segment_unload(local_buf);
+dealloc_gendisk:
+	blk_put_queue(dev_info->dcssblk_queue);
+	dev_info->gd->queue = NULL;
+	put_disk(dev_info->gd);
+free_dev_info:
+	kfree(dev_info);
+out:
+	kfree(local_buf);
+out_nobuf:
+	return rc;
+}
+
+/*
+ * device attribute for removing devices
+ */
+static ssize_t
+dcssblk_remove_store(struct device *dev, const char *buf, size_t count)
+{
+	struct dcssblk_dev_info *dev_info;
+	int rc, i;
+	char *local_buf;
+
+	if (dev != dcssblk_root_dev) {
+		return -EINVAL;
+	}
+	local_buf = kmalloc(count + 1, GFP_KERNEL);
+	if (local_buf == NULL) {
+		return -ENOMEM;
+	}
+	/*
+	 * parse input
+	 */
+	for (i = 0; ((*(buf+i)!='\0') && (*(buf+i)!='\n') && i < count); i++) {
+		local_buf[i] = toupper(buf[i]);
+	}
+	local_buf[i] = '\0';
+	if ((i == 0) || (i > 8)) {
+		rc = -ENAMETOOLONG;
+		goto out_buf;
+	}
+
+	down_write(&dcssblk_devices_sem);
+	dev_info = dcssblk_get_device_by_name(local_buf);
+	if (dev_info == NULL) {
+		up_write(&dcssblk_devices_sem);
+		PRINT_WARN("Segment %s is not loaded!\n", local_buf);
+		rc = -ENODEV;
+		goto out_buf;
+	}
+	if (atomic_read(&dev_info->use_count) != 0) {
+		up_write(&dcssblk_devices_sem);
+		PRINT_WARN("Segment %s is in use!\n", local_buf);
+		rc = -EBUSY;
+		goto out_buf;
+	}
+	list_del(&dev_info->lh);
+
+	del_gendisk(dev_info->gd);
+	blk_put_queue(dev_info->dcssblk_queue);
+	dev_info->gd->queue = NULL;
+	put_disk(dev_info->gd);
+	device_unregister(&dev_info->dev);
+	segment_unload(dev_info->segment_name);
+	PRINT_DEBUG("Segment %s unloaded successfully\n",
+			dev_info->segment_name);
+	put_device(&dev_info->dev);
+	up_write(&dcssblk_devices_sem);
+
+	rc = count;
+out_buf:
+	kfree(local_buf);
+	return rc;
+}
+
+static int
+dcssblk_open(struct inode *inode, struct file *filp)
+{
+	struct dcssblk_dev_info *dev_info;
+	int rc;
+
+	dev_info = inode->i_bdev->bd_disk->private_data;
+	if (NULL == dev_info) {
+		rc = -ENODEV;
+		goto out;
+	}
+	atomic_inc(&dev_info->use_count);
+	inode->i_bdev->bd_block_size = 4096;
+	rc = 0;
+out:
+	return rc;
+}
+
+static int
+dcssblk_release(struct inode *inode, struct file *filp)
+{
+	struct dcssblk_dev_info *dev_info;
+	int rc;
+
+	dev_info = inode->i_bdev->bd_disk->private_data;
+	if (NULL == dev_info) {
+		rc = -ENODEV;
+		goto out;
+	}
+	down_write(&dcssblk_devices_sem);
+	if (atomic_dec_and_test(&dev_info->use_count)
+	    && (dev_info->save_pending)) {
+		PRINT_INFO("Segment %s became idle and is being saved now\n",
+			    dev_info->segment_name);
+		segment_save(dev_info->segment_name);
+		dev_info->save_pending = 0;
+	}
+	up_write(&dcssblk_devices_sem);
+	rc = 0;
+out:
+	return rc;
+}
+
+static int
+dcssblk_make_request(request_queue_t *q, struct bio *bio)
+{
+	struct dcssblk_dev_info *dev_info;
+	struct bio_vec *bvec;
+	unsigned long index;
+	unsigned long page_addr;
+	unsigned long source_addr;
+	unsigned long bytes_done;
+	int i;
+
+	bytes_done = 0;
+	dev_info = bio->bi_bdev->bd_disk->private_data;
+	if (dev_info == NULL)
+		goto fail;
+	if ((bio->bi_sector & 7) != 0 || (bio->bi_size & 4095) != 0)
+		/* Request is not page-aligned. */
+		goto fail;
+	if (((bio->bi_size >> 9) + bio->bi_sector)
+			> get_capacity(bio->bi_bdev->bd_disk)) {
+		/* Request beyond end of DCSS segment. */
+		goto fail;
+	}
+	index = (bio->bi_sector >> 3);
+	bio_for_each_segment(bvec, bio, i) {
+		page_addr = (unsigned long)
+			page_address(bvec->bv_page) + bvec->bv_offset;
+		source_addr = dev_info->start + (index<<12) + bytes_done;
+		if (unlikely(page_addr & 4095) != 0 || (bvec->bv_len & 4095) != 0)
+			// More paranoia.
+			goto fail;
+		if (bio_data_dir(bio) == READ) {
+			memcpy((void*)page_addr, (void*)source_addr,
+				bvec->bv_len);
+		} else {
+			memcpy((void*)source_addr, (void*)page_addr,
+				bvec->bv_len);
+		}
+		bytes_done += bvec->bv_len;
+	}
+	bio_endio(bio, bytes_done, 0);
+	return 0;
+fail:
+	bio_io_error(bio, bytes_done);
+	return 0;
+}
+
+static void
+dcssblk_check_params(void)
+{
+	int rc, i, j, k;
+	char buf[9];
+	struct dcssblk_dev_info *dev_info;
+
+	for (i = 0; (i < DCSSBLK_PARM_LEN) && (dcssblk_segments[i] != '\0');
+	     i++) {
+		for (j = i; (dcssblk_segments[j] != ',')  &&
+			    (dcssblk_segments[j] != '\0') &&
+			    (dcssblk_segments[j] != '(')  &&
+			    (j - i) < 8; j++)
+		{
+			buf[j-i] = dcssblk_segments[j];
+		}
+		buf[j-i] = '\0';
+		rc = dcssblk_add_store(dcssblk_root_dev, buf, j-i);
+		if ((rc >= 0) && (dcssblk_segments[j] == '(')) {
+			for (k = 0; buf[k] != '\0'; k++)
+				buf[k] = toupper(buf[k]);
+			if (!strncmp(&dcssblk_segments[j], "(local)", 7)) {
+				down_read(&dcssblk_devices_sem);
+				dev_info = dcssblk_get_device_by_name(buf);
+				up_read(&dcssblk_devices_sem);
+				if (dev_info)
+					dcssblk_shared_store(&dev_info->dev,
+							     "0\n", 2);
+			}
+		}
+		while ((dcssblk_segments[j] != ',') &&
+		       (dcssblk_segments[j] != '\0'))
+		{
+			j++;
+		}
+		if (dcssblk_segments[j] == '\0')
+			break;
+		i = j;
+	}
+}
+
+/*
+ * The init/exit functions.
+ */
+static void __exit
+dcssblk_exit(void)
+{
+	int rc;
+
+	PRINT_DEBUG("DCSSBLOCK EXIT...\n");
+	s390_root_dev_unregister(dcssblk_root_dev);
+	rc = unregister_blkdev(dcssblk_major, DCSSBLK_NAME);
+	if (rc) {
+		PRINT_ERR("unregister_blkdev() failed!\n");
+	}
+	PRINT_DEBUG("...finished!\n");
+}
+
+static int __init
+dcssblk_init(void)
+{
+	int rc;
+
+	PRINT_DEBUG("DCSSBLOCK INIT...\n");
+	dcssblk_root_dev = s390_root_dev_register("dcssblk");
+	if (IS_ERR(dcssblk_root_dev)) {
+		PRINT_ERR("device_register() failed!\n");
+		return PTR_ERR(dcssblk_root_dev);
+	}
+	rc = device_create_file(dcssblk_root_dev, &dev_attr_add);
+	if (rc) {
+		PRINT_ERR("device_create_file(add) failed!\n");
+		s390_root_dev_unregister(dcssblk_root_dev);
+		return rc;
+	}
+	rc = device_create_file(dcssblk_root_dev, &dev_attr_remove);
+	if (rc) {
+		PRINT_ERR("device_create_file(remove) failed!\n");
+		s390_root_dev_unregister(dcssblk_root_dev);
+		return rc;
+	}
+	rc = register_blkdev(0, DCSSBLK_NAME);
+	if (rc < 0) {
+		PRINT_ERR("Can't get dynamic major!\n");
+		s390_root_dev_unregister(dcssblk_root_dev);
+		return rc;
+	}
+	dcssblk_major = rc;
+	init_rwsem(&dcssblk_devices_sem);
+
+	dcssblk_check_params();
+
+	PRINT_DEBUG("...finished!\n");
+	return 0;
+}
+
+module_init(dcssblk_init);
+module_exit(dcssblk_exit);
+
+module_param_string(segments, dcssblk_segments, DCSSBLK_PARM_LEN, 0444);
+MODULE_PARM_DESC(segments, "Name of DCSS segment(s) to be loaded, "
+		 "comma-separated list, each name max. 8 chars.\n"
+		 "Adding \"(local)\" to segment name equals echoing 0 to "
+		 "/sys/devices/dcssblk/<segment name>/shared after loading "
+		 "the segment - \n"
+		 "e.g. segments=\"mydcss1,mydcss2,mydcss3(local)\"");
+
+MODULE_LICENSE("GPL");

drivers/s390/block/xpram.c

@@ -0,0 +1,539 @@
+/*
+ * Xpram.c -- the S/390 expanded memory RAM-disk
+ *           
+ * significant parts of this code are based on
+ * the sbull device driver presented in
+ * A. Rubini: Linux Device Drivers
+ *
+ * Author of XPRAM specific coding: Reinhard Buendgen
+ *                                  buendgen@de.ibm.com
+ * Rewrite for 2.5: Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *
+ * External interfaces:
+ *   Interfaces to linux kernel
+ *        xpram_setup: read kernel parameters
+ *   Device specific file operations
+ *        xpram_iotcl
+ *        xpram_open
+ *
+ * "ad-hoc" partitioning:
+ *    the expanded memory can be partitioned among several devices 
+ *    (with different minors). The partitioning set up can be
+ *    set by kernel or module parameters (int devs & int sizes[])
+ *
+ * Potential future improvements:
+ *   generic hard disk support to replace ad-hoc partitioning
+ */
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/ctype.h>  /* isdigit, isxdigit */
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/blkpg.h>
+#include <linux/hdreg.h>  /* HDIO_GETGEO */
+#include <linux/sysdev.h>
+#include <linux/bio.h>
+#include <linux/devfs_fs_kernel.h>
+#include <asm/uaccess.h>
+
+#define XPRAM_NAME	"xpram"
+#define XPRAM_DEVS	1	/* one partition */
+#define XPRAM_MAX_DEVS	32	/* maximal number of devices (partitions) */
+
+#define PRINT_DEBUG(x...)	printk(KERN_DEBUG XPRAM_NAME " debug:" x)
+#define PRINT_INFO(x...)	printk(KERN_INFO XPRAM_NAME " info:" x)
+#define PRINT_WARN(x...)	printk(KERN_WARNING XPRAM_NAME " warning:" x)
+#define PRINT_ERR(x...)		printk(KERN_ERR XPRAM_NAME " error:" x)
+
+
+static struct sysdev_class xpram_sysclass = {
+	set_kset_name("xpram"),
+};
+
+static struct sys_device xpram_sys_device = {
+	.id	= 0,
+	.cls	= &xpram_sysclass,
+}; 
+
+typedef struct {
+	unsigned int	size;		/* size of xpram segment in pages */
+	unsigned int	offset;		/* start page of xpram segment */
+} xpram_device_t;
+
+static xpram_device_t xpram_devices[XPRAM_MAX_DEVS];
+static unsigned int xpram_sizes[XPRAM_MAX_DEVS];
+static struct gendisk *xpram_disks[XPRAM_MAX_DEVS];
+static unsigned int xpram_pages;
+static int xpram_devs;
+
+/*
+ * Parameter parsing functions.
+ */
+static int devs = XPRAM_DEVS;
+static unsigned int sizes[XPRAM_MAX_DEVS];
+
+module_param(devs, int, 0);
+module_param_array(sizes, int, NULL, 0);
+
+MODULE_PARM_DESC(devs, "number of devices (\"partitions\"), " \
+		 "the default is " __MODULE_STRING(XPRAM_DEVS) "\n");
+MODULE_PARM_DESC(sizes, "list of device (partition) sizes " \
+		 "the defaults are 0s \n" \
+		 "All devices with size 0 equally partition the "
+		 "remaining space on the expanded strorage not "
+		 "claimed by explicit sizes\n");
+MODULE_LICENSE("GPL");
+
+#ifndef MODULE
+/*
+ * Parses the kernel parameters given in the kernel parameter line.
+ * The expected format is
+ *           <number_of_partitions>[","<partition_size>]*
+ * where
+ *           devices is a positive integer that initializes xpram_devs
+ *           each size is a non-negative integer possibly followed by a
+ *           magnitude (k,K,m,M,g,G), the list of sizes initialises
+ *           xpram_sizes
+ *
+ * Arguments
+ *           str: substring of kernel parameter line that contains xprams
+ *                kernel parameters.
+ *
+ * Result    0 on success, -EINVAL else -- only for Version > 2.3
+ *
+ * Side effects
+ *           the global variabls devs is set to the value of
+ *           <number_of_partitions> and sizes[i] is set to the i-th
+ *           partition size (if provided). A parsing error of a value
+ *           results in this value being set to -EINVAL.
+ */
+static int __init xpram_setup (char *str)
+{
+	char *cp;
+	int i;
+
+	devs = simple_strtoul(str, &cp, 10);
+	if (cp <= str || devs > XPRAM_MAX_DEVS)
+		return 0;
+	for (i = 0; (i < devs) && (*cp++ == ','); i++) {
+		sizes[i] = simple_strtoul(cp, &cp, 10);
+		if (*cp == 'g' || *cp == 'G') {
+			sizes[i] <<= 20;
+			cp++;
+		} else if (*cp == 'm' || *cp == 'M') {
+			sizes[i] <<= 10;
+			cp++;
+		} else if (*cp == 'k' || *cp == 'K')
+			cp++;
+		while (isspace(*cp)) cp++;
+	}
+	if (*cp == ',' && i >= devs)
+		PRINT_WARN("partition sizes list has too many entries.\n");
+	else if (*cp != 0)
+		PRINT_WARN("ignored '%s' at end of parameter string.\n", cp);
+	return 1;
+}
+
+__setup("xpram_parts=", xpram_setup);
+#endif
+
+/*
+ * Copy expanded memory page (4kB) into main memory                  
+ * Arguments                                                         
+ *           page_addr:    address of target page                    
+ *           xpage_index:  index of expandeded memory page           
+ * Return value                                                      
+ *           0:            if operation succeeds
+ *           -EIO:         if pgin failed
+ *           -ENXIO:       if xpram has vanished
+ */
+static int xpram_page_in (unsigned long page_addr, unsigned int xpage_index)
+{
+	int cc;
+
+	__asm__ __volatile__ (
+		"   lhi   %0,2\n"  /* return unused cc 2 if pgin traps */
+		"   .insn rre,0xb22e0000,%1,%2\n"  /* pgin %1,%2 */
+                "0: ipm   %0\n"
+		"   srl   %0,28\n"
+		"1:\n"
+#ifndef CONFIG_ARCH_S390X
+		".section __ex_table,\"a\"\n"
+		"   .align 4\n"
+		"   .long  0b,1b\n"
+		".previous"
+#else
+                ".section __ex_table,\"a\"\n"
+                "   .align 8\n"
+                "   .quad 0b,1b\n"
+                ".previous"
+#endif
+		: "=&d" (cc) 
+		: "a" (__pa(page_addr)), "a" (xpage_index) 
+		: "cc" );
+	if (cc == 3)
+		return -ENXIO;
+	if (cc == 2) {
+		PRINT_ERR("expanded storage lost!\n");
+		return -ENXIO;
+	}
+	if (cc == 1) {
+		PRINT_ERR("page in failed for page index %u.\n",
+			  xpage_index);
+		return -EIO;
+	}
+	return 0;
+}
+
+/*
+ * Copy a 4kB page of main memory to an expanded memory page          
+ * Arguments                                                          
+ *           page_addr:    address of source page                     
+ *           xpage_index:  index of expandeded memory page            
+ * Return value                                                       
+ *           0:            if operation succeeds
+ *           -EIO:         if pgout failed
+ *           -ENXIO:       if xpram has vanished
+ */
+static long xpram_page_out (unsigned long page_addr, unsigned int xpage_index)
+{
+	int cc;
+
+	__asm__ __volatile__ (
+		"   lhi   %0,2\n"  /* return unused cc 2 if pgout traps */
+		"   .insn rre,0xb22f0000,%1,%2\n"  /* pgout %1,%2 */
+                "0: ipm   %0\n"
+		"   srl   %0,28\n"
+		"1:\n"
+#ifndef CONFIG_ARCH_S390X
+		".section __ex_table,\"a\"\n"
+		"   .align 4\n"
+		"   .long  0b,1b\n"
+		".previous"
+#else
+                ".section __ex_table,\"a\"\n"
+                "   .align 8\n"
+                "   .quad 0b,1b\n"
+                ".previous"
+#endif
+		: "=&d" (cc) 
+		: "a" (__pa(page_addr)), "a" (xpage_index) 
+		: "cc" );
+	if (cc == 3)
+		return -ENXIO;
+	if (cc == 2) {
+		PRINT_ERR("expanded storage lost!\n");
+		return -ENXIO;
+	}
+	if (cc == 1) {
+		PRINT_ERR("page out failed for page index %u.\n",
+			  xpage_index);
+		return -EIO;
+	}
+	return 0;
+}
+
+/*
+ * Check if xpram is available.
+ */
+static int __init xpram_present(void)
+{
+	unsigned long mem_page;
+	int rc;
+
+	mem_page = (unsigned long) __get_free_page(GFP_KERNEL);
+	if (!mem_page)
+		return -ENOMEM;
+	rc = xpram_page_in(mem_page, 0);
+	free_page(mem_page);
+	return rc ? -ENXIO : 0;
+}
+
+/*
+ * Return index of the last available xpram page.
+ */
+static unsigned long __init xpram_highest_page_index(void)
+{
+	unsigned int page_index, add_bit;
+	unsigned long mem_page;
+
+	mem_page = (unsigned long) __get_free_page(GFP_KERNEL);
+	if (!mem_page)
+		return 0;
+
+	page_index = 0;
+	add_bit = 1ULL << (sizeof(unsigned int)*8 - 1);
+	while (add_bit > 0) {
+		if (xpram_page_in(mem_page, page_index | add_bit) == 0)
+			page_index |= add_bit;
+		add_bit >>= 1;
+	}
+
+	free_page (mem_page);
+
+	return page_index;
+}
+
+/*
+ * Block device make request function.
+ */
+static int xpram_make_request(request_queue_t *q, struct bio *bio)
+{
+	xpram_device_t *xdev = bio->bi_bdev->bd_disk->private_data;
+	struct bio_vec *bvec;
+	unsigned int index;
+	unsigned long page_addr;
+	unsigned long bytes;
+	int i;
+
+	if ((bio->bi_sector & 7) != 0 || (bio->bi_size & 4095) != 0)
+		/* Request is not page-aligned. */
+		goto fail;
+	if ((bio->bi_size >> 12) > xdev->size)
+		/* Request size is no page-aligned. */
+		goto fail;
+	if ((bio->bi_sector >> 3) > 0xffffffffU - xdev->offset)
+		goto fail;
+	index = (bio->bi_sector >> 3) + xdev->offset;
+	bio_for_each_segment(bvec, bio, i) {
+		page_addr = (unsigned long)
+			kmap(bvec->bv_page) + bvec->bv_offset;
+		bytes = bvec->bv_len;
+		if ((page_addr & 4095) != 0 || (bytes & 4095) != 0)
+			/* More paranoia. */
+			goto fail;
+		while (bytes > 0) {
+			if (bio_data_dir(bio) == READ) {
+				if (xpram_page_in(page_addr, index) != 0)
+					goto fail;
+			} else {
+				if (xpram_page_out(page_addr, index) != 0)
+					goto fail;
+			}
+			page_addr += 4096;
+			bytes -= 4096;
+			index++;
+		}
+	}
+	set_bit(BIO_UPTODATE, &bio->bi_flags);
+	bytes = bio->bi_size;
+	bio->bi_size = 0;
+	bio->bi_end_io(bio, bytes, 0);
+	return 0;
+fail:
+	bio_io_error(bio, bio->bi_size);
+	return 0;
+}
+
+static int xpram_ioctl (struct inode *inode, struct file *filp,
+		 unsigned int cmd, unsigned long arg)
+{
+	struct hd_geometry __user *geo;
+	unsigned long size;
+	if (cmd != HDIO_GETGEO)
+		return -EINVAL;
+	/*
+	 * get geometry: we have to fake one...  trim the size to a
+	 * multiple of 64 (32k): tell we have 16 sectors, 4 heads,
+	 * whatever cylinders. Tell also that data starts at sector. 4.
+	 */
+	geo = (struct hd_geometry __user *) arg;
+	size = (xpram_pages * 8) & ~0x3f;
+	put_user(size >> 6, &geo->cylinders);
+	put_user(4, &geo->heads);
+	put_user(16, &geo->sectors);
+	put_user(4, &geo->start);
+	return 0;
+}
+
+static struct block_device_operations xpram_devops =
+{
+	.owner	= THIS_MODULE,
+	.ioctl	= xpram_ioctl,
+};
+
+/*
+ * Setup xpram_sizes array.
+ */
+static int __init xpram_setup_sizes(unsigned long pages)
+{
+	unsigned long mem_needed;
+	unsigned long mem_auto;
+	int mem_auto_no;
+	int i;
+
+	/* Check number of devices. */
+	if (devs <= 0 || devs > XPRAM_MAX_DEVS) {
+		PRINT_ERR("invalid number %d of devices\n",devs);
+		return -EINVAL;
+	}
+	xpram_devs = devs;
+
+	/*
+	 * Copy sizes array to xpram_sizes and align partition
+	 * sizes to page boundary.
+	 */
+	mem_needed = 0;
+	mem_auto_no = 0;
+	for (i = 0; i < xpram_devs; i++) {
+		xpram_sizes[i] = (sizes[i] + 3) & -4UL;
+		if (xpram_sizes[i])
+			mem_needed += xpram_sizes[i];
+		else
+			mem_auto_no++;
+	}
+	
+	PRINT_INFO("  number of devices (partitions): %d \n", xpram_devs);
+	for (i = 0; i < xpram_devs; i++) {
+		if (xpram_sizes[i])
+			PRINT_INFO("  size of partition %d: %u kB\n",
+				   i, xpram_sizes[i]);
+		else
+			PRINT_INFO("  size of partition %d to be set "
+				   "automatically\n",i);
+	}
+	PRINT_DEBUG("  memory needed (for sized partitions): %lu kB\n",
+		    mem_needed);
+	PRINT_DEBUG("  partitions to be sized automatically: %d\n",
+		    mem_auto_no);
+
+	if (mem_needed > pages * 4) {
+		PRINT_ERR("Not enough expanded memory available\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * partitioning:
+	 * xpram_sizes[i] != 0; partition i has size xpram_sizes[i] kB
+	 * else:             ; all partitions with zero xpram_sizes[i]
+	 *                     partition equally the remaining space
+	 */
+	if (mem_auto_no) {
+		mem_auto = ((pages - mem_needed / 4) / mem_auto_no) * 4;
+		PRINT_INFO("  automatically determined "
+			   "partition size: %lu kB\n", mem_auto);
+		for (i = 0; i < xpram_devs; i++)
+			if (xpram_sizes[i] == 0)
+				xpram_sizes[i] = mem_auto;
+	}
+	return 0;
+}
+
+static struct request_queue *xpram_queue;
+
+static int __init xpram_setup_blkdev(void)
+{
+	unsigned long offset;
+	int i, rc = -ENOMEM;
+
+	for (i = 0; i < xpram_devs; i++) {
+		struct gendisk *disk = alloc_disk(1);
+		if (!disk)
+			goto out;
+		xpram_disks[i] = disk;
+	}
+
+	/*
+	 * Register xpram major.
+	 */
+	rc = register_blkdev(XPRAM_MAJOR, XPRAM_NAME);
+	if (rc < 0)
+		goto out;
+
+	devfs_mk_dir("slram");
+
+	/*
+	 * Assign the other needed values: make request function, sizes and
+	 * hardsect size. All the minor devices feature the same value.
+	 */
+	xpram_queue = blk_alloc_queue(GFP_KERNEL);
+	if (!xpram_queue) {
+		rc = -ENOMEM;
+		goto out_unreg;
+	}
+	blk_queue_make_request(xpram_queue, xpram_make_request);
+	blk_queue_hardsect_size(xpram_queue, 4096);
+
+	/*
+	 * Setup device structures.
+	 */
+	offset = 0;
+	for (i = 0; i < xpram_devs; i++) {
+		struct gendisk *disk = xpram_disks[i];
+
+		xpram_devices[i].size = xpram_sizes[i] / 4;
+		xpram_devices[i].offset = offset;
+		offset += xpram_devices[i].size;
+		disk->major = XPRAM_MAJOR;
+		disk->first_minor = i;
+		disk->fops = &xpram_devops;
+		disk->private_data = &xpram_devices[i];
+		disk->queue = xpram_queue;
+		sprintf(disk->disk_name, "slram%d", i);
+		sprintf(disk->devfs_name, "slram/%d", i);
+		set_capacity(disk, xpram_sizes[i] << 1);
+		add_disk(disk);
+	}
+
+	return 0;
+out_unreg:
+	devfs_remove("slram");
+	unregister_blkdev(XPRAM_MAJOR, XPRAM_NAME);
+out:
+	while (i--)
+		put_disk(xpram_disks[i]);
+	return rc;
+}
+
+/*
+ * Finally, the init/exit functions.
+ */
+static void __exit xpram_exit(void)
+{
+	int i;
+	for (i = 0; i < xpram_devs; i++) {
+		del_gendisk(xpram_disks[i]);
+		put_disk(xpram_disks[i]);
+	}
+	unregister_blkdev(XPRAM_MAJOR, XPRAM_NAME);
+	devfs_remove("slram");
+	blk_cleanup_queue(xpram_queue);
+	sysdev_unregister(&xpram_sys_device);
+	sysdev_class_unregister(&xpram_sysclass);
+}
+
+static int __init xpram_init(void)
+{
+	int rc;
+
+	/* Find out size of expanded memory. */
+	if (xpram_present() != 0) {
+		PRINT_WARN("No expanded memory available\n");
+		return -ENODEV;
+	}
+	xpram_pages = xpram_highest_page_index();
+	PRINT_INFO("  %u pages expanded memory found (%lu KB).\n",
+		   xpram_pages, (unsigned long) xpram_pages*4);
+	rc = xpram_setup_sizes(xpram_pages);
+	if (rc)
+		return rc;
+	rc = sysdev_class_register(&xpram_sysclass);
+	if (rc)
+		return rc;
+
+	rc = sysdev_register(&xpram_sys_device);
+	if (rc) {
+		sysdev_class_unregister(&xpram_sysclass);
+		return rc;
+	}
+	rc = xpram_setup_blkdev();
+	if (rc)
+		sysdev_unregister(&xpram_sys_device);
+	return rc;
+}
+
+module_init(xpram_init);
+module_exit(xpram_exit);

drivers/s390/char/Makefile

@@ -0,0 +1,28 @@
+#
+# S/390 character devices
+#
+
+obj-y += ctrlchar.o keyboard.o defkeymap.o
+
+obj-$(CONFIG_TN3270) += raw3270.o
+obj-$(CONFIG_TN3270_CONSOLE) += con3270.o
+obj-$(CONFIG_TN3270_TTY) += tty3270.o
+obj-$(CONFIG_TN3270_FS) += fs3270.o
+
+obj-$(CONFIG_TN3215) += con3215.o
+
+obj-$(CONFIG_SCLP) += sclp.o sclp_rw.o sclp_quiesce.o
+obj-$(CONFIG_SCLP_TTY) += sclp_tty.o
+obj-$(CONFIG_SCLP_CONSOLE) += sclp_con.o
+obj-$(CONFIG_SCLP_VT220_TTY) += sclp_vt220.o
+obj-$(CONFIG_SCLP_CPI) += sclp_cpi.o
+
+obj-$(CONFIG_ZVM_WATCHDOG) += vmwatchdog.o
+obj-$(CONFIG_VMLOGRDR) += vmlogrdr.o
+
+tape-$(CONFIG_S390_TAPE_BLOCK) += tape_block.o
+tape-$(CONFIG_PROC_FS) += tape_proc.o
+tape-objs := tape_core.o tape_std.o tape_char.o $(tape-y)
+obj-$(CONFIG_S390_TAPE) += tape.o tape_class.o
+obj-$(CONFIG_S390_TAPE_34XX) += tape_34xx.o
+obj-$(CONFIG_MONREADER) += monreader.o

drivers/s390/char/con3270.c

@@ -0,0 +1,638 @@
+/*
+ *  drivers/s390/char/con3270.c
+ *    IBM/3270 Driver - console view.
+ *
+ *  Author(s):
+ *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ */
+
+#include <linux/config.h>
+#include <linux/bootmem.h>
+#include <linux/console.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/types.h>
+
+#include <asm/ccwdev.h>
+#include <asm/cio.h>
+#include <asm/cpcmd.h>
+#include <asm/ebcdic.h>
+
+#include "raw3270.h"
+#include "ctrlchar.h"
+
+#define CON3270_OUTPUT_BUFFER_SIZE 1024
+#define CON3270_STRING_PAGES 4
+
+static struct raw3270_fn con3270_fn;
+
+/*
+ * Main 3270 console view data structure.
+ */
+struct con3270 {
+	struct raw3270_view view;
+	spinlock_t lock;
+	struct list_head freemem;	/* list of free memory for strings. */
+
+	/* Output stuff. */
+	struct list_head lines;		/* list of lines. */
+	struct list_head update;	/* list of lines to update. */
+	int line_nr;			/* line number for next update. */
+	int nr_lines;			/* # lines in list. */
+	int nr_up;			/* # lines up in history. */
+	unsigned long update_flags;	/* Update indication bits. */
+	struct string *cline;		/* current output line. */
+	struct string *status;		/* last line of display. */
+	struct raw3270_request *write;	/* single write request. */
+	struct timer_list timer;
+
+	/* Input stuff. */
+	struct string *input;		/* input string for read request. */
+	struct raw3270_request *read;	/* single read request. */
+	struct raw3270_request *kreset;	/* single keyboard reset request. */
+	struct tasklet_struct readlet;	/* tasklet to issue read request. */
+};
+
+static struct con3270 *condev;
+
+/* con3270->update_flags. See con3270_update for details. */
+#define CON_UPDATE_ERASE	1	/* Use EWRITEA instead of WRITE. */
+#define CON_UPDATE_LIST		2	/* Update lines in tty3270->update. */
+#define CON_UPDATE_STATUS	4	/* Update status line. */
+#define CON_UPDATE_ALL		7
+
+static void con3270_update(struct con3270 *);
+
+/*
+ * Setup timeout for a device. On timeout trigger an update.
+ */
+void
+con3270_set_timer(struct con3270 *cp, int expires)
+{
+	if (expires == 0) {
+		if (timer_pending(&cp->timer))
+			del_timer(&cp->timer);
+		return;
+	}
+	if (timer_pending(&cp->timer) &&
+	    mod_timer(&cp->timer, jiffies + expires))
+		return;
+	cp->timer.function = (void (*)(unsigned long)) con3270_update;
+	cp->timer.data = (unsigned long) cp;
+	cp->timer.expires = jiffies + expires;
+	add_timer(&cp->timer);
+}
+
+/*
+ * The status line is the last line of the screen. It shows the string
+ * "console view" in the lower left corner and "Running"/"More..."/"Holding"
+ * in the lower right corner of the screen.
+ */
+static void
+con3270_update_status(struct con3270 *cp)
+{
+	char *str;
+
+	str = (cp->nr_up != 0) ? "History" : "Running";
+	memcpy(cp->status->string + 24, str, 7);
+	codepage_convert(cp->view.ascebc, cp->status->string + 24, 7);
+	cp->update_flags |= CON_UPDATE_STATUS;
+}
+
+static void
+con3270_create_status(struct con3270 *cp)
+{
+	static const unsigned char blueprint[] =
+		{ TO_SBA, 0, 0, TO_SF,TF_LOG,TO_SA,TAT_COLOR, TAC_GREEN,
+		  'c','o','n','s','o','l','e',' ','v','i','e','w',
+		  TO_RA,0,0,0,'R','u','n','n','i','n','g',TO_SF,TF_LOG };
+
+	cp->status = alloc_string(&cp->freemem, sizeof(blueprint));
+	/* Copy blueprint to status line */
+	memcpy(cp->status->string, blueprint, sizeof(blueprint));
+	/* Set TO_RA addresses. */
+	raw3270_buffer_address(cp->view.dev, cp->status->string + 1,
+			       cp->view.cols * (cp->view.rows - 1));
+	raw3270_buffer_address(cp->view.dev, cp->status->string + 21,
+			       cp->view.cols * cp->view.rows - 8);
+	/* Convert strings to ebcdic. */
+	codepage_convert(cp->view.ascebc, cp->status->string + 8, 12);
+	codepage_convert(cp->view.ascebc, cp->status->string + 24, 7);
+}
+
+/*
+ * Set output offsets to 3270 datastream fragment of a console string.
+ */
+static void
+con3270_update_string(struct con3270 *cp, struct string *s, int nr)
+{
+	if (s->len >= cp->view.cols - 5)
+		return;
+	raw3270_buffer_address(cp->view.dev, s->string + s->len - 3,
+			       cp->view.cols * (nr + 1));
+}
+
+/*
+ * Rebuild update list to print all lines.
+ */
+static void
+con3270_rebuild_update(struct con3270 *cp)
+{
+	struct string *s, *n;
+	int nr;
+
+	/* 
+	 * Throw away update list and create a new one,
+	 * containing all lines that will fit on the screen.
+	 */
+	list_for_each_entry_safe(s, n, &cp->update, update)
+		list_del_init(&s->update);
+	nr = cp->view.rows - 2 + cp->nr_up;
+	list_for_each_entry_reverse(s, &cp->lines, list) {
+		if (nr < cp->view.rows - 1)
+			list_add(&s->update, &cp->update);
+		if (--nr < 0)
+			break;
+	}
+	cp->line_nr = 0;
+	cp->update_flags |= CON_UPDATE_LIST;
+}
+
+/*
+ * Alloc string for size bytes. Free strings from history if necessary.
+ */
+static struct string *
+con3270_alloc_string(struct con3270 *cp, size_t size)
+{
+	struct string *s, *n;
+
+	s = alloc_string(&cp->freemem, size);
+	if (s)
+		return s;
+	list_for_each_entry_safe(s, n, &cp->lines, list) {
+		list_del(&s->list);
+		if (!list_empty(&s->update))
+			list_del(&s->update);
+		cp->nr_lines--;
+		if (free_string(&cp->freemem, s) >= size)
+			break;
+	}
+	s = alloc_string(&cp->freemem, size);
+	BUG_ON(!s);
+	if (cp->nr_up != 0 && cp->nr_up + cp->view.rows > cp->nr_lines) {
+		cp->nr_up = cp->nr_lines - cp->view.rows + 1;
+		con3270_rebuild_update(cp);
+		con3270_update_status(cp);
+	}
+	return s;
+}
+
+/*
+ * Write completion callback.
+ */
+static void
+con3270_write_callback(struct raw3270_request *rq, void *data)
+{
+	raw3270_request_reset(rq);
+	xchg(&((struct con3270 *) rq->view)->write, rq);
+}
+
+/*
+ * Update console display.
+ */
+static void
+con3270_update(struct con3270 *cp)
+{
+	struct raw3270_request *wrq;
+	char wcc, prolog[6];
+	unsigned long flags;
+	unsigned long updated;
+	struct string *s, *n;
+	int rc;
+
+	wrq = xchg(&cp->write, 0);
+	if (!wrq) {
+		con3270_set_timer(cp, 1);
+		return;
+	}
+
+	spin_lock_irqsave(&cp->view.lock, flags);
+	updated = 0;
+	if (cp->update_flags & CON_UPDATE_ERASE) {
+		/* Use erase write alternate to initialize display. */
+		raw3270_request_set_cmd(wrq, TC_EWRITEA);
+		updated |= CON_UPDATE_ERASE;
+	} else
+		raw3270_request_set_cmd(wrq, TC_WRITE);
+
+	wcc = TW_NONE;
+	raw3270_request_add_data(wrq, &wcc, 1);
+
+	/*
+	 * Update status line.
+	 */
+	if (cp->update_flags & CON_UPDATE_STATUS)
+		if (raw3270_request_add_data(wrq, cp->status->string,
+					     cp->status->len) == 0)
+			updated |= CON_UPDATE_STATUS;
+
+	if (cp->update_flags & CON_UPDATE_LIST) {
+		prolog[0] = TO_SBA;
+		prolog[3] = TO_SA;
+		prolog[4] = TAT_COLOR;
+		prolog[5] = TAC_TURQ;
+		raw3270_buffer_address(cp->view.dev, prolog + 1,
+				       cp->view.cols * cp->line_nr);
+		raw3270_request_add_data(wrq, prolog, 6);
+		/* Write strings in the update list to the screen. */
+		list_for_each_entry_safe(s, n, &cp->update, update) {
+			if (s != cp->cline)
+				con3270_update_string(cp, s, cp->line_nr);
+			if (raw3270_request_add_data(wrq, s->string,
+						     s->len) != 0)
+				break;
+			list_del_init(&s->update);
+			if (s != cp->cline)
+				cp->line_nr++;
+		}
+		if (list_empty(&cp->update))
+			updated |= CON_UPDATE_LIST;
+	}
+	wrq->callback = con3270_write_callback;
+	rc = raw3270_start(&cp->view, wrq);
+	if (rc == 0) {
+		cp->update_flags &= ~updated;
+		if (cp->update_flags)
+			con3270_set_timer(cp, 1);
+	} else {
+		raw3270_request_reset(wrq);
+		xchg(&cp->write, wrq);
+	}
+	spin_unlock_irqrestore(&cp->view.lock, flags);
+}
+
+/*
+ * Read tasklet.
+ */
+static void
+con3270_read_tasklet(struct raw3270_request *rrq)
+{
+	static char kreset_data = TW_KR;
+	struct con3270 *cp;
+	unsigned long flags;
+	int nr_up, deactivate;
+
+	cp = (struct con3270 *) rrq->view;
+	spin_lock_irqsave(&cp->view.lock, flags);
+	nr_up = cp->nr_up;
+	deactivate = 0;
+	/* Check aid byte. */
+	switch (cp->input->string[0]) {
+	case 0x7d:	/* enter: jump to bottom. */
+		nr_up = 0;
+		break;
+	case 0xf3:	/* PF3: deactivate the console view. */
+		deactivate = 1;
+		break;
+	case 0x6d:	/* clear: start from scratch. */
+		con3270_rebuild_update(cp);
+		cp->update_flags = CON_UPDATE_ALL;
+		con3270_set_timer(cp, 1);
+		break;
+	case 0xf7:	/* PF7: do a page up in the console log. */
+		nr_up += cp->view.rows - 2;
+		if (nr_up + cp->view.rows - 1 > cp->nr_lines) {
+			nr_up = cp->nr_lines - cp->view.rows + 1;
+			if (nr_up < 0)
+				nr_up = 0;
+		}
+		break;
+	case 0xf8:	/* PF8: do a page down in the console log. */
+		nr_up -= cp->view.rows - 2;
+		if (nr_up < 0)
+			nr_up = 0;
+		break;
+	}
+	if (nr_up != cp->nr_up) {
+		cp->nr_up = nr_up;
+		con3270_rebuild_update(cp);
+		con3270_update_status(cp);
+		con3270_set_timer(cp, 1);
+	}
+	spin_unlock_irqrestore(&cp->view.lock, flags);
+
+	/* Start keyboard reset command. */
+	raw3270_request_reset(cp->kreset);
+	raw3270_request_set_cmd(cp->kreset, TC_WRITE);
+	raw3270_request_add_data(cp->kreset, &kreset_data, 1);
+	raw3270_start(&cp->view, cp->kreset);
+
+	if (deactivate)
+		raw3270_deactivate_view(&cp->view);
+
+	raw3270_request_reset(rrq);
+	xchg(&cp->read, rrq);
+	raw3270_put_view(&cp->view);
+}
+
+/*
+ * Read request completion callback.
+ */
+static void
+con3270_read_callback(struct raw3270_request *rq, void *data)
+{
+	raw3270_get_view(rq->view);
+	/* Schedule tasklet to pass input to tty. */
+	tasklet_schedule(&((struct con3270 *) rq->view)->readlet);
+}
+
+/*
+ * Issue a read request. Called only from interrupt function.
+ */
+static void
+con3270_issue_read(struct con3270 *cp)
+{
+	struct raw3270_request *rrq;
+	int rc;
+
+	rrq = xchg(&cp->read, 0);
+	if (!rrq)
+		/* Read already scheduled. */
+		return;
+	rrq->callback = con3270_read_callback;
+	rrq->callback_data = cp;
+	raw3270_request_set_cmd(rrq, TC_READMOD);
+	raw3270_request_set_data(rrq, cp->input->string, cp->input->len);
+	/* Issue the read modified request. */
+	rc = raw3270_start_irq(&cp->view, rrq);
+	if (rc)
+		raw3270_request_reset(rrq);
+}
+
+/*
+ * Switch to the console view.
+ */
+static int
+con3270_activate(struct raw3270_view *view)
+{
+	unsigned long flags;
+	struct con3270 *cp;
+
+	cp = (struct con3270 *) view;
+	spin_lock_irqsave(&cp->view.lock, flags);
+	cp->nr_up = 0;
+	con3270_rebuild_update(cp);
+	con3270_update_status(cp);
+	cp->update_flags = CON_UPDATE_ALL;
+	con3270_set_timer(cp, 1);
+	spin_unlock_irqrestore(&cp->view.lock, flags);
+	return 0;
+}
+
+static void
+con3270_deactivate(struct raw3270_view *view)
+{
+	unsigned long flags;
+	struct con3270 *cp;
+
+	cp = (struct con3270 *) view;
+	spin_lock_irqsave(&cp->view.lock, flags);
+	del_timer(&cp->timer);
+	spin_unlock_irqrestore(&cp->view.lock, flags);
+}
+
+static int
+con3270_irq(struct con3270 *cp, struct raw3270_request *rq, struct irb *irb)
+{
+	/* Handle ATTN. Schedule tasklet to read aid. */
+	if (irb->scsw.dstat & DEV_STAT_ATTENTION)
+		con3270_issue_read(cp);
+
+	if (rq) {
+		if (irb->scsw.dstat & DEV_STAT_UNIT_CHECK)
+			rq->rc = -EIO;
+		else
+			/* Normal end. Copy residual count. */
+			rq->rescnt = irb->scsw.count;
+	}
+	return RAW3270_IO_DONE;
+}
+
+/* Console view to a 3270 device. */
+static struct raw3270_fn con3270_fn = {
+	.activate = con3270_activate,
+	.deactivate = con3270_deactivate,
+	.intv = (void *) con3270_irq
+};
+
+static inline void
+con3270_cline_add(struct con3270 *cp)
+{
+	if (!list_empty(&cp->cline->list))
+		/* Already added. */
+		return;
+	list_add_tail(&cp->cline->list, &cp->lines);
+	cp->nr_lines++;
+	con3270_rebuild_update(cp);
+}
+
+static inline void
+con3270_cline_insert(struct con3270 *cp, unsigned char c)
+{
+	cp->cline->string[cp->cline->len++] = 
+		cp->view.ascebc[(c < ' ') ? ' ' : c];
+	if (list_empty(&cp->cline->update)) {
+		list_add_tail(&cp->cline->update, &cp->update);
+		cp->update_flags |= CON_UPDATE_LIST;
+	}
+}
+
+static inline void
+con3270_cline_end(struct con3270 *cp)
+{
+	struct string *s;
+	unsigned int size;
+
+	/* Copy cline. */
+	size = (cp->cline->len < cp->view.cols - 5) ?
+		cp->cline->len + 4 : cp->view.cols;
+	s = con3270_alloc_string(cp, size);
+	memcpy(s->string, cp->cline->string, cp->cline->len);
+	if (s->len < cp->view.cols - 5) {
+		s->string[s->len - 4] = TO_RA;
+		s->string[s->len - 1] = 0;
+	} else {
+		while (--size > cp->cline->len)
+			s->string[size] = cp->view.ascebc[' '];
+	}
+	/* Replace cline with allocated line s and reset cline. */
+	list_add(&s->list, &cp->cline->list);
+	list_del_init(&cp->cline->list);
+	if (!list_empty(&cp->cline->update)) {
+		list_add(&s->update, &cp->cline->update);
+		list_del_init(&cp->cline->update);
+	}
+	cp->cline->len = 0;
+}
+
+/*
+ * Write a string to the 3270 console
+ */
+static void
+con3270_write(struct console *co, const char *str, unsigned int count)
+{
+	struct con3270 *cp;
+	unsigned long flags;
+	unsigned char c;
+
+	cp = condev;
+	if (cp->view.dev)
+		raw3270_activate_view(&cp->view);
+	spin_lock_irqsave(&cp->view.lock, flags);
+	while (count-- > 0) {
+		c = *str++;
+		if (cp->cline->len == 0)
+			con3270_cline_add(cp);
+		if (c != '\n')
+			con3270_cline_insert(cp, c);
+		if (c == '\n' || cp->cline->len >= cp->view.cols)
+			con3270_cline_end(cp);
+	}
+	/* Setup timer to output current console buffer after 1/10 second */
+	if (cp->view.dev && !timer_pending(&cp->timer))
+		con3270_set_timer(cp, HZ/10);
+	spin_unlock_irqrestore(&cp->view.lock,flags);
+}
+
+extern struct tty_driver *tty3270_driver;
+
+static struct tty_driver *
+con3270_device(struct console *c, int *index)
+{
+	*index = c->index;
+	return tty3270_driver;
+}
+
+/*
+ * Wait for end of write request.
+ */
+static void
+con3270_wait_write(struct con3270 *cp)
+{
+	while (!cp->write) {
+		raw3270_wait_cons_dev(cp->view.dev);
+		barrier();
+	}
+}
+
+/*
+ * panic() calls console_unblank before the system enters a
+ * disabled, endless loop.
+ */
+static void
+con3270_unblank(void)
+{
+	struct con3270 *cp;
+	unsigned long flags;
+
+	cp = condev;
+	if (!cp->view.dev)
+		return;
+	spin_lock_irqsave(&cp->view.lock, flags);
+	con3270_wait_write(cp);
+	cp->nr_up = 0;
+	con3270_rebuild_update(cp);
+	con3270_update_status(cp);
+	while (cp->update_flags != 0) {
+		spin_unlock_irqrestore(&cp->view.lock, flags);
+		con3270_update(cp);
+		spin_lock_irqsave(&cp->view.lock, flags);
+		con3270_wait_write(cp);
+	}
+	spin_unlock_irqrestore(&cp->view.lock, flags);
+}
+
+static int __init 
+con3270_consetup(struct console *co, char *options)
+{
+	return 0;
+}
+
+/*
+ *  The console structure for the 3270 console
+ */
+static struct console con3270 = {
+	.name	 = "tty3270",
+	.write	 = con3270_write,
+	.device	 = con3270_device,
+	.unblank = con3270_unblank,
+	.setup	 = con3270_consetup,
+	.flags	 = CON_PRINTBUFFER,
+};
+
+/*
+ * 3270 console initialization code called from console_init().
+ * NOTE: This is called before kmalloc is available.
+ */
+static int __init
+con3270_init(void)
+{
+	struct ccw_device *cdev;
+	struct raw3270 *rp;
+	void *cbuf;
+	int i;
+
+	/* Check if 3270 is to be the console */
+	if (!CONSOLE_IS_3270)
+		return -ENODEV;
+
+	/* Set the console mode for VM */
+	if (MACHINE_IS_VM) {
+		cpcmd("TERM CONMODE 3270", 0, 0);
+		cpcmd("TERM AUTOCR OFF", 0, 0);
+	}
+
+	cdev = ccw_device_probe_console();
+	if (!cdev)
+		return -ENODEV;
+	rp = raw3270_setup_console(cdev);
+	if (IS_ERR(rp))
+		return PTR_ERR(rp);
+
+	condev = (struct con3270 *) alloc_bootmem_low(sizeof(struct con3270));
+	memset(condev, 0, sizeof(struct con3270));
+	condev->view.dev = rp;
+
+	condev->read = raw3270_request_alloc_bootmem(0);
+	condev->read->callback = con3270_read_callback;
+	condev->read->callback_data = condev;
+	condev->write = 
+		raw3270_request_alloc_bootmem(CON3270_OUTPUT_BUFFER_SIZE);
+	condev->kreset = raw3270_request_alloc_bootmem(1);
+
+	INIT_LIST_HEAD(&condev->lines);
+	INIT_LIST_HEAD(&condev->update);
+	init_timer(&condev->timer);
+	tasklet_init(&condev->readlet, 
+		     (void (*)(unsigned long)) con3270_read_tasklet,
+		     (unsigned long) condev->read);
+
+	raw3270_add_view(&condev->view, &con3270_fn, 0);
+
+	INIT_LIST_HEAD(&condev->freemem);
+	for (i = 0; i < CON3270_STRING_PAGES; i++) {
+		cbuf = (void *) alloc_bootmem_low_pages(PAGE_SIZE);
+		add_string_memory(&condev->freemem, cbuf, PAGE_SIZE);
+	}
+	condev->cline = alloc_string(&condev->freemem, condev->view.cols);
+	condev->cline->len = 0;
+	con3270_create_status(condev);
+	condev->input = alloc_string(&condev->freemem, 80);
+	register_console(&con3270);
+	return 0;
+}
+
+console_initcall(con3270_init);

drivers/s390/char/ctrlchar.c

@@ -0,0 +1,75 @@
+/*
+ *  drivers/s390/char/ctrlchar.c
+ *  Unified handling of special chars.
+ *
+ *    Copyright (C) 2001 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Fritz Elfert <felfert@millenux.com> <elfert@de.ibm.com>
+ *
+ */
+
+#include <linux/config.h>
+#include <linux/stddef.h>
+#include <asm/errno.h>
+#include <linux/sysrq.h>
+#include <linux/ctype.h>
+
+#include "ctrlchar.h"
+
+#ifdef CONFIG_MAGIC_SYSRQ
+static int ctrlchar_sysrq_key;
+
+static void
+ctrlchar_handle_sysrq(void *tty)
+{
+	handle_sysrq(ctrlchar_sysrq_key, NULL, (struct tty_struct *) tty);
+}
+
+static DECLARE_WORK(ctrlchar_work, ctrlchar_handle_sysrq, 0);
+#endif
+
+
+/**
+ * Check for special chars at start of input.
+ *
+ * @param buf Console input buffer.
+ * @param len Length of valid data in buffer.
+ * @param tty The tty struct for this console.
+ * @return CTRLCHAR_NONE, if nothing matched,
+ *         CTRLCHAR_SYSRQ, if sysrq was encountered
+ *         otherwise char to be inserted logically or'ed
+ *         with CTRLCHAR_CTRL
+ */
+unsigned int
+ctrlchar_handle(const unsigned char *buf, int len, struct tty_struct *tty)
+{
+	if ((len < 2) || (len > 3))
+		return CTRLCHAR_NONE;
+
+	/* hat is 0xb1 in codepage 037 (US etc.) and thus */
+	/* converted to 0x5e in ascii ('^') */
+	if ((buf[0] != '^') && (buf[0] != '\252'))
+		return CTRLCHAR_NONE;
+
+#ifdef CONFIG_MAGIC_SYSRQ
+	/* racy */
+	if (len == 3 && buf[1] == '-') {
+		ctrlchar_sysrq_key = buf[2];
+		ctrlchar_work.data = tty;
+		schedule_work(&ctrlchar_work);
+		return CTRLCHAR_SYSRQ;
+	}
+#endif
+
+	if (len != 2)
+		return CTRLCHAR_NONE;
+
+	switch (tolower(buf[1])) {
+	case 'c':
+		return INTR_CHAR(tty) | CTRLCHAR_CTRL;
+	case 'd':
+		return EOF_CHAR(tty)  | CTRLCHAR_CTRL;
+	case 'z':
+		return SUSP_CHAR(tty) | CTRLCHAR_CTRL;
+	}
+	return CTRLCHAR_NONE;
+}

drivers/s390/char/ctrlchar.h

@@ -0,0 +1,20 @@
+/*
+ *  drivers/s390/char/ctrlchar.c
+ *  Unified handling of special chars.
+ *
+ *    Copyright (C) 2001 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Fritz Elfert <felfert@millenux.com> <elfert@de.ibm.com>
+ *
+ */
+
+#include <linux/tty.h>
+
+extern unsigned int
+ctrlchar_handle(const unsigned char *buf, int len, struct tty_struct *tty);
+
+
+#define CTRLCHAR_NONE  (1 << 8)
+#define CTRLCHAR_CTRL  (2 << 8)
+#define CTRLCHAR_SYSRQ (3 << 8)
+
+#define CTRLCHAR_MASK (~0xffu)

drivers/s390/char/defkeymap.c

@@ -0,0 +1,156 @@
+
+/* Do not edit this file! It was automatically generated by   */
+/*    loadkeys --mktable defkeymap.map > defkeymap.c          */
+
+#include <linux/types.h>
+#include <linux/keyboard.h>
+#include <linux/kd.h>
+
+u_short plain_map[NR_KEYS] = {
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,	0xf000,
+	0xf020,	0xf000,	0xf0e2,	0xf0e4,	0xf0e0,	0xf0e1,	0xf0e3,	0xf0e5,
+	0xf0e7,	0xf0f1,	0xf0a2,	0xf02e,	0xf03c,	0xf028,	0xf02b,	0xf07c,
+	0xf026,	0xf0e9,	0xf0e2,	0xf0eb,	0xf0e8,	0xf0ed,	0xf0ee,	0xf0ef,
+	0xf0ec,	0xf0df,	0xf021,	0xf024,	0xf02a,	0xf029,	0xf03b,	0xf0ac,
+	0xf02d,	0xf02f,	0xf0c2,	0xf0c4,	0xf0c0,	0xf0c1,	0xf0c3,	0xf0c5,
+	0xf0c7,	0xf0d1,	0xf0a6,	0xf02c,	0xf025,	0xf05f,	0xf03e,	0xf03f,
+	0xf0f8,	0xf0c9,	0xf0ca,	0xf0cb,	0xf0c8,	0xf0cd,	0xf0ce,	0xf0cf,
+	0xf0cc,	0xf060,	0xf03a,	0xf023,	0xf040,	0xf027,	0xf03d,	0xf022,
+};
+
+static u_short shift_map[NR_KEYS] = {
+	0xf0d8,	0xf061,	0xf062,	0xf063,	0xf064,	0xf065,	0xf066,	0xf067,
+	0xf068,	0xf069,	0xf0ab,	0xf0bb,	0xf0f0,	0xf0fd,	0xf0fe,	0xf0b1,
+	0xf0b0,	0xf06a,	0xf06b,	0xf06c,	0xf06d,	0xf06e,	0xf06f,	0xf070,
+	0xf071,	0xf072,	0xf000,	0xf000,	0xf0e6,	0xf0b8,	0xf0c6,	0xf0a4,
+	0xf0b5,	0xf07e,	0xf073,	0xf074,	0xf075,	0xf076,	0xf077,	0xf078,
+	0xf079,	0xf07a,	0xf0a1,	0xf0bf,	0xf0d0,	0xf0dd,	0xf0de,	0xf0ae,
+	0xf402,	0xf0a3,	0xf0a5,	0xf0b7,	0xf0a9,	0xf0a7,	0xf0b6,	0xf0bc,
+	0xf0bd,	0xf0be,	0xf05b,	0xf05d,	0xf000,	0xf0a8,	0xf0b4,	0xf0d7,
+	0xf07b,	0xf041,	0xf042,	0xf043,	0xf044,	0xf045,	0xf046,	0xf047,
+	0xf048,	0xf049,	0xf000,	0xf0f4,	0xf0f6,	0xf0f2,	0xf0f3,	0xf0f5,
+	0xf07d,	0xf04a,	0xf04b,	0xf04c,	0xf04d,	0xf04e,	0xf04f,	0xf050,
+	0xf051,	0xf052,	0xf0b9,	0xf0fb,	0xf0fc,	0xf0f9,	0xf0fa,	0xf0ff,
+	0xf05c,	0xf0f7,	0xf053,	0xf054,	0xf055,	0xf056,	0xf057,	0xf058,
+	0xf059,	0xf05a,	0xf0b2,	0xf0d4,	0xf0d6,	0xf0d2,	0xf0d3,	0xf0d5,
+	0xf030,	0xf031,	0xf032,	0xf033,	0xf034,	0xf035,	0xf036,	0xf037,
+	0xf038,	0xf039,	0xf0b3,	0xf0db,	0xf0dc,	0xf0d9,	0xf0da,	0xf000,
+};
+
+static u_short ctrl_map[NR_KEYS] = {
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf11f,	0xf120,	0xf121,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf01a,	0xf003,	0xf212,	0xf004,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf109,	0xf10a,	0xf206,	0xf00a,	0xf200,	0xf200,
+};
+
+static u_short shift_ctrl_map[NR_KEYS] = {
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf10c,	0xf10d,	0xf10e,	0xf10f,	0xf110,	0xf111,	0xf112,
+	0xf113,	0xf11e,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+	0xf200,	0xf100,	0xf101,	0xf211,	0xf103,	0xf104,	0xf105,	0xf20b,
+	0xf20a,	0xf108,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,	0xf200,
+};
+
+ushort *key_maps[MAX_NR_KEYMAPS] = {
+	plain_map, shift_map, 0, 0,
+	ctrl_map, shift_ctrl_map,	0
+};
+
+unsigned int keymap_count = 4;
+
+
+/*
+ * Philosophy: most people do not define more strings, but they who do
+ * often want quite a lot of string space. So, we statically allocate
+ * the default and allocate dynamically in chunks of 512 bytes.
+ */
+
+char func_buf[] = {
+	'\033', '[', '[', 'A', 0, 
+	'\033', '[', '[', 'B', 0, 
+	'\033', '[', '[', 'C', 0, 
+	'\033', '[', '[', 'D', 0, 
+	'\033', '[', '[', 'E', 0, 
+	'\033', '[', '1', '7', '~', 0, 
+	'\033', '[', '1', '8', '~', 0, 
+	'\033', '[', '1', '9', '~', 0, 
+	'\033', '[', '2', '0', '~', 0, 
+	'\033', '[', '2', '1', '~', 0, 
+	'\033', '[', '2', '3', '~', 0, 
+	'\033', '[', '2', '4', '~', 0, 
+	'\033', '[', '2', '5', '~', 0, 
+	'\033', '[', '2', '6', '~', 0, 
+	'\033', '[', '2', '8', '~', 0, 
+	'\033', '[', '2', '9', '~', 0, 
+	'\033', '[', '3', '1', '~', 0, 
+	'\033', '[', '3', '2', '~', 0, 
+	'\033', '[', '3', '3', '~', 0, 
+	'\033', '[', '3', '4', '~', 0, 
+};
+
+
+char *funcbufptr = func_buf;
+int funcbufsize = sizeof(func_buf);
+int funcbufleft = 0;          /* space left */
+
+char *func_table[MAX_NR_FUNC] = {
+	func_buf + 0,
+	func_buf + 5,
+	func_buf + 10,
+	func_buf + 15,
+	func_buf + 20,
+	func_buf + 25,
+	func_buf + 31,
+	func_buf + 37,
+	func_buf + 43,
+	func_buf + 49,
+	func_buf + 55,
+	func_buf + 61,
+	func_buf + 67,
+	func_buf + 73,
+	func_buf + 79,
+	func_buf + 85,
+	func_buf + 91,
+	func_buf + 97,
+	func_buf + 103,
+	func_buf + 109,
+	0,
+};
+
+struct kbdiacr accent_table[MAX_DIACR] = {
+	{'^', 'c', '\003'},	{'^', 'd', '\004'},
+	{'^', 'z', '\032'},	{'^', '\012', '\000'},
+};
+
+unsigned int accent_table_size = 4;

drivers/s390/char/defkeymap.map

@@ -0,0 +1,191 @@
+# Default keymap for 3270 (ebcdic codepage 037).
+keymaps 0-1,4-5
+
+keycode   0 = nul		Oslash
+keycode   1 = nul		a
+keycode   2 = nul		b
+keycode   3 = nul		c
+keycode   4 = nul		d
+keycode   5 = nul		e
+keycode   6 = nul		f
+keycode   7 = nul		g
+keycode   8 = nul		h
+keycode   9 = nul		i
+keycode  10 = nul		guillemotleft
+keycode  11 = nul		guillemotright
+keycode  12 = nul		eth
+keycode  13 = nul		yacute
+keycode  14 = nul		thorn
+keycode  15 = nul		plusminus
+keycode  16 = nul		degree
+keycode  17 = nul		j
+keycode  18 = nul		k
+keycode  19 = nul		l
+keycode  20 = nul		m
+keycode  21 = nul		n
+keycode  22 = nul		o
+keycode  23 = nul		p
+keycode  24 = nul		q
+keycode  25 = nul		r
+keycode  26 = nul		nul
+keycode  27 = nul		nul
+keycode  28 = nul		ae
+keycode  29 = nul		cedilla
+keycode  30 = nul		AE
+keycode  31 = nul		currency
+keycode  32 = nul		mu
+keycode  33 = nul		tilde
+keycode  34 = nul		s
+keycode  35 = nul		t
+keycode  36 = nul		u
+keycode  37 = nul		v
+keycode  38 = nul		w
+keycode  39 = nul		x
+keycode  40 = nul		y
+keycode  41 = nul		z
+keycode  42 = nul		exclamdown
+keycode  43 = nul		questiondown
+keycode  44 = nul		ETH
+keycode  45 = nul		Yacute
+keycode  46 = nul		THORN
+keycode  47 = nul		registered
+keycode  48 = nul		dead_circumflex
+keycode  49 = nul		sterling
+keycode  50 = nul		yen
+keycode  51 = nul		periodcentered
+keycode  52 = nul		copyright
+keycode  53 = nul		section
+keycode  54 = nul		paragraph
+keycode  55 = nul		onequarter
+keycode  56 = nul		onehalf
+keycode  57 = nul		threequarters
+keycode  58 = nul		bracketleft
+keycode  59 = nul		bracketright
+keycode  60 = nul		nul
+keycode  61 = nul		diaeresis
+keycode  62 = nul		acute
+keycode  63 = nul		multiply
+keycode  64 = space		braceleft
+keycode  65 = nul		A
+keycode  66 = acircumflex	B
+keycode  67 = adiaeresis	C
+keycode  68 = agrave		D
+keycode  69 = aacute		E
+keycode  70 = atilde		F
+keycode  71 = aring		G
+keycode  72 = ccedilla		H
+keycode  73 = ntilde		I
+keycode  74 = cent		nul
+keycode  75 = period		ocircumflex
+keycode  76 = less		odiaeresis
+keycode  77 = parenleft		ograve
+keycode  78 = plus		oacute
+keycode  79 = bar		otilde
+keycode  80 = ampersand		braceright
+keycode  81 = eacute		J
+keycode  82 = acircumflex	K
+keycode  83 = ediaeresis	L
+keycode  84 = egrave		M
+keycode  85 = iacute		N
+keycode  86 = icircumflex	O
+keycode  87 = idiaeresis	P
+keycode  88 = igrave		Q
+keycode  89 = ssharp		R
+keycode  90 = exclam		onesuperior
+keycode  91 = dollar		ucircumflex
+keycode  92 = asterisk		udiaeresis
+keycode  93 = parenright	ugrave
+keycode  94 = semicolon		uacute
+keycode  95 = notsign		ydiaeresis
+keycode  96 = minus		backslash
+keycode  97 = slash		division
+keycode  98 = Acircumflex	S
+keycode  99 = Adiaeresis	T
+keycode 100 = Agrave		U
+keycode 101 = Aacute		V
+keycode 102 = Atilde		W
+keycode 103 = Aring		X
+keycode 104 = Ccedilla		Y
+keycode 105 = Ntilde		Z
+keycode 106 = brokenbar		twosuperior
+keycode 107 = comma		Ocircumflex
+keycode 108 = percent		Odiaeresis
+keycode 109 = underscore	Ograve
+keycode 110 = greater		Oacute
+keycode 111 = question		Otilde
+keycode 112 = oslash		zero
+keycode 113 = Eacute		one
+keycode 114 = Ecircumflex	two
+keycode 115 = Ediaeresis	three
+keycode 116 = Egrave		four
+keycode 117 = Iacute		five
+keycode 118 = Icircumflex	six
+keycode 119 = Idiaeresis	seven
+keycode 120 = Igrave		eight
+keycode 121 = grave		nine
+keycode 122 = colon		threesuperior
+keycode 123 = numbersign	Ucircumflex
+keycode 124 = at		Udiaeresis
+keycode 125 = apostrophe	Ugrave
+keycode 126 = equal		Uacute
+keycode 127 = quotedbl		nul
+
+# AID keys
+control keycode  74 = F22
+control keycode  75 = F23
+control keycode  76 = F24
+control keycode 107 = Control_z		# PA3
+control keycode 108 = Control_c		# PA1
+control keycode 109 = KeyboardSignal	# Clear
+control keycode 110 = Control_d		# PA2
+control keycode 122 = F10
+control keycode 123 = F11		# F11
+control keycode 124 = Last_Console	# F12
+control keycode 125 = Linefeed
+shift control keycode  65 = F13
+shift control keycode  66 = F14
+shift control keycode  67 = F15
+shift control keycode  68 = F16
+shift control keycode  69 = F17
+shift control keycode  70 = F18
+shift control keycode  71 = F19
+shift control keycode  72 = F20
+shift control keycode  73 = F21
+shift control keycode 113 = F1
+shift control keycode 114 = F2
+shift control keycode 115 = Incr_Console
+shift control keycode 116 = F4
+shift control keycode 117 = F5
+shift control keycode 118 = F6
+shift control keycode 119 = Scroll_Backward
+shift control keycode 120 = Scroll_Forward
+shift control keycode 121 = F9
+
+string F1 = "\033[[A"
+string F2 = "\033[[B"
+string F3 = "\033[[C"
+string F4 = "\033[[D"
+string F5 = "\033[[E"
+string F6 = "\033[17~"
+string F7 = "\033[18~"
+string F8 = "\033[19~"
+string F9 = "\033[20~"
+string F10 = "\033[21~"
+string F11 = "\033[23~"
+string F12 = "\033[24~"
+string F13 = "\033[25~"
+string F14 = "\033[26~"
+string F15 = "\033[28~"
+string F16 = "\033[29~"
+string F17 = "\033[31~"
+string F18 = "\033[32~"
+string F19 = "\033[33~"
+string F20 = "\033[34~"
+# string F21 ??
+# string F22 ??
+# string F23 ??
+# string F24 ??
+compose '^' 'c' to Control_c
+compose '^' 'd' to Control_d
+compose '^' 'z' to Control_z
+compose '^' '\012' to nul

drivers/s390/char/fs3270.c

@@ -0,0 +1,373 @@
+/*
+ *  drivers/s390/char/fs3270.c
+ *    IBM/3270 Driver - fullscreen driver.
+ *
+ *  Author(s):
+ *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *    Rewritten for 2.5/2.6 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ */
+
+#include <linux/config.h>
+#include <linux/bootmem.h>
+#include <linux/console.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/types.h>
+
+#include <asm/ccwdev.h>
+#include <asm/cio.h>
+#include <asm/cpcmd.h>
+#include <asm/ebcdic.h>
+#include <asm/idals.h>
+
+#include "raw3270.h"
+#include "ctrlchar.h"
+
+struct raw3270_fn fs3270_fn;
+
+struct fs3270 {
+	struct raw3270_view view;
+	pid_t fs_pid;			/* Pid of controlling program. */
+	int read_command;		/* ccw command to use for reads. */
+	int write_command;		/* ccw command to use for writes. */
+	int attention;			/* Got attention. */
+	struct raw3270_request *clear;	/* single clear request. */
+	wait_queue_head_t attn_wait;	/* Attention wait queue. */
+};
+
+static void
+fs3270_wake_up(struct raw3270_request *rq, void *data)
+{
+	wake_up((wait_queue_head_t *) data);
+}
+
+static int
+fs3270_do_io(struct raw3270_view *view, struct raw3270_request *rq)
+{
+	wait_queue_head_t wq;
+	int rc;
+
+	init_waitqueue_head(&wq);
+	rq->callback = fs3270_wake_up;
+	rq->callback_data = &wq;
+	rc = raw3270_start(view, rq);
+	if (rc)
+		return rc;
+	/* Started sucessfully. Now wait for completion. */
+	wait_event(wq, raw3270_request_final(rq));
+	return rq->rc;
+}
+
+static void
+fs3270_reset_callback(struct raw3270_request *rq, void *data)
+{
+	raw3270_request_reset(rq);
+}
+
+/*
+ * Switch to the fullscreen view.
+ */
+static int
+fs3270_activate(struct raw3270_view *view)
+{
+	struct fs3270 *fp;
+
+	fp = (struct fs3270 *) view;
+	raw3270_request_set_cmd(fp->clear, TC_EWRITEA);
+	fp->clear->callback = fs3270_reset_callback;
+	return raw3270_start(view, fp->clear);
+}
+
+/*
+ * Shutdown fullscreen view.
+ */
+static void
+fs3270_deactivate(struct raw3270_view *view)
+{
+	// FIXME: is this a good idea? The user program using fullscreen 3270
+	// will die just because a console message appeared. On the other
+	// hand the fullscreen device is unoperational now.
+	struct fs3270 *fp;
+
+	fp = (struct fs3270 *) view;
+	if (fp->fs_pid != 0)
+		kill_proc(fp->fs_pid, SIGHUP, 1);
+	fp->fs_pid = 0;
+}
+
+static int
+fs3270_irq(struct fs3270 *fp, struct raw3270_request *rq, struct irb *irb)
+{
+	/* Handle ATTN. Set indication and wake waiters for attention. */
+	if (irb->scsw.dstat & DEV_STAT_ATTENTION) {
+		fp->attention = 1;
+		wake_up(&fp->attn_wait);
+	}
+
+	if (rq) {
+		if (irb->scsw.dstat & DEV_STAT_UNIT_CHECK)
+			rq->rc = -EIO;
+		else
+			/* Normal end. Copy residual count. */
+			rq->rescnt = irb->scsw.count;
+	}
+	return RAW3270_IO_DONE;
+}
+
+/*
+ * Process reads from fullscreen 3270.
+ */
+static ssize_t
+fs3270_read(struct file *filp, char *data, size_t count, loff_t *off)
+{
+	struct fs3270 *fp;
+	struct raw3270_request *rq;
+	struct idal_buffer *ib;
+	int rc;
+	
+	if (count == 0 || count > 65535)
+		return -EINVAL;
+	fp = filp->private_data;
+	if (!fp)
+		return -ENODEV;
+	ib = idal_buffer_alloc(count, 0);
+	if (!ib)
+		return -ENOMEM;
+	rq = raw3270_request_alloc(0);
+	if (!IS_ERR(rq)) {
+		if (fp->read_command == 0 && fp->write_command != 0)
+			fp->read_command = 6;
+		raw3270_request_set_cmd(rq, fp->read_command ? : 2);
+		raw3270_request_set_idal(rq, ib);
+		wait_event(fp->attn_wait, fp->attention);
+		rc = fs3270_do_io(&fp->view, rq);
+		if (rc == 0 && idal_buffer_to_user(ib, data, count))
+			rc = -EFAULT;
+		raw3270_request_free(rq);
+	} else
+		rc = PTR_ERR(rq);
+	idal_buffer_free(ib);
+	return rc;
+}
+
+/*
+ * Process writes to fullscreen 3270.
+ */
+static ssize_t
+fs3270_write(struct file *filp, const char *data, size_t count, loff_t *off)
+{
+	struct fs3270 *fp;
+	struct raw3270_request *rq;
+	struct idal_buffer *ib;
+	int write_command;
+	int rc;
+
+	fp = filp->private_data;
+	if (!fp)
+		return -ENODEV;
+	ib = idal_buffer_alloc(count, 0);
+	if (!ib)
+		return -ENOMEM;
+	rq = raw3270_request_alloc(0);
+	if (!IS_ERR(rq)) {
+		if (idal_buffer_from_user(ib, data, count) == 0) {
+			write_command = fp->write_command ? : 1;
+			if (write_command == 5)
+				write_command = 13;
+			raw3270_request_set_cmd(rq, write_command);
+			raw3270_request_set_idal(rq, ib);
+			rc = fs3270_do_io(&fp->view, rq);
+		} else
+			rc = -EFAULT;
+		raw3270_request_free(rq);
+	} else
+		rc = PTR_ERR(rq);
+	idal_buffer_free(ib);
+	return rc;
+}
+
+/*
+ * process ioctl commands for the tube driver
+ */
+static int
+fs3270_ioctl(struct inode *inode, struct file *filp,
+	     unsigned int cmd, unsigned long arg)
+{
+	struct fs3270 *fp;
+	struct raw3270_iocb iocb;
+	int rc;
+
+	fp = filp->private_data;
+	if (!fp)
+		return -ENODEV;
+	rc = 0;
+	switch (cmd) {
+	case TUBICMD:
+		fp->read_command = arg;
+		break;
+	case TUBOCMD:
+		fp->write_command = arg;
+		break;
+	case TUBGETI:
+		rc = put_user(fp->read_command, (char *) arg);
+		break;
+	case TUBGETO:
+		rc = put_user(fp->write_command,(char *) arg);
+		break;
+	case TUBGETMOD:
+		iocb.model = fp->view.model;
+		iocb.line_cnt = fp->view.rows;
+		iocb.col_cnt = fp->view.cols;
+		iocb.pf_cnt = 24;
+		iocb.re_cnt = 20;
+		iocb.map = 0;
+		if (copy_to_user((char *) arg, &iocb,
+				 sizeof(struct raw3270_iocb)))
+			rc = -EFAULT;
+		break;
+	}
+	return rc;
+}
+
+/*
+ * Allocate tty3270 structure.
+ */
+static struct fs3270 *
+fs3270_alloc_view(void)
+{
+	struct fs3270 *fp;
+
+	fp = (struct fs3270 *) kmalloc(sizeof(struct fs3270),GFP_KERNEL);
+	if (!fp)
+		return ERR_PTR(-ENOMEM);
+	memset(fp, 0, sizeof(struct fs3270));
+	fp->clear = raw3270_request_alloc(0);
+	if (!IS_ERR(fp->clear)) {
+		kfree(fp);
+		return ERR_PTR(-ENOMEM);
+	}
+	return fp;
+}
+
+/*
+ * Free tty3270 structure.
+ */
+static void
+fs3270_free_view(struct raw3270_view *view)
+{
+	raw3270_request_free(((struct fs3270 *) view)->clear);
+	kfree(view);
+}
+
+/*
+ * Unlink fs3270 data structure from filp.
+ */
+static void
+fs3270_release(struct raw3270_view *view)
+{
+}
+
+/* View to a 3270 device. Can be console, tty or fullscreen. */
+struct raw3270_fn fs3270_fn = {
+	.activate = fs3270_activate,
+	.deactivate = fs3270_deactivate,
+	.intv = (void *) fs3270_irq,
+	.release = fs3270_release,
+	.free = fs3270_free_view
+};
+
+/*
+ * This routine is called whenever a 3270 fullscreen device is opened.
+ */
+static int
+fs3270_open(struct inode *inode, struct file *filp)
+{
+	struct fs3270 *fp;
+	int minor, rc;
+
+	if (imajor(filp->f_dentry->d_inode) != IBM_FS3270_MAJOR)
+		return -ENODEV;
+	minor = iminor(filp->f_dentry->d_inode);
+	/* Check if some other program is already using fullscreen mode. */
+	fp = (struct fs3270 *) raw3270_find_view(&fs3270_fn, minor);
+	if (!IS_ERR(fp)) {
+		raw3270_put_view(&fp->view);
+		return -EBUSY;
+	}
+	/* Allocate fullscreen view structure. */
+	fp = fs3270_alloc_view();
+	if (IS_ERR(fp))
+		return PTR_ERR(fp);
+
+	init_waitqueue_head(&fp->attn_wait);
+	fp->fs_pid = current->pid;
+	rc = raw3270_add_view(&fp->view, &fs3270_fn, minor);
+	if (rc) {
+		fs3270_free_view(&fp->view);
+		return rc;
+	}
+
+	rc = raw3270_activate_view(&fp->view);
+	if (rc) {
+		raw3270_del_view(&fp->view);
+		return rc;
+	}
+	filp->private_data = fp;
+	return 0;
+}
+
+/*
+ * This routine is called when the 3270 tty is closed. We wait
+ * for the remaining request to be completed. Then we clean up.
+ */
+static int
+fs3270_close(struct inode *inode, struct file *filp)
+{
+	struct fs3270 *fp;
+
+	fp = filp->private_data;
+	filp->private_data = 0;
+	if (fp)
+		raw3270_del_view(&fp->view);
+	return 0;
+}
+
+static struct file_operations fs3270_fops = {
+	.owner	 = THIS_MODULE,		/* owner */
+	.read	 = fs3270_read,		/* read */
+	.write	 = fs3270_write,	/* write */
+	.ioctl	 = fs3270_ioctl,	/* ioctl */
+	.open	 = fs3270_open,		/* open */
+	.release = fs3270_close,	/* release */
+};
+
+/*
+ * 3270 fullscreen driver initialization.
+ */
+static int __init
+fs3270_init(void)
+{
+	int rc;
+
+	rc = register_chrdev(IBM_FS3270_MAJOR, "fs3270", &fs3270_fops);
+	if (rc) {
+		printk(KERN_ERR "fs3270 can't get major number %d: errno %d\n",
+		       IBM_FS3270_MAJOR, rc);
+		return rc;
+	}
+	return 0;
+}
+
+static void __exit
+fs3270_exit(void)
+{
+	unregister_chrdev(IBM_FS3270_MAJOR, "fs3270");
+}
+
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_CHARDEV_MAJOR(IBM_FS3270_MAJOR);
+
+module_init(fs3270_init);
+module_exit(fs3270_exit);

drivers/s390/char/keyboard.c

@@ -0,0 +1,519 @@
+/*
+ *  drivers/s390/char/keyboard.c
+ *    ebcdic keycode functions for s390 console drivers
+ *
+ *  S390 version
+ *    Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/sysrq.h>
+
+#include <linux/kbd_kern.h>
+#include <linux/kbd_diacr.h>
+#include <asm/uaccess.h>
+
+#include "keyboard.h"
+
+/*
+ * Handler Tables.
+ */
+#define K_HANDLERS\
+	k_self,		k_fn,		k_spec,		k_ignore,\
+	k_dead,		k_ignore,	k_ignore,	k_ignore,\
+	k_ignore,	k_ignore,	k_ignore,	k_ignore,\
+	k_ignore,	k_ignore,	k_ignore,	k_ignore
+
+typedef void (k_handler_fn)(struct kbd_data *, unsigned char);
+static k_handler_fn K_HANDLERS;
+static k_handler_fn *k_handler[16] = { K_HANDLERS };
+
+/* maximum values each key_handler can handle */
+static const int kbd_max_vals[] = {
+	255, ARRAY_SIZE(func_table) - 1, NR_FN_HANDLER - 1, 0,
+	NR_DEAD - 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
+};
+static const int KBD_NR_TYPES = ARRAY_SIZE(kbd_max_vals);
+
+static unsigned char ret_diacr[NR_DEAD] = {
+	'`', '\'', '^', '~', '"', ','
+};
+
+/*
+ * Alloc/free of kbd_data structures.
+ */
+struct kbd_data *
+kbd_alloc(void) {
+	struct kbd_data *kbd;
+	int i, len;
+
+	kbd = kmalloc(sizeof(struct kbd_data), GFP_KERNEL);
+	if (!kbd)
+		goto out;
+	memset(kbd, 0, sizeof(struct kbd_data));
+	kbd->key_maps = kmalloc(sizeof(key_maps), GFP_KERNEL);
+	if (!key_maps)
+		goto out_kbd;
+	memset(kbd->key_maps, 0, sizeof(key_maps));
+	for (i = 0; i < ARRAY_SIZE(key_maps); i++) {
+		if (key_maps[i]) {
+			kbd->key_maps[i] =
+				kmalloc(sizeof(u_short)*NR_KEYS, GFP_KERNEL);
+			if (!kbd->key_maps[i])
+				goto out_maps;
+			memcpy(kbd->key_maps[i], key_maps[i],
+			       sizeof(u_short)*NR_KEYS);
+		}
+	}
+	kbd->func_table = kmalloc(sizeof(func_table), GFP_KERNEL);
+	if (!kbd->func_table)
+		goto out_maps;
+	memset(kbd->func_table, 0, sizeof(func_table));
+	for (i = 0; i < ARRAY_SIZE(func_table); i++) {
+		if (func_table[i]) {
+			len = strlen(func_table[i]) + 1;
+			kbd->func_table[i] = kmalloc(len, GFP_KERNEL);
+			if (!kbd->func_table[i])
+				goto out_func;
+			memcpy(kbd->func_table[i], func_table[i], len);
+		}
+	}
+	kbd->fn_handler =
+		kmalloc(sizeof(fn_handler_fn *) * NR_FN_HANDLER, GFP_KERNEL);
+	if (!kbd->fn_handler)
+		goto out_func;
+	memset(kbd->fn_handler, 0, sizeof(fn_handler_fn *) * NR_FN_HANDLER);
+	kbd->accent_table =
+		kmalloc(sizeof(struct kbdiacr)*MAX_DIACR, GFP_KERNEL);
+	if (!kbd->accent_table)
+		goto out_fn_handler;
+	memcpy(kbd->accent_table, accent_table,
+	       sizeof(struct kbdiacr)*MAX_DIACR);
+	kbd->accent_table_size = accent_table_size;
+	return kbd;
+
+out_fn_handler:
+	kfree(kbd->fn_handler);
+out_func:
+	for (i = 0; i < ARRAY_SIZE(func_table); i++)
+		if (kbd->func_table[i])
+			kfree(kbd->func_table[i]);
+	kfree(kbd->func_table);
+out_maps:
+	for (i = 0; i < ARRAY_SIZE(key_maps); i++)
+		if (kbd->key_maps[i])
+			kfree(kbd->key_maps[i]);
+	kfree(kbd->key_maps);
+out_kbd:
+	kfree(kbd);
+out:
+	return 0;
+}
+
+void
+kbd_free(struct kbd_data *kbd)
+{
+	int i;
+
+	kfree(kbd->accent_table);
+	kfree(kbd->fn_handler);
+	for (i = 0; i < ARRAY_SIZE(func_table); i++)
+		if (kbd->func_table[i])
+			kfree(kbd->func_table[i]);
+	kfree(kbd->func_table);
+	for (i = 0; i < ARRAY_SIZE(key_maps); i++)
+		if (kbd->key_maps[i])
+			kfree(kbd->key_maps[i]);
+	kfree(kbd->key_maps);
+	kfree(kbd);
+}
+
+/*
+ * Generate ascii -> ebcdic translation table from kbd_data.
+ */
+void
+kbd_ascebc(struct kbd_data *kbd, unsigned char *ascebc)
+{
+	unsigned short *keymap, keysym;
+	int i, j, k;
+
+	memset(ascebc, 0x40, 256);
+	for (i = 0; i < ARRAY_SIZE(key_maps); i++) {
+		keymap = kbd->key_maps[i];
+		if (!keymap)
+			continue;
+		for (j = 0; j < NR_KEYS; j++) {
+			k = ((i & 1) << 7) + j;
+			keysym = keymap[j];
+			if (KTYP(keysym) == (KT_LATIN | 0xf0) ||
+			    KTYP(keysym) == (KT_LETTER | 0xf0))
+				ascebc[KVAL(keysym)] = k;
+			else if (KTYP(keysym) == (KT_DEAD | 0xf0))
+				ascebc[ret_diacr[KVAL(keysym)]] = k;
+		}
+	}
+}
+
+/*
+ * Generate ebcdic -> ascii translation table from kbd_data.
+ */
+void
+kbd_ebcasc(struct kbd_data *kbd, unsigned char *ebcasc)
+{
+	unsigned short *keymap, keysym;
+	int i, j, k;
+
+	memset(ebcasc, ' ', 256);
+	for (i = 0; i < ARRAY_SIZE(key_maps); i++) {
+		keymap = kbd->key_maps[i];
+		if (!keymap)
+			continue;
+		for (j = 0; j < NR_KEYS; j++) {
+			keysym = keymap[j];
+			k = ((i & 1) << 7) + j;
+			if (KTYP(keysym) == (KT_LATIN | 0xf0) ||
+			    KTYP(keysym) == (KT_LETTER | 0xf0))
+				ebcasc[k] = KVAL(keysym);
+			else if (KTYP(keysym) == (KT_DEAD | 0xf0))
+				ebcasc[k] = ret_diacr[KVAL(keysym)];
+		}
+	}
+}
+
+/*
+ * We have a combining character DIACR here, followed by the character CH.
+ * If the combination occurs in the table, return the corresponding value.
+ * Otherwise, if CH is a space or equals DIACR, return DIACR.
+ * Otherwise, conclude that DIACR was not combining after all,
+ * queue it and return CH.
+ */
+static unsigned char
+handle_diacr(struct kbd_data *kbd, unsigned char ch)
+{
+	int i, d;
+
+	d = kbd->diacr;
+	kbd->diacr = 0;
+
+	for (i = 0; i < kbd->accent_table_size; i++) {
+		if (kbd->accent_table[i].diacr == d &&
+		    kbd->accent_table[i].base == ch)
+			return kbd->accent_table[i].result;
+	}
+
+	if (ch == ' ' || ch == d)
+		return d;
+
+	kbd_put_queue(kbd->tty, d);
+	return ch;
+}
+
+/*
+ * Handle dead key.
+ */
+static void
+k_dead(struct kbd_data *kbd, unsigned char value)
+{
+	value = ret_diacr[value];
+	kbd->diacr = (kbd->diacr ? handle_diacr(kbd, value) : value);
+}
+
+/*
+ * Normal character handler.
+ */
+static void
+k_self(struct kbd_data *kbd, unsigned char value)
+{
+	if (kbd->diacr)
+		value = handle_diacr(kbd, value);
+	kbd_put_queue(kbd->tty, value);
+}
+
+/*
+ * Special key handlers
+ */
+static void
+k_ignore(struct kbd_data *kbd, unsigned char value)
+{
+}
+
+/*
+ * Function key handler.
+ */
+static void
+k_fn(struct kbd_data *kbd, unsigned char value)
+{
+	if (kbd->func_table[value])
+		kbd_puts_queue(kbd->tty, kbd->func_table[value]);
+}
+
+static void
+k_spec(struct kbd_data *kbd, unsigned char value)
+{
+	if (value >= NR_FN_HANDLER)
+		return;
+	if (kbd->fn_handler[value])
+		kbd->fn_handler[value](kbd);
+}
+
+/*
+ * Put utf8 character to tty flip buffer.
+ * UTF-8 is defined for words of up to 31 bits,
+ * but we need only 16 bits here
+ */
+static void
+to_utf8(struct tty_struct *tty, ushort c) 
+{
+	if (c < 0x80)
+		/*  0******* */
+		kbd_put_queue(tty, c);
+	else if (c < 0x800) {
+		/* 110***** 10****** */
+		kbd_put_queue(tty, 0xc0 | (c >> 6));
+		kbd_put_queue(tty, 0x80 | (c & 0x3f));
+	} else {
+		/* 1110**** 10****** 10****** */
+		kbd_put_queue(tty, 0xe0 | (c >> 12));
+		kbd_put_queue(tty, 0x80 | ((c >> 6) & 0x3f));
+		kbd_put_queue(tty, 0x80 | (c & 0x3f));
+	}
+}
+
+/*
+ * Process keycode.
+ */
+void
+kbd_keycode(struct kbd_data *kbd, unsigned int keycode)
+{
+	unsigned short keysym;
+	unsigned char type, value;
+
+	if (!kbd || !kbd->tty)
+		return;
+
+	if (keycode >= 384)
+		keysym = kbd->key_maps[5][keycode - 384];
+	else if (keycode >= 256)
+		keysym = kbd->key_maps[4][keycode - 256];
+	else if (keycode >= 128)
+		keysym = kbd->key_maps[1][keycode - 128];
+	else
+		keysym = kbd->key_maps[0][keycode];
+
+	type = KTYP(keysym);
+	if (type >= 0xf0) {
+		type -= 0xf0;
+		if (type == KT_LETTER)
+			type = KT_LATIN;
+		value = KVAL(keysym);
+#ifdef CONFIG_MAGIC_SYSRQ	       /* Handle the SysRq Hack */
+		if (kbd->sysrq) {
+			if (kbd->sysrq == K(KT_LATIN, '-')) {
+				kbd->sysrq = 0;
+				handle_sysrq(value, 0, kbd->tty);
+				return;
+			}
+			if (value == '-') {
+				kbd->sysrq = K(KT_LATIN, '-');
+				return;
+			}
+			/* Incomplete sysrq sequence. */
+			(*k_handler[KTYP(kbd->sysrq)])(kbd, KVAL(kbd->sysrq));
+			kbd->sysrq = 0;
+		} else if ((type == KT_LATIN && value == '^') ||
+			   (type == KT_DEAD && ret_diacr[value] == '^')) {
+			kbd->sysrq = K(type, value);
+			return;
+		}
+#endif
+		(*k_handler[type])(kbd, value);
+	} else
+		to_utf8(kbd->tty, keysym);
+}
+
+/*
+ * Ioctl stuff.
+ */
+static int
+do_kdsk_ioctl(struct kbd_data *kbd, struct kbentry __user *user_kbe,
+	      int cmd, int perm)
+{
+	struct kbentry tmp;
+	ushort *key_map, val, ov;
+
+	if (copy_from_user(&tmp, user_kbe, sizeof(struct kbentry)))
+		return -EFAULT;
+#if NR_KEYS < 256
+	if (tmp.kb_index >= NR_KEYS)
+		return -EINVAL;
+#endif
+#if MAX_NR_KEYMAPS < 256
+	if (tmp.kb_table >= MAX_NR_KEYMAPS)
+		return -EINVAL;	
+#endif
+
+	switch (cmd) {
+	case KDGKBENT:
+		key_map = kbd->key_maps[tmp.kb_table];
+		if (key_map) {
+		    val = U(key_map[tmp.kb_index]);
+		    if (KTYP(val) >= KBD_NR_TYPES)
+			val = K_HOLE;
+		} else
+		    val = (tmp.kb_index ? K_HOLE : K_NOSUCHMAP);
+		return put_user(val, &user_kbe->kb_value);
+	case KDSKBENT:
+		if (!perm)
+			return -EPERM;
+		if (!tmp.kb_index && tmp.kb_value == K_NOSUCHMAP) {
+			/* disallocate map */
+			key_map = kbd->key_maps[tmp.kb_table];
+			if (key_map) {
+			    kbd->key_maps[tmp.kb_table] = 0;
+			    kfree(key_map);
+			}
+			break;
+		}
+
+		if (KTYP(tmp.kb_value) >= KBD_NR_TYPES)
+			return -EINVAL;
+		if (KVAL(tmp.kb_value) > kbd_max_vals[KTYP(tmp.kb_value)])
+			return -EINVAL;
+
+		if (!(key_map = kbd->key_maps[tmp.kb_table])) {
+			int j;
+
+			key_map = (ushort *) kmalloc(sizeof(plain_map),
+						     GFP_KERNEL);
+			if (!key_map)
+				return -ENOMEM;
+			kbd->key_maps[tmp.kb_table] = key_map;
+			for (j = 0; j < NR_KEYS; j++)
+				key_map[j] = U(K_HOLE);
+		}
+		ov = U(key_map[tmp.kb_index]);
+		if (tmp.kb_value == ov)
+			break;	/* nothing to do */
+		/*
+		 * Attention Key.
+		 */
+		if (((ov == K_SAK) || (tmp.kb_value == K_SAK)) &&
+		    !capable(CAP_SYS_ADMIN))
+			return -EPERM;
+		key_map[tmp.kb_index] = U(tmp.kb_value);
+		break;
+	}
+	return 0;
+}
+
+static int
+do_kdgkb_ioctl(struct kbd_data *kbd, struct kbsentry __user *u_kbs,
+	       int cmd, int perm)
+{
+	unsigned char kb_func;
+	char *p;
+	int len;
+
+	/* Get u_kbs->kb_func. */
+	if (get_user(kb_func, &u_kbs->kb_func))
+		return -EFAULT;
+#if MAX_NR_FUNC < 256
+	if (kb_func >= MAX_NR_FUNC)
+		return -EINVAL;
+#endif
+
+	switch (cmd) {
+	case KDGKBSENT:
+		p = kbd->func_table[kb_func];
+		if (p) {
+			len = strlen(p);
+			if (len >= sizeof(u_kbs->kb_string))
+				len = sizeof(u_kbs->kb_string) - 1;
+			if (copy_to_user(u_kbs->kb_string, p, len))
+				return -EFAULT;
+		} else
+			len = 0;
+		if (put_user('\0', u_kbs->kb_string + len))
+			return -EFAULT;
+		break;
+	case KDSKBSENT:
+		if (!perm)
+			return -EPERM;
+		len = strnlen_user(u_kbs->kb_string,
+				   sizeof(u_kbs->kb_string) - 1);
+		p = kmalloc(len, GFP_KERNEL);
+		if (!p)
+			return -ENOMEM;
+		if (copy_from_user(p, u_kbs->kb_string, len)) {
+			kfree(p);
+			return -EFAULT;
+		}
+		p[len] = 0;
+		if (kbd->func_table[kb_func])
+			kfree(kbd->func_table[kb_func]);
+		kbd->func_table[kb_func] = p;
+		break;
+	}
+	return 0;
+}
+
+int
+kbd_ioctl(struct kbd_data *kbd, struct file *file,
+	  unsigned int cmd, unsigned long arg)
+{
+	struct kbdiacrs __user *a;
+	void __user *argp;
+	int ct, perm;
+
+	argp = (void __user *)arg;
+
+	/*
+	 * To have permissions to do most of the vt ioctls, we either have
+	 * to be the owner of the tty, or have CAP_SYS_TTY_CONFIG.
+	 */
+	perm = current->signal->tty == kbd->tty || capable(CAP_SYS_TTY_CONFIG);
+	switch (cmd) {
+	case KDGKBTYPE:
+		return put_user(KB_101, (char __user *)argp);
+	case KDGKBENT:
+	case KDSKBENT:
+		return do_kdsk_ioctl(kbd, argp, cmd, perm);
+	case KDGKBSENT:
+	case KDSKBSENT:
+		return do_kdgkb_ioctl(kbd, argp, cmd, perm);
+	case KDGKBDIACR:
+		a = argp;
+
+		if (put_user(kbd->accent_table_size, &a->kb_cnt))
+			return -EFAULT;
+		ct = kbd->accent_table_size;
+		if (copy_to_user(a->kbdiacr, kbd->accent_table,
+				 ct * sizeof(struct kbdiacr)))
+			return -EFAULT;
+		return 0;
+	case KDSKBDIACR:
+		a = argp;
+		if (!perm)
+			return -EPERM;
+		if (get_user(ct, &a->kb_cnt))
+			return -EFAULT;
+		if (ct >= MAX_DIACR)
+			return -EINVAL;
+		kbd->accent_table_size = ct;
+		if (copy_from_user(kbd->accent_table, a->kbdiacr,
+				   ct * sizeof(struct kbdiacr)))
+			return -EFAULT;
+		return 0;
+	default:
+		return -ENOIOCTLCMD;
+	}
+}
+
+EXPORT_SYMBOL(kbd_ioctl);
+EXPORT_SYMBOL(kbd_ascebc);
+EXPORT_SYMBOL(kbd_free);
+EXPORT_SYMBOL(kbd_alloc);
+EXPORT_SYMBOL(kbd_keycode);

drivers/s390/char/keyboard.h

@@ -0,0 +1,57 @@
+/*
+ *  drivers/s390/char/keyboard.h
+ *    ebcdic keycode functions for s390 console drivers
+ *
+ *    Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com),
+ */
+
+#include <linux/tty.h>
+#include <linux/tty_flip.h>
+#include <linux/keyboard.h>
+
+#define NR_FN_HANDLER	20
+
+struct kbd_data;
+
+typedef void (fn_handler_fn)(struct kbd_data *);
+
+/*
+ * FIXME: explain key_maps tricks.
+ */
+
+struct kbd_data {
+	struct tty_struct *tty;
+	unsigned short **key_maps;
+	char **func_table;
+	fn_handler_fn **fn_handler;
+	struct kbdiacr *accent_table;
+	unsigned int accent_table_size;
+	unsigned char diacr;
+	unsigned short sysrq;
+};
+
+struct kbd_data *kbd_alloc(void);
+void kbd_free(struct kbd_data *);
+void kbd_ascebc(struct kbd_data *, unsigned char *);
+
+void kbd_keycode(struct kbd_data *, unsigned int);
+int kbd_ioctl(struct kbd_data *, struct file *, unsigned int, unsigned long);
+
+/*
+ * Helper Functions.
+ */
+extern inline void
+kbd_put_queue(struct tty_struct *tty, int ch)
+{
+	tty_insert_flip_char(tty, ch, 0);
+	tty_schedule_flip(tty);
+}
+
+extern inline void
+kbd_puts_queue(struct tty_struct *tty, char *cp)
+{
+	while (*cp)
+		tty_insert_flip_char(tty, *cp++, 0);
+	tty_schedule_flip(tty);
+}

drivers/s390/char/monreader.c

@@ -0,0 +1,662 @@
+/*
+ * drivers/s390/char/monreader.c
+ *
+ * Character device driver for reading z/VM *MONITOR service records.
+ *
+ * Copyright (C) 2004 IBM Corporation, IBM Deutschland Entwicklung GmbH.
+ *
+ * Author: Gerald Schaefer <geraldsc@de.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/miscdevice.h>
+#include <linux/ctype.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <asm/uaccess.h>
+#include <asm/ebcdic.h>
+#include <asm/extmem.h>
+#include <linux/poll.h>
+#include "../net/iucv.h"
+
+
+//#define MON_DEBUG			/* Debug messages on/off */
+
+#define MON_NAME "monreader"
+
+#define P_INFO(x...)	printk(KERN_INFO MON_NAME " info: " x)
+#define P_ERROR(x...)	printk(KERN_ERR MON_NAME " error: " x)
+#define P_WARNING(x...)	printk(KERN_WARNING MON_NAME " warning: " x)
+
+#ifdef MON_DEBUG
+#define P_DEBUG(x...)   printk(KERN_DEBUG MON_NAME " debug: " x)
+#else
+#define P_DEBUG(x...)   do {} while (0)
+#endif
+
+#define MON_COLLECT_SAMPLE 0x80
+#define MON_COLLECT_EVENT  0x40
+#define MON_SERVICE	   "*MONITOR"
+#define MON_IN_USE	   0x01
+#define MON_MSGLIM	   255
+
+static char mon_dcss_name[9] = "MONDCSS\0";
+
+struct mon_msg {
+	u32 pos;
+	u32 mca_offset;
+	iucv_MessagePending local_eib;
+	char msglim_reached;
+	char replied_msglim;
+};
+
+struct mon_private {
+	u16 pathid;
+	iucv_handle_t iucv_handle;
+	struct mon_msg *msg_array[MON_MSGLIM];
+	unsigned int   write_index;
+	unsigned int   read_index;
+	atomic_t msglim_count;
+	atomic_t read_ready;
+	atomic_t iucv_connected;
+	atomic_t iucv_severed;
+};
+
+static unsigned long mon_in_use = 0;
+
+static unsigned long mon_dcss_start;
+static unsigned long mon_dcss_end;
+
+static DECLARE_WAIT_QUEUE_HEAD(mon_read_wait_queue);
+static DECLARE_WAIT_QUEUE_HEAD(mon_conn_wait_queue);
+
+static u8 iucv_host[8] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00};
+
+static u8 user_data_connect[16] = {
+	/* Version code, must be 0x01 for shared mode */
+	0x01,
+	/* what to collect */
+	MON_COLLECT_SAMPLE | MON_COLLECT_EVENT,
+	/* DCSS name in EBCDIC, 8 bytes padded with blanks */
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+};
+
+static u8 user_data_sever[16] = {
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+};
+
+
+/******************************************************************************
+ *                             helper functions                               *
+ *****************************************************************************/
+/*
+ * Create the 8 bytes EBCDIC DCSS segment name from
+ * an ASCII name, incl. padding
+ */
+static inline void
+dcss_mkname(char *ascii_name, char *ebcdic_name)
+{
+	int i;
+
+	for (i = 0; i < 8; i++) {
+		if (ascii_name[i] == '\0')
+			break;
+		ebcdic_name[i] = toupper(ascii_name[i]);
+	};
+	for (; i < 8; i++)
+		ebcdic_name[i] = ' ';
+	ASCEBC(ebcdic_name, 8);
+}
+
+/*
+ * print appropriate error message for segment_load()/segment_type()
+ * return code
+ */
+static void
+mon_segment_warn(int rc, char* seg_name)
+{
+	switch (rc) {
+	case -ENOENT:
+		P_WARNING("cannot load/query segment %s, does not exist\n",
+			  seg_name);
+		break;
+	case -ENOSYS:
+		P_WARNING("cannot load/query segment %s, not running on VM\n",
+			  seg_name);
+		break;
+	case -EIO:
+		P_WARNING("cannot load/query segment %s, hardware error\n",
+			  seg_name);
+		break;
+	case -ENOTSUPP:
+		P_WARNING("cannot load/query segment %s, is a multi-part "
+			  "segment\n", seg_name);
+		break;
+	case -ENOSPC:
+		P_WARNING("cannot load/query segment %s, overlaps with "
+			  "storage\n", seg_name);
+		break;
+	case -EBUSY:
+		P_WARNING("cannot load/query segment %s, overlaps with "
+			  "already loaded dcss\n", seg_name);
+		break;
+	case -EPERM:
+		P_WARNING("cannot load/query segment %s, already loaded in "
+			  "incompatible mode\n", seg_name);
+		break;
+	case -ENOMEM:
+		P_WARNING("cannot load/query segment %s, out of memory\n",
+			  seg_name);
+		break;
+	case -ERANGE:
+		P_WARNING("cannot load/query segment %s, exceeds kernel "
+			  "mapping range\n", seg_name);
+		break;
+	default:
+		P_WARNING("cannot load/query segment %s, return value %i\n",
+			  seg_name, rc);
+		break;
+	}
+}
+
+static inline unsigned long
+mon_mca_start(struct mon_msg *monmsg)
+{
+	return monmsg->local_eib.ln1msg1.iprmmsg1_u32;
+}
+
+static inline unsigned long
+mon_mca_end(struct mon_msg *monmsg)
+{
+	return monmsg->local_eib.ln1msg2.ipbfln1f;
+}
+
+static inline u8
+mon_mca_type(struct mon_msg *monmsg, u8 index)
+{
+	return *((u8 *) mon_mca_start(monmsg) + monmsg->mca_offset + index);
+}
+
+static inline u32
+mon_mca_size(struct mon_msg *monmsg)
+{
+	return mon_mca_end(monmsg) - mon_mca_start(monmsg) + 1;
+}
+
+static inline u32
+mon_rec_start(struct mon_msg *monmsg)
+{
+	return *((u32 *) (mon_mca_start(monmsg) + monmsg->mca_offset + 4));
+}
+
+static inline u32
+mon_rec_end(struct mon_msg *monmsg)
+{
+	return *((u32 *) (mon_mca_start(monmsg) + monmsg->mca_offset + 8));
+}
+
+static inline int
+mon_check_mca(struct mon_msg *monmsg)
+{
+	if ((mon_rec_end(monmsg) <= mon_rec_start(monmsg)) ||
+	    (mon_rec_start(monmsg) < mon_dcss_start) ||
+	    (mon_rec_end(monmsg) > mon_dcss_end) ||
+	    (mon_mca_type(monmsg, 0) == 0) ||
+	    (mon_mca_size(monmsg) % 12 != 0) ||
+	    (mon_mca_end(monmsg) <= mon_mca_start(monmsg)) ||
+	    (mon_mca_end(monmsg) > mon_dcss_end) ||
+	    (mon_mca_start(monmsg) < mon_dcss_start) ||
+	    ((mon_mca_type(monmsg, 1) == 0) && (mon_mca_type(monmsg, 2) == 0)))
+	{
+		P_DEBUG("READ, IGNORED INVALID MCA\n\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static inline int
+mon_send_reply(struct mon_msg *monmsg, struct mon_private *monpriv)
+{
+	u8 prmmsg[8];
+	int rc;
+
+	P_DEBUG("read, REPLY: pathid = 0x%04X, msgid = 0x%08X, trgcls = "
+		"0x%08X\n\n",
+		monmsg->local_eib.ippathid, monmsg->local_eib.ipmsgid,
+		monmsg->local_eib.iptrgcls);
+	rc = iucv_reply_prmmsg(monmsg->local_eib.ippathid,
+				monmsg->local_eib.ipmsgid,
+				monmsg->local_eib.iptrgcls,
+				0, prmmsg);
+	atomic_dec(&monpriv->msglim_count);
+	if (likely(!monmsg->msglim_reached)) {
+		monmsg->pos = 0;
+		monmsg->mca_offset = 0;
+		monpriv->read_index = (monpriv->read_index + 1) %
+				      MON_MSGLIM;
+		atomic_dec(&monpriv->read_ready);
+	} else
+		monmsg->replied_msglim = 1;
+	if (rc) {
+		P_ERROR("read, IUCV reply failed with rc = %i\n\n", rc);
+		return -EIO;
+	}
+	return 0;
+}
+
+static inline struct mon_private *
+mon_alloc_mem(void)
+{
+	int i,j;
+	struct mon_private *monpriv;
+
+	monpriv = kmalloc(sizeof(struct mon_private), GFP_KERNEL);
+	if (!monpriv) {
+		P_ERROR("no memory for monpriv\n");
+		return NULL;
+	}
+	memset(monpriv, 0, sizeof(struct mon_private));
+	for (i = 0; i < MON_MSGLIM; i++) {
+		monpriv->msg_array[i] = kmalloc(sizeof(struct mon_msg),
+						    GFP_KERNEL);
+		if (!monpriv->msg_array[i]) {
+			P_ERROR("open, no memory for msg_array\n");
+			for (j = 0; j < i; j++)
+				kfree(monpriv->msg_array[j]);
+			return NULL;
+		}
+		memset(monpriv->msg_array[i], 0, sizeof(struct mon_msg));
+	}
+	return monpriv;
+}
+
+static inline void
+mon_read_debug(struct mon_msg *monmsg, struct mon_private *monpriv)
+{
+#ifdef MON_DEBUG
+	u8 msg_type[2], mca_type;
+	unsigned long records_len;
+
+	records_len = mon_rec_end(monmsg) - mon_rec_start(monmsg) + 1;
+
+	memcpy(msg_type, &monmsg->local_eib.iptrgcls, 2);
+	EBCASC(msg_type, 2);
+	mca_type = mon_mca_type(monmsg, 0);
+	EBCASC(&mca_type, 1);
+
+	P_DEBUG("read, mon_read_index = %i, mon_write_index = %i\n",
+		monpriv->read_index, monpriv->write_index);
+	P_DEBUG("read, pathid = 0x%04X, msgid = 0x%08X, trgcls = 0x%08X\n",
+		monmsg->local_eib.ippathid, monmsg->local_eib.ipmsgid,
+		monmsg->local_eib.iptrgcls);
+	P_DEBUG("read, msg_type = '%c%c', mca_type = '%c' / 0x%X / 0x%X\n",
+		msg_type[0], msg_type[1], mca_type ? mca_type : 'X',
+		mon_mca_type(monmsg, 1), mon_mca_type(monmsg, 2));
+	P_DEBUG("read, MCA: start = 0x%lX, end = 0x%lX\n",
+		mon_mca_start(monmsg), mon_mca_end(monmsg));
+	P_DEBUG("read, REC: start = 0x%X, end = 0x%X, len = %lu\n\n",
+		mon_rec_start(monmsg), mon_rec_end(monmsg), records_len);
+	if (mon_mca_size(monmsg) > 12)
+		P_DEBUG("READ, MORE THAN ONE MCA\n\n");
+#endif
+}
+
+static inline void
+mon_next_mca(struct mon_msg *monmsg)
+{
+	if (likely((mon_mca_size(monmsg) - monmsg->mca_offset) == 12))
+		return;
+	P_DEBUG("READ, NEXT MCA\n\n");
+	monmsg->mca_offset += 12;
+	monmsg->pos = 0;
+}
+
+static inline struct mon_msg *
+mon_next_message(struct mon_private *monpriv)
+{
+	struct mon_msg *monmsg;
+
+	if (!atomic_read(&monpriv->read_ready))
+		return NULL;
+	monmsg = monpriv->msg_array[monpriv->read_index];
+	if (unlikely(monmsg->replied_msglim)) {
+		monmsg->replied_msglim = 0;
+		monmsg->msglim_reached = 0;
+		monmsg->pos = 0;
+		monmsg->mca_offset = 0;
+		P_WARNING("read, message limit reached\n");
+		monpriv->read_index = (monpriv->read_index + 1) %
+				      MON_MSGLIM;
+		atomic_dec(&monpriv->read_ready);
+		return ERR_PTR(-EOVERFLOW);
+	}
+	return monmsg;
+}
+
+
+/******************************************************************************
+ *                               IUCV handler                                 *
+ *****************************************************************************/
+static void
+mon_iucv_ConnectionComplete(iucv_ConnectionComplete *eib, void *pgm_data)
+{
+	struct mon_private *monpriv = (struct mon_private *) pgm_data;
+
+	P_DEBUG("IUCV connection completed\n");
+	P_DEBUG("IUCV ACCEPT (from *MONITOR): Version = 0x%02X, Event = "
+		"0x%02X, Sample = 0x%02X\n",
+		eib->ipuser[0], eib->ipuser[1], eib->ipuser[2]);
+	atomic_set(&monpriv->iucv_connected, 1);
+	wake_up(&mon_conn_wait_queue);
+}
+
+static void
+mon_iucv_ConnectionSevered(iucv_ConnectionSevered *eib, void *pgm_data)
+{
+	struct mon_private *monpriv = (struct mon_private *) pgm_data;
+
+	P_ERROR("IUCV connection severed with rc = 0x%X\n",
+		(u8) eib->ipuser[0]);
+	atomic_set(&monpriv->iucv_severed, 1);
+	wake_up(&mon_conn_wait_queue);
+	wake_up_interruptible(&mon_read_wait_queue);
+}
+
+static void
+mon_iucv_MessagePending(iucv_MessagePending *eib, void *pgm_data)
+{
+	struct mon_private *monpriv = (struct mon_private *) pgm_data;
+
+	P_DEBUG("IUCV message pending\n");
+	memcpy(&monpriv->msg_array[monpriv->write_index]->local_eib, eib,
+	       sizeof(iucv_MessagePending));
+	if (atomic_inc_return(&monpriv->msglim_count) == MON_MSGLIM) {
+		P_WARNING("IUCV message pending, message limit (%i) reached\n",
+			  MON_MSGLIM);
+		monpriv->msg_array[monpriv->write_index]->msglim_reached = 1;
+	}
+	monpriv->write_index = (monpriv->write_index + 1) % MON_MSGLIM;
+	atomic_inc(&monpriv->read_ready);
+	wake_up_interruptible(&mon_read_wait_queue);
+}
+
+static iucv_interrupt_ops_t mon_iucvops = {
+	.ConnectionComplete = mon_iucv_ConnectionComplete,
+	.ConnectionSevered  = mon_iucv_ConnectionSevered,
+	.MessagePending     = mon_iucv_MessagePending,
+};
+
+/******************************************************************************
+ *                               file operations                              *
+ *****************************************************************************/
+static int
+mon_open(struct inode *inode, struct file *filp)
+{
+	int rc, i;
+	struct mon_private *monpriv;
+
+	/*
+	 * only one user allowed
+	 */
+	if (test_and_set_bit(MON_IN_USE, &mon_in_use))
+		return -EBUSY;
+
+	monpriv = mon_alloc_mem();
+	if (!monpriv)
+		return -ENOMEM;
+
+	/*
+	 * Register with IUCV and connect to *MONITOR service
+	 */
+	monpriv->iucv_handle = iucv_register_program("my_monreader    ",
+							MON_SERVICE,
+							NULL,
+							&mon_iucvops,
+							monpriv);
+	if (!monpriv->iucv_handle) {
+		P_ERROR("failed to register with iucv driver\n");
+		rc = -EIO;
+		goto out_error;
+	}
+	P_INFO("open, registered with IUCV\n");
+
+	rc = iucv_connect(&monpriv->pathid, MON_MSGLIM, user_data_connect,
+			  MON_SERVICE, iucv_host, IPRMDATA, NULL, NULL,
+			  monpriv->iucv_handle, NULL);
+	if (rc) {
+		P_ERROR("iucv connection to *MONITOR failed with "
+			"IPUSER SEVER code = %i\n", rc);
+		rc = -EIO;
+		goto out_unregister;
+	}
+	/*
+	 * Wait for connection confirmation
+	 */
+	wait_event(mon_conn_wait_queue,
+		   atomic_read(&monpriv->iucv_connected) ||
+		   atomic_read(&monpriv->iucv_severed));
+	if (atomic_read(&monpriv->iucv_severed)) {
+		atomic_set(&monpriv->iucv_severed, 0);
+		atomic_set(&monpriv->iucv_connected, 0);
+		rc = -EIO;
+		goto out_unregister;
+	}
+	P_INFO("open, established connection to *MONITOR service\n\n");
+	filp->private_data = monpriv;
+	return nonseekable_open(inode, filp);
+
+out_unregister:
+	iucv_unregister_program(monpriv->iucv_handle);
+out_error:
+	for (i = 0; i < MON_MSGLIM; i++)
+		kfree(monpriv->msg_array[i]);
+	kfree(monpriv);
+	clear_bit(MON_IN_USE, &mon_in_use);
+	return rc;
+}
+
+static int
+mon_close(struct inode *inode, struct file *filp)
+{
+	int rc, i;
+	struct mon_private *monpriv = filp->private_data;
+
+	/*
+	 * Close IUCV connection and unregister
+	 */
+	rc = iucv_sever(monpriv->pathid, user_data_sever);
+	if (rc)
+		P_ERROR("close, iucv_sever failed with rc = %i\n", rc);
+	else
+		P_INFO("close, terminated connection to *MONITOR service\n");
+
+	rc = iucv_unregister_program(monpriv->iucv_handle);
+	if (rc)
+		P_ERROR("close, iucv_unregister failed with rc = %i\n", rc);
+	else
+		P_INFO("close, unregistered with IUCV\n");
+
+	atomic_set(&monpriv->iucv_severed, 0);
+	atomic_set(&monpriv->iucv_connected, 0);
+	atomic_set(&monpriv->read_ready, 0);
+	atomic_set(&monpriv->msglim_count, 0);
+	monpriv->write_index  = 0;
+	monpriv->read_index   = 0;
+
+	for (i = 0; i < MON_MSGLIM; i++)
+		kfree(monpriv->msg_array[i]);
+	kfree(monpriv);
+	clear_bit(MON_IN_USE, &mon_in_use);
+	return 0;
+}
+
+static ssize_t
+mon_read(struct file *filp, char __user *data, size_t count, loff_t *ppos)
+{
+	struct mon_private *monpriv = filp->private_data;
+	struct mon_msg *monmsg;
+	int ret;
+	u32 mce_start;
+
+	monmsg = mon_next_message(monpriv);
+	if (IS_ERR(monmsg))
+		return PTR_ERR(monmsg);
+
+	if (!monmsg) {
+		if (filp->f_flags & O_NONBLOCK)
+			return -EAGAIN;
+		ret = wait_event_interruptible(mon_read_wait_queue,
+					atomic_read(&monpriv->read_ready) ||
+					atomic_read(&monpriv->iucv_severed));
+		if (ret)
+			return ret;
+		if (unlikely(atomic_read(&monpriv->iucv_severed)))
+			return -EIO;
+		monmsg = monpriv->msg_array[monpriv->read_index];
+	}
+
+	if (!monmsg->pos) {
+		monmsg->pos = mon_mca_start(monmsg) + monmsg->mca_offset;
+		mon_read_debug(monmsg, monpriv);
+	}
+	if (mon_check_mca(monmsg))
+		goto reply;
+
+	/* read monitor control element (12 bytes) first */
+	mce_start = mon_mca_start(monmsg) + monmsg->mca_offset;
+	if ((monmsg->pos >= mce_start) && (monmsg->pos < mce_start + 12)) {
+		count = min(count, (size_t) mce_start + 12 - monmsg->pos);
+		ret = copy_to_user(data, (void *) (unsigned long) monmsg->pos,
+				   count);
+		if (ret)
+			return -EFAULT;
+		monmsg->pos += count;
+		if (monmsg->pos == mce_start + 12)
+			monmsg->pos = mon_rec_start(monmsg);
+		goto out_copy;
+	}
+
+	/* read records */
+	if (monmsg->pos <= mon_rec_end(monmsg)) {
+		count = min(count, (size_t) mon_rec_end(monmsg) - monmsg->pos
+					    + 1);
+		ret = copy_to_user(data, (void *) (unsigned long) monmsg->pos,
+				   count);
+		if (ret)
+			return -EFAULT;
+		monmsg->pos += count;
+		if (monmsg->pos > mon_rec_end(monmsg))
+			mon_next_mca(monmsg);
+		goto out_copy;
+	}
+reply:
+	ret = mon_send_reply(monmsg, monpriv);
+	return ret;
+
+out_copy:
+	*ppos += count;
+	return count;
+}
+
+static unsigned int
+mon_poll(struct file *filp, struct poll_table_struct *p)
+{
+	struct mon_private *monpriv = filp->private_data;
+
+	poll_wait(filp, &mon_read_wait_queue, p);
+	if (unlikely(atomic_read(&monpriv->iucv_severed)))
+		return POLLERR;
+	if (atomic_read(&monpriv->read_ready))
+		return POLLIN | POLLRDNORM;
+	return 0;
+}
+
+static struct file_operations mon_fops = {
+	.owner   = THIS_MODULE,
+	.open    = &mon_open,
+	.release = &mon_close,
+	.read    = &mon_read,
+	.poll    = &mon_poll,
+};
+
+static struct miscdevice mon_dev = {
+	.name       = "monreader",
+	.devfs_name = "monreader",
+	.fops       = &mon_fops,
+	.minor      = MISC_DYNAMIC_MINOR,
+};
+
+/******************************************************************************
+ *                              module init/exit                              *
+ *****************************************************************************/
+static int __init
+mon_init(void)
+{
+	int rc;
+
+	if (!MACHINE_IS_VM) {
+		P_ERROR("not running under z/VM, driver not loaded\n");
+		return -ENODEV;
+	}
+
+	rc = segment_type(mon_dcss_name);
+	if (rc < 0) {
+		mon_segment_warn(rc, mon_dcss_name);
+		return rc;
+	}
+	if (rc != SEG_TYPE_SC) {
+		P_ERROR("segment %s has unsupported type, should be SC\n",
+			mon_dcss_name);
+		return -EINVAL;
+	}
+
+	rc = segment_load(mon_dcss_name, SEGMENT_SHARED,
+			  &mon_dcss_start, &mon_dcss_end);
+	if (rc < 0) {
+		mon_segment_warn(rc, mon_dcss_name);
+		return -EINVAL;
+	}
+	dcss_mkname(mon_dcss_name, &user_data_connect[8]);
+
+	rc = misc_register(&mon_dev);
+	if (rc < 0 ) {
+		P_ERROR("misc_register failed, rc = %i\n", rc);
+		goto out;
+	}
+	P_INFO("Loaded segment %s from %p to %p, size = %lu Byte\n",
+		mon_dcss_name, (void *) mon_dcss_start, (void *) mon_dcss_end,
+		mon_dcss_end - mon_dcss_start + 1);
+	return 0;
+
+out:
+	segment_unload(mon_dcss_name);
+	return rc;
+}
+
+static void __exit
+mon_exit(void)
+{
+	segment_unload(mon_dcss_name);
+	WARN_ON(misc_deregister(&mon_dev) != 0);
+	return;
+}
+
+
+module_init(mon_init);
+module_exit(mon_exit);
+
+module_param_string(mondcss, mon_dcss_name, 9, 0444);
+MODULE_PARM_DESC(mondcss, "Name of DCSS segment to be used for *MONITOR "
+		 "service, max. 8 chars. Default is MONDCSS");
+
+MODULE_AUTHOR("Gerald Schaefer <[email protected]>");
+MODULE_DESCRIPTION("Character device driver for reading z/VM "
+		   "monitor service records.");
+MODULE_LICENSE("GPL");

drivers/s390/char/raw3270.h

@@ -0,0 +1,274 @@
+/*
+ *  drivers/s390/char/raw3270.h
+ *    IBM/3270 Driver
+ *
+ *  Author(s):
+ *    Original 3270 Code for 2.4 written by Richard Hitt (UTS Global)
+ *    Rewritten for 2.5 by Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *	-- Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ */
+
+#include <asm/idals.h>
+#include <asm/ioctl.h>
+
+/* ioctls for fullscreen 3270 */
+#define TUBICMD		_IO('3', 3)	/* set ccw command for fs reads. */
+#define TUBOCMD		_IO('3', 4)	/* set ccw command for fs writes. */
+#define TUBGETI		_IO('3', 7)	/* get ccw command for fs reads. */
+#define TUBGETO		_IO('3', 8)	/* get ccw command for fs writes. */
+#define TUBSETMOD	_IO('3',12)	/* FIXME: what does it do ?*/
+#define TUBGETMOD	_IO('3',13)	/* FIXME: what does it do ?*/
+
+/* Local Channel Commands */
+#define TC_WRITE	0x01		/* Write */
+#define TC_EWRITE	0x05		/* Erase write */
+#define TC_READMOD	0x06		/* Read modified */
+#define TC_EWRITEA	0x0d		/* Erase write alternate */
+#define TC_WRITESF	0x11		/* Write structured field */
+
+/* Buffer Control Orders */
+#define TO_SF		0x1d		/* Start field */
+#define TO_SBA		0x11		/* Set buffer address */
+#define TO_IC		0x13		/* Insert cursor */
+#define TO_PT		0x05		/* Program tab */
+#define TO_RA		0x3c		/* Repeat to address */
+#define TO_SFE		0x29		/* Start field extended */
+#define TO_EUA		0x12		/* Erase unprotected to address */
+#define TO_MF		0x2c		/* Modify field */
+#define TO_SA		0x28		/* Set attribute */
+
+/* Field Attribute Bytes */
+#define TF_INPUT	0x40		/* Visible input */
+#define TF_INPUTN	0x4c		/* Invisible input */
+#define TF_INMDT	0xc1		/* Visible, Set-MDT */
+#define TF_LOG		0x60
+
+/* Character Attribute Bytes */
+#define TAT_RESET	0x00
+#define TAT_FIELD	0xc0
+#define TAT_EXTHI	0x41
+#define TAT_COLOR	0x42
+#define TAT_CHARS	0x43
+#define TAT_TRANS	0x46
+
+/* Extended-Highlighting Bytes */
+#define TAX_RESET	0x00
+#define TAX_BLINK	0xf1
+#define TAX_REVER	0xf2
+#define TAX_UNDER	0xf4
+
+/* Reset value */
+#define TAR_RESET	0x00
+
+/* Color values */
+#define TAC_RESET	0x00
+#define TAC_BLUE	0xf1
+#define TAC_RED		0xf2
+#define TAC_PINK	0xf3
+#define TAC_GREEN	0xf4
+#define TAC_TURQ	0xf5
+#define TAC_YELLOW	0xf6
+#define TAC_WHITE	0xf7
+#define TAC_DEFAULT	0x00
+
+/* Write Control Characters */
+#define TW_NONE		0x40		/* No particular action */
+#define TW_KR		0xc2		/* Keyboard restore */
+#define TW_PLUSALARM	0x04		/* Add this bit for alarm */
+
+#define RAW3270_MAXDEVS	256
+
+/* For TUBGETMOD and TUBSETMOD. Should include. */
+struct raw3270_iocb {
+	short model;
+	short line_cnt;
+	short col_cnt;
+	short pf_cnt;
+	short re_cnt;
+	short map;
+};
+
+struct raw3270;
+struct raw3270_view;
+
+/* 3270 CCW request */
+struct raw3270_request {
+	struct list_head list;		/* list head for request queueing. */
+	struct raw3270_view *view;	/* view of this request */
+	struct ccw1 ccw;		/* single ccw. */
+	void *buffer;			/* output buffer. */
+	size_t size;			/* size of output buffer. */
+	int rescnt;			/* residual count from devstat. */
+	int rc;				/* return code for this request. */
+
+	/* Callback for delivering final status. */
+	void (*callback)(struct raw3270_request *, void *);
+	void *callback_data;
+};
+
+struct raw3270_request *raw3270_request_alloc(size_t size);
+struct raw3270_request *raw3270_request_alloc_bootmem(size_t size);
+void raw3270_request_free(struct raw3270_request *);
+void raw3270_request_reset(struct raw3270_request *);
+void raw3270_request_set_cmd(struct raw3270_request *, u8 cmd);
+int  raw3270_request_add_data(struct raw3270_request *, void *, size_t);
+void raw3270_request_set_data(struct raw3270_request *, void *, size_t);
+void raw3270_request_set_idal(struct raw3270_request *, struct idal_buffer *);
+
+static inline int
+raw3270_request_final(struct raw3270_request *rq)
+{
+	return list_empty(&rq->list);
+}
+
+void raw3270_buffer_address(struct raw3270 *, char *, unsigned short);
+
+/* Return value of *intv (see raw3270_fn below) can be one of the following: */
+#define RAW3270_IO_DONE		0	/* request finished */
+#define RAW3270_IO_BUSY		1	/* request still active */
+#define RAW3270_IO_RETRY	2	/* retry current request */
+#define RAW3270_IO_STOP		3	/* kill current request */
+
+/*
+ * Functions of a 3270 view.
+ */
+struct raw3270_fn {
+	int  (*activate)(struct raw3270_view *);
+	void (*deactivate)(struct raw3270_view *);
+	int  (*intv)(struct raw3270_view *,
+		     struct raw3270_request *, struct irb *);
+	void (*release)(struct raw3270_view *);
+	void (*free)(struct raw3270_view *);
+};
+
+/*
+ * View structure chaining. The raw3270_view structure is meant to
+ * be embedded at the start of the real view data structure, e.g.:
+ *   struct example {
+ *     struct raw3270_view view;
+ *     ...
+ *   };
+ */
+struct raw3270_view {
+	struct list_head list;
+	spinlock_t lock;
+	atomic_t ref_count;
+	struct raw3270 *dev;
+	struct raw3270_fn *fn;
+	unsigned int model;
+	unsigned int rows, cols;	/* # of rows & colums of the view */
+	unsigned char *ascebc;		/* ascii -> ebcdic table */
+};
+
+int raw3270_add_view(struct raw3270_view *, struct raw3270_fn *, int);
+int raw3270_activate_view(struct raw3270_view *);
+void raw3270_del_view(struct raw3270_view *);
+void raw3270_deactivate_view(struct raw3270_view *);
+struct raw3270_view *raw3270_find_view(struct raw3270_fn *, int);
+int raw3270_start(struct raw3270_view *, struct raw3270_request *);
+int raw3270_start_irq(struct raw3270_view *, struct raw3270_request *);
+
+/* Reference count inliner for view structures. */
+static inline void
+raw3270_get_view(struct raw3270_view *view)
+{
+	atomic_inc(&view->ref_count);
+}
+
+extern wait_queue_head_t raw3270_wait_queue;
+
+static inline void
+raw3270_put_view(struct raw3270_view *view)
+{
+	if (atomic_dec_return(&view->ref_count) == 0)
+		wake_up(&raw3270_wait_queue);
+}
+
+struct raw3270 *raw3270_setup_console(struct ccw_device *cdev);
+void raw3270_wait_cons_dev(struct raw3270 *);
+
+/* Notifier for device addition/removal */
+int raw3270_register_notifier(void (*notifier)(int, int));
+void raw3270_unregister_notifier(void (*notifier)(int, int));
+
+/*
+ * Little memory allocator for string objects. 
+ */
+struct string
+{
+	struct list_head list;
+	struct list_head update;
+	unsigned long size;
+	unsigned long len;
+	char string[0];
+} __attribute__ ((aligned(8)));
+
+static inline struct string *
+alloc_string(struct list_head *free_list, unsigned long len)
+{
+	struct string *cs, *tmp;
+	unsigned long size;
+
+	size = (len + 7L) & -8L;
+	list_for_each_entry(cs, free_list, list) {
+		if (cs->size < size)
+			continue;
+		if (cs->size > size + sizeof(struct string)) {
+			char *endaddr = (char *) (cs + 1) + cs->size;
+			tmp = (struct string *) (endaddr - size) - 1;
+			tmp->size = size;
+			cs->size -= size + sizeof(struct string);
+			cs = tmp;
+		} else
+			list_del(&cs->list);
+		cs->len = len;
+		INIT_LIST_HEAD(&cs->list);
+		INIT_LIST_HEAD(&cs->update);
+		return cs;
+	}
+	return 0;
+}
+
+static inline unsigned long
+free_string(struct list_head *free_list, struct string *cs)
+{
+	struct string *tmp;
+	struct list_head *p, *left;
+
+	/* Find out the left neighbour in free memory list. */
+	left = free_list;
+	list_for_each(p, free_list) {
+		if (list_entry(p, struct string, list) > cs)
+			break;
+		left = p;
+	}
+	/* Try to merge with right neighbour = next element from left. */
+	if (left->next != free_list) {
+		tmp = list_entry(left->next, struct string, list);
+		if ((char *) (cs + 1) + cs->size == (char *) tmp) {
+			list_del(&tmp->list);
+			cs->size += tmp->size + sizeof(struct string);
+		}
+	}
+	/* Try to merge with left neighbour. */
+	if (left != free_list) {
+		tmp = list_entry(left, struct string, list);
+		if ((char *) (tmp + 1) + tmp->size == (char *) cs) {
+			tmp->size += cs->size + sizeof(struct string);
+			return tmp->size;
+		}
+	}
+	__list_add(&cs->list, left, left->next);
+	return cs->size;
+}
+
+static inline void
+add_string_memory(struct list_head *free_list, void *mem, unsigned long size)
+{
+	struct string *cs;
+
+	cs = (struct string *) mem;
+	cs->size = size - sizeof(struct string);
+	free_string(free_list, cs);
+}
+

drivers/s390/char/sclp.c

@@ -0,0 +1,915 @@
+/*
+ *  drivers/s390/char/sclp.c
+ *     core function to access sclp interface
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <linux/timer.h>
+#include <linux/reboot.h>
+#include <linux/jiffies.h>
+#include <asm/types.h>
+#include <asm/s390_ext.h>
+
+#include "sclp.h"
+
+#define SCLP_HEADER		"sclp: "
+
+/* Structure for register_early_external_interrupt. */
+static ext_int_info_t ext_int_info_hwc;
+
+/* Lock to protect internal data consistency. */
+static DEFINE_SPINLOCK(sclp_lock);
+
+/* Mask of events that we can receive from the sclp interface. */
+static sccb_mask_t sclp_receive_mask;
+
+/* Mask of events that we can send to the sclp interface. */
+static sccb_mask_t sclp_send_mask;
+
+/* List of registered event listeners and senders. */
+static struct list_head sclp_reg_list;
+
+/* List of queued requests. */
+static struct list_head sclp_req_queue;
+
+/* Data for read and and init requests. */
+static struct sclp_req sclp_read_req;
+static struct sclp_req sclp_init_req;
+static char sclp_read_sccb[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
+static char sclp_init_sccb[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
+
+/* Timer for request retries. */
+static struct timer_list sclp_request_timer;
+
+/* Internal state: is the driver initialized? */
+static volatile enum sclp_init_state_t {
+	sclp_init_state_uninitialized,
+	sclp_init_state_initializing,
+	sclp_init_state_initialized
+} sclp_init_state = sclp_init_state_uninitialized;
+
+/* Internal state: is a request active at the sclp? */
+static volatile enum sclp_running_state_t {
+	sclp_running_state_idle,
+	sclp_running_state_running
+} sclp_running_state = sclp_running_state_idle;
+
+/* Internal state: is a read request pending? */
+static volatile enum sclp_reading_state_t {
+	sclp_reading_state_idle,
+	sclp_reading_state_reading
+} sclp_reading_state = sclp_reading_state_idle;
+
+/* Internal state: is the driver currently serving requests? */
+static volatile enum sclp_activation_state_t {
+	sclp_activation_state_active,
+	sclp_activation_state_deactivating,
+	sclp_activation_state_inactive,
+	sclp_activation_state_activating
+} sclp_activation_state = sclp_activation_state_active;
+
+/* Internal state: is an init mask request pending? */
+static volatile enum sclp_mask_state_t {
+	sclp_mask_state_idle,
+	sclp_mask_state_initializing
+} sclp_mask_state = sclp_mask_state_idle;
+
+/* Maximum retry counts */
+#define SCLP_INIT_RETRY		3
+#define SCLP_MASK_RETRY		3
+#define SCLP_REQUEST_RETRY	3
+
+/* Timeout intervals in seconds.*/
+#define SCLP_BUSY_INTERVAL	2
+#define SCLP_RETRY_INTERVAL	5
+
+static void sclp_process_queue(void);
+static int sclp_init_mask(int calculate);
+static int sclp_init(void);
+
+/* Perform service call. Return 0 on success, non-zero otherwise. */
+static int
+service_call(sclp_cmdw_t command, void *sccb)
+{
+	int cc;
+
+	__asm__ __volatile__(
+		"   .insn rre,0xb2200000,%1,%2\n"  /* servc %1,%2 */
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=&d" (cc)
+		: "d" (command), "a" (__pa(sccb))
+		: "cc", "memory" );
+	if (cc == 3)
+		return -EIO;
+	if (cc == 2)
+		return -EBUSY;
+	return 0;
+}
+
+/* Request timeout handler. Restart the request queue. If DATA is non-zero,
+ * force restart of running request. */
+static void
+sclp_request_timeout(unsigned long data)
+{
+	unsigned long flags;
+
+	if (data) {
+		spin_lock_irqsave(&sclp_lock, flags);
+		sclp_running_state = sclp_running_state_idle;
+		spin_unlock_irqrestore(&sclp_lock, flags);
+	}
+	sclp_process_queue();
+}
+
+/* Set up request retry timer. Called while sclp_lock is locked. */
+static inline void
+__sclp_set_request_timer(unsigned long time, void (*function)(unsigned long),
+			 unsigned long data)
+{
+	del_timer(&sclp_request_timer);
+	sclp_request_timer.function = function;
+	sclp_request_timer.data = data;
+	sclp_request_timer.expires = jiffies + time;
+	add_timer(&sclp_request_timer);
+}
+
+/* Try to start a request. Return zero if the request was successfully
+ * started or if it will be started at a later time. Return non-zero otherwise.
+ * Called while sclp_lock is locked. */
+static int
+__sclp_start_request(struct sclp_req *req)
+{
+	int rc;
+
+	if (sclp_running_state != sclp_running_state_idle)
+		return 0;
+	del_timer(&sclp_request_timer);
+	if (req->start_count <= SCLP_REQUEST_RETRY) {
+		rc = service_call(req->command, req->sccb);
+		req->start_count++;
+	} else
+		rc = -EIO;
+	if (rc == 0) {
+		/* Sucessfully started request */
+		req->status = SCLP_REQ_RUNNING;
+		sclp_running_state = sclp_running_state_running;
+		__sclp_set_request_timer(SCLP_RETRY_INTERVAL * HZ,
+					 sclp_request_timeout, 1);
+		return 0;
+	} else if (rc == -EBUSY) {
+		/* Try again later */
+		__sclp_set_request_timer(SCLP_BUSY_INTERVAL * HZ,
+					 sclp_request_timeout, 0);
+		return 0;
+	}
+	/* Request failed */
+	req->status = SCLP_REQ_FAILED;
+	return rc;
+}
+
+/* Try to start queued requests. */
+static void
+sclp_process_queue(void)
+{
+	struct sclp_req *req;
+	int rc;
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (sclp_running_state != sclp_running_state_idle) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return;
+	}
+	del_timer(&sclp_request_timer);
+	while (!list_empty(&sclp_req_queue)) {
+		req = list_entry(sclp_req_queue.next, struct sclp_req, list);
+		rc = __sclp_start_request(req);
+		if (rc == 0)
+			break;
+		/* Request failed. */
+		list_del(&req->list);
+		if (req->callback) {
+			spin_unlock_irqrestore(&sclp_lock, flags);
+			req->callback(req, req->callback_data);
+			spin_lock_irqsave(&sclp_lock, flags);
+		}
+	}
+	spin_unlock_irqrestore(&sclp_lock, flags);
+}
+
+/* Queue a new request. Return zero on success, non-zero otherwise. */
+int
+sclp_add_request(struct sclp_req *req)
+{
+	unsigned long flags;
+	int rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	if ((sclp_init_state != sclp_init_state_initialized ||
+	     sclp_activation_state != sclp_activation_state_active) &&
+	    req != &sclp_init_req) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return -EIO;
+	}
+	req->status = SCLP_REQ_QUEUED;
+	req->start_count = 0;
+	list_add_tail(&req->list, &sclp_req_queue);
+	rc = 0;
+	/* Start if request is first in list */
+	if (req->list.prev == &sclp_req_queue) {
+		rc = __sclp_start_request(req);
+		if (rc)
+			list_del(&req->list);
+	}
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
+}
+
+EXPORT_SYMBOL(sclp_add_request);
+
+/* Dispatch events found in request buffer to registered listeners. Return 0
+ * if all events were dispatched, non-zero otherwise. */
+static int
+sclp_dispatch_evbufs(struct sccb_header *sccb)
+{
+	unsigned long flags;
+	struct evbuf_header *evbuf;
+	struct list_head *l;
+	struct sclp_register *reg;
+	int offset;
+	int rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	rc = 0;
+	for (offset = sizeof(struct sccb_header); offset < sccb->length;
+	     offset += evbuf->length) {
+		/* Search for event handler */
+		evbuf = (struct evbuf_header *) ((addr_t) sccb + offset);
+		reg = NULL;
+		list_for_each(l, &sclp_reg_list) {
+			reg = list_entry(l, struct sclp_register, list);
+			if (reg->receive_mask & (1 << (32 - evbuf->type)))
+				break;
+			else
+				reg = NULL;
+		}
+		if (reg && reg->receiver_fn) {
+			spin_unlock_irqrestore(&sclp_lock, flags);
+			reg->receiver_fn(evbuf);
+			spin_lock_irqsave(&sclp_lock, flags);
+		} else if (reg == NULL)
+			rc = -ENOSYS;
+	}
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
+}
+
+/* Read event data request callback. */
+static void
+sclp_read_cb(struct sclp_req *req, void *data)
+{
+	unsigned long flags;
+	struct sccb_header *sccb;
+
+	sccb = (struct sccb_header *) req->sccb;
+	if (req->status == SCLP_REQ_DONE && (sccb->response_code == 0x20 ||
+	    sccb->response_code == 0x220))
+		sclp_dispatch_evbufs(sccb);
+	spin_lock_irqsave(&sclp_lock, flags);
+	sclp_reading_state = sclp_reading_state_idle;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+}
+
+/* Prepare read event data request. Called while sclp_lock is locked. */
+static inline void
+__sclp_make_read_req(void)
+{
+	struct sccb_header *sccb;
+
+	sccb = (struct sccb_header *) sclp_read_sccb;
+	clear_page(sccb);
+	memset(&sclp_read_req, 0, sizeof(struct sclp_req));
+	sclp_read_req.command = SCLP_CMDW_READDATA;
+	sclp_read_req.status = SCLP_REQ_QUEUED;
+	sclp_read_req.start_count = 0;
+	sclp_read_req.callback = sclp_read_cb;
+	sclp_read_req.sccb = sccb;
+	sccb->length = PAGE_SIZE;
+	sccb->function_code = 0;
+	sccb->control_mask[2] = 0x80;
+}
+
+/* Search request list for request with matching sccb. Return request if found,
+ * NULL otherwise. Called while sclp_lock is locked. */
+static inline struct sclp_req *
+__sclp_find_req(u32 sccb)
+{
+	struct list_head *l;
+	struct sclp_req *req;
+
+	list_for_each(l, &sclp_req_queue) {
+		req = list_entry(l, struct sclp_req, list);
+		if (sccb == (u32) (addr_t) req->sccb)
+				return req;
+	}
+	return NULL;
+}
+
+/* Handler for external interruption. Perform request post-processing.
+ * Prepare read event data request if necessary. Start processing of next
+ * request on queue. */
+static void
+sclp_interrupt_handler(struct pt_regs *regs, __u16 code)
+{
+	struct sclp_req *req;
+	u32 finished_sccb;
+	u32 evbuf_pending;
+
+	spin_lock(&sclp_lock);
+	finished_sccb = S390_lowcore.ext_params & 0xfffffff8;
+	evbuf_pending = S390_lowcore.ext_params & 0x3;
+	if (finished_sccb) {
+		req = __sclp_find_req(finished_sccb);
+		if (req) {
+			/* Request post-processing */
+			list_del(&req->list);
+			req->status = SCLP_REQ_DONE;
+			if (req->callback) {
+				spin_unlock(&sclp_lock);
+				req->callback(req, req->callback_data);
+				spin_lock(&sclp_lock);
+			}
+		}
+		sclp_running_state = sclp_running_state_idle;
+	}
+	if (evbuf_pending && sclp_receive_mask != 0 &&
+	    sclp_reading_state == sclp_reading_state_idle &&
+	    sclp_activation_state == sclp_activation_state_active ) {
+		sclp_reading_state = sclp_reading_state_reading;
+		__sclp_make_read_req();
+		/* Add request to head of queue */
+		list_add(&sclp_read_req.list, &sclp_req_queue);
+	}
+	spin_unlock(&sclp_lock);
+	sclp_process_queue();
+}
+
+/* Return current Time-Of-Day clock. */
+static inline u64
+sclp_get_clock(void)
+{
+	u64 result;
+
+	asm volatile ("STCK 0(%1)" : "=m" (result) : "a" (&(result)) : "cc");
+	return result;
+}
+
+/* Convert interval in jiffies to TOD ticks. */
+static inline u64
+sclp_tod_from_jiffies(unsigned long jiffies)
+{
+	return (u64) (jiffies / HZ) << 32;
+}
+
+/* Wait until a currently running request finished. Note: while this function
+ * is running, no timers are served on the calling CPU. */
+void
+sclp_sync_wait(void)
+{
+	unsigned long psw_mask;
+	unsigned long cr0, cr0_sync;
+	u64 timeout;
+
+	/* We'll be disabling timer interrupts, so we need a custom timeout
+	 * mechanism */
+	timeout = 0;
+	if (timer_pending(&sclp_request_timer)) {
+		/* Get timeout TOD value */
+		timeout = sclp_get_clock() +
+			  sclp_tod_from_jiffies(sclp_request_timer.expires -
+						jiffies);
+	}
+	/* Prevent bottom half from executing once we force interrupts open */
+	local_bh_disable();
+	/* Enable service-signal interruption, disable timer interrupts */
+	__ctl_store(cr0, 0, 0);
+	cr0_sync = cr0;
+	cr0_sync |= 0x00000200;
+	cr0_sync &= 0xFFFFF3AC;
+	__ctl_load(cr0_sync, 0, 0);
+	asm volatile ("STOSM 0(%1),0x01"
+		      : "=m" (psw_mask) : "a" (&psw_mask) : "memory");
+	/* Loop until driver state indicates finished request */
+	while (sclp_running_state != sclp_running_state_idle) {
+		/* Check for expired request timer */
+		if (timer_pending(&sclp_request_timer) &&
+		    sclp_get_clock() > timeout &&
+		    del_timer(&sclp_request_timer))
+			sclp_request_timer.function(sclp_request_timer.data);
+		barrier();
+		cpu_relax();
+	}
+	/* Restore interrupt settings */
+	asm volatile ("SSM 0(%0)"
+		      : : "a" (&psw_mask) : "memory");
+	__ctl_load(cr0, 0, 0);
+	__local_bh_enable();
+}
+
+EXPORT_SYMBOL(sclp_sync_wait);
+
+/* Dispatch changes in send and receive mask to registered listeners. */
+static inline void
+sclp_dispatch_state_change(void)
+{
+	struct list_head *l;
+	struct sclp_register *reg;
+	unsigned long flags;
+	sccb_mask_t receive_mask;
+	sccb_mask_t send_mask;
+
+	do {
+		spin_lock_irqsave(&sclp_lock, flags);
+		reg = NULL;
+		list_for_each(l, &sclp_reg_list) {
+			reg = list_entry(l, struct sclp_register, list);
+			receive_mask = reg->receive_mask & sclp_receive_mask;
+			send_mask = reg->send_mask & sclp_send_mask;
+			if (reg->sclp_receive_mask != receive_mask ||
+			    reg->sclp_send_mask != send_mask) {
+				reg->sclp_receive_mask = receive_mask;
+				reg->sclp_send_mask = send_mask;
+				break;
+			} else
+				reg = NULL;
+		}
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		if (reg && reg->state_change_fn)
+			reg->state_change_fn(reg);
+	} while (reg);
+}
+
+struct sclp_statechangebuf {
+	struct evbuf_header	header;
+	u8		validity_sclp_active_facility_mask : 1;
+	u8		validity_sclp_receive_mask : 1;
+	u8		validity_sclp_send_mask : 1;
+	u8		validity_read_data_function_mask : 1;
+	u16		_zeros : 12;
+	u16		mask_length;
+	u64		sclp_active_facility_mask;
+	sccb_mask_t	sclp_receive_mask;
+	sccb_mask_t	sclp_send_mask;
+	u32		read_data_function_mask;
+} __attribute__((packed));
+
+
+/* State change event callback. Inform listeners of changes. */
+static void
+sclp_state_change_cb(struct evbuf_header *evbuf)
+{
+	unsigned long flags;
+	struct sclp_statechangebuf *scbuf;
+
+	scbuf = (struct sclp_statechangebuf *) evbuf;
+	if (scbuf->mask_length != sizeof(sccb_mask_t))
+		return;
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (scbuf->validity_sclp_receive_mask)
+		sclp_receive_mask = scbuf->sclp_receive_mask;
+	if (scbuf->validity_sclp_send_mask)
+		sclp_send_mask = scbuf->sclp_send_mask;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	sclp_dispatch_state_change();
+}
+
+static struct sclp_register sclp_state_change_event = {
+	.receive_mask = EvTyp_StateChange_Mask,
+	.receiver_fn = sclp_state_change_cb
+};
+
+/* Calculate receive and send mask of currently registered listeners.
+ * Called while sclp_lock is locked. */
+static inline void
+__sclp_get_mask(sccb_mask_t *receive_mask, sccb_mask_t *send_mask)
+{
+	struct list_head *l;
+	struct sclp_register *t;
+
+	*receive_mask = 0;
+	*send_mask = 0;
+	list_for_each(l, &sclp_reg_list) {
+		t = list_entry(l, struct sclp_register, list);
+		*receive_mask |= t->receive_mask;
+		*send_mask |= t->send_mask;
+	}
+}
+
+/* Register event listener. Return 0 on success, non-zero otherwise. */
+int
+sclp_register(struct sclp_register *reg)
+{
+	unsigned long flags;
+	sccb_mask_t receive_mask;
+	sccb_mask_t send_mask;
+	int rc;
+
+	rc = sclp_init();
+	if (rc)
+		return rc;
+	spin_lock_irqsave(&sclp_lock, flags);
+	/* Check event mask for collisions */
+	__sclp_get_mask(&receive_mask, &send_mask);
+	if (reg->receive_mask & receive_mask || reg->send_mask & send_mask) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return -EBUSY;
+	}
+	/* Trigger initial state change callback */
+	reg->sclp_receive_mask = 0;
+	reg->sclp_send_mask = 0;
+	list_add(&reg->list, &sclp_reg_list);
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	rc = sclp_init_mask(1);
+	if (rc) {
+		spin_lock_irqsave(&sclp_lock, flags);
+		list_del(&reg->list);
+		spin_unlock_irqrestore(&sclp_lock, flags);
+	}
+	return rc;
+}
+
+EXPORT_SYMBOL(sclp_register);
+
+/* Unregister event listener. */
+void
+sclp_unregister(struct sclp_register *reg)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	list_del(&reg->list);
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	sclp_init_mask(1);
+}
+
+EXPORT_SYMBOL(sclp_unregister);
+
+/* Remove event buffers which are marked processed. Return the number of
+ * remaining event buffers. */
+int
+sclp_remove_processed(struct sccb_header *sccb)
+{
+	struct evbuf_header *evbuf;
+	int unprocessed;
+	u16 remaining;
+
+	evbuf = (struct evbuf_header *) (sccb + 1);
+	unprocessed = 0;
+	remaining = sccb->length - sizeof(struct sccb_header);
+	while (remaining > 0) {
+		remaining -= evbuf->length;
+		if (evbuf->flags & 0x80) {
+			sccb->length -= evbuf->length;
+			memcpy(evbuf, (void *) ((addr_t) evbuf + evbuf->length),
+			       remaining);
+		} else {
+			unprocessed++;
+			evbuf = (struct evbuf_header *)
+					((addr_t) evbuf + evbuf->length);
+		}
+	}
+	return unprocessed;
+}
+
+EXPORT_SYMBOL(sclp_remove_processed);
+
+struct init_sccb {
+	struct sccb_header header;
+	u16 _reserved;
+	u16 mask_length;
+	sccb_mask_t receive_mask;
+	sccb_mask_t send_mask;
+	sccb_mask_t sclp_send_mask;
+	sccb_mask_t sclp_receive_mask;
+} __attribute__((packed));
+
+/* Prepare init mask request. Called while sclp_lock is locked. */
+static inline void
+__sclp_make_init_req(u32 receive_mask, u32 send_mask)
+{
+	struct init_sccb *sccb;
+
+	sccb = (struct init_sccb *) sclp_init_sccb;
+	clear_page(sccb);
+	memset(&sclp_init_req, 0, sizeof(struct sclp_req));
+	sclp_init_req.command = SCLP_CMDW_WRITEMASK;
+	sclp_init_req.status = SCLP_REQ_FILLED;
+	sclp_init_req.start_count = 0;
+	sclp_init_req.callback = NULL;
+	sclp_init_req.callback_data = NULL;
+	sclp_init_req.sccb = sccb;
+	sccb->header.length = sizeof(struct init_sccb);
+	sccb->mask_length = sizeof(sccb_mask_t);
+	sccb->receive_mask = receive_mask;
+	sccb->send_mask = send_mask;
+	sccb->sclp_receive_mask = 0;
+	sccb->sclp_send_mask = 0;
+}
+
+/* Start init mask request. If calculate is non-zero, calculate the mask as
+ * requested by registered listeners. Use zero mask otherwise. Return 0 on
+ * success, non-zero otherwise. */
+static int
+sclp_init_mask(int calculate)
+{
+	unsigned long flags;
+	struct init_sccb *sccb = (struct init_sccb *) sclp_init_sccb;
+	sccb_mask_t receive_mask;
+	sccb_mask_t send_mask;
+	int retry;
+	int rc;
+	unsigned long wait;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	/* Check if interface is in appropriate state */
+	if (sclp_mask_state != sclp_mask_state_idle) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return -EBUSY;
+	}
+	if (sclp_activation_state == sclp_activation_state_inactive) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return -EINVAL;
+	}
+	sclp_mask_state = sclp_mask_state_initializing;
+	/* Determine mask */
+	if (calculate)
+		__sclp_get_mask(&receive_mask, &send_mask);
+	else {
+		receive_mask = 0;
+		send_mask = 0;
+	}
+	rc = -EIO;
+	for (retry = 0; retry <= SCLP_MASK_RETRY; retry++) {
+		/* Prepare request */
+		__sclp_make_init_req(receive_mask, send_mask);
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		if (sclp_add_request(&sclp_init_req)) {
+			/* Try again later */
+			wait = jiffies + SCLP_BUSY_INTERVAL * HZ;
+			while (time_before(jiffies, wait))
+				sclp_sync_wait();
+			spin_lock_irqsave(&sclp_lock, flags);
+			continue;
+		}
+		while (sclp_init_req.status != SCLP_REQ_DONE &&
+		       sclp_init_req.status != SCLP_REQ_FAILED)
+			sclp_sync_wait();
+		spin_lock_irqsave(&sclp_lock, flags);
+		if (sclp_init_req.status == SCLP_REQ_DONE &&
+		    sccb->header.response_code == 0x20) {
+			/* Successful request */
+			if (calculate) {
+				sclp_receive_mask = sccb->sclp_receive_mask;
+				sclp_send_mask = sccb->sclp_send_mask;
+			} else {
+				sclp_receive_mask = 0;
+				sclp_send_mask = 0;
+			}
+			spin_unlock_irqrestore(&sclp_lock, flags);
+			sclp_dispatch_state_change();
+			spin_lock_irqsave(&sclp_lock, flags);
+			rc = 0;
+			break;
+		}
+	}
+	sclp_mask_state = sclp_mask_state_idle;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
+}
+
+/* Deactivate SCLP interface. On success, new requests will be rejected,
+ * events will no longer be dispatched. Return 0 on success, non-zero
+ * otherwise. */
+int
+sclp_deactivate(void)
+{
+	unsigned long flags;
+	int rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	/* Deactivate can only be called when active */
+	if (sclp_activation_state != sclp_activation_state_active) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return -EINVAL;
+	}
+	sclp_activation_state = sclp_activation_state_deactivating;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	rc = sclp_init_mask(0);
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (rc == 0)
+		sclp_activation_state = sclp_activation_state_inactive;
+	else
+		sclp_activation_state = sclp_activation_state_active;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
+}
+
+EXPORT_SYMBOL(sclp_deactivate);
+
+/* Reactivate SCLP interface after sclp_deactivate. On success, new
+ * requests will be accepted, events will be dispatched again. Return 0 on
+ * success, non-zero otherwise. */
+int
+sclp_reactivate(void)
+{
+	unsigned long flags;
+	int rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	/* Reactivate can only be called when inactive */
+	if (sclp_activation_state != sclp_activation_state_inactive) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return -EINVAL;
+	}
+	sclp_activation_state = sclp_activation_state_activating;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	rc = sclp_init_mask(1);
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (rc == 0)
+		sclp_activation_state = sclp_activation_state_active;
+	else
+		sclp_activation_state = sclp_activation_state_inactive;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
+}
+
+EXPORT_SYMBOL(sclp_reactivate);
+
+/* Handler for external interruption used during initialization. Modify
+ * request state to done. */
+static void
+sclp_check_handler(struct pt_regs *regs, __u16 code)
+{
+	u32 finished_sccb;
+
+	finished_sccb = S390_lowcore.ext_params & 0xfffffff8;
+	/* Is this the interrupt we are waiting for? */
+	if (finished_sccb == 0)
+		return;
+	if (finished_sccb != (u32) (addr_t) sclp_init_sccb) {
+		printk(KERN_WARNING SCLP_HEADER "unsolicited interrupt "
+		       "for buffer at 0x%x\n", finished_sccb);
+		return;
+	}
+	spin_lock(&sclp_lock);
+	if (sclp_running_state == sclp_running_state_running) {
+		sclp_init_req.status = SCLP_REQ_DONE;
+		sclp_running_state = sclp_running_state_idle;
+	}
+	spin_unlock(&sclp_lock);
+}
+
+/* Initial init mask request timed out. Modify request state to failed. */
+static void
+sclp_check_timeout(unsigned long data)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (sclp_running_state == sclp_running_state_running) {
+		sclp_init_req.status = SCLP_REQ_FAILED;
+		sclp_running_state = sclp_running_state_idle;
+	}
+	spin_unlock_irqrestore(&sclp_lock, flags);
+}
+
+/* Perform a check of the SCLP interface. Return zero if the interface is
+ * available and there are no pending requests from a previous instance.
+ * Return non-zero otherwise. */
+static int
+sclp_check_interface(void)
+{
+	struct init_sccb *sccb;
+	unsigned long flags;
+	int retry;
+	int rc;
+
+	spin_lock_irqsave(&sclp_lock, flags);
+	/* Prepare init mask command */
+	rc = register_early_external_interrupt(0x2401, sclp_check_handler,
+					       &ext_int_info_hwc);
+	if (rc) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return rc;
+	}
+	for (retry = 0; retry <= SCLP_INIT_RETRY; retry++) {
+		__sclp_make_init_req(0, 0);
+		sccb = (struct init_sccb *) sclp_init_req.sccb;
+		rc = service_call(sclp_init_req.command, sccb);
+		if (rc == -EIO)
+			break;
+		sclp_init_req.status = SCLP_REQ_RUNNING;
+		sclp_running_state = sclp_running_state_running;
+		__sclp_set_request_timer(SCLP_RETRY_INTERVAL * HZ,
+					 sclp_check_timeout, 0);
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		/* Enable service-signal interruption - needs to happen
+		 * with IRQs enabled. */
+		ctl_set_bit(0, 9);
+		/* Wait for signal from interrupt or timeout */
+		sclp_sync_wait();
+		/* Disable service-signal interruption - needs to happen
+		 * with IRQs enabled. */
+		ctl_clear_bit(0,9);
+		spin_lock_irqsave(&sclp_lock, flags);
+		del_timer(&sclp_request_timer);
+		if (sclp_init_req.status == SCLP_REQ_DONE &&
+		    sccb->header.response_code == 0x20) {
+			rc = 0;
+			break;
+		} else
+			rc = -EBUSY;
+	}
+	unregister_early_external_interrupt(0x2401, sclp_check_handler,
+					    &ext_int_info_hwc);
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	return rc;
+}
+
+/* Reboot event handler. Reset send and receive mask to prevent pending SCLP
+ * events from interfering with rebooted system. */
+static int
+sclp_reboot_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	sclp_deactivate();
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block sclp_reboot_notifier = {
+	.notifier_call = sclp_reboot_event
+};
+
+/* Initialize SCLP driver. Return zero if driver is operational, non-zero
+ * otherwise. */
+static int
+sclp_init(void)
+{
+	unsigned long flags;
+	int rc;
+
+	if (!MACHINE_HAS_SCLP)
+		return -ENODEV;
+	spin_lock_irqsave(&sclp_lock, flags);
+	/* Check for previous or running initialization */
+	if (sclp_init_state != sclp_init_state_uninitialized) {
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return 0;
+	}
+	sclp_init_state = sclp_init_state_initializing;
+	/* Set up variables */
+	INIT_LIST_HEAD(&sclp_req_queue);
+	INIT_LIST_HEAD(&sclp_reg_list);
+	list_add(&sclp_state_change_event.list, &sclp_reg_list);
+	init_timer(&sclp_request_timer);
+	/* Check interface */
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	rc = sclp_check_interface();
+	spin_lock_irqsave(&sclp_lock, flags);
+	if (rc) {
+		sclp_init_state = sclp_init_state_uninitialized;
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return rc;
+	}
+	/* Register reboot handler */
+	rc = register_reboot_notifier(&sclp_reboot_notifier);
+	if (rc) {
+		sclp_init_state = sclp_init_state_uninitialized;
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return rc;
+	}
+	/* Register interrupt handler */
+	rc = register_early_external_interrupt(0x2401, sclp_interrupt_handler,
+					       &ext_int_info_hwc);
+	if (rc) {
+		unregister_reboot_notifier(&sclp_reboot_notifier);
+		sclp_init_state = sclp_init_state_uninitialized;
+		spin_unlock_irqrestore(&sclp_lock, flags);
+		return rc;
+	}
+	sclp_init_state = sclp_init_state_initialized;
+	spin_unlock_irqrestore(&sclp_lock, flags);
+	/* Enable service-signal external interruption - needs to happen with
+	 * IRQs enabled. */
+	ctl_set_bit(0, 9);
+	sclp_init_mask(1);
+	return 0;
+}

drivers/s390/char/sclp.h

@@ -0,0 +1,159 @@
+/*
+ *  drivers/s390/char/sclp.h
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#ifndef __SCLP_H__
+#define __SCLP_H__
+
+#include <linux/types.h>
+#include <linux/list.h>
+
+#include <asm/ebcdic.h>
+
+/* maximum number of pages concerning our own memory management */
+#define MAX_KMEM_PAGES (sizeof(unsigned long) << 3)
+#define MAX_CONSOLE_PAGES	4
+
+#define EvTyp_OpCmd		0x01
+#define EvTyp_Msg		0x02
+#define EvTyp_StateChange	0x08
+#define EvTyp_PMsgCmd		0x09
+#define EvTyp_CntlProgOpCmd	0x20
+#define EvTyp_CntlProgIdent	0x0B
+#define EvTyp_SigQuiesce	0x1D
+#define EvTyp_VT220Msg		0x1A
+
+#define EvTyp_OpCmd_Mask	0x80000000
+#define EvTyp_Msg_Mask		0x40000000
+#define EvTyp_StateChange_Mask	0x01000000
+#define EvTyp_PMsgCmd_Mask	0x00800000
+#define EvTyp_CtlProgOpCmd_Mask	0x00000001
+#define EvTyp_CtlProgIdent_Mask	0x00200000
+#define EvTyp_SigQuiesce_Mask	0x00000008
+#define EvTyp_VT220Msg_Mask	0x00000040
+
+#define GnrlMsgFlgs_DOM		0x8000
+#define GnrlMsgFlgs_SndAlrm	0x4000
+#define GnrlMsgFlgs_HoldMsg	0x2000
+
+#define LnTpFlgs_CntlText	0x8000
+#define LnTpFlgs_LabelText	0x4000
+#define LnTpFlgs_DataText	0x2000
+#define LnTpFlgs_EndText	0x1000
+#define LnTpFlgs_PromptText	0x0800
+
+typedef unsigned int sclp_cmdw_t;
+
+#define SCLP_CMDW_READDATA	0x00770005
+#define SCLP_CMDW_WRITEDATA	0x00760005
+#define SCLP_CMDW_WRITEMASK	0x00780005
+
+#define GDS_ID_MDSMU		0x1310
+#define GDS_ID_MDSRouteInfo	0x1311
+#define GDS_ID_AgUnWrkCorr	0x1549
+#define GDS_ID_SNACondReport	0x1532
+#define GDS_ID_CPMSU		0x1212
+#define GDS_ID_RoutTargInstr	0x154D
+#define GDS_ID_OpReq		0x8070
+#define GDS_ID_TextCmd		0x1320
+
+#define GDS_KEY_SelfDefTextMsg	0x31
+
+typedef u32 sccb_mask_t;	/* ATTENTION: assumes 32bit mask !!! */
+
+struct sccb_header {
+	u16	length;
+	u8	function_code;
+	u8	control_mask[3];
+	u16	response_code;
+} __attribute__((packed));
+
+struct gds_subvector {
+	u8	length;
+	u8	key;
+} __attribute__((packed));
+
+struct gds_vector {
+	u16	length;
+	u16	gds_id;
+} __attribute__((packed));
+
+struct evbuf_header {
+	u16	length;
+	u8	type;
+	u8	flags;
+	u16	_reserved;
+} __attribute__((packed));
+
+struct sclp_req {
+	struct list_head list;		/* list_head for request queueing. */
+	sclp_cmdw_t command;		/* sclp command to execute */
+	void	*sccb;			/* pointer to the sccb to execute */
+	char	status;			/* status of this request */
+	int     start_count;		/* number of SVCs done for this req */
+	/* Callback that is called after reaching final status. */
+	void (*callback)(struct sclp_req *, void *data);
+	void *callback_data;
+};
+
+#define SCLP_REQ_FILLED	  0x00	/* request is ready to be processed */
+#define SCLP_REQ_QUEUED	  0x01	/* request is queued to be processed */
+#define SCLP_REQ_RUNNING  0x02	/* request is currently running */
+#define SCLP_REQ_DONE	  0x03	/* request is completed successfully */
+#define SCLP_REQ_FAILED	  0x05	/* request is finally failed */
+
+/* function pointers that a high level driver has to use for registration */
+/* of some routines it wants to be called from the low level driver */
+struct sclp_register {
+	struct list_head list;
+	/* event masks this user is registered for */
+	sccb_mask_t receive_mask;
+	sccb_mask_t send_mask;
+	/* actually present events */
+	sccb_mask_t sclp_receive_mask;
+	sccb_mask_t sclp_send_mask;
+	/* called if event type availability changes */
+	void (*state_change_fn)(struct sclp_register *);
+	/* called for events in cp_receive_mask/sclp_receive_mask */
+	void (*receiver_fn)(struct evbuf_header *);
+};
+
+/* externals from sclp.c */
+int sclp_add_request(struct sclp_req *req);
+void sclp_sync_wait(void);
+int sclp_register(struct sclp_register *reg);
+void sclp_unregister(struct sclp_register *reg);
+int sclp_remove_processed(struct sccb_header *sccb);
+int sclp_deactivate(void);
+int sclp_reactivate(void);
+
+/* useful inlines */
+
+/* VM uses EBCDIC 037, LPAR+native(SE+HMC) use EBCDIC 500 */
+/* translate single character from ASCII to EBCDIC */
+static inline unsigned char
+sclp_ascebc(unsigned char ch)
+{
+	return (MACHINE_IS_VM) ? _ascebc[ch] : _ascebc_500[ch];
+}
+
+/* translate string from EBCDIC to ASCII */
+static inline void
+sclp_ebcasc_str(unsigned char *str, int nr)
+{
+	(MACHINE_IS_VM) ? EBCASC(str, nr) : EBCASC_500(str, nr);
+}
+
+/* translate string from ASCII to EBCDIC */
+static inline void
+sclp_ascebc_str(unsigned char *str, int nr)
+{
+	(MACHINE_IS_VM) ? ASCEBC(str, nr) : ASCEBC_500(str, nr);
+}
+
+#endif	 /* __SCLP_H__ */

drivers/s390/char/sclp_con.c

@@ -0,0 +1,252 @@
+/*
+ *  drivers/s390/char/sclp_con.c
+ *    SCLP line mode console driver
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/kmod.h>
+#include <linux/console.h>
+#include <linux/init.h>
+#include <linux/timer.h>
+#include <linux/jiffies.h>
+#include <linux/bootmem.h>
+#include <linux/err.h>
+
+#include "sclp.h"
+#include "sclp_rw.h"
+#include "sclp_tty.h"
+
+#define SCLP_CON_PRINT_HEADER "sclp console driver: "
+
+#define sclp_console_major 4		/* TTYAUX_MAJOR */
+#define sclp_console_minor 64
+#define sclp_console_name  "ttyS"
+
+/* Lock to guard over changes to global variables */
+static spinlock_t sclp_con_lock;
+/* List of free pages that can be used for console output buffering */
+static struct list_head sclp_con_pages;
+/* List of full struct sclp_buffer structures ready for output */
+static struct list_head sclp_con_outqueue;
+/* Counter how many buffers are emitted (max 1) and how many */
+/* are on the output queue. */
+static int sclp_con_buffer_count;
+/* Pointer to current console buffer */
+static struct sclp_buffer *sclp_conbuf;
+/* Timer for delayed output of console messages */
+static struct timer_list sclp_con_timer;
+
+/* Output format for console messages */
+static unsigned short sclp_con_columns;
+static unsigned short sclp_con_width_htab;
+
+static void
+sclp_conbuf_callback(struct sclp_buffer *buffer, int rc)
+{
+	unsigned long flags;
+	void *page;
+
+	do {
+		page = sclp_unmake_buffer(buffer);
+		spin_lock_irqsave(&sclp_con_lock, flags);
+		/* Remove buffer from outqueue */
+		list_del(&buffer->list);
+		sclp_con_buffer_count--;
+		list_add_tail((struct list_head *) page, &sclp_con_pages);
+		/* Check if there is a pending buffer on the out queue. */
+		buffer = NULL;
+		if (!list_empty(&sclp_con_outqueue))
+			buffer = list_entry(sclp_con_outqueue.next,
+					    struct sclp_buffer, list);
+		spin_unlock_irqrestore(&sclp_con_lock, flags);
+	} while (buffer && sclp_emit_buffer(buffer, sclp_conbuf_callback));
+}
+
+static inline void
+sclp_conbuf_emit(void)
+{
+	struct sclp_buffer* buffer;
+	unsigned long flags;
+	int count;
+	int rc;
+
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	buffer = sclp_conbuf;
+	sclp_conbuf = NULL;
+	if (buffer == NULL) {
+		spin_unlock_irqrestore(&sclp_con_lock, flags);
+		return;
+	}
+	list_add_tail(&buffer->list, &sclp_con_outqueue);
+	count = sclp_con_buffer_count++;
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
+	if (count)
+		return;
+	rc = sclp_emit_buffer(buffer, sclp_conbuf_callback);
+	if (rc)
+		sclp_conbuf_callback(buffer, rc);
+}
+
+/*
+ * When this routine is called from the timer then we flush the
+ * temporary write buffer without further waiting on a final new line.
+ */
+static void
+sclp_console_timeout(unsigned long data)
+{
+	sclp_conbuf_emit();
+}
+
+/*
+ * Writes the given message to S390 system console
+ */
+static void
+sclp_console_write(struct console *console, const char *message,
+		   unsigned int count)
+{
+	unsigned long flags;
+	void *page;
+	int written;
+
+	if (count == 0)
+		return;
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	/*
+	 * process escape characters, write message into buffer,
+	 * send buffer to SCLP
+	 */
+	do {
+		/* make sure we have a console output buffer */
+		if (sclp_conbuf == NULL) {
+			while (list_empty(&sclp_con_pages)) {
+				spin_unlock_irqrestore(&sclp_con_lock, flags);
+				sclp_sync_wait();
+				spin_lock_irqsave(&sclp_con_lock, flags);
+			}
+			page = sclp_con_pages.next;
+			list_del((struct list_head *) page);
+			sclp_conbuf = sclp_make_buffer(page, sclp_con_columns,
+						       sclp_con_width_htab);
+		}
+		/* try to write the string to the current output buffer */
+		written = sclp_write(sclp_conbuf, (const unsigned char *)
+				     message, count);
+		if (written == count)
+			break;
+		/*
+		 * Not all characters could be written to the current
+		 * output buffer. Emit the buffer, create a new buffer
+		 * and then output the rest of the string.
+		 */
+		spin_unlock_irqrestore(&sclp_con_lock, flags);
+		sclp_conbuf_emit();
+		spin_lock_irqsave(&sclp_con_lock, flags);
+		message += written;
+		count -= written;
+	} while (count > 0);
+	/* Setup timer to output current console buffer after 1/10 second */
+	if (sclp_conbuf != NULL && sclp_chars_in_buffer(sclp_conbuf) != 0 &&
+	    !timer_pending(&sclp_con_timer)) {
+		init_timer(&sclp_con_timer);
+		sclp_con_timer.function = sclp_console_timeout;
+		sclp_con_timer.data = 0UL;
+		sclp_con_timer.expires = jiffies + HZ/10;
+		add_timer(&sclp_con_timer);
+	}
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
+}
+
+static struct tty_driver *
+sclp_console_device(struct console *c, int *index)
+{
+	*index = c->index;
+	return sclp_tty_driver;
+}
+
+/*
+ * This routine is called from panic when the kernel
+ * is going to give up. We have to make sure that all buffers
+ * will be flushed to the SCLP.
+ */
+static void
+sclp_console_unblank(void)
+{
+	unsigned long flags;
+
+	sclp_conbuf_emit();
+	spin_lock_irqsave(&sclp_con_lock, flags);
+	if (timer_pending(&sclp_con_timer))
+		del_timer(&sclp_con_timer);
+	while (sclp_con_buffer_count > 0) {
+		spin_unlock_irqrestore(&sclp_con_lock, flags);
+		sclp_sync_wait();
+		spin_lock_irqsave(&sclp_con_lock, flags);
+	}
+	spin_unlock_irqrestore(&sclp_con_lock, flags);
+}
+
+/*
+ * used to register the SCLP console to the kernel and to
+ * give printk necessary information
+ */
+static struct console sclp_console =
+{
+	.name = sclp_console_name,
+	.write = sclp_console_write,
+	.device = sclp_console_device,
+	.unblank = sclp_console_unblank,
+	.flags = CON_PRINTBUFFER,
+	.index = 0 /* ttyS0 */
+};
+
+/*
+ * called by console_init() in drivers/char/tty_io.c at boot-time.
+ */
+static int __init
+sclp_console_init(void)
+{
+	void *page;
+	int i;
+	int rc;
+
+	if (!CONSOLE_IS_SCLP)
+		return 0;
+	rc = sclp_rw_init();
+	if (rc)
+		return rc;
+	/* Allocate pages for output buffering */
+	INIT_LIST_HEAD(&sclp_con_pages);
+	for (i = 0; i < MAX_CONSOLE_PAGES; i++) {
+		page = alloc_bootmem_low_pages(PAGE_SIZE);
+		if (page == NULL)
+			return -ENOMEM;
+		list_add_tail((struct list_head *) page, &sclp_con_pages);
+	}
+	INIT_LIST_HEAD(&sclp_con_outqueue);
+	spin_lock_init(&sclp_con_lock);
+	sclp_con_buffer_count = 0;
+	sclp_conbuf = NULL;
+	init_timer(&sclp_con_timer);
+
+	/* Set output format */
+	if (MACHINE_IS_VM)
+		/*
+		 * save 4 characters for the CPU number
+		 * written at start of each line by VM/CP
+		 */
+		sclp_con_columns = 76;
+	else
+		sclp_con_columns = 80;
+	sclp_con_width_htab = 8;
+
+	/* enable printk-access to this driver */
+	register_console(&sclp_console);
+	return 0;
+}
+
+console_initcall(sclp_console_init);

drivers/s390/char/sclp_cpi.c

@@ -0,0 +1,254 @@
+/*
+ * Author: Martin Peschke <mpeschke@de.ibm.com>
+ * Copyright (C) 2001 IBM Entwicklung GmbH, IBM Corporation
+ *
+ * SCLP Control-Program Identification.
+ */
+
+#include <linux/config.h>
+#include <linux/version.h>
+#include <linux/kmod.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/timer.h>
+#include <linux/string.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <asm/ebcdic.h>
+#include <asm/semaphore.h>
+
+#include "sclp.h"
+#include "sclp_rw.h"
+
+#define CPI_LENGTH_SYSTEM_TYPE	8
+#define CPI_LENGTH_SYSTEM_NAME	8
+#define CPI_LENGTH_SYSPLEX_NAME	8
+
+struct cpi_evbuf {
+	struct evbuf_header header;
+	u8	id_format;
+	u8	reserved0;
+	u8	system_type[CPI_LENGTH_SYSTEM_TYPE];
+	u64	reserved1;
+	u8	system_name[CPI_LENGTH_SYSTEM_NAME];
+	u64	reserved2;
+	u64	system_level;
+	u64	reserved3;
+	u8	sysplex_name[CPI_LENGTH_SYSPLEX_NAME];
+	u8	reserved4[16];
+} __attribute__((packed));
+
+struct cpi_sccb {
+	struct sccb_header header;
+	struct cpi_evbuf cpi_evbuf;
+} __attribute__((packed));
+
+/* Event type structure for write message and write priority message */
+static struct sclp_register sclp_cpi_event =
+{
+	.send_mask = EvTyp_CtlProgIdent_Mask
+};
+
+MODULE_AUTHOR(
+	"Martin Peschke, IBM Deutschland Entwicklung GmbH "
+	"<[email protected]>");
+
+MODULE_DESCRIPTION(
+	"identify this operating system instance to the S/390 "
+	"or zSeries hardware");
+
+static char *system_name = NULL;
+module_param(system_name, charp, 0);
+MODULE_PARM_DESC(system_name, "e.g. hostname - max. 8 characters");
+
+static char *sysplex_name = NULL;
+#ifdef ALLOW_SYSPLEX_NAME
+module_param(sysplex_name, charp, 0);
+MODULE_PARM_DESC(sysplex_name, "if applicable - max. 8 characters");
+#endif
+
+/* use default value for this field (as well as for system level) */
+static char *system_type = "LINUX";
+
+static int
+cpi_check_parms(void)
+{
+	/* reject if no system type specified */
+	if (!system_type) {
+		printk("cpi: bug: no system type specified\n");
+		return -EINVAL;
+	}
+
+	/* reject if system type larger than 8 characters */
+	if (strlen(system_type) > CPI_LENGTH_SYSTEM_NAME) {
+		printk("cpi: bug: system type has length of %li characters - "
+		       "only %i characters supported\n",
+		       strlen(system_type), CPI_LENGTH_SYSTEM_TYPE);
+		return -EINVAL;
+	}
+
+	/* reject if no system name specified */
+	if (!system_name) {
+		printk("cpi: no system name specified\n");
+		return -EINVAL;
+	}
+
+	/* reject if system name larger than 8 characters */
+	if (strlen(system_name) > CPI_LENGTH_SYSTEM_NAME) {
+		printk("cpi: system name has length of %li characters - "
+		       "only %i characters supported\n",
+		       strlen(system_name), CPI_LENGTH_SYSTEM_NAME);
+		return -EINVAL;
+	}
+
+	/* reject if specified sysplex name larger than 8 characters */
+	if (sysplex_name && strlen(sysplex_name) > CPI_LENGTH_SYSPLEX_NAME) {
+		printk("cpi: sysplex name has length of %li characters"
+		       " - only %i characters supported\n",
+		       strlen(sysplex_name), CPI_LENGTH_SYSPLEX_NAME);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static void
+cpi_callback(struct sclp_req *req, void *data)
+{
+	struct semaphore *sem;
+
+	sem = (struct semaphore *) data;
+	up(sem);
+}
+
+static struct sclp_req *
+cpi_prepare_req(void)
+{
+	struct sclp_req *req;
+	struct cpi_sccb *sccb;
+	struct cpi_evbuf *evb;
+
+	req = (struct sclp_req *) kmalloc(sizeof(struct sclp_req), GFP_KERNEL);
+	if (req == NULL)
+		return ERR_PTR(-ENOMEM);
+	sccb = (struct cpi_sccb *) __get_free_page(GFP_KERNEL | GFP_DMA);
+	if (sccb == NULL) {
+		kfree(req);
+		return ERR_PTR(-ENOMEM);
+	}
+	memset(sccb, 0, sizeof(struct cpi_sccb));
+
+	/* setup SCCB for Control-Program Identification */
+	sccb->header.length = sizeof(struct cpi_sccb);
+	sccb->cpi_evbuf.header.length = sizeof(struct cpi_evbuf);
+	sccb->cpi_evbuf.header.type = 0x0B;
+	evb = &sccb->cpi_evbuf;
+
+	/* set system type */
+	memset(evb->system_type, ' ', CPI_LENGTH_SYSTEM_TYPE);
+	memcpy(evb->system_type, system_type, strlen(system_type));
+	sclp_ascebc_str(evb->system_type, CPI_LENGTH_SYSTEM_TYPE);
+	EBC_TOUPPER(evb->system_type, CPI_LENGTH_SYSTEM_TYPE);
+
+	/* set system name */
+	memset(evb->system_name, ' ', CPI_LENGTH_SYSTEM_NAME);
+	memcpy(evb->system_name, system_name, strlen(system_name));
+	sclp_ascebc_str(evb->system_name, CPI_LENGTH_SYSTEM_NAME);
+	EBC_TOUPPER(evb->system_name, CPI_LENGTH_SYSTEM_NAME);
+
+	/* set sytem level */
+	evb->system_level = LINUX_VERSION_CODE;
+
+	/* set sysplex name */
+	if (sysplex_name) {
+		memset(evb->sysplex_name, ' ', CPI_LENGTH_SYSPLEX_NAME);
+		memcpy(evb->sysplex_name, sysplex_name, strlen(sysplex_name));
+		sclp_ascebc_str(evb->sysplex_name, CPI_LENGTH_SYSPLEX_NAME);
+		EBC_TOUPPER(evb->sysplex_name, CPI_LENGTH_SYSPLEX_NAME);
+	}
+
+	/* prepare request data structure presented to SCLP driver */
+	req->command = SCLP_CMDW_WRITEDATA;
+	req->sccb = sccb;
+	req->status = SCLP_REQ_FILLED;
+	req->callback = cpi_callback;
+	return req;
+}
+
+static void
+cpi_free_req(struct sclp_req *req)
+{
+	free_page((unsigned long) req->sccb);
+	kfree(req);
+}
+
+static int __init
+cpi_module_init(void)
+{
+	struct semaphore sem;
+	struct sclp_req *req;
+	int rc;
+
+	rc = cpi_check_parms();
+	if (rc)
+		return rc;
+
+	rc = sclp_register(&sclp_cpi_event);
+	if (rc) {
+		/* could not register sclp event. Die. */
+		printk(KERN_WARNING "cpi: could not register to hardware "
+		       "console.\n");
+		return -EINVAL;
+	}
+	if (!(sclp_cpi_event.sclp_send_mask & EvTyp_CtlProgIdent_Mask)) {
+		printk(KERN_WARNING "cpi: no control program identification "
+		       "support\n");
+		sclp_unregister(&sclp_cpi_event);
+		return -ENOTSUPP;
+	}
+
+	req = cpi_prepare_req();
+	if (IS_ERR(req)) {
+		printk(KERN_WARNING "cpi: couldn't allocate request\n");
+		sclp_unregister(&sclp_cpi_event);
+		return PTR_ERR(req);
+	}
+
+	/* Prepare semaphore */
+	sema_init(&sem, 0);
+	req->callback_data = &sem;
+	/* Add request to sclp queue */
+	rc = sclp_add_request(req);
+	if (rc) {
+		printk(KERN_WARNING "cpi: could not start request\n");
+		cpi_free_req(req);
+		sclp_unregister(&sclp_cpi_event);
+		return rc;
+	}
+	/* make "insmod" sleep until callback arrives */
+	down(&sem);
+
+	rc = ((struct cpi_sccb *) req->sccb)->header.response_code;
+	if (rc != 0x0020) {
+		printk(KERN_WARNING "cpi: failed with response code 0x%x\n",
+		       rc);
+		rc = -ECOMM;
+	} else
+		rc = 0;
+
+	cpi_free_req(req);
+	sclp_unregister(&sclp_cpi_event);
+
+	return rc;
+}
+
+
+static void __exit cpi_module_exit(void)
+{
+}
+
+
+/* declare driver module init/cleanup functions */
+module_init(cpi_module_init);
+module_exit(cpi_module_exit);
+

drivers/s390/char/sclp_quiesce.c

@@ -0,0 +1,99 @@
+/*
+ *  drivers/s390/char/sclp_quiesce.c
+ *     signal quiesce handler
+ *
+ *  (C) Copyright IBM Corp. 1999,2004
+ *  Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *             Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/cpumask.h>
+#include <linux/smp.h>
+#include <linux/init.h>
+#include <asm/atomic.h>
+#include <asm/ptrace.h>
+#include <asm/sigp.h>
+
+#include "sclp.h"
+
+
+#ifdef CONFIG_SMP
+/* Signal completion of shutdown process. All CPUs except the first to enter
+ * this function: go to stopped state. First CPU: wait until all other
+ * CPUs are in stopped or check stop state. Afterwards, load special PSW
+ * to indicate completion. */
+static void
+do_load_quiesce_psw(void * __unused)
+{
+	static atomic_t cpuid = ATOMIC_INIT(-1);
+	psw_t quiesce_psw;
+	int cpu;
+
+	if (atomic_compare_and_swap(-1, smp_processor_id(), &cpuid))
+		signal_processor(smp_processor_id(), sigp_stop);
+	/* Wait for all other cpus to enter stopped state */
+	for_each_online_cpu(cpu) {
+		if (cpu == smp_processor_id())
+			continue;
+		while(!smp_cpu_not_running(cpu))
+			cpu_relax();
+	}
+	/* Quiesce the last cpu with the special psw */
+	quiesce_psw.mask = PSW_BASE_BITS | PSW_MASK_WAIT;
+	quiesce_psw.addr = 0xfff;
+	__load_psw(quiesce_psw);
+}
+
+/* Shutdown handler. Perform shutdown function on all CPUs. */
+static void
+do_machine_quiesce(void)
+{
+	on_each_cpu(do_load_quiesce_psw, NULL, 0, 0);
+}
+#else
+/* Shutdown handler. Signal completion of shutdown by loading special PSW. */
+static void
+do_machine_quiesce(void)
+{
+	psw_t quiesce_psw;
+
+	quiesce_psw.mask = PSW_BASE_BITS | PSW_MASK_WAIT;
+	quiesce_psw.addr = 0xfff;
+	__load_psw(quiesce_psw);
+}
+#endif
+
+extern void ctrl_alt_del(void);
+
+/* Handler for quiesce event. Start shutdown procedure. */
+static void
+sclp_quiesce_handler(struct evbuf_header *evbuf)
+{
+	_machine_restart = (void *) do_machine_quiesce;
+	_machine_halt = do_machine_quiesce;
+	_machine_power_off = do_machine_quiesce;
+	ctrl_alt_del();
+}
+
+static struct sclp_register sclp_quiesce_event = {
+	.receive_mask = EvTyp_SigQuiesce_Mask,
+	.receiver_fn = sclp_quiesce_handler
+};
+
+/* Initialize quiesce driver. */
+static int __init
+sclp_quiesce_init(void)
+{
+	int rc;
+
+	rc = sclp_register(&sclp_quiesce_event);
+	if (rc)
+		printk(KERN_WARNING "sclp: could not register quiesce handler "
+		       "(rc=%d)\n", rc);
+	return rc;
+}
+
+module_init(sclp_quiesce_init);

drivers/s390/char/sclp_rw.c

@@ -0,0 +1,471 @@
+/*
+ *  drivers/s390/char/sclp_rw.c
+ *     driver: reading from and writing to system console on S/390 via SCLP
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/kmod.h>
+#include <linux/types.h>
+#include <linux/err.h>
+#include <linux/string.h>
+#include <linux/spinlock.h>
+#include <linux/ctype.h>
+#include <asm/uaccess.h>
+
+#include "sclp.h"
+#include "sclp_rw.h"
+
+#define SCLP_RW_PRINT_HEADER "sclp low level driver: "
+
+/*
+ * The room for the SCCB (only for writing) is not equal to a pages size
+ * (as it is specified as the maximum size in the the SCLP ducumentation)
+ * because of the additional data structure described above.
+ */
+#define MAX_SCCB_ROOM (PAGE_SIZE - sizeof(struct sclp_buffer))
+
+/* Event type structure for write message and write priority message */
+static struct sclp_register sclp_rw_event = {
+	.send_mask = EvTyp_Msg_Mask | EvTyp_PMsgCmd_Mask
+};
+
+/*
+ * Setup a sclp write buffer. Gets a page as input (4K) and returns
+ * a pointer to a struct sclp_buffer structure that is located at the
+ * end of the input page. This reduces the buffer space by a few
+ * bytes but simplifies things.
+ */
+struct sclp_buffer *
+sclp_make_buffer(void *page, unsigned short columns, unsigned short htab)
+{
+	struct sclp_buffer *buffer;
+	struct write_sccb *sccb;
+
+	sccb = (struct write_sccb *) page;
+	/*
+	 * We keep the struct sclp_buffer structure at the end
+	 * of the sccb page.
+	 */
+	buffer = ((struct sclp_buffer *) ((addr_t) sccb + PAGE_SIZE)) - 1;
+	buffer->sccb = sccb;
+	buffer->retry_count = 0;
+	buffer->mto_number = 0;
+	buffer->mto_char_sum = 0;
+	buffer->current_line = NULL;
+	buffer->current_length = 0;
+	buffer->columns = columns;
+	buffer->htab = htab;
+
+	/* initialize sccb */
+	memset(sccb, 0, sizeof(struct write_sccb));
+	sccb->header.length = sizeof(struct write_sccb);
+	sccb->msg_buf.header.length = sizeof(struct msg_buf);
+	sccb->msg_buf.header.type = EvTyp_Msg;
+	sccb->msg_buf.mdb.header.length = sizeof(struct mdb);
+	sccb->msg_buf.mdb.header.type = 1;
+	sccb->msg_buf.mdb.header.tag = 0xD4C4C240;	/* ebcdic "MDB " */
+	sccb->msg_buf.mdb.header.revision_code = 1;
+	sccb->msg_buf.mdb.go.length = sizeof(struct go);
+	sccb->msg_buf.mdb.go.type = 1;
+
+	return buffer;
+}
+
+/*
+ * Return a pointer to the orignal page that has been used to create
+ * the buffer.
+ */
+void *
+sclp_unmake_buffer(struct sclp_buffer *buffer)
+{
+	return buffer->sccb;
+}
+
+/*
+ * Initialize a new Message Text Object (MTO) at the end of the provided buffer
+ * with enough room for max_len characters. Return 0 on success.
+ */
+static int
+sclp_initialize_mto(struct sclp_buffer *buffer, int max_len)
+{
+	struct write_sccb *sccb;
+	struct mto *mto;
+	int mto_size;
+
+	/* max size of new Message Text Object including message text  */
+	mto_size = sizeof(struct mto) + max_len;
+
+	/* check if current buffer sccb can contain the mto */
+	sccb = buffer->sccb;
+	if ((MAX_SCCB_ROOM - sccb->header.length) < mto_size)
+		return -ENOMEM;
+
+	/* find address of new message text object */
+	mto = (struct mto *)(((addr_t) sccb) + sccb->header.length);
+
+	/*
+	 * fill the new Message-Text Object,
+	 * starting behind the former last byte of the SCCB
+	 */
+	memset(mto, 0, sizeof(struct mto));
+	mto->length = sizeof(struct mto);
+	mto->type = 4;	/* message text object */
+	mto->line_type_flags = LnTpFlgs_EndText; /* end text */
+
+	/* set pointer to first byte after struct mto. */
+	buffer->current_line = (char *) (mto + 1);
+	buffer->current_length = 0;
+
+	return 0;
+}
+
+/*
+ * Finalize MTO initialized by sclp_initialize_mto(), updating the sizes of
+ * MTO, enclosing MDB, event buffer and SCCB.
+ */
+static void
+sclp_finalize_mto(struct sclp_buffer *buffer)
+{
+	struct write_sccb *sccb;
+	struct mto *mto;
+	int str_len, mto_size;
+
+	str_len = buffer->current_length;
+	buffer->current_line = NULL;
+	buffer->current_length = 0;
+
+	/* real size of new Message Text Object including message text	*/
+	mto_size = sizeof(struct mto) + str_len;
+
+	/* find address of new message text object */
+	sccb = buffer->sccb;
+	mto = (struct mto *)(((addr_t) sccb) + sccb->header.length);
+
+	/* set size of message text object */
+	mto->length = mto_size;
+
+	/*
+	 * update values of sizes
+	 * (SCCB, Event(Message) Buffer, Message Data Block)
+	 */
+	sccb->header.length += mto_size;
+	sccb->msg_buf.header.length += mto_size;
+	sccb->msg_buf.mdb.header.length += mto_size;
+
+	/*
+	 * count number of buffered messages (= number of Message Text
+	 * Objects) and number of buffered characters
+	 * for the SCCB currently used for buffering and at all
+	 */
+	buffer->mto_number++;
+	buffer->mto_char_sum += str_len;
+}
+
+/*
+ * processing of a message including escape characters,
+ * returns number of characters written to the output sccb
+ * ("processed" means that is not guaranteed that the character have already
+ *  been sent to the SCLP but that it will be done at least next time the SCLP
+ *  is not busy)
+ */
+int
+sclp_write(struct sclp_buffer *buffer, const unsigned char *msg, int count)
+{
+	int spaces, i_msg;
+	int rc;
+
+	/*
+	 * parse msg for escape sequences (\t,\v ...) and put formated
+	 * msg into an mto (created by sclp_initialize_mto).
+	 *
+	 * We have to do this work ourselfs because there is no support for
+	 * these characters on the native machine and only partial support
+	 * under VM (Why does VM interpret \n but the native machine doesn't ?)
+	 *
+	 * Depending on i/o-control setting the message is always written
+	 * immediately or we wait for a final new line maybe coming with the
+	 * next message. Besides we avoid a buffer overrun by writing its
+	 * content.
+	 *
+	 * RESTRICTIONS:
+	 *
+	 * \r and \b work within one line because we are not able to modify
+	 * previous output that have already been accepted by the SCLP.
+	 *
+	 * \t combined with following \r is not correctly represented because
+	 * \t is expanded to some spaces but \r does not know about a
+	 * previous \t and decreases the current position by one column.
+	 * This is in order to a slim and quick implementation.
+	 */
+	for (i_msg = 0; i_msg < count; i_msg++) {
+		switch (msg[i_msg]) {
+		case '\n':	/* new line, line feed (ASCII)	*/
+			/* check if new mto needs to be created */
+			if (buffer->current_line == NULL) {
+				rc = sclp_initialize_mto(buffer, 0);
+				if (rc)
+					return i_msg;
+			}
+			sclp_finalize_mto(buffer);
+			break;
+		case '\a':	/* bell, one for several times	*/
+			/* set SCLP sound alarm bit in General Object */
+			buffer->sccb->msg_buf.mdb.go.general_msg_flags |=
+				GnrlMsgFlgs_SndAlrm;
+			break;
+		case '\t':	/* horizontal tabulator	 */
+			/* check if new mto needs to be created */
+			if (buffer->current_line == NULL) {
+				rc = sclp_initialize_mto(buffer,
+							 buffer->columns);
+				if (rc)
+					return i_msg;
+			}
+			/* "go to (next htab-boundary + 1, same line)" */
+			do {
+				if (buffer->current_length >= buffer->columns)
+					break;
+				/* ok, add a blank */
+				*buffer->current_line++ = 0x40;
+				buffer->current_length++;
+			} while (buffer->current_length % buffer->htab);
+			break;
+		case '\f':	/* form feed  */
+		case '\v':	/* vertical tabulator  */
+			/* "go to (actual column, actual line + 1)" */
+			/* = new line, leading spaces */
+			if (buffer->current_line != NULL) {
+				spaces = buffer->current_length;
+				sclp_finalize_mto(buffer);
+				rc = sclp_initialize_mto(buffer,
+							 buffer->columns);
+				if (rc)
+					return i_msg;
+				memset(buffer->current_line, 0x40, spaces);
+				buffer->current_line += spaces;
+				buffer->current_length = spaces;
+			} else {
+				/* one an empty line this is the same as \n */
+				rc = sclp_initialize_mto(buffer,
+							 buffer->columns);
+				if (rc)
+					return i_msg;
+				sclp_finalize_mto(buffer);
+			}
+			break;
+		case '\b':	/* backspace  */
+			/* "go to (actual column - 1, actual line)" */
+			/* decrement counter indicating position, */
+			/* do not remove last character */
+			if (buffer->current_line != NULL &&
+			    buffer->current_length > 0) {
+				buffer->current_length--;
+				buffer->current_line--;
+			}
+			break;
+		case 0x00:	/* end of string  */
+			/* transfer current line to SCCB */
+			if (buffer->current_line != NULL)
+				sclp_finalize_mto(buffer);
+			/* skip the rest of the message including the 0 byte */
+			i_msg = count - 1;
+			break;
+		default:	/* no escape character	*/
+			/* do not output unprintable characters */
+			if (!isprint(msg[i_msg]))
+				break;
+			/* check if new mto needs to be created */
+			if (buffer->current_line == NULL) {
+				rc = sclp_initialize_mto(buffer,
+							 buffer->columns);
+				if (rc)
+					return i_msg;
+			}
+			*buffer->current_line++ = sclp_ascebc(msg[i_msg]);
+			buffer->current_length++;
+			break;
+		}
+		/* check if current mto is full */
+		if (buffer->current_line != NULL &&
+		    buffer->current_length >= buffer->columns)
+			sclp_finalize_mto(buffer);
+	}
+
+	/* return number of processed characters */
+	return i_msg;
+}
+
+/*
+ * Return the number of free bytes in the sccb
+ */
+int
+sclp_buffer_space(struct sclp_buffer *buffer)
+{
+	int count;
+
+	count = MAX_SCCB_ROOM - buffer->sccb->header.length;
+	if (buffer->current_line != NULL)
+		count -= sizeof(struct mto) + buffer->current_length;
+	return count;
+}
+
+/*
+ * Return number of characters in buffer
+ */
+int
+sclp_chars_in_buffer(struct sclp_buffer *buffer)
+{
+	int count;
+
+	count = buffer->mto_char_sum;
+	if (buffer->current_line != NULL)
+		count += buffer->current_length;
+	return count;
+}
+
+/*
+ * sets or provides some values that influence the drivers behaviour
+ */
+void
+sclp_set_columns(struct sclp_buffer *buffer, unsigned short columns)
+{
+	buffer->columns = columns;
+	if (buffer->current_line != NULL &&
+	    buffer->current_length > buffer->columns)
+		sclp_finalize_mto(buffer);
+}
+
+void
+sclp_set_htab(struct sclp_buffer *buffer, unsigned short htab)
+{
+	buffer->htab = htab;
+}
+
+/*
+ * called by sclp_console_init and/or sclp_tty_init
+ */
+int
+sclp_rw_init(void)
+{
+	static int init_done = 0;
+	int rc;
+
+	if (init_done)
+		return 0;
+
+	rc = sclp_register(&sclp_rw_event);
+	if (rc == 0)
+		init_done = 1;
+	return rc;
+}
+
+#define SCLP_BUFFER_MAX_RETRY		1
+
+/*
+ * second half of Write Event Data-function that has to be done after
+ * interruption indicating completion of Service Call.
+ */
+static void
+sclp_writedata_callback(struct sclp_req *request, void *data)
+{
+	int rc;
+	struct sclp_buffer *buffer;
+	struct write_sccb *sccb;
+
+	buffer = (struct sclp_buffer *) data;
+	sccb = buffer->sccb;
+
+	if (request->status == SCLP_REQ_FAILED) {
+		if (buffer->callback != NULL)
+			buffer->callback(buffer, -EIO);
+		return;
+	}
+	/* check SCLP response code and choose suitable action	*/
+	switch (sccb->header.response_code) {
+	case 0x0020 :
+		/* Normal completion, buffer processed, message(s) sent */
+		rc = 0;
+		break;
+
+	case 0x0340: /* Contained SCLP equipment check */
+		if (++buffer->retry_count > SCLP_BUFFER_MAX_RETRY) {
+			rc = -EIO;
+			break;
+		}
+		/* remove processed buffers and requeue rest */
+		if (sclp_remove_processed((struct sccb_header *) sccb) > 0) {
+			/* not all buffers were processed */
+			sccb->header.response_code = 0x0000;
+			buffer->request.status = SCLP_REQ_FILLED;
+			rc = sclp_add_request(request);
+			if (rc == 0)
+				return;
+		} else
+			rc = 0;
+		break;
+
+	case 0x0040: /* SCLP equipment check */
+	case 0x05f0: /* Target resource in improper state */
+		if (++buffer->retry_count > SCLP_BUFFER_MAX_RETRY) {
+			rc = -EIO;
+			break;
+		}
+		/* retry request */
+		sccb->header.response_code = 0x0000;
+		buffer->request.status = SCLP_REQ_FILLED;
+		rc = sclp_add_request(request);
+		if (rc == 0)
+			return;
+		break;
+	default:
+		if (sccb->header.response_code == 0x71f0)
+			rc = -ENOMEM;
+		else
+			rc = -EINVAL;
+		break;
+	}
+	if (buffer->callback != NULL)
+		buffer->callback(buffer, rc);
+}
+
+/*
+ * Setup the request structure in the struct sclp_buffer to do SCLP Write
+ * Event Data and pass the request to the core SCLP loop. Return zero on
+ * success, non-zero otherwise.
+ */
+int
+sclp_emit_buffer(struct sclp_buffer *buffer,
+		 void (*callback)(struct sclp_buffer *, int))
+{
+	struct write_sccb *sccb;
+
+	/* add current line if there is one */
+	if (buffer->current_line != NULL)
+		sclp_finalize_mto(buffer);
+
+	/* Are there messages in the output buffer ? */
+	if (buffer->mto_number == 0)
+		return -EIO;
+
+	sccb = buffer->sccb;
+	if (sclp_rw_event.sclp_send_mask & EvTyp_Msg_Mask)
+		/* Use normal write message */
+		sccb->msg_buf.header.type = EvTyp_Msg;
+	else if (sclp_rw_event.sclp_send_mask & EvTyp_PMsgCmd_Mask)
+		/* Use write priority message */
+		sccb->msg_buf.header.type = EvTyp_PMsgCmd;
+	else
+		return -ENOSYS;
+	buffer->request.command = SCLP_CMDW_WRITEDATA;
+	buffer->request.status = SCLP_REQ_FILLED;
+	buffer->request.callback = sclp_writedata_callback;
+	buffer->request.callback_data = buffer;
+	buffer->request.sccb = sccb;
+	buffer->callback = callback;
+	return sclp_add_request(&buffer->request);
+}

drivers/s390/char/sclp_rw.h

@@ -0,0 +1,96 @@
+/*
+ *  drivers/s390/char/sclp_rw.h
+ *    interface to the SCLP-read/write driver
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#ifndef __SCLP_RW_H__
+#define __SCLP_RW_H__
+
+#include <linux/list.h>
+
+struct mto {
+	u16 length;
+	u16 type;
+	u16 line_type_flags;
+	u8  alarm_control;
+	u8  _reserved[3];
+} __attribute__((packed));
+
+struct go {
+	u16 length;
+	u16 type;
+	u32 domid;
+	u8  hhmmss_time[8];
+	u8  th_time[3];
+	u8  reserved_0;
+	u8  dddyyyy_date[7];
+	u8  _reserved_1;
+	u16 general_msg_flags;
+	u8  _reserved_2[10];
+	u8  originating_system_name[8];
+	u8  job_guest_name[8];
+} __attribute__((packed));
+
+struct mdb_header {
+	u16 length;
+	u16 type;
+	u32 tag;
+	u32 revision_code;
+} __attribute__((packed));
+
+struct mdb {
+	struct mdb_header header;
+	struct go go;
+} __attribute__((packed));
+
+struct msg_buf {
+	struct evbuf_header header;
+	struct mdb mdb;
+} __attribute__((packed));
+
+struct write_sccb {
+	struct sccb_header header;
+	struct msg_buf msg_buf;
+} __attribute__((packed));
+
+/* The number of empty mto buffers that can be contained in a single sccb. */
+#define NR_EMPTY_MTO_PER_SCCB ((PAGE_SIZE - sizeof(struct sclp_buffer) - \
+			sizeof(struct write_sccb)) / sizeof(struct mto))
+
+/*
+ * data structure for information about list of SCCBs (only for writing),
+ * will be located at the end of a SCCBs page
+ */
+struct sclp_buffer {
+	struct list_head list;		/* list_head for sccb_info chain */
+	struct sclp_req request;
+	struct write_sccb *sccb;
+	char *current_line;
+	int current_length;
+	int retry_count;
+	/* output format settings */
+	unsigned short columns;
+	unsigned short htab;
+	/* statistics about this buffer */
+	unsigned int mto_char_sum;	/* # chars in sccb */
+	unsigned int mto_number;	/* # mtos in sccb */
+	/* Callback that is called after reaching final status. */
+	void (*callback)(struct sclp_buffer *, int);
+};
+
+int sclp_rw_init(void);
+struct sclp_buffer *sclp_make_buffer(void *, unsigned short, unsigned short);
+void *sclp_unmake_buffer(struct sclp_buffer *);
+int sclp_buffer_space(struct sclp_buffer *);
+int sclp_write(struct sclp_buffer *buffer, const unsigned char *, int);
+int sclp_emit_buffer(struct sclp_buffer *,void (*)(struct sclp_buffer *,int));
+void sclp_set_columns(struct sclp_buffer *, unsigned short);
+void sclp_set_htab(struct sclp_buffer *, unsigned short);
+int sclp_chars_in_buffer(struct sclp_buffer *);
+
+#endif	/* __SCLP_RW_H__ */

drivers/s390/char/sclp_tty.c

@@ -0,0 +1,813 @@
+/*
+ *  drivers/s390/char/sclp_tty.c
+ *    SCLP line mode terminal driver.
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/kmod.h>
+#include <linux/tty.h>
+#include <linux/tty_driver.h>
+#include <linux/sched.h>
+#include <linux/wait.h>
+#include <linux/slab.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <asm/uaccess.h>
+
+#include "ctrlchar.h"
+#include "sclp.h"
+#include "sclp_rw.h"
+#include "sclp_tty.h"
+
+#define SCLP_TTY_PRINT_HEADER "sclp tty driver: "
+
+/*
+ * size of a buffer that collects single characters coming in
+ * via sclp_tty_put_char()
+ */
+#define SCLP_TTY_BUF_SIZE 512
+
+/*
+ * There is exactly one SCLP terminal, so we can keep things simple
+ * and allocate all variables statically.
+ */
+
+/* Lock to guard over changes to global variables. */
+static spinlock_t sclp_tty_lock;
+/* List of free pages that can be used for console output buffering. */
+static struct list_head sclp_tty_pages;
+/* List of full struct sclp_buffer structures ready for output. */
+static struct list_head sclp_tty_outqueue;
+/* Counter how many buffers are emitted. */
+static int sclp_tty_buffer_count;
+/* Pointer to current console buffer. */
+static struct sclp_buffer *sclp_ttybuf;
+/* Timer for delayed output of console messages. */
+static struct timer_list sclp_tty_timer;
+/* Waitqueue to wait for buffers to get empty. */
+static wait_queue_head_t sclp_tty_waitq;
+
+static struct tty_struct *sclp_tty;
+static unsigned char sclp_tty_chars[SCLP_TTY_BUF_SIZE];
+static unsigned short int sclp_tty_chars_count;
+
+struct tty_driver *sclp_tty_driver;
+
+extern struct termios  tty_std_termios;
+
+static struct sclp_ioctls sclp_ioctls;
+static struct sclp_ioctls sclp_ioctls_init =
+{
+	8,			/* 1 hor. tab. = 8 spaces */
+	0,			/* no echo of input by this driver */
+	80,			/* 80 characters/line */
+	1,			/* write after 1/10 s without final new line */
+	MAX_KMEM_PAGES,		/* quick fix: avoid __alloc_pages */
+	MAX_KMEM_PAGES,		/* take 32/64 pages from kernel memory, */
+	0,			/* do not convert to lower case */
+	0x6c			/* to seprate upper and lower case */
+				/* ('%' in EBCDIC) */
+};
+
+/* This routine is called whenever we try to open a SCLP terminal. */
+static int
+sclp_tty_open(struct tty_struct *tty, struct file *filp)
+{
+	sclp_tty = tty;
+	tty->driver_data = NULL;
+	tty->low_latency = 0;
+	return 0;
+}
+
+/* This routine is called when the SCLP terminal is closed. */
+static void
+sclp_tty_close(struct tty_struct *tty, struct file *filp)
+{
+	if (tty->count > 1)
+		return;
+	sclp_tty = NULL;
+}
+
+/* execute commands to control the i/o behaviour of the SCLP tty at runtime */
+static int
+sclp_tty_ioctl(struct tty_struct *tty, struct file * file,
+	       unsigned int cmd, unsigned long arg)
+{
+	unsigned long flags;
+	unsigned int obuf;
+	int check;
+	int rc;
+
+	if (tty->flags & (1 << TTY_IO_ERROR))
+		return -EIO;
+	rc = 0;
+	check = 0;
+	switch (cmd) {
+	case TIOCSCLPSHTAB:
+		/* set width of horizontal tab	*/
+		if (get_user(sclp_ioctls.htab, (unsigned short __user *) arg))
+			rc = -EFAULT;
+		else
+			check = 1;
+		break;
+	case TIOCSCLPGHTAB:
+		/* get width of horizontal tab	*/
+		if (put_user(sclp_ioctls.htab, (unsigned short __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSECHO:
+		/* enable/disable echo of input */
+		if (get_user(sclp_ioctls.echo, (unsigned char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPGECHO:
+		/* Is echo of input enabled ?  */
+		if (put_user(sclp_ioctls.echo, (unsigned char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSCOLS:
+		/* set number of columns for output  */
+		if (get_user(sclp_ioctls.columns, (unsigned short __user *) arg))
+			rc = -EFAULT;
+		else
+			check = 1;
+		break;
+	case TIOCSCLPGCOLS:
+		/* get number of columns for output  */
+		if (put_user(sclp_ioctls.columns, (unsigned short __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSNL:
+		/* enable/disable writing without final new line character  */
+		if (get_user(sclp_ioctls.final_nl, (signed char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPGNL:
+		/* Is writing without final new line character enabled ?  */
+		if (put_user(sclp_ioctls.final_nl, (signed char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSOBUF:
+		/*
+		 * set the maximum buffers size for output, will be rounded
+		 * up to next 4kB boundary and stored as number of SCCBs
+		 * (4kB Buffers) limitation: 256 x 4kB
+		 */
+		if (get_user(obuf, (unsigned int __user *) arg) == 0) {
+			if (obuf & 0xFFF)
+				sclp_ioctls.max_sccb = (obuf >> 12) + 1;
+			else
+				sclp_ioctls.max_sccb = (obuf >> 12);
+		} else
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPGOBUF:
+		/* get the maximum buffers size for output  */
+		obuf = sclp_ioctls.max_sccb << 12;
+		if (put_user(obuf, (unsigned int __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPGKBUF:
+		/* get the number of buffers got from kernel at startup */
+		if (put_user(sclp_ioctls.kmem_sccb, (unsigned short __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSCASE:
+		/* enable/disable conversion from upper to lower case */
+		if (get_user(sclp_ioctls.tolower, (unsigned char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPGCASE:
+		/* Is conversion from upper to lower case of input enabled? */
+		if (put_user(sclp_ioctls.tolower, (unsigned char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSDELIM:
+		/*
+		 * set special character used for separating upper and
+		 * lower case, 0x00 disables this feature
+		 */
+		if (get_user(sclp_ioctls.delim, (unsigned char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPGDELIM:
+		/*
+		 * get special character used for separating upper and
+		 * lower case, 0x00 disables this feature
+		 */
+		if (put_user(sclp_ioctls.delim, (unsigned char __user *) arg))
+			rc = -EFAULT;
+		break;
+	case TIOCSCLPSINIT:
+		/* set initial (default) sclp ioctls  */
+		sclp_ioctls = sclp_ioctls_init;
+		check = 1;
+		break;
+	default:
+		rc = -ENOIOCTLCMD;
+		break;
+	}
+	if (check) {
+		spin_lock_irqsave(&sclp_tty_lock, flags);
+		if (sclp_ttybuf != NULL) {
+			sclp_set_htab(sclp_ttybuf, sclp_ioctls.htab);
+			sclp_set_columns(sclp_ttybuf, sclp_ioctls.columns);
+		}
+		spin_unlock_irqrestore(&sclp_tty_lock, flags);
+	}
+	return rc;
+}
+
+/*
+ * This routine returns the numbers of characters the tty driver
+ * will accept for queuing to be written.  This number is subject
+ * to change as output buffers get emptied, or if the output flow
+ * control is acted. This is not an exact number because not every
+ * character needs the same space in the sccb. The worst case is
+ * a string of newlines. Every newlines creates a new mto which
+ * needs 8 bytes.
+ */
+static int
+sclp_tty_write_room (struct tty_struct *tty)
+{
+	unsigned long flags;
+	struct list_head *l;
+	int count;
+
+	spin_lock_irqsave(&sclp_tty_lock, flags);
+	count = 0;
+	if (sclp_ttybuf != NULL)
+		count = sclp_buffer_space(sclp_ttybuf) / sizeof(struct mto);
+	list_for_each(l, &sclp_tty_pages)
+		count += NR_EMPTY_MTO_PER_SCCB;
+	spin_unlock_irqrestore(&sclp_tty_lock, flags);
+	return count;
+}
+
+static void
+sclp_ttybuf_callback(struct sclp_buffer *buffer, int rc)
+{
+	unsigned long flags;
+	void *page;
+
+	do {
+		page = sclp_unmake_buffer(buffer);
+		spin_lock_irqsave(&sclp_tty_lock, flags);
+		/* Remove buffer from outqueue */
+		list_del(&buffer->list);
+		sclp_tty_buffer_count--;
+		list_add_tail((struct list_head *) page, &sclp_tty_pages);
+		/* Check if there is a pending buffer on the out queue. */
+		buffer = NULL;
+		if (!list_empty(&sclp_tty_outqueue))
+			buffer = list_entry(sclp_tty_outqueue.next,
+					    struct sclp_buffer, list);
+		spin_unlock_irqrestore(&sclp_tty_lock, flags);
+	} while (buffer && sclp_emit_buffer(buffer, sclp_ttybuf_callback));
+	wake_up(&sclp_tty_waitq);
+	/* check if the tty needs a wake up call */
+	if (sclp_tty != NULL) {
+		tty_wakeup(sclp_tty);
+	}
+}
+
+static inline void
+__sclp_ttybuf_emit(struct sclp_buffer *buffer)
+{
+	unsigned long flags;
+	int count;
+	int rc;
+
+	spin_lock_irqsave(&sclp_tty_lock, flags);
+	list_add_tail(&buffer->list, &sclp_tty_outqueue);
+	count = sclp_tty_buffer_count++;
+	spin_unlock_irqrestore(&sclp_tty_lock, flags);
+	if (count)
+		return;
+	rc = sclp_emit_buffer(buffer, sclp_ttybuf_callback);
+	if (rc)
+		sclp_ttybuf_callback(buffer, rc);
+}
+
+/*
+ * When this routine is called from the timer then we flush the
+ * temporary write buffer.
+ */
+static void
+sclp_tty_timeout(unsigned long data)
+{
+	unsigned long flags;
+	struct sclp_buffer *buf;
+
+	spin_lock_irqsave(&sclp_tty_lock, flags);
+	buf = sclp_ttybuf;
+	sclp_ttybuf = NULL;
+	spin_unlock_irqrestore(&sclp_tty_lock, flags);
+
+	if (buf != NULL) {
+		__sclp_ttybuf_emit(buf);
+	}
+}
+
+/*
+ * Write a string to the sclp tty.
+ */
+static void
+sclp_tty_write_string(const unsigned char *str, int count)
+{
+	unsigned long flags;
+	void *page;
+	int written;
+	struct sclp_buffer *buf;
+
+	if (count <= 0)
+		return;
+	spin_lock_irqsave(&sclp_tty_lock, flags);
+	do {
+		/* Create a sclp output buffer if none exists yet */
+		if (sclp_ttybuf == NULL) {
+			while (list_empty(&sclp_tty_pages)) {
+				spin_unlock_irqrestore(&sclp_tty_lock, flags);
+				if (in_interrupt())
+					sclp_sync_wait();
+				else
+					wait_event(sclp_tty_waitq,
+						!list_empty(&sclp_tty_pages));
+				spin_lock_irqsave(&sclp_tty_lock, flags);
+			}
+			page = sclp_tty_pages.next;
+			list_del((struct list_head *) page);
+			sclp_ttybuf = sclp_make_buffer(page,
+						       sclp_ioctls.columns,
+						       sclp_ioctls.htab);
+		}
+		/* try to write the string to the current output buffer */
+		written = sclp_write(sclp_ttybuf, str, count);
+		if (written == count)
+			break;
+		/*
+		 * Not all characters could be written to the current
+		 * output buffer. Emit the buffer, create a new buffer
+		 * and then output the rest of the string.
+		 */
+		buf = sclp_ttybuf;
+		sclp_ttybuf = NULL;
+		spin_unlock_irqrestore(&sclp_tty_lock, flags);
+		__sclp_ttybuf_emit(buf);
+		spin_lock_irqsave(&sclp_tty_lock, flags);
+		str += written;
+		count -= written;
+	} while (count > 0);
+	/* Setup timer to output current console buffer after 1/10 second */
+	if (sclp_ioctls.final_nl) {
+		if (sclp_ttybuf != NULL &&
+		    sclp_chars_in_buffer(sclp_ttybuf) != 0 &&
+		    !timer_pending(&sclp_tty_timer)) {
+			init_timer(&sclp_tty_timer);
+			sclp_tty_timer.function = sclp_tty_timeout;
+			sclp_tty_timer.data = 0UL;
+			sclp_tty_timer.expires = jiffies + HZ/10;
+			add_timer(&sclp_tty_timer);
+		}
+	} else {
+		if (sclp_ttybuf != NULL &&
+		    sclp_chars_in_buffer(sclp_ttybuf) != 0) {
+			buf = sclp_ttybuf;
+			sclp_ttybuf = NULL;
+			spin_unlock_irqrestore(&sclp_tty_lock, flags);
+			__sclp_ttybuf_emit(buf);
+			spin_lock_irqsave(&sclp_tty_lock, flags);
+		}
+	}
+	spin_unlock_irqrestore(&sclp_tty_lock, flags);
+}
+
+/*
+ * This routine is called by the kernel to write a series of characters to the
+ * tty device. The characters may come from user space or kernel space. This
+ * routine will return the number of characters actually accepted for writing.
+ */
+static int
+sclp_tty_write(struct tty_struct *tty, const unsigned char *buf, int count)
+{
+	if (sclp_tty_chars_count > 0) {
+		sclp_tty_write_string(sclp_tty_chars, sclp_tty_chars_count);
+		sclp_tty_chars_count = 0;
+	}
+	sclp_tty_write_string(buf, count);
+	return count;
+}
+
+/*
+ * This routine is called by the kernel to write a single character to the tty
+ * device. If the kernel uses this routine, it must call the flush_chars()
+ * routine (if defined) when it is done stuffing characters into the driver.
+ *
+ * Characters provided to sclp_tty_put_char() are buffered by the SCLP driver.
+ * If the given character is a '\n' the contents of the SCLP write buffer
+ * - including previous characters from sclp_tty_put_char() and strings from
+ * sclp_write() without final '\n' - will be written.
+ */
+static void
+sclp_tty_put_char(struct tty_struct *tty, unsigned char ch)
+{
+	sclp_tty_chars[sclp_tty_chars_count++] = ch;
+	if (ch == '\n' || sclp_tty_chars_count >= SCLP_TTY_BUF_SIZE) {
+		sclp_tty_write_string(sclp_tty_chars, sclp_tty_chars_count);
+		sclp_tty_chars_count = 0;
+	}
+}
+
+/*
+ * This routine is called by the kernel after it has written a series of
+ * characters to the tty device using put_char().
+ */
+static void
+sclp_tty_flush_chars(struct tty_struct *tty)
+{
+	if (sclp_tty_chars_count > 0) {
+		sclp_tty_write_string(sclp_tty_chars, sclp_tty_chars_count);
+		sclp_tty_chars_count = 0;
+	}
+}
+
+/*
+ * This routine returns the number of characters in the write buffer of the
+ * SCLP driver. The provided number includes all characters that are stored
+ * in the SCCB (will be written next time the SCLP is not busy) as well as
+ * characters in the write buffer (will not be written as long as there is a
+ * final line feed missing).
+ */
+static int
+sclp_tty_chars_in_buffer(struct tty_struct *tty)
+{
+	unsigned long flags;
+	struct list_head *l;
+	struct sclp_buffer *t;
+	int count;
+
+	spin_lock_irqsave(&sclp_tty_lock, flags);
+	count = 0;
+	if (sclp_ttybuf != NULL)
+		count = sclp_chars_in_buffer(sclp_ttybuf);
+	list_for_each(l, &sclp_tty_outqueue) {
+		t = list_entry(l, struct sclp_buffer, list);
+		count += sclp_chars_in_buffer(t);
+	}
+	spin_unlock_irqrestore(&sclp_tty_lock, flags);
+	return count;
+}
+
+/*
+ * removes all content from buffers of low level driver
+ */
+static void
+sclp_tty_flush_buffer(struct tty_struct *tty)
+{
+	if (sclp_tty_chars_count > 0) {
+		sclp_tty_write_string(sclp_tty_chars, sclp_tty_chars_count);
+		sclp_tty_chars_count = 0;
+	}
+}
+
+/*
+ * push input to tty
+ */
+static void
+sclp_tty_input(unsigned char* buf, unsigned int count)
+{
+	unsigned int cchar;
+
+	/*
+	 * If this tty driver is currently closed
+	 * then throw the received input away.
+	 */
+	if (sclp_tty == NULL)
+		return;
+	cchar = ctrlchar_handle(buf, count, sclp_tty);
+	switch (cchar & CTRLCHAR_MASK) {
+	case CTRLCHAR_SYSRQ:
+		break;
+	case CTRLCHAR_CTRL:
+		sclp_tty->flip.count++;
+		*sclp_tty->flip.flag_buf_ptr++ = TTY_NORMAL;
+		*sclp_tty->flip.char_buf_ptr++ = cchar;
+		tty_flip_buffer_push(sclp_tty);
+		break;
+	case CTRLCHAR_NONE:
+		/* send (normal) input to line discipline */
+		memcpy(sclp_tty->flip.char_buf_ptr, buf, count);
+		if (count < 2 ||
+		    (strncmp ((const char *) buf + count - 2, "^n", 2) &&
+		     strncmp ((const char *) buf + count - 2, "\0252n", 2))) {
+			sclp_tty->flip.char_buf_ptr[count] = '\n';
+			count++;
+		} else
+			count -= 2;
+		memset(sclp_tty->flip.flag_buf_ptr, TTY_NORMAL, count);
+		sclp_tty->flip.char_buf_ptr += count;
+		sclp_tty->flip.flag_buf_ptr += count;
+		sclp_tty->flip.count += count;
+		tty_flip_buffer_push(sclp_tty);
+		break;
+	}
+}
+
+/*
+ * get a EBCDIC string in upper/lower case,
+ * find out characters in lower/upper case separated by a special character,
+ * modifiy original string,
+ * returns length of resulting string
+ */
+static int
+sclp_switch_cases(unsigned char *buf, int count,
+		  unsigned char delim, int tolower)
+{
+	unsigned char *ip, *op;
+	int toggle;
+
+	/* initially changing case is off */
+	toggle = 0;
+	ip = op = buf;
+	while (count-- > 0) {
+		/* compare with special character */
+		if (*ip == delim) {
+			/* followed by another special character? */
+			if (count && ip[1] == delim) {
+				/*
+				 * ... then put a single copy of the special
+				 * character to the output string
+				 */
+				*op++ = *ip++;
+				count--;
+			} else
+				/*
+				 * ... special character follower by a normal
+				 * character toggles the case change behaviour
+				 */
+				toggle = ~toggle;
+			/* skip special character */
+			ip++;
+		} else
+			/* not the special character */
+			if (toggle)
+				/* but case switching is on */
+				if (tolower)
+					/* switch to uppercase */
+					*op++ = _ebc_toupper[(int) *ip++];
+				else
+					/* switch to lowercase */
+					*op++ = _ebc_tolower[(int) *ip++];
+			else
+				/* no case switching, copy the character */
+				*op++ = *ip++;
+	}
+	/* return length of reformatted string. */
+	return op - buf;
+}
+
+static void
+sclp_get_input(unsigned char *start, unsigned char *end)
+{
+	int count;
+
+	count = end - start;
+	/*
+	 * if set in ioctl convert EBCDIC to lower case
+	 * (modify original input in SCCB)
+	 */
+	if (sclp_ioctls.tolower)
+		EBC_TOLOWER(start, count);
+
+	/*
+	 * if set in ioctl find out characters in lower or upper case
+	 * (depends on current case) separated by a special character,
+	 * works on EBCDIC
+	 */
+	if (sclp_ioctls.delim)
+		count = sclp_switch_cases(start, count,
+					  sclp_ioctls.delim,
+					  sclp_ioctls.tolower);
+
+	/* convert EBCDIC to ASCII (modify original input in SCCB) */
+	sclp_ebcasc_str(start, count);
+
+	/* if set in ioctl write operators input to console  */
+	if (sclp_ioctls.echo)
+		sclp_tty_write(sclp_tty, start, count);
+
+	/* transfer input to high level driver */
+	sclp_tty_input(start, count);
+}
+
+static inline struct gds_vector *
+find_gds_vector(struct gds_vector *start, struct gds_vector *end, u16 id)
+{
+	struct gds_vector *vec;
+
+	for (vec = start; vec < end; vec = (void *) vec + vec->length)
+		if (vec->gds_id == id)
+			return vec;
+	return NULL;
+}
+
+static inline struct gds_subvector *
+find_gds_subvector(struct gds_subvector *start,
+		   struct gds_subvector *end, u8 key)
+{
+	struct gds_subvector *subvec;
+
+	for (subvec = start; subvec < end;
+	     subvec = (void *) subvec + subvec->length)
+		if (subvec->key == key)
+			return subvec;
+	return NULL;
+}
+
+static inline void
+sclp_eval_selfdeftextmsg(struct gds_subvector *start,
+			 struct gds_subvector *end)
+{
+	struct gds_subvector *subvec;
+
+	subvec = start;
+	while (subvec < end) {
+		subvec = find_gds_subvector(subvec, end, 0x30);
+		if (!subvec)
+			break;
+		sclp_get_input((unsigned char *)(subvec + 1),
+			       (unsigned char *) subvec + subvec->length);
+		subvec = (void *) subvec + subvec->length;
+	}
+}
+
+static inline void
+sclp_eval_textcmd(struct gds_subvector *start,
+		  struct gds_subvector *end)
+{
+	struct gds_subvector *subvec;
+
+	subvec = start;
+	while (subvec < end) {
+		subvec = find_gds_subvector(subvec, end,
+					    GDS_KEY_SelfDefTextMsg);
+		if (!subvec)
+			break;
+		sclp_eval_selfdeftextmsg((struct gds_subvector *)(subvec + 1),
+					 (void *)subvec + subvec->length);
+		subvec = (void *) subvec + subvec->length;
+	}
+}
+
+static inline void
+sclp_eval_cpmsu(struct gds_vector *start, struct gds_vector *end)
+{
+	struct gds_vector *vec;
+
+	vec = start;
+	while (vec < end) {
+		vec = find_gds_vector(vec, end, GDS_ID_TextCmd);
+		if (!vec)
+			break;
+		sclp_eval_textcmd((struct gds_subvector *)(vec + 1),
+				  (void *) vec + vec->length);
+		vec = (void *) vec + vec->length;
+	}
+}
+
+
+static inline void
+sclp_eval_mdsmu(struct gds_vector *start, void *end)
+{
+	struct gds_vector *vec;
+
+	vec = find_gds_vector(start, end, GDS_ID_CPMSU);
+	if (vec)
+		sclp_eval_cpmsu(vec + 1, (void *) vec + vec->length);
+}
+
+static void
+sclp_tty_receiver(struct evbuf_header *evbuf)
+{
+	struct gds_vector *start, *end, *vec;
+
+	start = (struct gds_vector *)(evbuf + 1);
+	end = (void *) evbuf + evbuf->length;
+	vec = find_gds_vector(start, end, GDS_ID_MDSMU);
+	if (vec)
+		sclp_eval_mdsmu(vec + 1, (void *) vec + vec->length);
+}
+
+static void
+sclp_tty_state_change(struct sclp_register *reg)
+{
+}
+
+static struct sclp_register sclp_input_event =
+{
+	.receive_mask = EvTyp_OpCmd_Mask | EvTyp_PMsgCmd_Mask,
+	.state_change_fn = sclp_tty_state_change,
+	.receiver_fn = sclp_tty_receiver
+};
+
+static struct tty_operations sclp_ops = {
+	.open = sclp_tty_open,
+	.close = sclp_tty_close,
+	.write = sclp_tty_write,
+	.put_char = sclp_tty_put_char,
+	.flush_chars = sclp_tty_flush_chars,
+	.write_room = sclp_tty_write_room,
+	.chars_in_buffer = sclp_tty_chars_in_buffer,
+	.flush_buffer = sclp_tty_flush_buffer,
+	.ioctl = sclp_tty_ioctl,
+};
+
+int __init
+sclp_tty_init(void)
+{
+	struct tty_driver *driver;
+	void *page;
+	int i;
+	int rc;
+
+	if (!CONSOLE_IS_SCLP)
+		return 0;
+	driver = alloc_tty_driver(1);
+	if (!driver)
+		return -ENOMEM;
+
+	rc = sclp_rw_init();
+	if (rc) {
+		printk(KERN_ERR SCLP_TTY_PRINT_HEADER
+		       "could not register tty - "
+		       "sclp_rw_init returned %d\n", rc);
+		put_tty_driver(driver);
+		return rc;
+	}
+	/* Allocate pages for output buffering */
+	INIT_LIST_HEAD(&sclp_tty_pages);
+	for (i = 0; i < MAX_KMEM_PAGES; i++) {
+		page = (void *) get_zeroed_page(GFP_KERNEL | GFP_DMA);
+		if (page == NULL) {
+			put_tty_driver(driver);
+			return -ENOMEM;
+		}
+		list_add_tail((struct list_head *) page, &sclp_tty_pages);
+	}
+	INIT_LIST_HEAD(&sclp_tty_outqueue);
+	spin_lock_init(&sclp_tty_lock);
+	init_waitqueue_head(&sclp_tty_waitq);
+	init_timer(&sclp_tty_timer);
+	sclp_ttybuf = NULL;
+	sclp_tty_buffer_count = 0;
+	if (MACHINE_IS_VM) {
+		/*
+		 * save 4 characters for the CPU number
+		 * written at start of each line by VM/CP
+		 */
+		sclp_ioctls_init.columns = 76;
+		/* case input lines to lowercase */
+		sclp_ioctls_init.tolower = 1;
+	}
+	sclp_ioctls = sclp_ioctls_init;
+	sclp_tty_chars_count = 0;
+	sclp_tty = NULL;
+
+	rc = sclp_register(&sclp_input_event);
+	if (rc) {
+		put_tty_driver(driver);
+		return rc;
+	}
+
+	driver->owner = THIS_MODULE;
+	driver->driver_name = "sclp_line";
+	driver->name = "sclp_line";
+	driver->major = TTY_MAJOR;
+	driver->minor_start = 64;
+	driver->type = TTY_DRIVER_TYPE_SYSTEM;
+	driver->subtype = SYSTEM_TYPE_TTY;
+	driver->init_termios = tty_std_termios;
+	driver->init_termios.c_iflag = IGNBRK | IGNPAR;
+	driver->init_termios.c_oflag = ONLCR | XTABS;
+	driver->init_termios.c_lflag = ISIG | ECHO;
+	driver->flags = TTY_DRIVER_REAL_RAW;
+	tty_set_operations(driver, &sclp_ops);
+	rc = tty_register_driver(driver);
+	if (rc) {
+		printk(KERN_ERR SCLP_TTY_PRINT_HEADER
+		       "could not register tty - "
+		       "tty_register_driver returned %d\n", rc);
+		put_tty_driver(driver);
+		return rc;
+	}
+	sclp_tty_driver = driver;
+	return 0;
+}
+module_init(sclp_tty_init);

drivers/s390/char/sclp_tty.h

@@ -0,0 +1,71 @@
+/*
+ *  drivers/s390/char/sclp_tty.h
+ *    interface to the SCLP-read/write driver
+ *
+ *  S390 version
+ *    Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Martin Peschke <mpeschke@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#ifndef __SCLP_TTY_H__
+#define __SCLP_TTY_H__
+
+#include <linux/ioctl.h>
+#include <linux/termios.h>
+#include <linux/tty_driver.h>
+
+/* This is the type of data structures storing sclp ioctl setting. */
+struct sclp_ioctls {
+	unsigned short htab;
+	unsigned char echo;
+	unsigned short columns;
+	unsigned char final_nl;
+	unsigned short max_sccb;
+	unsigned short kmem_sccb;	/* can't be modified at run time */
+	unsigned char tolower;
+	unsigned char delim;
+};
+
+/* must be unique, FIXME: must be added in Documentation/ioctl_number.txt */
+#define SCLP_IOCTL_LETTER 'B'
+
+/* set width of horizontal tabulator */
+#define TIOCSCLPSHTAB	_IOW(SCLP_IOCTL_LETTER, 0, unsigned short)
+/* enable/disable echo of input (independent from line discipline) */
+#define TIOCSCLPSECHO	_IOW(SCLP_IOCTL_LETTER, 1, unsigned char)
+/* set number of colums for output */
+#define TIOCSCLPSCOLS	_IOW(SCLP_IOCTL_LETTER, 2, unsigned short)
+/* enable/disable writing without final new line character */
+#define TIOCSCLPSNL	_IOW(SCLP_IOCTL_LETTER, 4, signed char)
+/* set the maximum buffers size for output, rounded up to next 4kB boundary */
+#define TIOCSCLPSOBUF	_IOW(SCLP_IOCTL_LETTER, 5, unsigned short)
+/* set initial (default) sclp ioctls */
+#define TIOCSCLPSINIT	_IO(SCLP_IOCTL_LETTER, 6)
+/* enable/disable conversion from upper to lower case of input */
+#define TIOCSCLPSCASE	_IOW(SCLP_IOCTL_LETTER, 7, unsigned char)
+/* set special character used for separating upper and lower case, */
+/* 0x00 disables this feature */
+#define TIOCSCLPSDELIM	_IOW(SCLP_IOCTL_LETTER, 9, unsigned char)
+
+/* get width of horizontal tabulator */
+#define TIOCSCLPGHTAB	_IOR(SCLP_IOCTL_LETTER, 10, unsigned short)
+/* Is echo of input enabled ? (independent from line discipline) */
+#define TIOCSCLPGECHO	_IOR(SCLP_IOCTL_LETTER, 11, unsigned char)
+/* get number of colums for output */
+#define TIOCSCLPGCOLS	_IOR(SCLP_IOCTL_LETTER, 12, unsigned short)
+/* Is writing without final new line character enabled ? */
+#define TIOCSCLPGNL	_IOR(SCLP_IOCTL_LETTER, 14, signed char)
+/* get the maximum buffers size for output */
+#define TIOCSCLPGOBUF	_IOR(SCLP_IOCTL_LETTER, 15, unsigned short)
+/* Is conversion from upper to lower case of input enabled ? */
+#define TIOCSCLPGCASE	_IOR(SCLP_IOCTL_LETTER, 17, unsigned char)
+/* get special character used for separating upper and lower case, */
+/* 0x00 disables this feature */
+#define TIOCSCLPGDELIM	_IOR(SCLP_IOCTL_LETTER, 19, unsigned char)
+/* get the number of buffers/pages got from kernel at startup */
+#define TIOCSCLPGKBUF	_IOR(SCLP_IOCTL_LETTER, 20, unsigned short)
+
+extern struct tty_driver *sclp_tty_driver;
+
+#endif	/* __SCLP_TTY_H__ */

drivers/s390/char/sclp_vt220.c

@@ -0,0 +1,785 @@
+/*
+ *  drivers/s390/char/sclp_vt220.c
+ *    SCLP VT220 terminal driver.
+ *
+ *  S390 version
+ *    Copyright (C) 2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Peter Oberparleiter <Peter.Oberparleiter@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/list.h>
+#include <linux/wait.h>
+#include <linux/timer.h>
+#include <linux/kernel.h>
+#include <linux/tty.h>
+#include <linux/tty_driver.h>
+#include <linux/sched.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/major.h>
+#include <linux/console.h>
+#include <linux/kdev_t.h>
+#include <linux/bootmem.h>
+#include <linux/interrupt.h>
+#include <linux/init.h>
+#include <asm/uaccess.h>
+#include "sclp.h"
+
+#define SCLP_VT220_PRINT_HEADER 	"sclp vt220 tty driver: "
+#define SCLP_VT220_MAJOR		TTY_MAJOR
+#define SCLP_VT220_MINOR		65
+#define SCLP_VT220_DRIVER_NAME		"sclp_vt220"
+#define SCLP_VT220_DEVICE_NAME		"ttysclp"
+#define SCLP_VT220_CONSOLE_NAME		"ttyS"
+#define SCLP_VT220_CONSOLE_INDEX	1	/* console=ttyS1 */
+#define SCLP_VT220_BUF_SIZE		80
+
+/* Representation of a single write request */
+struct sclp_vt220_request {
+	struct list_head list;
+	struct sclp_req sclp_req;
+	int retry_count;
+};
+
+/* VT220 SCCB */
+struct sclp_vt220_sccb {
+	struct sccb_header header;
+	struct evbuf_header evbuf;
+};
+
+#define SCLP_VT220_MAX_CHARS_PER_BUFFER	(PAGE_SIZE - \
+					 sizeof(struct sclp_vt220_request) - \
+					 sizeof(struct sclp_vt220_sccb))
+
+/* Structures and data needed to register tty driver */
+static struct tty_driver *sclp_vt220_driver;
+
+/* The tty_struct that the kernel associated with us */
+static struct tty_struct *sclp_vt220_tty;
+
+/* Lock to protect internal data from concurrent access */
+static spinlock_t sclp_vt220_lock;
+
+/* List of empty pages to be used as write request buffers */
+static struct list_head sclp_vt220_empty;
+
+/* List of pending requests */
+static struct list_head sclp_vt220_outqueue;
+
+/* Number of requests in outqueue */
+static int sclp_vt220_outqueue_count;
+
+/* Wait queue used to delay write requests while we've run out of buffers */
+static wait_queue_head_t sclp_vt220_waitq;
+
+/* Timer used for delaying write requests to merge subsequent messages into
+ * a single buffer */
+static struct timer_list sclp_vt220_timer;
+
+/* Pointer to current request buffer which has been partially filled but not
+ * yet sent */
+static struct sclp_vt220_request *sclp_vt220_current_request;
+
+/* Number of characters in current request buffer */
+static int sclp_vt220_buffered_chars;
+
+/* Flag indicating whether this driver has already been initialized */
+static int sclp_vt220_initialized = 0;
+
+/* Flag indicating that sclp_vt220_current_request should really
+ * have been already queued but wasn't because the SCLP was processing
+ * another buffer */
+static int sclp_vt220_flush_later;
+
+static void sclp_vt220_receiver_fn(struct evbuf_header *evbuf);
+static int __sclp_vt220_emit(struct sclp_vt220_request *request);
+static void sclp_vt220_emit_current(void);
+
+/* Registration structure for our interest in SCLP event buffers */
+static struct sclp_register sclp_vt220_register = {
+	.send_mask		= EvTyp_VT220Msg_Mask,
+	.receive_mask		= EvTyp_VT220Msg_Mask,
+	.state_change_fn	= NULL,
+	.receiver_fn		= sclp_vt220_receiver_fn
+};
+
+
+/*
+ * Put provided request buffer back into queue and check emit pending
+ * buffers if necessary.
+ */
+static void
+sclp_vt220_process_queue(struct sclp_vt220_request *request)
+{
+	unsigned long flags;
+	void *page;
+
+	do {
+		/* Put buffer back to list of empty buffers */
+		page = request->sclp_req.sccb;
+		spin_lock_irqsave(&sclp_vt220_lock, flags);
+		/* Move request from outqueue to empty queue */
+		list_del(&request->list);
+		sclp_vt220_outqueue_count--;
+		list_add_tail((struct list_head *) page, &sclp_vt220_empty);
+		/* Check if there is a pending buffer on the out queue. */
+		request = NULL;
+		if (!list_empty(&sclp_vt220_outqueue))
+			request = list_entry(sclp_vt220_outqueue.next,
+					     struct sclp_vt220_request, list);
+		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	} while (request && __sclp_vt220_emit(request));
+	if (request == NULL && sclp_vt220_flush_later)
+		sclp_vt220_emit_current();
+	wake_up(&sclp_vt220_waitq);
+	/* Check if the tty needs a wake up call */
+	if (sclp_vt220_tty != NULL) {
+		tty_wakeup(sclp_vt220_tty);
+	}
+}
+
+#define SCLP_BUFFER_MAX_RETRY		1
+
+/*
+ * Callback through which the result of a write request is reported by the
+ * SCLP.
+ */
+static void
+sclp_vt220_callback(struct sclp_req *request, void *data)
+{
+	struct sclp_vt220_request *vt220_request;
+	struct sclp_vt220_sccb *sccb;
+
+	vt220_request = (struct sclp_vt220_request *) data;
+	if (request->status == SCLP_REQ_FAILED) {
+		sclp_vt220_process_queue(vt220_request);
+		return;
+	}
+	sccb = (struct sclp_vt220_sccb *) vt220_request->sclp_req.sccb;
+
+	/* Check SCLP response code and choose suitable action	*/
+	switch (sccb->header.response_code) {
+	case 0x0020 :
+		break;
+
+	case 0x05f0: /* Target resource in improper state */
+		break;
+
+	case 0x0340: /* Contained SCLP equipment check */
+		if (++vt220_request->retry_count > SCLP_BUFFER_MAX_RETRY)
+			break;
+		/* Remove processed buffers and requeue rest */
+		if (sclp_remove_processed((struct sccb_header *) sccb) > 0) {
+			/* Not all buffers were processed */
+			sccb->header.response_code = 0x0000;
+			vt220_request->sclp_req.status = SCLP_REQ_FILLED;
+			if (sclp_add_request(request) == 0)
+				return;
+		}
+		break;
+
+	case 0x0040: /* SCLP equipment check */
+		if (++vt220_request->retry_count > SCLP_BUFFER_MAX_RETRY)
+			break;
+		sccb->header.response_code = 0x0000;
+		vt220_request->sclp_req.status = SCLP_REQ_FILLED;
+		if (sclp_add_request(request) == 0)
+			return;
+		break;
+
+	default:
+		break;
+	}
+	sclp_vt220_process_queue(vt220_request);
+}
+
+/*
+ * Emit vt220 request buffer to SCLP. Return zero on success, non-zero
+ * otherwise.
+ */
+static int
+__sclp_vt220_emit(struct sclp_vt220_request *request)
+{
+	if (!(sclp_vt220_register.sclp_send_mask & EvTyp_VT220Msg_Mask)) {
+		request->sclp_req.status = SCLP_REQ_FAILED;
+		return -EIO;
+	}
+	request->sclp_req.command = SCLP_CMDW_WRITEDATA;
+	request->sclp_req.status = SCLP_REQ_FILLED;
+	request->sclp_req.callback = sclp_vt220_callback;
+	request->sclp_req.callback_data = (void *) request;
+
+	return sclp_add_request(&request->sclp_req);
+}
+
+/*
+ * Queue and emit given request.
+ */
+static void
+sclp_vt220_emit(struct sclp_vt220_request *request)
+{
+	unsigned long flags;
+	int count;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	list_add_tail(&request->list, &sclp_vt220_outqueue);
+	count = sclp_vt220_outqueue_count++;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	/* Emit only the first buffer immediately - callback takes care of
+	 * the rest */
+	if (count == 0 && __sclp_vt220_emit(request))
+		sclp_vt220_process_queue(request);
+}
+
+/*
+ * Queue and emit current request. Return zero on success, non-zero otherwise.
+ */
+static void
+sclp_vt220_emit_current(void)
+{
+	unsigned long flags;
+	struct sclp_vt220_request *request;
+	struct sclp_vt220_sccb *sccb;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	request = NULL;
+	if (sclp_vt220_current_request != NULL) {
+		sccb = (struct sclp_vt220_sccb *) 
+				sclp_vt220_current_request->sclp_req.sccb;
+		/* Only emit buffers with content */
+		if (sccb->header.length != sizeof(struct sclp_vt220_sccb)) {
+			request = sclp_vt220_current_request;
+			sclp_vt220_current_request = NULL;
+			if (timer_pending(&sclp_vt220_timer))
+				del_timer(&sclp_vt220_timer);
+		}
+		sclp_vt220_flush_later = 0;
+	}
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	if (request != NULL)
+		sclp_vt220_emit(request);
+}
+
+#define SCLP_NORMAL_WRITE	0x00
+
+/*
+ * Helper function to initialize a page with the sclp request structure.
+ */
+static struct sclp_vt220_request *
+sclp_vt220_initialize_page(void *page)
+{
+	struct sclp_vt220_request *request;
+	struct sclp_vt220_sccb *sccb;
+
+	/* Place request structure at end of page */
+	request = ((struct sclp_vt220_request *)
+			((addr_t) page + PAGE_SIZE)) - 1;
+	request->retry_count = 0;
+	request->sclp_req.sccb = page;
+	/* SCCB goes at start of page */
+	sccb = (struct sclp_vt220_sccb *) page;
+	memset((void *) sccb, 0, sizeof(struct sclp_vt220_sccb));
+	sccb->header.length = sizeof(struct sclp_vt220_sccb);
+	sccb->header.function_code = SCLP_NORMAL_WRITE;
+	sccb->header.response_code = 0x0000;
+	sccb->evbuf.type = EvTyp_VT220Msg;
+	sccb->evbuf.length = sizeof(struct evbuf_header);
+
+	return request;
+}
+
+static inline unsigned int
+sclp_vt220_space_left(struct sclp_vt220_request *request)
+{
+	struct sclp_vt220_sccb *sccb;
+	sccb = (struct sclp_vt220_sccb *) request->sclp_req.sccb;
+	return PAGE_SIZE - sizeof(struct sclp_vt220_request) -
+	       sccb->header.length;
+}
+
+static inline unsigned int
+sclp_vt220_chars_stored(struct sclp_vt220_request *request)
+{
+	struct sclp_vt220_sccb *sccb;
+	sccb = (struct sclp_vt220_sccb *) request->sclp_req.sccb;
+	return sccb->evbuf.length - sizeof(struct evbuf_header);
+}
+
+/*
+ * Add msg to buffer associated with request. Return the number of characters
+ * added.
+ */
+static int
+sclp_vt220_add_msg(struct sclp_vt220_request *request,
+		   const unsigned char *msg, int count, int convertlf)
+{
+	struct sclp_vt220_sccb *sccb;
+	void *buffer;
+	unsigned char c;
+	int from;
+	int to;
+
+	if (count > sclp_vt220_space_left(request))
+		count = sclp_vt220_space_left(request);
+	if (count <= 0)
+		return 0;
+
+	sccb = (struct sclp_vt220_sccb *) request->sclp_req.sccb;
+	buffer = (void *) ((addr_t) sccb + sccb->header.length);
+
+	if (convertlf) {
+		/* Perform Linefeed conversion (0x0a -> 0x0a 0x0d)*/
+		for (from=0, to=0;
+		     (from < count) && (to < sclp_vt220_space_left(request));
+		     from++) {
+			/* Retrieve character */
+			c = msg[from];
+			/* Perform conversion */
+			if (c == 0x0a) {
+				if (to + 1 < sclp_vt220_space_left(request)) {
+					((unsigned char *) buffer)[to++] = c;
+					((unsigned char *) buffer)[to++] = 0x0d;
+				} else
+					break;
+
+			} else
+				((unsigned char *) buffer)[to++] = c;
+		}
+		sccb->header.length += to;
+		sccb->evbuf.length += to;
+		return from;
+	} else {
+		memcpy(buffer, (const void *) msg, count);
+		sccb->header.length += count;
+		sccb->evbuf.length += count;
+		return count;
+	}
+}
+
+/*
+ * Emit buffer after having waited long enough for more data to arrive.
+ */
+static void
+sclp_vt220_timeout(unsigned long data)
+{
+	sclp_vt220_emit_current();
+}
+
+#define BUFFER_MAX_DELAY	HZ/2
+
+/* 
+ * Internal implementation of the write function. Write COUNT bytes of data
+ * from memory at BUF
+ * to the SCLP interface. In case that the data does not fit into the current
+ * write buffer, emit the current one and allocate a new one. If there are no
+ * more empty buffers available, wait until one gets emptied. If DO_SCHEDULE
+ * is non-zero, the buffer will be scheduled for emitting after a timeout -
+ * otherwise the user has to explicitly call the flush function.
+ * A non-zero CONVERTLF parameter indicates that 0x0a characters in the message
+ * buffer should be converted to 0x0a 0x0d. After completion, return the number
+ * of bytes written.
+ */
+static int
+__sclp_vt220_write(const unsigned char *buf, int count, int do_schedule,
+		   int convertlf)
+{
+	unsigned long flags;
+	void *page;
+	int written;
+	int overall_written;
+
+	if (count <= 0)
+		return 0;
+	overall_written = 0;
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	do {
+		/* Create a sclp output buffer if none exists yet */
+		if (sclp_vt220_current_request == NULL) {
+			while (list_empty(&sclp_vt220_empty)) {
+				spin_unlock_irqrestore(&sclp_vt220_lock,
+						       flags);
+				if (in_interrupt())
+					sclp_sync_wait();
+				else
+					wait_event(sclp_vt220_waitq,
+						!list_empty(&sclp_vt220_empty));
+				spin_lock_irqsave(&sclp_vt220_lock, flags);
+			}
+			page = (void *) sclp_vt220_empty.next;
+			list_del((struct list_head *) page);
+			sclp_vt220_current_request =
+				sclp_vt220_initialize_page(page);
+		}
+		/* Try to write the string to the current request buffer */
+		written = sclp_vt220_add_msg(sclp_vt220_current_request,
+					     buf, count, convertlf);
+		overall_written += written;
+		if (written == count)
+			break;
+		/*
+		 * Not all characters could be written to the current
+		 * output buffer. Emit the buffer, create a new buffer
+		 * and then output the rest of the string.
+		 */
+		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+		sclp_vt220_emit_current();
+		spin_lock_irqsave(&sclp_vt220_lock, flags);
+		buf += written;
+		count -= written;
+	} while (count > 0);
+	/* Setup timer to output current console buffer after some time */
+	if (sclp_vt220_current_request != NULL &&
+	    !timer_pending(&sclp_vt220_timer) && do_schedule) {
+		sclp_vt220_timer.function = sclp_vt220_timeout;
+		sclp_vt220_timer.data = 0UL;
+		sclp_vt220_timer.expires = jiffies + BUFFER_MAX_DELAY;
+		add_timer(&sclp_vt220_timer);
+	}
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	return overall_written;
+}
+
+/*
+ * This routine is called by the kernel to write a series of
+ * characters to the tty device.  The characters may come from
+ * user space or kernel space.  This routine will return the
+ * number of characters actually accepted for writing.
+ */
+static int
+sclp_vt220_write(struct tty_struct *tty, const unsigned char *buf, int count)
+{
+	return __sclp_vt220_write(buf, count, 1, 0);
+}
+
+#define SCLP_VT220_SESSION_ENDED	0x01
+#define	SCLP_VT220_SESSION_STARTED	0x80
+#define SCLP_VT220_SESSION_DATA		0x00
+
+/*
+ * Called by the SCLP to report incoming event buffers.
+ */
+static void
+sclp_vt220_receiver_fn(struct evbuf_header *evbuf)
+{
+	char *buffer;
+	unsigned int count;
+
+	/* Ignore input if device is not open */
+	if (sclp_vt220_tty == NULL)
+		return;
+
+	buffer = (char *) ((addr_t) evbuf + sizeof(struct evbuf_header));
+	count = evbuf->length - sizeof(struct evbuf_header);
+
+	switch (*buffer) {
+	case SCLP_VT220_SESSION_ENDED:
+	case SCLP_VT220_SESSION_STARTED:
+		break;
+	case SCLP_VT220_SESSION_DATA:
+		/* Send input to line discipline */
+		buffer++;
+		count--;
+		/* Prevent buffer overrun by discarding input. Note that
+		 * because buffer_push works asynchronously, we cannot wait
+		 * for the buffer to be emptied. */
+		if (count + sclp_vt220_tty->flip.count > TTY_FLIPBUF_SIZE)
+			count = TTY_FLIPBUF_SIZE - sclp_vt220_tty->flip.count;
+		memcpy(sclp_vt220_tty->flip.char_buf_ptr, buffer, count);
+		memset(sclp_vt220_tty->flip.flag_buf_ptr, TTY_NORMAL, count);
+		sclp_vt220_tty->flip.char_buf_ptr += count;
+		sclp_vt220_tty->flip.flag_buf_ptr += count;
+		sclp_vt220_tty->flip.count += count;
+		tty_flip_buffer_push(sclp_vt220_tty);
+		break;
+	}
+}
+
+/*
+ * This routine is called when a particular tty device is opened.
+ */
+static int
+sclp_vt220_open(struct tty_struct *tty, struct file *filp)
+{
+	if (tty->count == 1) {
+		sclp_vt220_tty = tty;
+		tty->driver_data = kmalloc(SCLP_VT220_BUF_SIZE, GFP_KERNEL);
+		if (tty->driver_data == NULL)
+			return -ENOMEM;
+		tty->low_latency = 0;
+	}
+	return 0;
+}
+
+/*
+ * This routine is called when a particular tty device is closed.
+ */
+static void
+sclp_vt220_close(struct tty_struct *tty, struct file *filp)
+{
+	if (tty->count == 1) {
+		sclp_vt220_tty = NULL;
+		kfree(tty->driver_data);
+		tty->driver_data = NULL;
+	}
+}
+
+/*
+ * This routine is called by the kernel to write a single
+ * character to the tty device.  If the kernel uses this routine,
+ * it must call the flush_chars() routine (if defined) when it is
+ * done stuffing characters into the driver.
+ *
+ * NOTE: include/linux/tty_driver.h specifies that a character should be
+ * ignored if there is no room in the queue. This driver implements a different
+ * semantic in that it will block when there is no more room left.
+ */
+static void
+sclp_vt220_put_char(struct tty_struct *tty, unsigned char ch)
+{
+	__sclp_vt220_write(&ch, 1, 0, 0);
+}
+
+/*
+ * This routine is called by the kernel after it has written a
+ * series of characters to the tty device using put_char().  
+ */
+static void
+sclp_vt220_flush_chars(struct tty_struct *tty)
+{
+	if (sclp_vt220_outqueue_count == 0)
+		sclp_vt220_emit_current();
+	else
+		sclp_vt220_flush_later = 1;
+}
+
+/*
+ * This routine returns the numbers of characters the tty driver
+ * will accept for queuing to be written.  This number is subject
+ * to change as output buffers get emptied, or if the output flow
+ * control is acted.
+ */
+static int
+sclp_vt220_write_room(struct tty_struct *tty)
+{
+	unsigned long flags;
+	struct list_head *l;
+	int count;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	count = 0;
+	if (sclp_vt220_current_request != NULL)
+		count = sclp_vt220_space_left(sclp_vt220_current_request);
+	list_for_each(l, &sclp_vt220_empty)
+		count += SCLP_VT220_MAX_CHARS_PER_BUFFER;
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	return count;
+}
+
+/*
+ * Return number of buffered chars.
+ */
+static int
+sclp_vt220_chars_in_buffer(struct tty_struct *tty)
+{
+	unsigned long flags;
+	struct list_head *l;
+	struct sclp_vt220_request *r;
+	int count;
+
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	count = 0;
+	if (sclp_vt220_current_request != NULL)
+		count = sclp_vt220_chars_stored(sclp_vt220_current_request);
+	list_for_each(l, &sclp_vt220_outqueue) {
+		r = list_entry(l, struct sclp_vt220_request, list);
+		count += sclp_vt220_chars_stored(r);
+	}
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+	return count;
+}
+
+static void
+__sclp_vt220_flush_buffer(void)
+{
+	unsigned long flags;
+
+	sclp_vt220_emit_current();
+	spin_lock_irqsave(&sclp_vt220_lock, flags);
+	if (timer_pending(&sclp_vt220_timer))
+		del_timer(&sclp_vt220_timer);
+	while (sclp_vt220_outqueue_count > 0) {
+		spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+		sclp_sync_wait();
+		spin_lock_irqsave(&sclp_vt220_lock, flags);
+	}
+	spin_unlock_irqrestore(&sclp_vt220_lock, flags);
+}
+
+/*
+ * Pass on all buffers to the hardware. Return only when there are no more
+ * buffers pending.
+ */
+static void
+sclp_vt220_flush_buffer(struct tty_struct *tty)
+{
+	sclp_vt220_emit_current();
+}
+
+/*
+ * Initialize all relevant components and register driver with system.
+ */
+static int
+__sclp_vt220_init(int early)
+{
+	void *page;
+	int i;
+
+	if (sclp_vt220_initialized)
+		return 0;
+	sclp_vt220_initialized = 1;
+	spin_lock_init(&sclp_vt220_lock);
+	INIT_LIST_HEAD(&sclp_vt220_empty);
+	INIT_LIST_HEAD(&sclp_vt220_outqueue);
+	init_waitqueue_head(&sclp_vt220_waitq);
+	init_timer(&sclp_vt220_timer);
+	sclp_vt220_current_request = NULL;
+	sclp_vt220_buffered_chars = 0;
+	sclp_vt220_outqueue_count = 0;
+	sclp_vt220_tty = NULL;
+	sclp_vt220_flush_later = 0;
+
+	/* Allocate pages for output buffering */
+	for (i = 0; i < (early ? MAX_CONSOLE_PAGES : MAX_KMEM_PAGES); i++) {
+		if (early)
+			page = alloc_bootmem_low_pages(PAGE_SIZE);
+		else
+			page = (void *) get_zeroed_page(GFP_KERNEL | GFP_DMA);
+		if (!page)
+			return -ENOMEM;
+		list_add_tail((struct list_head *) page, &sclp_vt220_empty);
+	}
+	return 0;
+}
+
+static struct tty_operations sclp_vt220_ops = {
+	.open = sclp_vt220_open,
+	.close = sclp_vt220_close,
+	.write = sclp_vt220_write,
+	.put_char = sclp_vt220_put_char,
+	.flush_chars = sclp_vt220_flush_chars,
+	.write_room = sclp_vt220_write_room,
+	.chars_in_buffer = sclp_vt220_chars_in_buffer,
+	.flush_buffer = sclp_vt220_flush_buffer
+};
+
+/*
+ * Register driver with SCLP and Linux and initialize internal tty structures.
+ */
+int __init
+sclp_vt220_tty_init(void)
+{
+	struct tty_driver *driver;
+	int rc;
+
+	/* Note: we're not testing for CONSOLE_IS_SCLP here to preserve
+	 * symmetry between VM and LPAR systems regarding ttyS1. */
+	driver = alloc_tty_driver(1);
+	if (!driver)
+		return -ENOMEM;
+	rc = __sclp_vt220_init(0);
+	if (rc) {
+		put_tty_driver(driver);
+		return rc;
+	}
+	rc = sclp_register(&sclp_vt220_register);
+	if (rc) {
+		printk(KERN_ERR SCLP_VT220_PRINT_HEADER
+		       "could not register tty - "
+		       "sclp_register returned %d\n", rc);
+		put_tty_driver(driver);
+		return rc;
+	}
+
+	driver->owner = THIS_MODULE;
+	driver->driver_name = SCLP_VT220_DRIVER_NAME;
+	driver->name = SCLP_VT220_DEVICE_NAME;
+	driver->major = SCLP_VT220_MAJOR;
+	driver->minor_start = SCLP_VT220_MINOR;
+	driver->type = TTY_DRIVER_TYPE_SYSTEM;
+	driver->subtype = SYSTEM_TYPE_TTY;
+	driver->init_termios = tty_std_termios;
+	driver->flags = TTY_DRIVER_REAL_RAW;
+	tty_set_operations(driver, &sclp_vt220_ops);
+
+	rc = tty_register_driver(driver);
+	if (rc) {
+		printk(KERN_ERR SCLP_VT220_PRINT_HEADER
+		       "could not register tty - "
+		       "tty_register_driver returned %d\n", rc);
+		put_tty_driver(driver);
+		return rc;
+	}
+	sclp_vt220_driver = driver;
+	return 0;
+}
+
+module_init(sclp_vt220_tty_init);
+
+#ifdef CONFIG_SCLP_VT220_CONSOLE
+
+static void
+sclp_vt220_con_write(struct console *con, const char *buf, unsigned int count)
+{
+	__sclp_vt220_write((const unsigned char *) buf, count, 1, 1);
+}
+
+static struct tty_driver *
+sclp_vt220_con_device(struct console *c, int *index)
+{
+	*index = 0;
+	return sclp_vt220_driver;
+}
+
+/*
+ * This routine is called from panic when the kernel is going to give up.
+ * We have to make sure that all buffers will be flushed to the SCLP.
+ * Note that this function may be called from within an interrupt context.
+ */
+static void
+sclp_vt220_con_unblank(void)
+{
+	__sclp_vt220_flush_buffer();
+}
+
+/* Structure needed to register with printk */
+static struct console sclp_vt220_console =
+{
+	.name = SCLP_VT220_CONSOLE_NAME,
+	.write = sclp_vt220_con_write,
+	.device = sclp_vt220_con_device,
+	.unblank = sclp_vt220_con_unblank,
+	.flags = CON_PRINTBUFFER,
+	.index = SCLP_VT220_CONSOLE_INDEX
+};
+
+static int __init
+sclp_vt220_con_init(void)
+{
+	int rc;
+
+	if (!CONSOLE_IS_SCLP)
+		return 0;
+	rc = __sclp_vt220_init(1);
+	if (rc)
+		return rc;
+	/* Attach linux console */
+	register_console(&sclp_vt220_console);
+	return 0;
+}
+
+console_initcall(sclp_vt220_con_init);
+#endif /* CONFIG_SCLP_VT220_CONSOLE */
+

drivers/s390/char/tape.h

@@ -0,0 +1,384 @@
+/*
+ *  drivers/s390/char/tape.h
+ *    tape device driver for 3480/3490E/3590 tapes.
+ *
+ *  S390 and zSeries version
+ *    Copyright (C) 2001,2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Carsten Otte <cotte@de.ibm.com>
+ *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#ifndef _TAPE_H
+#define _TAPE_H
+
+#include <asm/ccwdev.h>
+#include <asm/debug.h>
+#include <asm/idals.h>
+#include <linux/config.h>
+#include <linux/blkdev.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/mtio.h>
+#include <linux/interrupt.h>
+#include <linux/workqueue.h>
+
+struct gendisk;
+
+/*
+ * Define DBF_LIKE_HELL for lots of messages in the debug feature.
+ */
+#define DBF_LIKE_HELL
+#ifdef  DBF_LIKE_HELL
+#define DBF_LH(level, str, ...) \
+do { \
+	debug_sprintf_event(TAPE_DBF_AREA, level, str, ## __VA_ARGS__); \
+} while (0)
+#else
+#define DBF_LH(level, str, ...) do {} while(0)
+#endif
+
+/*
+ * macros s390 debug feature (dbf)
+ */
+#define DBF_EVENT(d_level, d_str...) \
+do { \
+	debug_sprintf_event(TAPE_DBF_AREA, d_level, d_str); \
+} while (0)
+
+#define DBF_EXCEPTION(d_level, d_str...) \
+do { \
+	debug_sprintf_exception(TAPE_DBF_AREA, d_level, d_str); \
+} while (0)
+
+#define TAPE_VERSION_MAJOR 2
+#define TAPE_VERSION_MINOR 0
+#define TAPE_MAGIC "tape"
+
+#define TAPE_MINORS_PER_DEV 2	    /* two minors per device */
+#define TAPEBLOCK_HSEC_SIZE	2048
+#define TAPEBLOCK_HSEC_S2B	2
+#define TAPEBLOCK_RETRIES	5
+
+enum tape_medium_state {
+	MS_UNKNOWN,
+	MS_LOADED,
+	MS_UNLOADED,
+	MS_SIZE
+};
+
+enum tape_state {
+	TS_UNUSED=0,
+	TS_IN_USE,
+	TS_BLKUSE,
+	TS_INIT,
+	TS_NOT_OPER,
+	TS_SIZE
+};
+
+enum tape_op {
+	TO_BLOCK,	/* Block read */
+	TO_BSB,		/* Backward space block */
+	TO_BSF,		/* Backward space filemark */
+	TO_DSE,		/* Data security erase */
+	TO_FSB,		/* Forward space block */
+	TO_FSF,		/* Forward space filemark */
+	TO_LBL,		/* Locate block label */
+	TO_NOP,		/* No operation */
+	TO_RBA,		/* Read backward */
+	TO_RBI,		/* Read block information */
+	TO_RFO,		/* Read forward */
+	TO_REW,		/* Rewind tape */
+	TO_RUN,		/* Rewind and unload tape */
+	TO_WRI,		/* Write block */
+	TO_WTM,		/* Write tape mark */
+	TO_MSEN,	/* Medium sense */
+	TO_LOAD,	/* Load tape */
+	TO_READ_CONFIG, /* Read configuration data */
+	TO_READ_ATTMSG, /* Read attention message */
+	TO_DIS,		/* Tape display */
+	TO_ASSIGN,	/* Assign tape to channel path */
+	TO_UNASSIGN,	/* Unassign tape from channel path */
+	TO_SIZE		/* #entries in tape_op_t */
+};
+
+/* Forward declaration */
+struct tape_device;
+
+/* tape_request->status can be: */
+enum tape_request_status {
+	TAPE_REQUEST_INIT,	/* request is ready to be processed */
+	TAPE_REQUEST_QUEUED,	/* request is queued to be processed */
+	TAPE_REQUEST_IN_IO,	/* request is currently in IO */
+	TAPE_REQUEST_DONE,	/* request is completed. */
+};
+
+/* Tape CCW request */
+struct tape_request {
+	struct list_head list;		/* list head for request queueing. */
+	struct tape_device *device;	/* tape device of this request */
+	struct ccw1 *cpaddr;		/* address of the channel program. */
+	void *cpdata;			/* pointer to ccw data. */
+	enum tape_request_status status;/* status of this request */
+	int options;			/* options for execution. */
+	int retries;			/* retry counter for error recovery. */
+	int rescnt;			/* residual count from devstat. */
+
+	/* Callback for delivering final status. */
+	void (*callback)(struct tape_request *, void *);
+	void *callback_data;
+
+	enum tape_op op;
+	int rc;
+};
+
+/* Function type for magnetic tape commands */
+typedef int (*tape_mtop_fn)(struct tape_device *, int);
+
+/* Size of the array containing the mtops for a discipline */
+#define TAPE_NR_MTOPS (MTMKPART+1)
+
+/* Tape Discipline */
+struct tape_discipline {
+	struct module *owner;
+	int  (*setup_device)(struct tape_device *);
+	void (*cleanup_device)(struct tape_device *);
+	int (*irq)(struct tape_device *, struct tape_request *, struct irb *);
+	struct tape_request *(*read_block)(struct tape_device *, size_t);
+	struct tape_request *(*write_block)(struct tape_device *, size_t);
+	void (*process_eov)(struct tape_device*);
+#ifdef CONFIG_S390_TAPE_BLOCK
+	/* Block device stuff. */
+	struct tape_request *(*bread)(struct tape_device *, struct request *);
+	void (*check_locate)(struct tape_device *, struct tape_request *);
+	void (*free_bread)(struct tape_request *);
+#endif
+	/* ioctl function for additional ioctls. */
+	int (*ioctl_fn)(struct tape_device *, unsigned int, unsigned long);
+	/* Array of tape commands with TAPE_NR_MTOPS entries */
+	tape_mtop_fn *mtop_array;
+};
+
+/*
+ * The discipline irq function either returns an error code (<0) which
+ * means that the request has failed with an error or one of the following:
+ */
+#define TAPE_IO_SUCCESS 0	/* request successful */
+#define TAPE_IO_PENDING 1	/* request still running */
+#define TAPE_IO_RETRY	2	/* retry to current request */
+#define TAPE_IO_STOP	3	/* stop the running request */
+
+/* Char Frontend Data */
+struct tape_char_data {
+	struct idal_buffer *idal_buf;	/* idal buffer for user char data */
+	int block_size;			/*   of size block_size. */
+};
+
+#ifdef CONFIG_S390_TAPE_BLOCK
+/* Block Frontend Data */
+struct tape_blk_data
+{
+	/* Block device request queue. */
+	request_queue_t *	request_queue;
+	spinlock_t		request_queue_lock;
+
+	/* Task to move entries from block request to CCS request queue. */
+	struct work_struct	requeue_task;
+	atomic_t		requeue_scheduled;
+
+	/* Current position on the tape. */
+	long			block_position;
+	int			medium_changed;
+	struct gendisk *	disk;
+};
+#endif
+
+/* Tape Info */
+struct tape_device {
+	/* entry in tape_device_list */
+	struct list_head		node;
+
+	int				cdev_id;
+	struct ccw_device *		cdev;
+	struct tape_class_device *	nt;
+	struct tape_class_device *	rt;
+
+	/* Device discipline information. */
+	struct tape_discipline *	discipline;
+	void *				discdata;
+
+	/* Generic status flags */
+	long				tape_generic_status;
+
+	/* Device state information. */
+	wait_queue_head_t		state_change_wq;
+	enum tape_state			tape_state;
+	enum tape_medium_state		medium_state;
+	unsigned char *			modeset_byte;
+
+	/* Reference count. */
+	atomic_t			ref_count;
+
+	/* Request queue. */
+	struct list_head		req_queue;
+
+	/* Each tape device has (currently) two minor numbers. */
+	int				first_minor;
+
+	/* Number of tapemarks required for correct termination. */
+	int				required_tapemarks;
+
+	/* Block ID of the BOF */
+	unsigned int			bof;
+
+	/* Character device frontend data */
+	struct tape_char_data		char_data;
+#ifdef CONFIG_S390_TAPE_BLOCK
+	/* Block dev frontend data */
+	struct tape_blk_data		blk_data;
+#endif
+};
+
+/* Externals from tape_core.c */
+extern struct tape_request *tape_alloc_request(int cplength, int datasize);
+extern void tape_free_request(struct tape_request *);
+extern int tape_do_io(struct tape_device *, struct tape_request *);
+extern int tape_do_io_async(struct tape_device *, struct tape_request *);
+extern int tape_do_io_interruptible(struct tape_device *, struct tape_request *);
+void tape_hotplug_event(struct tape_device *, int major, int action);
+
+static inline int
+tape_do_io_free(struct tape_device *device, struct tape_request *request)
+{
+	int rc;
+
+	rc = tape_do_io(device, request);
+	tape_free_request(request);
+	return rc;
+}
+
+extern int tape_oper_handler(int irq, int status);
+extern void tape_noper_handler(int irq, int status);
+extern int tape_open(struct tape_device *);
+extern int tape_release(struct tape_device *);
+extern int tape_mtop(struct tape_device *, int, int);
+extern void tape_state_set(struct tape_device *, enum tape_state);
+
+extern int tape_generic_online(struct tape_device *, struct tape_discipline *);
+extern int tape_generic_offline(struct tape_device *device);
+
+/* Externals from tape_devmap.c */
+extern int tape_generic_probe(struct ccw_device *);
+extern void tape_generic_remove(struct ccw_device *);
+
+extern struct tape_device *tape_get_device(int devindex);
+extern struct tape_device *tape_get_device_reference(struct tape_device *);
+extern struct tape_device *tape_put_device(struct tape_device *);
+
+/* Externals from tape_char.c */
+extern int tapechar_init(void);
+extern void tapechar_exit(void);
+extern int  tapechar_setup_device(struct tape_device *);
+extern void tapechar_cleanup_device(struct tape_device *);
+
+/* Externals from tape_block.c */
+#ifdef CONFIG_S390_TAPE_BLOCK
+extern int tapeblock_init (void);
+extern void tapeblock_exit(void);
+extern int tapeblock_setup_device(struct tape_device *);
+extern void tapeblock_cleanup_device(struct tape_device *);
+#else
+static inline int tapeblock_init (void) {return 0;}
+static inline void tapeblock_exit (void) {;}
+static inline int tapeblock_setup_device(struct tape_device *t) {return 0;}
+static inline void tapeblock_cleanup_device (struct tape_device *t) {;}
+#endif
+
+/* tape initialisation functions */
+#ifdef CONFIG_PROC_FS
+extern void tape_proc_init (void);
+extern void tape_proc_cleanup (void);
+#else
+static inline void tape_proc_init (void) {;}
+static inline void tape_proc_cleanup (void) {;}
+#endif
+
+/* a function for dumping device sense info */
+extern void tape_dump_sense(struct tape_device *, struct tape_request *,
+			    struct irb *);
+extern void tape_dump_sense_dbf(struct tape_device *, struct tape_request *,
+				struct irb *);
+
+/* functions for handling the status of a device */
+extern void tape_med_state_set(struct tape_device *, enum tape_medium_state);
+
+/* The debug area */
+extern debug_info_t *TAPE_DBF_AREA;
+
+/* functions for building ccws */
+static inline struct ccw1 *
+tape_ccw_cc(struct ccw1 *ccw, __u8 cmd_code, __u16 memsize, void *cda)
+{
+	ccw->cmd_code = cmd_code;
+	ccw->flags = CCW_FLAG_CC;
+	ccw->count = memsize;
+	ccw->cda = (__u32)(addr_t) cda;
+	return ccw + 1;
+}
+
+static inline struct ccw1 *
+tape_ccw_end(struct ccw1 *ccw, __u8 cmd_code, __u16 memsize, void *cda)
+{
+	ccw->cmd_code = cmd_code;
+	ccw->flags = 0;
+	ccw->count = memsize;
+	ccw->cda = (__u32)(addr_t) cda;
+	return ccw + 1;
+}
+
+static inline struct ccw1 *
+tape_ccw_cmd(struct ccw1 *ccw, __u8 cmd_code)
+{
+	ccw->cmd_code = cmd_code;
+	ccw->flags = 0;
+	ccw->count = 0;
+	ccw->cda = (__u32)(addr_t) &ccw->cmd_code;
+	return ccw + 1;
+}
+
+static inline struct ccw1 *
+tape_ccw_repeat(struct ccw1 *ccw, __u8 cmd_code, int count)
+{
+	while (count-- > 0) {
+		ccw->cmd_code = cmd_code;
+		ccw->flags = CCW_FLAG_CC;
+		ccw->count = 0;
+		ccw->cda = (__u32)(addr_t) &ccw->cmd_code;
+		ccw++;
+	}
+	return ccw;
+}
+
+static inline struct ccw1 *
+tape_ccw_cc_idal(struct ccw1 *ccw, __u8 cmd_code, struct idal_buffer *idal)
+{
+	ccw->cmd_code = cmd_code;
+	ccw->flags    = CCW_FLAG_CC;
+	idal_buffer_set_cda(idal, ccw);
+	return ccw++;
+}
+
+static inline struct ccw1 *
+tape_ccw_end_idal(struct ccw1 *ccw, __u8 cmd_code, struct idal_buffer *idal)
+{
+	ccw->cmd_code = cmd_code;
+	ccw->flags    = 0;
+	idal_buffer_set_cda(idal, ccw);
+	return ccw++;
+}
+
+/* Global vars */
+extern const char *tape_state_verbose[];
+extern const char *tape_op_verbose[];
+
+#endif /* for ifdef tape.h */

drivers/s390/char/tape_block.c

@@ -0,0 +1,492 @@
+/*
+ *  drivers/s390/char/tape_block.c
+ *    block device frontend for tape device driver
+ *
+ *  S390 and zSeries version
+ *    Copyright (C) 2001,2003 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Carsten Otte <cotte@de.ibm.com>
+ *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *		 Stefan Bader <shbader@de.ibm.com>
+ */
+
+#include <linux/fs.h>
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/blkdev.h>
+#include <linux/interrupt.h>
+#include <linux/buffer_head.h>
+
+#include <asm/debug.h>
+
+#define TAPE_DBF_AREA	tape_core_dbf
+
+#include "tape.h"
+
+#define PRINTK_HEADER "TAPE_BLOCK: "
+
+#define TAPEBLOCK_MAX_SEC	100
+#define TAPEBLOCK_MIN_REQUEUE	3
+
+/*
+ * 2003/11/25  Stefan Bader <shbader@de.ibm.com>
+ *
+ * In 2.5/2.6 the block device request function is very likely to be called
+ * with disabled interrupts (e.g. generic_unplug_device). So the driver can't
+ * just call any function that tries to allocate CCW requests from that con-
+ * text since it might sleep. There are two choices to work around this:
+ *	a) do not allocate with kmalloc but use its own memory pool
+ *      b) take requests from the queue outside that context, knowing that
+ *         allocation might sleep
+ */
+
+/*
+ * file operation structure for tape block frontend
+ */
+static int tapeblock_open(struct inode *, struct file *);
+static int tapeblock_release(struct inode *, struct file *);
+static int tapeblock_ioctl(struct inode *, struct file *, unsigned int,
+				unsigned long);
+static int tapeblock_medium_changed(struct gendisk *);
+static int tapeblock_revalidate_disk(struct gendisk *);
+
+static struct block_device_operations tapeblock_fops = {
+	.owner		 = THIS_MODULE,
+	.open		 = tapeblock_open,
+	.release	 = tapeblock_release,
+	.ioctl           = tapeblock_ioctl,
+	.media_changed   = tapeblock_medium_changed,
+	.revalidate_disk = tapeblock_revalidate_disk,
+};
+
+static int tapeblock_major = 0;
+
+static void
+tapeblock_trigger_requeue(struct tape_device *device)
+{
+	/* Protect against rescheduling. */
+	if (atomic_compare_and_swap(0, 1, &device->blk_data.requeue_scheduled))
+		return;
+	schedule_work(&device->blk_data.requeue_task);
+}
+
+/*
+ * Post finished request.
+ */
+static inline void
+tapeblock_end_request(struct request *req, int uptodate)
+{
+	if (end_that_request_first(req, uptodate, req->hard_nr_sectors))
+		BUG();
+	end_that_request_last(req);
+}
+
+static void
+__tapeblock_end_request(struct tape_request *ccw_req, void *data)
+{
+	struct tape_device *device;
+	struct request *req;
+
+	DBF_LH(6, "__tapeblock_end_request()\n");
+
+	device = ccw_req->device;
+	req = (struct request *) data;
+	tapeblock_end_request(req, ccw_req->rc == 0);
+	if (ccw_req->rc == 0)
+		/* Update position. */
+		device->blk_data.block_position =
+			(req->sector + req->nr_sectors) >> TAPEBLOCK_HSEC_S2B;
+	else
+		/* We lost the position information due to an error. */
+		device->blk_data.block_position = -1;
+	device->discipline->free_bread(ccw_req);
+	if (!list_empty(&device->req_queue) ||
+	    elv_next_request(device->blk_data.request_queue))
+		tapeblock_trigger_requeue(device);
+}
+
+/*
+ * Feed the tape device CCW queue with requests supplied in a list.
+ */
+static inline int
+tapeblock_start_request(struct tape_device *device, struct request *req)
+{
+	struct tape_request *	ccw_req;
+	int			rc;
+
+	DBF_LH(6, "tapeblock_start_request(%p, %p)\n", device, req);
+
+	ccw_req = device->discipline->bread(device, req);
+	if (IS_ERR(ccw_req)) {
+		DBF_EVENT(1, "TBLOCK: bread failed\n");
+		tapeblock_end_request(req, 0);
+		return PTR_ERR(ccw_req);
+	}
+	ccw_req->callback = __tapeblock_end_request;
+	ccw_req->callback_data = (void *) req;
+	ccw_req->retries = TAPEBLOCK_RETRIES;
+
+	rc = tape_do_io_async(device, ccw_req);
+	if (rc) {
+		/*
+		 * Start/enqueueing failed. No retries in
+		 * this case.
+		 */
+		tapeblock_end_request(req, 0);
+		device->discipline->free_bread(ccw_req);
+	}
+
+	return rc;
+}
+
+/*
+ * Move requests from the block device request queue to the tape device ccw
+ * queue.
+ */
+static void
+tapeblock_requeue(void *data) {
+	struct tape_device *	device;
+	request_queue_t *	queue;
+	int			nr_queued;
+	struct request *	req;
+	struct list_head *	l;
+	int			rc;
+
+	device = (struct tape_device *) data;
+	if (!device)
+		return;
+
+	spin_lock_irq(get_ccwdev_lock(device->cdev));
+	queue  = device->blk_data.request_queue;
+
+	/* Count number of requests on ccw queue. */
+	nr_queued = 0;
+	list_for_each(l, &device->req_queue)
+		nr_queued++;
+	spin_unlock(get_ccwdev_lock(device->cdev));
+
+	spin_lock(&device->blk_data.request_queue_lock);
+	while (
+		!blk_queue_plugged(queue) &&
+		elv_next_request(queue)   &&
+		nr_queued < TAPEBLOCK_MIN_REQUEUE
+	) {
+		req = elv_next_request(queue);
+		if (rq_data_dir(req) == WRITE) {
+			DBF_EVENT(1, "TBLOCK: Rejecting write request\n");
+			blkdev_dequeue_request(req);
+			tapeblock_end_request(req, 0);
+			continue;
+		}
+		spin_unlock_irq(&device->blk_data.request_queue_lock);
+		rc = tapeblock_start_request(device, req);
+		spin_lock_irq(&device->blk_data.request_queue_lock);
+		blkdev_dequeue_request(req);
+		nr_queued++;
+	}
+	spin_unlock_irq(&device->blk_data.request_queue_lock);
+	atomic_set(&device->blk_data.requeue_scheduled, 0);
+}
+
+/*
+ * Tape request queue function. Called from ll_rw_blk.c
+ */
+static void
+tapeblock_request_fn(request_queue_t *queue)
+{
+	struct tape_device *device;
+
+	device = (struct tape_device *) queue->queuedata;
+	DBF_LH(6, "tapeblock_request_fn(device=%p)\n", device);
+	if (device == NULL)
+		BUG();
+
+	tapeblock_trigger_requeue(device);
+}
+
+/*
+ * This function is called for every new tapedevice
+ */
+int
+tapeblock_setup_device(struct tape_device * device)
+{
+	struct tape_blk_data *	blkdat;
+	struct gendisk *	disk;
+	int			rc;
+
+	blkdat = &device->blk_data;
+	spin_lock_init(&blkdat->request_queue_lock);
+	atomic_set(&blkdat->requeue_scheduled, 0);
+
+	blkdat->request_queue = blk_init_queue(
+		tapeblock_request_fn,
+		&blkdat->request_queue_lock
+	);
+	if (!blkdat->request_queue)
+		return -ENOMEM;
+
+	elevator_exit(blkdat->request_queue->elevator);
+	rc = elevator_init(blkdat->request_queue, "noop");
+	if (rc)
+		goto cleanup_queue;
+
+	blk_queue_hardsect_size(blkdat->request_queue, TAPEBLOCK_HSEC_SIZE);
+	blk_queue_max_sectors(blkdat->request_queue, TAPEBLOCK_MAX_SEC);
+	blk_queue_max_phys_segments(blkdat->request_queue, -1L);
+	blk_queue_max_hw_segments(blkdat->request_queue, -1L);
+	blk_queue_max_segment_size(blkdat->request_queue, -1L);
+	blk_queue_segment_boundary(blkdat->request_queue, -1L);
+
+	disk = alloc_disk(1);
+	if (!disk) {
+		rc = -ENOMEM;
+		goto cleanup_queue;
+	}
+
+	disk->major = tapeblock_major;
+	disk->first_minor = device->first_minor;
+	disk->fops = &tapeblock_fops;
+	disk->private_data = tape_get_device_reference(device);
+	disk->queue = blkdat->request_queue;
+	set_capacity(disk, 0);
+	sprintf(disk->disk_name, "btibm%d",
+		device->first_minor / TAPE_MINORS_PER_DEV);
+
+	blkdat->disk = disk;
+	blkdat->medium_changed = 1;
+	blkdat->request_queue->queuedata = tape_get_device_reference(device);
+
+	add_disk(disk);
+
+	INIT_WORK(&blkdat->requeue_task, tapeblock_requeue,
+		tape_get_device_reference(device));
+
+	return 0;
+
+cleanup_queue:
+	blk_cleanup_queue(blkdat->request_queue);
+	blkdat->request_queue = NULL;
+
+	return rc;
+}
+
+void
+tapeblock_cleanup_device(struct tape_device *device)
+{
+	flush_scheduled_work();
+	device->blk_data.requeue_task.data = tape_put_device(device);
+
+	if (!device->blk_data.disk) {
+		PRINT_ERR("(%s): No gendisk to clean up!\n",
+			device->cdev->dev.bus_id);
+		goto cleanup_queue;
+	}
+
+	del_gendisk(device->blk_data.disk);
+	device->blk_data.disk->private_data =
+		tape_put_device(device->blk_data.disk->private_data);
+	put_disk(device->blk_data.disk);
+
+	device->blk_data.disk = NULL;
+cleanup_queue:
+	device->blk_data.request_queue->queuedata = tape_put_device(device);
+
+	blk_cleanup_queue(device->blk_data.request_queue);
+	device->blk_data.request_queue = NULL;
+}
+
+/*
+ * Detect number of blocks of the tape.
+ * FIXME: can we extent this to detect the blocks size as well ?
+ */
+static int
+tapeblock_revalidate_disk(struct gendisk *disk)
+{
+	struct tape_device *	device;
+	unsigned int		nr_of_blks;
+	int			rc;
+
+	device = (struct tape_device *) disk->private_data;
+	if (!device)
+		BUG();
+
+	if (!device->blk_data.medium_changed)
+		return 0;
+
+	PRINT_INFO("Detecting media size...\n");
+	rc = tape_mtop(device, MTFSFM, 1);
+	if (rc)
+		return rc;
+
+	rc = tape_mtop(device, MTTELL, 1);
+	if (rc < 0)
+		return rc;
+
+	DBF_LH(3, "Image file ends at %d\n", rc);
+	nr_of_blks = rc;
+
+	/* This will fail for the first file. Catch the error by checking the
+	 * position. */
+	tape_mtop(device, MTBSF, 1);
+
+	rc = tape_mtop(device, MTTELL, 1);
+	if (rc < 0)
+		return rc;
+
+	if (rc > nr_of_blks)
+		return -EINVAL;
+
+	DBF_LH(3, "Image file starts at %d\n", rc);
+	device->bof = rc;
+	nr_of_blks -= rc;
+
+	PRINT_INFO("Found %i blocks on media\n", nr_of_blks);
+	set_capacity(device->blk_data.disk,
+		nr_of_blks*(TAPEBLOCK_HSEC_SIZE/512));
+
+	device->blk_data.block_position = 0;
+	device->blk_data.medium_changed = 0;
+	return 0;
+}
+
+static int
+tapeblock_medium_changed(struct gendisk *disk)
+{
+	struct tape_device *device;
+
+	device = (struct tape_device *) disk->private_data;
+	DBF_LH(6, "tapeblock_medium_changed(%p) = %d\n",
+		device, device->blk_data.medium_changed);
+
+	return device->blk_data.medium_changed;
+}
+
+/*
+ * Block frontend tape device open function.
+ */
+static int
+tapeblock_open(struct inode *inode, struct file *filp)
+{
+	struct gendisk *	disk;
+	struct tape_device *	device;
+	int			rc;
+
+	disk   = inode->i_bdev->bd_disk;
+	device = tape_get_device_reference(disk->private_data);
+
+	if (device->required_tapemarks) {
+		DBF_EVENT(2, "TBLOCK: missing tapemarks\n");
+		PRINT_ERR("TBLOCK: Refusing to open tape with missing"
+			" end of file marks.\n");
+		rc = -EPERM;
+		goto put_device;
+	}
+
+	rc = tape_open(device);
+	if (rc)
+		goto put_device;
+
+	rc = tapeblock_revalidate_disk(disk);
+	if (rc)
+		goto release;
+
+	/*
+	 * Note: The reference to <device> is hold until the release function
+	 *       is called.
+	 */
+	tape_state_set(device, TS_BLKUSE);
+	return 0;
+
+release:
+	tape_release(device);
+ put_device:
+	tape_put_device(device);
+	return rc;
+}
+
+/*
+ * Block frontend tape device release function.
+ *
+ * Note: One reference to the tape device was made by the open function. So
+ *       we just get the pointer here and release the reference.
+ */
+static int
+tapeblock_release(struct inode *inode, struct file *filp)
+{
+	struct gendisk *disk = inode->i_bdev->bd_disk;
+	struct tape_device *device = disk->private_data;
+
+	tape_state_set(device, TS_IN_USE);
+	tape_release(device);
+	tape_put_device(device);
+
+	return 0;
+}
+
+/*
+ * Support of some generic block device IOCTLs.
+ */
+static int
+tapeblock_ioctl(
+	struct inode *		inode,
+	struct file *		file,
+	unsigned int		command,
+	unsigned long		arg
+) {
+	int rc;
+	int minor;
+	struct gendisk *disk = inode->i_bdev->bd_disk;
+	struct tape_device *device = disk->private_data;
+
+	rc     = 0;
+	disk   = inode->i_bdev->bd_disk;
+	if (!disk)
+		BUG();
+	device = disk->private_data;
+	if (!device)
+		BUG();
+	minor  = iminor(inode);
+
+	DBF_LH(6, "tapeblock_ioctl(0x%0x)\n", command);
+	DBF_LH(6, "device = %d:%d\n", tapeblock_major, minor);
+
+	switch (command) {
+		/* Refuse some IOCTL calls without complaining (mount). */
+		case 0x5310:		/* CDROMMULTISESSION */
+			rc = -EINVAL;
+			break;
+		default:
+			PRINT_WARN("invalid ioctl 0x%x\n", command);
+			rc = -EINVAL;
+	}
+
+	return rc;
+}
+
+/*
+ * Initialize block device frontend.
+ */
+int
+tapeblock_init(void)
+{
+	int rc;
+
+	/* Register the tape major number to the kernel */
+	rc = register_blkdev(tapeblock_major, "tBLK");
+	if (rc < 0)
+		return rc;
+
+	if (tapeblock_major == 0)
+		tapeblock_major = rc;
+	PRINT_INFO("tape gets major %d for block device\n", tapeblock_major);
+	return 0;
+}
+
+/*
+ * Deregister major for block device frontend
+ */
+void
+tapeblock_exit(void)
+{
+	unregister_blkdev(tapeblock_major, "tBLK");
+}

drivers/s390/char/tape_char.c

@@ -0,0 +1,492 @@
+/*
+ *  drivers/s390/char/tape_char.c
+ *    character device frontend for tape device driver
+ *
+ *  S390 and zSeries version
+ *    Copyright (C) 2001,2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Carsten Otte <cotte@de.ibm.com>
+ *		 Michael Holzheu <holzheu@de.ibm.com>
+ *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/proc_fs.h>
+#include <linux/mtio.h>
+
+#include <asm/uaccess.h>
+
+#define TAPE_DBF_AREA	tape_core_dbf
+
+#include "tape.h"
+#include "tape_std.h"
+#include "tape_class.h"
+
+#define PRINTK_HEADER "TAPE_CHAR: "
+
+#define TAPECHAR_MAJOR		0	/* get dynamic major */
+
+/*
+ * file operation structure for tape character frontend
+ */
+static ssize_t tapechar_read(struct file *, char __user *, size_t, loff_t *);
+static ssize_t tapechar_write(struct file *, const char __user *, size_t, loff_t *);
+static int tapechar_open(struct inode *,struct file *);
+static int tapechar_release(struct inode *,struct file *);
+static int tapechar_ioctl(struct inode *, struct file *, unsigned int,
+			  unsigned long);
+
+static struct file_operations tape_fops =
+{
+	.owner = THIS_MODULE,
+	.read = tapechar_read,
+	.write = tapechar_write,
+	.ioctl = tapechar_ioctl,
+	.open = tapechar_open,
+	.release = tapechar_release,
+};
+
+static int tapechar_major = TAPECHAR_MAJOR;
+
+/*
+ * This function is called for every new tapedevice
+ */
+int
+tapechar_setup_device(struct tape_device * device)
+{
+	char	device_name[20];
+
+	sprintf(device_name, "ntibm%i", device->first_minor / 2);
+	device->nt = register_tape_dev(
+		&device->cdev->dev,
+		MKDEV(tapechar_major, device->first_minor),
+		&tape_fops,
+		device_name,
+		"non-rewinding"
+	);
+	device_name[0] = 'r';
+	device->rt = register_tape_dev(
+		&device->cdev->dev,
+		MKDEV(tapechar_major, device->first_minor + 1),
+		&tape_fops,
+		device_name,
+		"rewinding"
+	);
+
+	return 0;
+}
+
+void
+tapechar_cleanup_device(struct tape_device *device)
+{
+	unregister_tape_dev(device->rt);
+	device->rt = NULL;
+	unregister_tape_dev(device->nt);
+	device->nt = NULL;
+}
+
+/*
+ * Terminate write command (we write two TMs and skip backward over last)
+ * This ensures that the tape is always correctly terminated.
+ * When the user writes afterwards a new file, he will overwrite the
+ * second TM and therefore one TM will remain to separate the
+ * two files on the tape...
+ */
+static inline void
+tapechar_terminate_write(struct tape_device *device)
+{
+	if (tape_mtop(device, MTWEOF, 1) == 0 &&
+	    tape_mtop(device, MTWEOF, 1) == 0)
+		tape_mtop(device, MTBSR, 1);
+}
+
+static inline int
+tapechar_check_idalbuffer(struct tape_device *device, size_t block_size)
+{
+	struct idal_buffer *new;
+
+	if (device->char_data.idal_buf != NULL &&
+	    device->char_data.idal_buf->size == block_size)
+		return 0;
+
+	if (block_size > MAX_BLOCKSIZE) {
+		DBF_EVENT(3, "Invalid blocksize (%zd > %d)\n",
+			block_size, MAX_BLOCKSIZE);
+		PRINT_ERR("Invalid blocksize (%zd> %d)\n",
+			block_size, MAX_BLOCKSIZE);
+		return -EINVAL;
+	}
+
+	/* The current idal buffer is not correct. Allocate a new one. */
+	new = idal_buffer_alloc(block_size, 0);
+	if (new == NULL)
+		return -ENOMEM;
+
+	if (device->char_data.idal_buf != NULL)
+		idal_buffer_free(device->char_data.idal_buf);
+
+	device->char_data.idal_buf = new;
+
+	return 0;
+}
+
+/*
+ * Tape device read function
+ */
+ssize_t
+tapechar_read(struct file *filp, char __user *data, size_t count, loff_t *ppos)
+{
+	struct tape_device *device;
+	struct tape_request *request;
+	size_t block_size;
+	int rc;
+
+	DBF_EVENT(6, "TCHAR:read\n");
+	device = (struct tape_device *) filp->private_data;
+
+	/*
+	 * If the tape isn't terminated yet, do it now. And since we then
+	 * are at the end of the tape there wouldn't be anything to read
+	 * anyways. So we return immediatly.
+	 */
+	if(device->required_tapemarks) {
+		return tape_std_terminate_write(device);
+	}
+
+	/* Find out block size to use */
+	if (device->char_data.block_size != 0) {
+		if (count < device->char_data.block_size) {
+			DBF_EVENT(3, "TCHAR:read smaller than block "
+				  "size was requested\n");
+			return -EINVAL;
+		}
+		block_size = device->char_data.block_size;
+	} else {
+		block_size = count;
+	}
+
+	rc = tapechar_check_idalbuffer(device, block_size);
+	if (rc)
+		return rc;
+
+#ifdef CONFIG_S390_TAPE_BLOCK
+	/* Changes position. */
+	device->blk_data.medium_changed = 1;
+#endif
+
+	DBF_EVENT(6, "TCHAR:nbytes: %lx\n", block_size);
+	/* Let the discipline build the ccw chain. */
+	request = device->discipline->read_block(device, block_size);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	/* Execute it. */
+	rc = tape_do_io(device, request);
+	if (rc == 0) {
+		rc = block_size - request->rescnt;
+		DBF_EVENT(6, "TCHAR:rbytes:  %x\n", rc);
+		filp->f_pos += rc;
+		/* Copy data from idal buffer to user space. */
+		if (idal_buffer_to_user(device->char_data.idal_buf,
+					data, rc) != 0)
+			rc = -EFAULT;
+	}
+	tape_free_request(request);
+	return rc;
+}
+
+/*
+ * Tape device write function
+ */
+ssize_t
+tapechar_write(struct file *filp, const char __user *data, size_t count, loff_t *ppos)
+{
+	struct tape_device *device;
+	struct tape_request *request;
+	size_t block_size;
+	size_t written;
+	int nblocks;
+	int i, rc;
+
+	DBF_EVENT(6, "TCHAR:write\n");
+	device = (struct tape_device *) filp->private_data;
+	/* Find out block size and number of blocks */
+	if (device->char_data.block_size != 0) {
+		if (count < device->char_data.block_size) {
+			DBF_EVENT(3, "TCHAR:write smaller than block "
+				  "size was requested\n");
+			return -EINVAL;
+		}
+		block_size = device->char_data.block_size;
+		nblocks = count / block_size;
+	} else {
+		block_size = count;
+		nblocks = 1;
+	}
+
+	rc = tapechar_check_idalbuffer(device, block_size);
+	if (rc)
+		return rc;
+
+#ifdef CONFIG_S390_TAPE_BLOCK
+	/* Changes position. */
+	device->blk_data.medium_changed = 1;
+#endif
+
+	DBF_EVENT(6,"TCHAR:nbytes: %lx\n", block_size);
+	DBF_EVENT(6, "TCHAR:nblocks: %x\n", nblocks);
+	/* Let the discipline build the ccw chain. */
+	request = device->discipline->write_block(device, block_size);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	rc = 0;
+	written = 0;
+	for (i = 0; i < nblocks; i++) {
+		/* Copy data from user space to idal buffer. */
+		if (idal_buffer_from_user(device->char_data.idal_buf,
+					  data, block_size)) {
+			rc = -EFAULT;
+			break;
+		}
+		rc = tape_do_io(device, request);
+		if (rc)
+			break;
+		DBF_EVENT(6, "TCHAR:wbytes: %lx\n",
+			  block_size - request->rescnt);
+		filp->f_pos += block_size - request->rescnt;
+		written += block_size - request->rescnt;
+		if (request->rescnt != 0)
+			break;
+		data += block_size;
+	}
+	tape_free_request(request);
+	if (rc == -ENOSPC) {
+		/*
+		 * Ok, the device has no more space. It has NOT written
+		 * the block.
+		 */
+		if (device->discipline->process_eov)
+			device->discipline->process_eov(device);
+		if (written > 0)
+			rc = 0;
+
+	}
+
+	/*
+	 * After doing a write we always need two tapemarks to correctly
+	 * terminate the tape (one to terminate the file, the second to
+	 * flag the end of recorded data.
+	 * Since process_eov positions the tape in front of the written
+	 * tapemark it doesn't hurt to write two marks again.
+	 */
+	if (!rc)
+		device->required_tapemarks = 2;
+
+	return rc ? rc : written;
+}
+
+/*
+ * Character frontend tape device open function.
+ */
+int
+tapechar_open (struct inode *inode, struct file *filp)
+{
+	struct tape_device *device;
+	int minor, rc;
+
+	DBF_EVENT(6, "TCHAR:open: %i:%i\n",
+		imajor(filp->f_dentry->d_inode),
+		iminor(filp->f_dentry->d_inode));
+
+	if (imajor(filp->f_dentry->d_inode) != tapechar_major)
+		return -ENODEV;
+
+	minor = iminor(filp->f_dentry->d_inode);
+	device = tape_get_device(minor / TAPE_MINORS_PER_DEV);
+	if (IS_ERR(device)) {
+		DBF_EVENT(3, "TCHAR:open: tape_get_device() failed\n");
+		return PTR_ERR(device);
+	}
+
+
+	rc = tape_open(device);
+	if (rc == 0) {
+		filp->private_data = device;
+		return nonseekable_open(inode, filp);
+	}
+	tape_put_device(device);
+
+	return rc;
+}
+
+/*
+ * Character frontend tape device release function.
+ */
+
+int
+tapechar_release(struct inode *inode, struct file *filp)
+{
+	struct tape_device *device;
+
+	DBF_EVENT(6, "TCHAR:release: %x\n", iminor(inode));
+	device = (struct tape_device *) filp->private_data;
+
+	/*
+	 * If this is the rewinding tape minor then rewind. In that case we
+	 * write all required tapemarks. Otherwise only one to terminate the
+	 * file.
+	 */
+	if ((iminor(inode) & 1) != 0) {
+		if (device->required_tapemarks)
+			tape_std_terminate_write(device);
+		tape_mtop(device, MTREW, 1);
+	} else {
+		if (device->required_tapemarks > 1) {
+			if (tape_mtop(device, MTWEOF, 1) == 0)
+				device->required_tapemarks--;
+		}
+	}
+
+	if (device->char_data.idal_buf != NULL) {
+		idal_buffer_free(device->char_data.idal_buf);
+		device->char_data.idal_buf = NULL;
+	}
+	tape_release(device);
+	filp->private_data = tape_put_device(device);
+
+	return 0;
+}
+
+/*
+ * Tape device io controls.
+ */
+static int
+tapechar_ioctl(struct inode *inp, struct file *filp,
+	       unsigned int no, unsigned long data)
+{
+	struct tape_device *device;
+	int rc;
+
+	DBF_EVENT(6, "TCHAR:ioct\n");
+
+	device = (struct tape_device *) filp->private_data;
+
+	if (no == MTIOCTOP) {
+		struct mtop op;
+
+		if (copy_from_user(&op, (char __user *) data, sizeof(op)) != 0)
+			return -EFAULT;
+		if (op.mt_count < 0)
+			return -EINVAL;
+
+		/*
+		 * Operations that change tape position should write final
+		 * tapemarks.
+		 */
+		switch (op.mt_op) {
+			case MTFSF:
+			case MTBSF:
+			case MTFSR:
+			case MTBSR:
+			case MTREW:
+			case MTOFFL:
+			case MTEOM:
+			case MTRETEN:
+			case MTBSFM:
+			case MTFSFM:
+			case MTSEEK:
+#ifdef CONFIG_S390_TAPE_BLOCK
+				device->blk_data.medium_changed = 1;
+#endif
+				if (device->required_tapemarks)
+					tape_std_terminate_write(device);
+			default:
+				;
+		}
+		rc = tape_mtop(device, op.mt_op, op.mt_count);
+
+		if (op.mt_op == MTWEOF && rc == 0) {
+			if (op.mt_count > device->required_tapemarks)
+				device->required_tapemarks = 0;
+			else
+				device->required_tapemarks -= op.mt_count;
+		}
+		return rc;
+	}
+	if (no == MTIOCPOS) {
+		/* MTIOCPOS: query the tape position. */
+		struct mtpos pos;
+
+		rc = tape_mtop(device, MTTELL, 1);
+		if (rc < 0)
+			return rc;
+		pos.mt_blkno = rc;
+		if (copy_to_user((char __user *) data, &pos, sizeof(pos)) != 0)
+			return -EFAULT;
+		return 0;
+	}
+	if (no == MTIOCGET) {
+		/* MTIOCGET: query the tape drive status. */
+		struct mtget get;
+
+		memset(&get, 0, sizeof(get));
+		get.mt_type = MT_ISUNKNOWN;
+		get.mt_resid = 0 /* device->devstat.rescnt */;
+		get.mt_dsreg = device->tape_state;
+		/* FIXME: mt_gstat, mt_erreg, mt_fileno */
+		get.mt_gstat = 0;
+		get.mt_erreg = 0;
+		get.mt_fileno = 0;
+		get.mt_gstat  = device->tape_generic_status;
+
+		if (device->medium_state == MS_LOADED) {
+			rc = tape_mtop(device, MTTELL, 1);
+
+			if (rc < 0)
+				return rc;
+
+			if (rc == 0)
+				get.mt_gstat |= GMT_BOT(~0);
+
+			get.mt_blkno = rc;
+		}
+
+		if (copy_to_user((char __user *) data, &get, sizeof(get)) != 0)
+			return -EFAULT;
+
+		return 0;
+	}
+	/* Try the discipline ioctl function. */
+	if (device->discipline->ioctl_fn == NULL)
+		return -EINVAL;
+	return device->discipline->ioctl_fn(device, no, data);
+}
+
+/*
+ * Initialize character device frontend.
+ */
+int
+tapechar_init (void)
+{
+	dev_t	dev;
+
+	if (alloc_chrdev_region(&dev, 0, 256, "tape") != 0)
+		return -1;
+
+	tapechar_major = MAJOR(dev);
+	PRINT_INFO("tape gets major %d for character devices\n", MAJOR(dev));
+
+	return 0;
+}
+
+/*
+ * cleanup
+ */
+void
+tapechar_exit(void)
+{
+	PRINT_INFO("tape releases major %d for character devices\n",
+		tapechar_major);
+	unregister_chrdev_region(MKDEV(tapechar_major, 0), 256);
+}

drivers/s390/char/tape_class.c

@@ -0,0 +1,126 @@
+/*
+ * (C) Copyright IBM Corp. 2004
+ * tape_class.c ($Revision: 1.8 $)
+ *
+ * Tape class device support
+ *
+ * Author: Stefan Bader <shbader@de.ibm.com>
+ * Based on simple class device code by Greg K-H
+ */
+#include "tape_class.h"
+
+MODULE_AUTHOR("Stefan Bader <[email protected]>");
+MODULE_DESCRIPTION(
+	"(C) Copyright IBM Corp. 2004   All Rights Reserved.\n"
+	"tape_class.c ($Revision: 1.8 $)"
+);
+MODULE_LICENSE("GPL");
+
+struct class_simple *tape_class;
+
+/*
+ * Register a tape device and return a pointer to the cdev structure.
+ *
+ * device
+ *	The pointer to the struct device of the physical (base) device.
+ * drivername
+ *	The pointer to the drivers name for it's character devices.
+ * dev
+ *	The intended major/minor number. The major number may be 0 to
+ *	get a dynamic major number.
+ * fops
+ *	The pointer to the drivers file operations for the tape device.
+ * devname
+ *	The pointer to the name of the character device.
+ */
+struct tape_class_device *register_tape_dev(
+	struct device *		device,
+	dev_t			dev,
+	struct file_operations *fops,
+	char *			device_name,
+	char *			mode_name)
+{
+	struct tape_class_device *	tcd;
+	int		rc;
+	char *		s;
+
+	tcd = kmalloc(sizeof(struct tape_class_device), GFP_KERNEL);
+	if (!tcd)
+		return ERR_PTR(-ENOMEM);
+
+	memset(tcd, 0, sizeof(struct tape_class_device));
+	strncpy(tcd->device_name, device_name, TAPECLASS_NAME_LEN);
+	for (s = strchr(tcd->device_name, '/'); s; s = strchr(s, '/'))
+		*s = '!';
+	strncpy(tcd->mode_name, mode_name, TAPECLASS_NAME_LEN);
+	for (s = strchr(tcd->mode_name, '/'); s; s = strchr(s, '/'))
+		*s = '!';
+
+	tcd->char_device = cdev_alloc();
+	if (!tcd->char_device) {
+		rc = -ENOMEM;
+		goto fail_with_tcd;
+	}
+
+	tcd->char_device->owner = fops->owner;
+	tcd->char_device->ops   = fops;
+	tcd->char_device->dev   = dev;
+
+	rc = cdev_add(tcd->char_device, tcd->char_device->dev, 1);
+	if (rc)
+		goto fail_with_cdev;
+
+	tcd->class_device = class_simple_device_add(
+				tape_class,
+				tcd->char_device->dev,
+				device,
+				"%s", tcd->device_name
+			);
+	sysfs_create_link(
+		&device->kobj,
+		&tcd->class_device->kobj,
+		tcd->mode_name
+	);
+
+	return tcd;
+
+fail_with_cdev:
+	cdev_del(tcd->char_device);
+
+fail_with_tcd:
+	kfree(tcd);
+
+	return ERR_PTR(rc);
+}
+EXPORT_SYMBOL(register_tape_dev);
+
+void unregister_tape_dev(struct tape_class_device *tcd)
+{
+	if (tcd != NULL && !IS_ERR(tcd)) {
+		sysfs_remove_link(
+			&tcd->class_device->dev->kobj,
+			tcd->mode_name
+		);
+		class_simple_device_remove(tcd->char_device->dev);
+		cdev_del(tcd->char_device);
+		kfree(tcd);
+	}
+}
+EXPORT_SYMBOL(unregister_tape_dev);
+
+
+static int __init tape_init(void)
+{
+	tape_class = class_simple_create(THIS_MODULE, "tape390");
+
+	return 0;
+}
+
+static void __exit tape_exit(void)
+{
+	class_simple_destroy(tape_class);
+	tape_class = NULL;
+}
+
+postcore_initcall(tape_init);
+module_exit(tape_exit);

drivers/s390/char/tape_class.h

@@ -0,0 +1,61 @@
+/*
+ * (C) Copyright IBM Corp. 2004   All Rights Reserved.
+ * tape_class.h ($Revision: 1.4 $)
+ *
+ * Tape class device support
+ *
+ * Author: Stefan Bader <shbader@de.ibm.com>
+ * Based on simple class device code by Greg K-H
+ */
+#ifndef __TAPE_CLASS_H__
+#define __TAPE_CLASS_H__
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/fs.h>
+#include <linux/major.h>
+#include <linux/kobject.h>
+#include <linux/kobj_map.h>
+#include <linux/cdev.h>
+
+#include <linux/device.h>
+#include <linux/kdev_t.h>
+
+#define TAPECLASS_NAME_LEN	32
+
+struct tape_class_device {
+	struct cdev *		char_device;
+	struct class_device *	class_device;
+	char			device_name[TAPECLASS_NAME_LEN];
+	char			mode_name[TAPECLASS_NAME_LEN];
+};
+
+/*
+ * Register a tape device and return a pointer to the tape class device
+ * created by the call.
+ *
+ * device
+ *	The pointer to the struct device of the physical (base) device.
+ * dev
+ *	The intended major/minor number. The major number may be 0 to
+ *	get a dynamic major number.
+ * fops
+ *	The pointer to the drivers file operations for the tape device.
+ * device_name
+ *	Pointer to the logical device name (will also be used as kobject name
+ *	of the cdev). This can also be called the name of the tape class
+ *	device.
+ * mode_name
+ *	Points to the name of the tape mode. This creates a link with that
+ *	name from the physical device to the logical device (class).
+ */
+struct tape_class_device *register_tape_dev(
+	struct device *		device,
+	dev_t			dev,
+	struct file_operations *fops,
+	char *			device_name,
+	char *			node_name
+);
+void unregister_tape_dev(struct tape_class_device *tcd);
+
+#endif /* __TAPE_CLASS_H__ */

drivers/s390/char/tape_proc.c

@@ -0,0 +1,145 @@
+/*
+ *  drivers/s390/char/tape.c
+ *    tape device driver for S/390 and zSeries tapes.
+ *
+ *  S390 and zSeries version
+ *    Copyright (C) 2001 IBM Corporation
+ *    Author(s): Carsten Otte <cotte@de.ibm.com>
+ *		 Michael Holzheu <holzheu@de.ibm.com>
+ *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
+ *
+ * PROCFS Functions
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/seq_file.h>
+
+#define TAPE_DBF_AREA	tape_core_dbf
+
+#include "tape.h"
+
+#define PRINTK_HEADER "TAPE_PROC: "
+
+static const char *tape_med_st_verbose[MS_SIZE] =
+{
+	[MS_UNKNOWN] = "UNKNOWN ",
+	[MS_LOADED] = "LOADED  ",
+	[MS_UNLOADED] = "UNLOADED"
+};
+
+/* our proc tapedevices entry */
+static struct proc_dir_entry *tape_proc_devices;
+
+/*
+ * Show function for /proc/tapedevices
+ */
+static int tape_proc_show(struct seq_file *m, void *v)
+{
+	struct tape_device *device;
+	struct tape_request *request;
+	const char *str;
+	unsigned long n;
+
+	n = (unsigned long) v - 1;
+	if (!n) {
+		seq_printf(m, "TapeNo\tBusID      CuType/Model\t"
+			"DevType/Model\tBlkSize\tState\tOp\tMedState\n");
+	}
+	device = tape_get_device(n);
+	if (IS_ERR(device))
+		return 0;
+	spin_lock_irq(get_ccwdev_lock(device->cdev));
+	seq_printf(m, "%d\t", (int) n);
+	seq_printf(m, "%-10.10s ", device->cdev->dev.bus_id);
+	seq_printf(m, "%04X/", device->cdev->id.cu_type);
+	seq_printf(m, "%02X\t", device->cdev->id.cu_model);
+	seq_printf(m, "%04X/", device->cdev->id.dev_type);
+	seq_printf(m, "%02X\t\t", device->cdev->id.dev_model);
+	if (device->char_data.block_size == 0)
+		seq_printf(m, "auto\t");
+	else
+		seq_printf(m, "%i\t", device->char_data.block_size);
+	if (device->tape_state >= 0 &&
+	    device->tape_state < TS_SIZE)
+		str = tape_state_verbose[device->tape_state];
+	else
+		str = "UNKNOWN";
+	seq_printf(m, "%s\t", str);
+	if (!list_empty(&device->req_queue)) {
+		request = list_entry(device->req_queue.next,
+				     struct tape_request, list);
+		str = tape_op_verbose[request->op];
+	} else
+		str = "---";
+	seq_printf(m, "%s\t", str);
+	seq_printf(m, "%s\n", tape_med_st_verbose[device->medium_state]);
+	spin_unlock_irq(get_ccwdev_lock(device->cdev));
+	tape_put_device(device);
+        return 0;
+}
+
+static void *tape_proc_start(struct seq_file *m, loff_t *pos)
+{
+	if (*pos >= 256 / TAPE_MINORS_PER_DEV)
+		return NULL;
+	return (void *)((unsigned long) *pos + 1);
+}
+
+static void *tape_proc_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	++*pos;
+	return tape_proc_start(m, pos);
+}
+
+static void tape_proc_stop(struct seq_file *m, void *v)
+{
+}
+
+static struct seq_operations tape_proc_seq = {
+	.start		= tape_proc_start,
+	.next		= tape_proc_next,
+	.stop		= tape_proc_stop,
+	.show		= tape_proc_show,
+};
+
+static int tape_proc_open(struct inode *inode, struct file *file)
+{
+	return seq_open(file, &tape_proc_seq);
+}
+
+static struct file_operations tape_proc_ops =
+{
+	.open		= tape_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
+
+/*
+ * Initialize procfs stuff on startup
+ */
+void
+tape_proc_init(void)
+{
+	tape_proc_devices =
+		create_proc_entry ("tapedevices", S_IFREG | S_IRUGO | S_IWUSR,
+				   &proc_root);
+	if (tape_proc_devices == NULL) {
+		PRINT_WARN("tape: Cannot register procfs entry tapedevices\n");
+		return;
+	}
+	tape_proc_devices->proc_fops = &tape_proc_ops;
+	tape_proc_devices->owner = THIS_MODULE;
+}
+
+/*
+ * Cleanup all stuff registered to the procfs
+ */
+void
+tape_proc_cleanup(void)
+{
+	if (tape_proc_devices != NULL)
+		remove_proc_entry ("tapedevices", &proc_root);
+}

drivers/s390/char/tape_std.c

@@ -0,0 +1,765 @@
+/*
+ *  drivers/s390/char/tape_std.c
+ *    standard tape device functions for ibm tapes.
+ *
+ *  S390 and zSeries version
+ *    Copyright (C) 2001,2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Carsten Otte <cotte@de.ibm.com>
+ *		 Michael Holzheu <holzheu@de.ibm.com>
+ *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *		 Stefan Bader <shbader@de.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/stddef.h>
+#include <linux/kernel.h>
+#include <linux/bio.h>
+#include <linux/timer.h>
+
+#include <asm/types.h>
+#include <asm/idals.h>
+#include <asm/ebcdic.h>
+#include <asm/tape390.h>
+
+#define TAPE_DBF_AREA	tape_core_dbf
+
+#include "tape.h"
+#include "tape_std.h"
+
+#define PRINTK_HEADER "TAPE_STD: "
+
+/*
+ * tape_std_assign
+ */
+static void
+tape_std_assign_timeout(unsigned long data)
+{
+	struct tape_request *	request;
+	struct tape_device *	device;
+
+	request = (struct tape_request *) data;
+	if ((device = request->device) == NULL)
+		BUG();
+
+	spin_lock_irq(get_ccwdev_lock(device->cdev));
+	if (request->callback != NULL) {
+		DBF_EVENT(3, "%08x: Assignment timeout. Device busy.\n",
+			device->cdev_id);
+		PRINT_ERR("%s: Assignment timeout. Device busy.\n",
+			device->cdev->dev.bus_id);
+		ccw_device_clear(device->cdev, (long) request);
+	}
+	spin_unlock_irq(get_ccwdev_lock(device->cdev));
+}
+
+int
+tape_std_assign(struct tape_device *device)
+{
+	int                  rc;
+	struct timer_list    timeout;
+	struct tape_request *request;
+
+	request = tape_alloc_request(2, 11);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+
+	request->op = TO_ASSIGN;
+	tape_ccw_cc(request->cpaddr, ASSIGN, 11, request->cpdata);
+	tape_ccw_end(request->cpaddr + 1, NOP, 0, NULL);
+
+	/*
+	 * The assign command sometimes blocks if the device is assigned
+	 * to another host (actually this shouldn't happen but it does).
+	 * So we set up a timeout for this call.
+	 */
+	init_timer(&timeout);
+	timeout.function = tape_std_assign_timeout;
+	timeout.data     = (unsigned long) request;
+	timeout.expires  = jiffies + 2 * HZ;
+	add_timer(&timeout);
+
+	rc = tape_do_io_interruptible(device, request);
+
+	del_timer(&timeout);
+
+	if (rc != 0) {
+		PRINT_WARN("%s: assign failed - device might be busy\n",
+			device->cdev->dev.bus_id);
+		DBF_EVENT(3, "%08x: assign failed - device might be busy\n",
+			device->cdev_id);
+	} else {
+		DBF_EVENT(3, "%08x: Tape assigned\n", device->cdev_id);
+	}
+	tape_free_request(request);
+	return rc;
+}
+
+/*
+ * tape_std_unassign
+ */
+int
+tape_std_unassign (struct tape_device *device)
+{
+	int                  rc;
+	struct tape_request *request;
+
+	if (device->tape_state == TS_NOT_OPER) {
+		DBF_EVENT(3, "(%08x): Can't unassign device\n",
+			device->cdev_id);
+		PRINT_WARN("(%s): Can't unassign device - device gone\n",
+			device->cdev->dev.bus_id);
+		return -EIO;
+	}
+
+	request = tape_alloc_request(2, 11);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+
+	request->op = TO_UNASSIGN;
+	tape_ccw_cc(request->cpaddr, UNASSIGN, 11, request->cpdata);
+	tape_ccw_end(request->cpaddr + 1, NOP, 0, NULL);
+
+	if ((rc = tape_do_io(device, request)) != 0) {
+		DBF_EVENT(3, "%08x: Unassign failed\n", device->cdev_id);
+		PRINT_WARN("%s: Unassign failed\n", device->cdev->dev.bus_id);
+	} else {
+		DBF_EVENT(3, "%08x: Tape unassigned\n", device->cdev_id);
+	}
+	tape_free_request(request);
+	return rc;
+}
+
+/*
+ * TAPE390_DISPLAY: Show a string on the tape display.
+ */
+int
+tape_std_display(struct tape_device *device, struct display_struct *disp)
+{
+	struct tape_request *request;
+	int rc;
+
+	request = tape_alloc_request(2, 17);
+	if (IS_ERR(request)) {
+		DBF_EVENT(3, "TAPE: load display failed\n");
+		return PTR_ERR(request);
+	}
+	request->op = TO_DIS;
+
+	*(unsigned char *) request->cpdata = disp->cntrl;
+	DBF_EVENT(5, "TAPE: display cntrl=%04x\n", disp->cntrl);
+	memcpy(((unsigned char *) request->cpdata) + 1, disp->message1, 8);
+	memcpy(((unsigned char *) request->cpdata) + 9, disp->message2, 8);
+	ASCEBC(((unsigned char*) request->cpdata) + 1, 16);
+
+	tape_ccw_cc(request->cpaddr, LOAD_DISPLAY, 17, request->cpdata);
+	tape_ccw_end(request->cpaddr + 1, NOP, 0, NULL);
+
+	rc = tape_do_io_interruptible(device, request);
+	tape_free_request(request);
+	return rc;
+}
+
+/*
+ * Read block id.
+ */
+int
+tape_std_read_block_id(struct tape_device *device, __u64 *id)
+{
+	struct tape_request *request;
+	int rc;
+
+	request = tape_alloc_request(3, 8);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_RBI;
+	/* setup ccws */
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_cc(request->cpaddr + 1, READ_BLOCK_ID, 8, request->cpdata);
+	tape_ccw_end(request->cpaddr + 2, NOP, 0, NULL);
+	/* execute it */
+	rc = tape_do_io(device, request);
+	if (rc == 0)
+		/* Get result from read buffer. */
+		*id = *(__u64 *) request->cpdata;
+	tape_free_request(request);
+	return rc;
+}
+
+int
+tape_std_terminate_write(struct tape_device *device)
+{
+	int rc;
+
+	if(device->required_tapemarks == 0)
+		return 0;
+
+	DBF_LH(5, "tape%d: terminate write %dxEOF\n", device->first_minor,
+		device->required_tapemarks);
+
+	rc = tape_mtop(device, MTWEOF, device->required_tapemarks);
+	if (rc)
+		return rc;
+
+	device->required_tapemarks = 0;
+	return tape_mtop(device, MTBSR, 1);
+}
+
+/*
+ * MTLOAD: Loads the tape.
+ * The default implementation just wait until the tape medium state changes
+ * to MS_LOADED.
+ */
+int
+tape_std_mtload(struct tape_device *device, int count)
+{
+	return wait_event_interruptible(device->state_change_wq,
+		(device->medium_state == MS_LOADED));
+}
+
+/*
+ * MTSETBLK: Set block size.
+ */
+int
+tape_std_mtsetblk(struct tape_device *device, int count)
+{
+	struct idal_buffer *new;
+
+	DBF_LH(6, "tape_std_mtsetblk(%d)\n", count);
+	if (count <= 0) {
+		/*
+		 * Just set block_size to 0. tapechar_read/tapechar_write
+		 * will realloc the idal buffer if a bigger one than the
+		 * current is needed.
+		 */
+		device->char_data.block_size = 0;
+		return 0;
+	}
+	if (device->char_data.idal_buf != NULL &&
+	    device->char_data.idal_buf->size == count)
+		/* We already have a idal buffer of that size. */
+		return 0;
+
+	if (count > MAX_BLOCKSIZE) {
+		DBF_EVENT(3, "Invalid block size (%d > %d) given.\n",
+			count, MAX_BLOCKSIZE);
+		PRINT_ERR("Invalid block size (%d > %d) given.\n",
+			count, MAX_BLOCKSIZE);
+		return -EINVAL;
+	}
+
+	/* Allocate a new idal buffer. */
+	new = idal_buffer_alloc(count, 0);
+	if (new == NULL)
+		return -ENOMEM;
+	if (device->char_data.idal_buf != NULL)
+		idal_buffer_free(device->char_data.idal_buf);
+	device->char_data.idal_buf = new;
+	device->char_data.block_size = count;
+
+	DBF_LH(6, "new blocksize is %d\n", device->char_data.block_size);
+
+	return 0;
+}
+
+/*
+ * MTRESET: Set block size to 0.
+ */
+int
+tape_std_mtreset(struct tape_device *device, int count)
+{
+	DBF_EVENT(6, "TCHAR:devreset:\n");
+	device->char_data.block_size = 0;
+	return 0;
+}
+
+/*
+ * MTFSF: Forward space over 'count' file marks. The tape is positioned
+ * at the EOT (End of Tape) side of the file mark.
+ */
+int
+tape_std_mtfsf(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_FSF;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, FORSPACEFILE, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTFSR: Forward space over 'count' tape blocks (blocksize is set
+ * via MTSETBLK.
+ */
+int
+tape_std_mtfsr(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+	int rc;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_FSB;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, FORSPACEBLOCK, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+
+	/* execute it */
+	rc = tape_do_io(device, request);
+	if (rc == 0 && request->rescnt > 0) {
+		DBF_LH(3, "FSR over tapemark\n");
+		rc = 1;
+	}
+	tape_free_request(request);
+
+	return rc;
+}
+
+/*
+ * MTBSR: Backward space over 'count' tape blocks.
+ * (blocksize is set via MTSETBLK.
+ */
+int
+tape_std_mtbsr(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+	int rc;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_BSB;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, BACKSPACEBLOCK, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+
+	/* execute it */
+	rc = tape_do_io(device, request);
+	if (rc == 0 && request->rescnt > 0) {
+		DBF_LH(3, "BSR over tapemark\n");
+		rc = 1;
+	}
+	tape_free_request(request);
+
+	return rc;
+}
+
+/*
+ * MTWEOF: Write 'count' file marks at the current position.
+ */
+int
+tape_std_mtweof(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_WTM;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, WRITETAPEMARK, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTBSFM: Backward space over 'count' file marks.
+ * The tape is positioned at the BOT (Begin Of Tape) side of the
+ * last skipped file mark.
+ */
+int
+tape_std_mtbsfm(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_BSF;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, BACKSPACEFILE, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTBSF: Backward space over 'count' file marks. The tape is positioned at
+ * the EOT (End of Tape) side of the last skipped file mark.
+ */
+int
+tape_std_mtbsf(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+	int rc;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_BSF;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, BACKSPACEFILE, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+	/* execute it */
+	rc = tape_do_io_free(device, request);
+	if (rc == 0) {
+		rc = tape_mtop(device, MTFSR, 1);
+		if (rc > 0)
+			rc = 0;
+	}
+	return rc;
+}
+
+/*
+ * MTFSFM: Forward space over 'count' file marks.
+ * The tape is positioned at the BOT (Begin Of Tape) side
+ * of the last skipped file mark.
+ */
+int
+tape_std_mtfsfm(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	struct ccw1 *ccw;
+	int rc;
+
+	request = tape_alloc_request(mt_count + 2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_FSF;
+	/* setup ccws */
+	ccw = tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+			  device->modeset_byte);
+	ccw = tape_ccw_repeat(ccw, FORSPACEFILE, mt_count);
+	ccw = tape_ccw_end(ccw, NOP, 0, NULL);
+	/* execute it */
+	rc = tape_do_io_free(device, request);
+	if (rc == 0) {
+		rc = tape_mtop(device, MTBSR, 1);
+		if (rc > 0)
+			rc = 0;
+	}
+
+	return rc;
+}
+
+/*
+ * MTREW: Rewind the tape.
+ */
+int
+tape_std_mtrew(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+
+	request = tape_alloc_request(3, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_REW;
+	/* setup ccws */
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1,
+		    device->modeset_byte);
+	tape_ccw_cc(request->cpaddr + 1, REWIND, 0, NULL);
+	tape_ccw_end(request->cpaddr + 2, NOP, 0, NULL);
+
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTOFFL: Rewind the tape and put the drive off-line.
+ * Implement 'rewind unload'
+ */
+int
+tape_std_mtoffl(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+
+	request = tape_alloc_request(3, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_RUN;
+	/* setup ccws */
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_cc(request->cpaddr + 1, REWIND_UNLOAD, 0, NULL);
+	tape_ccw_end(request->cpaddr + 2, NOP, 0, NULL);
+
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTNOP: 'No operation'.
+ */
+int
+tape_std_mtnop(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+
+	request = tape_alloc_request(2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_NOP;
+	/* setup ccws */
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_end(request->cpaddr + 1, NOP, 0, NULL);
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTEOM: positions at the end of the portion of the tape already used
+ * for recordind data. MTEOM positions after the last file mark, ready for
+ * appending another file.
+ */
+int
+tape_std_mteom(struct tape_device *device, int mt_count)
+{
+	int rc;
+
+	/*
+	 * Seek from the beginning of tape (rewind).
+	 */
+	if ((rc = tape_mtop(device, MTREW, 1)) < 0)
+		return rc;
+
+	/*
+	 * The logical end of volume is given by two sewuential tapemarks.
+	 * Look for this by skipping to the next file (over one tapemark)
+	 * and then test for another one (fsr returns 1 if a tapemark was
+	 * encountered).
+	 */
+	do {
+		if ((rc = tape_mtop(device, MTFSF, 1)) < 0)
+			return rc;
+		if ((rc = tape_mtop(device, MTFSR, 1)) < 0)
+			return rc;
+	} while (rc == 0);
+
+	return tape_mtop(device, MTBSR, 1);
+}
+
+/*
+ * MTRETEN: Retension the tape, i.e. forward space to end of tape and rewind.
+ */
+int
+tape_std_mtreten(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+	int rc;
+
+	request = tape_alloc_request(4, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_FSF;
+	/* setup ccws */
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_cc(request->cpaddr + 1,FORSPACEFILE, 0, NULL);
+	tape_ccw_cc(request->cpaddr + 2, NOP, 0, NULL);
+	tape_ccw_end(request->cpaddr + 3, CCW_CMD_TIC, 0, request->cpaddr);
+	/* execute it, MTRETEN rc gets ignored */
+	rc = tape_do_io_interruptible(device, request);
+	tape_free_request(request);
+	return tape_mtop(device, MTREW, 1);
+}
+
+/*
+ * MTERASE: erases the tape.
+ */
+int
+tape_std_mterase(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+
+	request = tape_alloc_request(6, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_DSE;
+	/* setup ccws */
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_cc(request->cpaddr + 1, REWIND, 0, NULL);
+	tape_ccw_cc(request->cpaddr + 2, ERASE_GAP, 0, NULL);
+	tape_ccw_cc(request->cpaddr + 3, DATA_SEC_ERASE, 0, NULL);
+	tape_ccw_cc(request->cpaddr + 4, REWIND, 0, NULL);
+	tape_ccw_end(request->cpaddr + 5, NOP, 0, NULL);
+
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * MTUNLOAD: Rewind the tape and unload it.
+ */
+int
+tape_std_mtunload(struct tape_device *device, int mt_count)
+{
+	return tape_mtop(device, MTOFFL, mt_count);
+}
+
+/*
+ * MTCOMPRESSION: used to enable compression.
+ * Sets the IDRC on/off.
+ */
+int
+tape_std_mtcompression(struct tape_device *device, int mt_count)
+{
+	struct tape_request *request;
+
+	if (mt_count < 0 || mt_count > 1) {
+		DBF_EXCEPTION(6, "xcom parm\n");
+		if (*device->modeset_byte & 0x08)
+			PRINT_INFO("(%s) Compression is currently on\n",
+				   device->cdev->dev.bus_id);
+		else
+			PRINT_INFO("(%s) Compression is currently off\n",
+				   device->cdev->dev.bus_id);
+		PRINT_INFO("Use 1 to switch compression on, 0 to "
+			   "switch it off\n");
+		return -EINVAL;
+	}
+	request = tape_alloc_request(2, 0);
+	if (IS_ERR(request))
+		return PTR_ERR(request);
+	request->op = TO_NOP;
+	/* setup ccws */
+	*device->modeset_byte = (mt_count == 0) ? 0x00 : 0x08;
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_end(request->cpaddr + 1, NOP, 0, NULL);
+	/* execute it */
+	return tape_do_io_free(device, request);
+}
+
+/*
+ * Read Block
+ */
+struct tape_request *
+tape_std_read_block(struct tape_device *device, size_t count)
+{
+	struct tape_request *request;
+
+	/*
+	 * We have to alloc 4 ccws in order to be able to transform request
+	 * into a read backward request in error case.
+	 */
+	request = tape_alloc_request(4, 0);
+	if (IS_ERR(request)) {
+		DBF_EXCEPTION(6, "xrbl fail");
+		return request;
+	}
+	request->op = TO_RFO;
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_end_idal(request->cpaddr + 1, READ_FORWARD,
+			  device->char_data.idal_buf);
+	DBF_EVENT(6, "xrbl ccwg\n");
+	return request;
+}
+
+/*
+ * Read Block backward transformation function.
+ */
+void
+tape_std_read_backward(struct tape_device *device, struct tape_request *request)
+{
+	/*
+	 * We have allocated 4 ccws in tape_std_read, so we can now
+	 * transform the request to a read backward, followed by a
+	 * forward space block.
+	 */
+	request->op = TO_RBA;
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_cc_idal(request->cpaddr + 1, READ_BACKWARD,
+			 device->char_data.idal_buf);
+	tape_ccw_cc(request->cpaddr + 2, FORSPACEBLOCK, 0, NULL);
+	tape_ccw_end(request->cpaddr + 3, NOP, 0, NULL);
+	DBF_EVENT(6, "xrop ccwg");}
+
+/*
+ * Write Block
+ */
+struct tape_request *
+tape_std_write_block(struct tape_device *device, size_t count)
+{
+	struct tape_request *request;
+
+	request = tape_alloc_request(2, 0);
+	if (IS_ERR(request)) {
+		DBF_EXCEPTION(6, "xwbl fail\n");
+		return request;
+	}
+	request->op = TO_WRI;
+	tape_ccw_cc(request->cpaddr, MODE_SET_DB, 1, device->modeset_byte);
+	tape_ccw_end_idal(request->cpaddr + 1, WRITE_CMD,
+			  device->char_data.idal_buf);
+	DBF_EVENT(6, "xwbl ccwg\n");
+	return request;
+}
+
+/*
+ * This routine is called by frontend after an ENOSP on write
+ */
+void
+tape_std_process_eov(struct tape_device *device)
+{
+	/*
+	 * End of volume: We have to backspace the last written record, then
+	 * we TRY to write a tapemark and then backspace over the written TM
+	 */
+	if (tape_mtop(device, MTBSR, 1) == 0 &&
+	    tape_mtop(device, MTWEOF, 1) == 0) {
+		tape_mtop(device, MTBSR, 1);
+	}
+}
+
+EXPORT_SYMBOL(tape_std_assign);
+EXPORT_SYMBOL(tape_std_unassign);
+EXPORT_SYMBOL(tape_std_display);
+EXPORT_SYMBOL(tape_std_read_block_id);
+EXPORT_SYMBOL(tape_std_mtload);
+EXPORT_SYMBOL(tape_std_mtsetblk);
+EXPORT_SYMBOL(tape_std_mtreset);
+EXPORT_SYMBOL(tape_std_mtfsf);
+EXPORT_SYMBOL(tape_std_mtfsr);
+EXPORT_SYMBOL(tape_std_mtbsr);
+EXPORT_SYMBOL(tape_std_mtweof);
+EXPORT_SYMBOL(tape_std_mtbsfm);
+EXPORT_SYMBOL(tape_std_mtbsf);
+EXPORT_SYMBOL(tape_std_mtfsfm);
+EXPORT_SYMBOL(tape_std_mtrew);
+EXPORT_SYMBOL(tape_std_mtoffl);
+EXPORT_SYMBOL(tape_std_mtnop);
+EXPORT_SYMBOL(tape_std_mteom);
+EXPORT_SYMBOL(tape_std_mtreten);
+EXPORT_SYMBOL(tape_std_mterase);
+EXPORT_SYMBOL(tape_std_mtunload);
+EXPORT_SYMBOL(tape_std_mtcompression);
+EXPORT_SYMBOL(tape_std_read_block);
+EXPORT_SYMBOL(tape_std_read_backward);
+EXPORT_SYMBOL(tape_std_write_block);
+EXPORT_SYMBOL(tape_std_process_eov);

drivers/s390/char/tape_std.h

@@ -0,0 +1,152 @@
+/*
+ *  drivers/s390/char/tape_34xx.h
+ *    standard tape device functions for ibm tapes.
+ *
+ *  S390 and zSeries version
+ *    Copyright (C) 2001,2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ *    Author(s): Carsten Otte <cotte@de.ibm.com>
+ *		 Tuan Ngo-Anh <ngoanh@de.ibm.com>
+ *		 Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#ifndef _TAPE_STD_H
+#define _TAPE_STD_H
+
+#include <asm/tape390.h>
+
+/*
+ * Biggest block size to handle. Currently 64K because we only build
+ * channel programs without data chaining.
+ */
+#define MAX_BLOCKSIZE   65535
+
+/*
+ * The CCW commands for the Tape type of command.
+ */
+#define INVALID_00		0x00	/* Invalid cmd */
+#define BACKSPACEBLOCK		0x27	/* Back Space block */
+#define BACKSPACEFILE		0x2f	/* Back Space file */
+#define DATA_SEC_ERASE		0x97	/* Data security erase */
+#define ERASE_GAP		0x17	/* Erase Gap */
+#define FORSPACEBLOCK		0x37	/* Forward space block */
+#define FORSPACEFILE		0x3F	/* Forward Space file */
+#define FORCE_STREAM_CNT	0xEB	/* Forced streaming count # */
+#define NOP			0x03	/* No operation	*/
+#define READ_FORWARD		0x02	/* Read forward */
+#define REWIND			0x07	/* Rewind */
+#define REWIND_UNLOAD		0x0F	/* Rewind and Unload */
+#define SENSE			0x04	/* Sense */
+#define NEW_MODE_SET		0xEB	/* Guess it is Mode set */
+#define WRITE_CMD		0x01	/* Write */
+#define WRITETAPEMARK		0x1F	/* Write Tape Mark */
+
+#define ASSIGN			0xB7	/* 3420 REJECT,3480 OK	*/
+#define CONTROL_ACCESS		0xE3	/* Set high speed */
+#define DIAG_MODE_SET		0x0B	/* 3420 NOP, 3480 REJECT */
+#define LOAD_DISPLAY		0x9F	/* 3420 REJECT,3480 OK */
+#define LOCATE			0x4F	/* 3420 REJ, 3480 NOP */
+#define LOOP_WRITE_TO_READ	0x8B	/* 3480 REJECT */
+#define MODE_SET_DB		0xDB	/* 3420 REJECT,3480 OK */
+#define MODE_SET_C3		0xC3	/* for 3420 */
+#define MODE_SET_CB		0xCB	/* for 3420 */
+#define MODE_SET_D3		0xD3	/* for 3420 */
+#define READ_BACKWARD		0x0C	/* */
+#define READ_BLOCK_ID		0x22	/* 3420 REJECT,3480 OK */
+#define READ_BUFFER		0x12	/* 3420 REJECT,3480 OK */
+#define READ_BUFF_LOG		0x24	/* 3420 REJECT,3480 OK */
+#define RELEASE			0xD4	/* 3420 NOP, 3480 REJECT */
+#define REQ_TRK_IN_ERROR	0x1B	/* 3420 NOP, 3480 REJECT */
+#define RESERVE			0xF4	/* 3420 NOP, 3480 REJECT */
+#define SENSE_GROUP_ID		0x34	/* 3420 REJECT,3480 OK */
+#define SENSE_ID		0xE4	/* 3420 REJECT,3480 OK */
+#define READ_DEV_CHAR		0x64	/* Read device characteristics */
+#define SET_DIAGNOSE		0x4B	/* 3420 NOP, 3480 REJECT */
+#define SET_GROUP_ID		0xAF	/* 3420 REJECT,3480 OK */
+#define SET_TAPE_WRITE_IMMED	0xC3	/* for 3480 */
+#define SUSPEND			0x5B	/* 3420 REJ, 3480 NOP */
+#define SYNC			0x43	/* Synchronize (flush buffer) */
+#define UNASSIGN		0xC7	/* 3420 REJECT,3480 OK */
+#define PERF_SUBSYS_FUNC	0x77	/* 3490 CMD */
+#define READ_CONFIG_DATA	0xFA	/* 3490 CMD */
+#define READ_MESSAGE_ID		0x4E	/* 3490 CMD */
+#define READ_SUBSYS_DATA	0x3E	/* 3490 CMD */
+#define SET_INTERFACE_ID	0x73	/* 3490 CMD */
+
+#define SENSE_COMMAND_REJECT		0x80
+#define SENSE_INTERVENTION_REQUIRED	0x40
+#define SENSE_BUS_OUT_CHECK		0x20
+#define SENSE_EQUIPMENT_CHECK		0x10
+#define SENSE_DATA_CHECK		0x08
+#define SENSE_OVERRUN			0x04
+#define SENSE_DEFERRED_UNIT_CHECK	0x02
+#define SENSE_ASSIGNED_ELSEWHERE	0x01
+
+#define SENSE_LOCATE_FAILURE		0x80
+#define SENSE_DRIVE_ONLINE		0x40
+#define SENSE_RESERVED			0x20
+#define SENSE_RECORD_SEQUENCE_ERR	0x10
+#define SENSE_BEGINNING_OF_TAPE		0x08
+#define SENSE_WRITE_MODE		0x04
+#define SENSE_WRITE_PROTECT		0x02
+#define SENSE_NOT_CAPABLE		0x01
+
+#define SENSE_CHANNEL_ADAPTER_CODE	0xE0
+#define SENSE_CHANNEL_ADAPTER_LOC	0x10
+#define SENSE_REPORTING_CU		0x08
+#define SENSE_AUTOMATIC_LOADER		0x04
+#define SENSE_TAPE_SYNC_MODE		0x02
+#define SENSE_TAPE_POSITIONING		0x01
+
+/* discipline functions */
+struct tape_request *tape_std_read_block(struct tape_device *, size_t);
+void tape_std_read_backward(struct tape_device *device,
+			    struct tape_request *request);
+struct tape_request *tape_std_write_block(struct tape_device *, size_t);
+struct tape_request *tape_std_bread(struct tape_device *, struct request *);
+void tape_std_free_bread(struct tape_request *);
+void tape_std_check_locate(struct tape_device *, struct tape_request *);
+struct tape_request *tape_std_bwrite(struct request *,
+				     struct tape_device *, int);
+
+/* Some non-mtop commands. */
+int tape_std_assign(struct tape_device *);
+int tape_std_unassign(struct tape_device *);
+int tape_std_read_block_id(struct tape_device *device, __u64 *id);
+int tape_std_display(struct tape_device *, struct display_struct *disp);
+int tape_std_terminate_write(struct tape_device *);
+
+/* Standard magnetic tape commands. */
+int tape_std_mtbsf(struct tape_device *, int);
+int tape_std_mtbsfm(struct tape_device *, int);
+int tape_std_mtbsr(struct tape_device *, int);
+int tape_std_mtcompression(struct tape_device *, int);
+int tape_std_mteom(struct tape_device *, int);
+int tape_std_mterase(struct tape_device *, int);
+int tape_std_mtfsf(struct tape_device *, int);
+int tape_std_mtfsfm(struct tape_device *, int);
+int tape_std_mtfsr(struct tape_device *, int);
+int tape_std_mtload(struct tape_device *, int);
+int tape_std_mtnop(struct tape_device *, int);
+int tape_std_mtoffl(struct tape_device *, int);
+int tape_std_mtreset(struct tape_device *, int);
+int tape_std_mtreten(struct tape_device *, int);
+int tape_std_mtrew(struct tape_device *, int);
+int tape_std_mtsetblk(struct tape_device *, int);
+int tape_std_mtunload(struct tape_device *, int);
+int tape_std_mtweof(struct tape_device *, int);
+
+/* Event handlers */
+void tape_std_default_handler(struct tape_device *);
+void tape_std_unexpect_uchk_handler(struct tape_device *);
+void tape_std_irq(struct tape_device *);
+void tape_std_process_eov(struct tape_device *);
+
+// the error recovery stuff:
+void tape_std_error_recovery(struct tape_device *);
+void tape_std_error_recovery_has_failed(struct tape_device *,int error_id);
+void tape_std_error_recovery_succeded(struct tape_device *);
+void tape_std_error_recovery_do_retry(struct tape_device *);
+void tape_std_error_recovery_read_opposite(struct tape_device *);
+void tape_std_error_recovery_HWBUG(struct tape_device *, int condno);
+
+#endif // _TAPE_STD_H

drivers/s390/char/vmlogrdr.c

@@ -0,0 +1,920 @@
+/*
+ * drivers/s390/char/vmlogrdr.c
+ *	character device driver for reading z/VM system service records
+ *
+ *
+ *	Copyright (C) 2004 IBM Corporation
+ *	character device driver for reading z/VM system service records,
+ *	Version 1.0
+ *	Author(s): Xenia Tkatschow <xenia@us.ibm.com>
+ *		   Stefan Weinhuber <wein@de.ibm.com>
+ *
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+#include <asm/atomic.h>
+#include <asm/uaccess.h>
+#include <asm/cpcmd.h>
+#include <asm/debug.h>
+#include <asm/ebcdic.h>
+#include "../net/iucv.h"
+#include <linux/kmod.h>
+#include <linux/cdev.h>
+#include <linux/device.h>
+#include <linux/string.h>
+
+
+
+MODULE_AUTHOR
+	("(C) 2004 IBM Corporation by Xenia Tkatschow ([email protected])\n"
+	 "                            Stefan Weinhuber ([email protected])");
+MODULE_DESCRIPTION ("Character device driver for reading z/VM "
+		    "system service records.");
+MODULE_LICENSE("GPL");
+
+
+/*
+ * The size of the buffer for iucv data transfer is one page,
+ * but in addition to the data we read from iucv we also
+ * place an integer and some characters into that buffer,
+ * so the maximum size for record data is a little less then
+ * one page.
+ */
+#define NET_BUFFER_SIZE	(PAGE_SIZE - sizeof(int) - sizeof(FENCE))
+
+/*
+ * The elements that are concurrently accessed by bottom halves are
+ * connection_established, iucv_path_severed, local_interrupt_buffer
+ * and receive_ready. The first three can be protected by
+ * priv_lock.  receive_ready is atomic, so it can be incremented and
+ * decremented without holding a lock.
+ * The variable dev_in_use needs to be protected by the lock, since
+ * it's a flag used by open to make sure that the device is opened only
+ * by one user at the same time.
+ */
+struct vmlogrdr_priv_t {
+	char system_service[8];
+	char internal_name[8];
+	char recording_name[8];
+	u16 pathid;
+	int connection_established;
+	int iucv_path_severed;
+	iucv_MessagePending local_interrupt_buffer;
+	atomic_t receive_ready;
+	iucv_handle_t iucv_handle;
+	int minor_num;
+	char * buffer;
+	char * current_position;
+	int remaining;
+	ulong residual_length;
+	int buffer_free;
+	int dev_in_use; /* 1: already opened, 0: not opened*/
+	spinlock_t priv_lock;
+	struct device  *device;
+	struct class_device  *class_device;
+	int autorecording;
+	int autopurge;
+};
+
+
+/*
+ * File operation structure for vmlogrdr devices
+ */
+static int vmlogrdr_open(struct inode *, struct file *);
+static int vmlogrdr_release(struct inode *, struct file *);
+static ssize_t vmlogrdr_read (struct file *filp, char *data, size_t count,
+			       loff_t * ppos);
+
+static struct file_operations vmlogrdr_fops = {
+	.owner   = THIS_MODULE,
+	.open    = vmlogrdr_open,
+	.release = vmlogrdr_release,
+	.read    = vmlogrdr_read,
+};
+
+
+static u8 iucvMagic[16] = {
+	0xF0, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40,
+	0xF0, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40
+};
+
+
+static u8 mask[] = {
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff
+};
+
+
+static u8 iucv_host[8] = { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 };
+
+
+static void
+vmlogrdr_iucv_ConnectionComplete(iucv_ConnectionComplete *eib, void *pgm_data);
+static void
+vmlogrdr_iucv_ConnectionSevered(iucv_ConnectionSevered *eib, void *pgm_data);
+static void
+vmlogrdr_iucv_MessagePending(iucv_MessagePending *eib, void *pgm_data);
+
+
+static iucv_interrupt_ops_t vmlogrdr_iucvops = {
+	.ConnectionComplete = vmlogrdr_iucv_ConnectionComplete,
+	.ConnectionSevered  = vmlogrdr_iucv_ConnectionSevered,
+	.MessagePending     = vmlogrdr_iucv_MessagePending,
+};
+
+
+DECLARE_WAIT_QUEUE_HEAD(conn_wait_queue);
+DECLARE_WAIT_QUEUE_HEAD(read_wait_queue);
+
+/*
+ * pointer to system service private structure
+ * minor number 0 --> logrec
+ * minor number 1 --> account
+ * minor number 2 --> symptom
+ */
+
+static struct vmlogrdr_priv_t sys_ser[] = {
+	{ .system_service = "*LOGREC ",
+	  .internal_name  = "logrec",
+	  .recording_name = "EREP",
+	  .minor_num      = 0,
+	  .buffer_free    = 1,
+	  .priv_lock      = SPIN_LOCK_UNLOCKED,
+	  .autorecording  = 1,
+	  .autopurge      = 1,
+	},
+	{ .system_service = "*ACCOUNT",
+	  .internal_name  = "account",
+	  .recording_name = "ACCOUNT",
+	  .minor_num      = 1,
+	  .buffer_free    = 1,
+	  .priv_lock      = SPIN_LOCK_UNLOCKED,
+	  .autorecording  = 1,
+	  .autopurge      = 1,
+	},
+	{ .system_service = "*SYMPTOM",
+	  .internal_name  = "symptom",
+	  .recording_name = "SYMPTOM",
+	  .minor_num      = 2,
+	  .buffer_free    = 1,
+	  .priv_lock      = SPIN_LOCK_UNLOCKED,
+	  .autorecording  = 1,
+	  .autopurge      = 1,
+	}
+};
+
+#define MAXMINOR  (sizeof(sys_ser)/sizeof(struct vmlogrdr_priv_t))
+
+static char FENCE[] = {"EOR"};
+static int vmlogrdr_major = 0;
+static struct cdev  *vmlogrdr_cdev = NULL;
+static int recording_class_AB;
+
+
+static void
+vmlogrdr_iucv_ConnectionComplete (iucv_ConnectionComplete * eib,
+				   void * pgm_data)
+{
+	struct vmlogrdr_priv_t * logptr = pgm_data;
+	spin_lock(&logptr->priv_lock);
+	logptr->connection_established = 1;
+	spin_unlock(&logptr->priv_lock);
+	wake_up(&conn_wait_queue);
+	return;
+}
+
+
+static void
+vmlogrdr_iucv_ConnectionSevered (iucv_ConnectionSevered * eib, void * pgm_data)
+{
+	u8 reason = (u8) eib->ipuser[8];
+	struct vmlogrdr_priv_t * logptr = pgm_data;
+
+	printk (KERN_ERR "vmlogrdr: connection severed with"
+		" reason %i\n", reason);
+
+	spin_lock(&logptr->priv_lock);
+	logptr->connection_established = 0;
+	logptr->iucv_path_severed = 1;
+	spin_unlock(&logptr->priv_lock);
+
+	wake_up(&conn_wait_queue);
+	/* just in case we're sleeping waiting for a record */
+	wake_up_interruptible(&read_wait_queue);
+}
+
+
+static void
+vmlogrdr_iucv_MessagePending (iucv_MessagePending * eib, void * pgm_data)
+{
+	struct vmlogrdr_priv_t * logptr = pgm_data;
+
+	/*
+	 * This function is the bottom half so it should be quick.
+	 * Copy the external interrupt data into our local eib and increment
+	 * the usage count
+	 */
+	spin_lock(&logptr->priv_lock);
+	memcpy(&(logptr->local_interrupt_buffer), eib, sizeof(*eib));
+	atomic_inc(&logptr->receive_ready);
+	spin_unlock(&logptr->priv_lock);
+	wake_up_interruptible(&read_wait_queue);
+}
+
+
+static int
+vmlogrdr_get_recording_class_AB(void) {
+	char cp_command[]="QUERY COMMAND RECORDING ";
+	char cp_response[80];
+	char *tail;
+	int len,i;
+
+	printk (KERN_DEBUG "vmlogrdr: query command: %s\n", cp_command);
+	cpcmd(cp_command, cp_response, sizeof(cp_response));
+	printk (KERN_DEBUG "vmlogrdr: response: %s", cp_response);
+	len = strnlen(cp_response,sizeof(cp_response));
+	// now the parsing
+	tail=strnchr(cp_response,len,'=');
+	if (!tail)
+		return 0;
+	tail++;
+	if (!strncmp("ANY",tail,3))
+		return 1;
+	if (!strncmp("NONE",tail,4))
+		return 0;
+	/*
+	 * expect comma separated list of classes here, if one of them
+	 * is A or B return 1 otherwise 0
+	 */
+        for (i=tail-cp_response; i<len; i++)
+		if ( cp_response[i]=='A' || cp_response[i]=='B' )
+			return 1;
+	return 0;
+}
+
+
+static int
+vmlogrdr_recording(struct vmlogrdr_priv_t * logptr, int action, int purge) {
+
+	char cp_command[80];
+	char cp_response[160];
+	char *onoff, *qid_string;
+
+	memset(cp_command, 0x00, sizeof(cp_command));
+	memset(cp_response, 0x00, sizeof(cp_response));
+
+        onoff = ((action == 1) ? "ON" : "OFF");
+	qid_string = ((recording_class_AB == 1) ? " QID * " : "");
+
+        /*
+	 * The recording commands needs to be called with option QID
+	 * for guests that have previlege classes A or B.
+	 * Purging has to be done as separate step, because recording
+	 * can't be switched on as long as records are on the queue.
+	 * Doing both at the same time doesn't work.
+	 */
+
+	if (purge) {
+		snprintf(cp_command, sizeof(cp_command),
+			 "RECORDING %s PURGE %s",
+			 logptr->recording_name,
+			 qid_string);
+
+		printk (KERN_DEBUG "vmlogrdr: recording command: %s\n",
+			cp_command);
+		cpcmd(cp_command, cp_response, sizeof(cp_response));
+		printk (KERN_DEBUG "vmlogrdr: recording response: %s",
+			cp_response);
+	}
+
+	memset(cp_command, 0x00, sizeof(cp_command));
+	memset(cp_response, 0x00, sizeof(cp_response));
+	snprintf(cp_command, sizeof(cp_command), "RECORDING %s %s %s",
+		logptr->recording_name,
+		onoff,
+		qid_string);
+
+	printk (KERN_DEBUG "vmlogrdr: recording command: %s\n", cp_command);
+	cpcmd(cp_command, cp_response, sizeof(cp_response));
+	printk (KERN_DEBUG "vmlogrdr: recording response: %s",
+		cp_response);
+	/* The recording command will usually answer with 'Command complete'
+	 * on success, but when the specific service was never connected
+	 * before then there might be an additional informational message
+	 * 'HCPCRC8072I Recording entry not found' before the
+         * 'Command complete'. So I use strstr rather then the strncmp.
+	 */
+	if (strstr(cp_response,"Command complete"))
+		return 0;
+	else
+		return -EIO;
+
+}
+
+
+static int
+vmlogrdr_open (struct inode *inode, struct file *filp)
+{
+	int dev_num = 0;
+	struct vmlogrdr_priv_t * logptr = NULL;
+	int connect_rc = 0;
+	int ret;
+
+	dev_num = iminor(inode);
+	if (dev_num > MAXMINOR)
+		return -ENODEV;
+
+	logptr = &sys_ser[dev_num];
+	if (logptr == NULL)
+		return -ENODEV;
+
+	/*
+	 * only allow for blocking reads to be open
+	 */
+	if (filp->f_flags & O_NONBLOCK)
+		return -ENOSYS;
+
+	/* Besure this device hasn't already been opened */
+	spin_lock_bh(&logptr->priv_lock);
+	if (logptr->dev_in_use)	{
+		spin_unlock_bh(&logptr->priv_lock);
+		return -EBUSY;
+	} else {
+		logptr->dev_in_use = 1;
+		spin_unlock_bh(&logptr->priv_lock);
+	}
+
+	atomic_set(&logptr->receive_ready, 0);
+	logptr->buffer_free = 1;
+
+	/* set the file options */
+	filp->private_data = logptr;
+	filp->f_op = &vmlogrdr_fops;
+
+	/* start recording for this service*/
+	ret=0;
+	if (logptr->autorecording)
+		ret = vmlogrdr_recording(logptr,1,logptr->autopurge);
+	if (ret)
+		printk (KERN_WARNING "vmlogrdr: failed to start "
+			"recording automatically\n");
+
+	/* Register with iucv driver */
+	logptr->iucv_handle = iucv_register_program(iucvMagic,
+			logptr->system_service, mask, &vmlogrdr_iucvops,
+			logptr);
+
+	if (logptr->iucv_handle == NULL) {
+		printk (KERN_ERR "vmlogrdr: failed to register with"
+			"iucv driver\n");
+		goto not_registered;
+	}
+
+	/* create connection to the system service */
+	spin_lock_bh(&logptr->priv_lock);
+	logptr->connection_established = 0;
+	logptr->iucv_path_severed = 0;
+	spin_unlock_bh(&logptr->priv_lock);
+
+	connect_rc = iucv_connect (&(logptr->pathid), 10, iucvMagic,
+					logptr->system_service, iucv_host, 0,
+					NULL, NULL,
+					logptr->iucv_handle, NULL);
+	if (connect_rc) {
+		printk (KERN_ERR "vmlogrdr: iucv connection to %s "
+			"failed with rc %i \n", logptr->system_service,
+			connect_rc);
+		goto not_connected;
+	}
+
+	/* We've issued the connect and now we must wait for a
+	 * ConnectionComplete or ConnectinSevered Interrupt
+	 * before we can continue to process.
+	 */
+	wait_event(conn_wait_queue, (logptr->connection_established)
+		   || (logptr->iucv_path_severed));
+	if (logptr->iucv_path_severed) {
+		goto not_connected;
+	}
+
+ 	return nonseekable_open(inode, filp);
+
+not_connected:
+	iucv_unregister_program(logptr->iucv_handle);
+	logptr->iucv_handle = NULL;
+not_registered:
+	if (logptr->autorecording)
+		vmlogrdr_recording(logptr,0,logptr->autopurge);
+	logptr->dev_in_use = 0;
+	return -EIO;
+
+
+}
+
+
+static int
+vmlogrdr_release (struct inode *inode, struct file *filp)
+{
+	int ret;
+
+	struct vmlogrdr_priv_t * logptr = filp->private_data;
+
+	iucv_unregister_program(logptr->iucv_handle);
+	logptr->iucv_handle = NULL;
+
+	if (logptr->autorecording) {
+		ret = vmlogrdr_recording(logptr,0,logptr->autopurge);
+		if (ret)
+			printk (KERN_WARNING "vmlogrdr: failed to stop "
+				"recording automatically\n");
+	}
+	logptr->dev_in_use = 0;
+
+	return 0;
+}
+
+
+static int
+vmlogrdr_receive_data(struct vmlogrdr_priv_t *priv) {
+	int rc, *temp;
+	/* we need to keep track of two data sizes here:
+	 * The number of bytes we need to receive from iucv and
+	 * the total number of bytes we actually write into the buffer.
+	 */
+	int user_data_count, iucv_data_count;
+	char * buffer;
+
+	if (atomic_read(&priv->receive_ready)) {
+		spin_lock_bh(&priv->priv_lock);
+		if (priv->residual_length){
+			/* receive second half of a record */
+			iucv_data_count = priv->residual_length;
+			user_data_count = 0;
+			buffer = priv->buffer;
+		} else {
+			/* receive a new record:
+			 * We need to return the total length of the record
+                         * + size of FENCE in the first 4 bytes of the buffer.
+		         */
+			iucv_data_count =
+				priv->local_interrupt_buffer.ln1msg2.ipbfln1f;
+			user_data_count = sizeof(int);
+			temp = (int*)priv->buffer;
+			*temp= iucv_data_count + sizeof(FENCE);
+			buffer = priv->buffer + sizeof(int);
+		}
+		/*
+		 * If the record is bigger then our buffer, we receive only
+		 * a part of it. We can get the rest later.
+		 */
+		if (iucv_data_count > NET_BUFFER_SIZE)
+			iucv_data_count = NET_BUFFER_SIZE;
+		rc = iucv_receive(priv->pathid,
+				  priv->local_interrupt_buffer.ipmsgid,
+				  priv->local_interrupt_buffer.iptrgcls,
+				  buffer,
+				  iucv_data_count,
+				  NULL,
+				  NULL,
+				  &priv->residual_length);
+		spin_unlock_bh(&priv->priv_lock);
+		/* An rc of 5 indicates that the record was bigger then
+		 * the buffer, which is OK for us. A 9 indicates that the
+		 * record was purged befor we could receive it.
+		 */
+		if (rc == 5)
+			rc = 0;
+		if (rc == 9)
+			atomic_set(&priv->receive_ready, 0);
+	} else {
+		rc = 1;
+	}
+	if (!rc) {
+		priv->buffer_free = 0;
+ 		user_data_count += iucv_data_count;
+		priv->current_position = priv->buffer;
+		if (priv->residual_length == 0){
+			/* the whole record has been captured,
+			 * now add the fence */
+			atomic_dec(&priv->receive_ready);
+			buffer = priv->buffer + user_data_count;
+			memcpy(buffer, FENCE, sizeof(FENCE));
+			user_data_count += sizeof(FENCE);
+		}
+		priv->remaining = user_data_count;
+	}
+
+	return rc;
+}
+
+
+static ssize_t
+vmlogrdr_read (struct file *filp, char *data, size_t count, loff_t * ppos)
+{
+	int rc;
+	struct vmlogrdr_priv_t * priv = filp->private_data;
+
+	while (priv->buffer_free) {
+		rc = vmlogrdr_receive_data(priv);
+		if (rc) {
+			rc = wait_event_interruptible(read_wait_queue,
+					atomic_read(&priv->receive_ready));
+			if (rc)
+				return rc;
+		}
+	}
+	/* copy only up to end of record */
+	if (count > priv->remaining)
+		count = priv->remaining;
+
+	if (copy_to_user(data, priv->current_position, count))
+		return -EFAULT;
+
+	*ppos += count;
+	priv->current_position += count;
+	priv->remaining -= count;
+
+	/* if all data has been transferred, set buffer free */
+	if (priv->remaining == 0)
+		priv->buffer_free = 1;
+
+	return count;
+}
+
+static ssize_t
+vmlogrdr_autopurge_store(struct device * dev, const char * buf, size_t count) {
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+	ssize_t ret = count;
+
+	switch (buf[0]) {
+	case '0':
+		priv->autopurge=0;
+		break;
+	case '1':
+		priv->autopurge=1;
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	return ret;
+}
+
+
+static ssize_t
+vmlogrdr_autopurge_show(struct device *dev, char *buf) {
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+	return sprintf(buf, "%u\n", priv->autopurge);
+}
+
+
+static DEVICE_ATTR(autopurge, 0644, vmlogrdr_autopurge_show,
+		   vmlogrdr_autopurge_store);
+
+
+static ssize_t
+vmlogrdr_purge_store(struct device * dev, const char * buf, size_t count) {
+
+	char cp_command[80];
+	char cp_response[80];
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+
+	if (buf[0] != '1')
+		return -EINVAL;
+
+	memset(cp_command, 0x00, sizeof(cp_command));
+	memset(cp_response, 0x00, sizeof(cp_response));
+
+        /*
+	 * The recording command needs to be called with option QID
+	 * for guests that have previlege classes A or B.
+	 * Other guests will not recognize the command and we have to
+	 * issue the same command without the QID parameter.
+	 */
+
+	if (recording_class_AB)
+		snprintf(cp_command, sizeof(cp_command),
+			 "RECORDING %s PURGE QID * ",
+			 priv->recording_name);
+	else
+		snprintf(cp_command, sizeof(cp_command),
+			 "RECORDING %s PURGE ",
+			 priv->recording_name);
+
+	printk (KERN_DEBUG "vmlogrdr: recording command: %s\n", cp_command);
+	cpcmd(cp_command, cp_response, sizeof(cp_response));
+	printk (KERN_DEBUG "vmlogrdr: recording response: %s",
+		cp_response);
+
+	return count;
+}
+
+
+static DEVICE_ATTR(purge, 0200, NULL, vmlogrdr_purge_store);
+
+
+static ssize_t
+vmlogrdr_autorecording_store(struct device *dev, const char *buf,
+			     size_t count) {
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+	ssize_t ret = count;
+
+	switch (buf[0]) {
+	case '0':
+		priv->autorecording=0;
+		break;
+	case '1':
+		priv->autorecording=1;
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	return ret;
+}
+
+
+static ssize_t
+vmlogrdr_autorecording_show(struct device *dev, char *buf) {
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+	return sprintf(buf, "%u\n", priv->autorecording);
+}
+
+
+static DEVICE_ATTR(autorecording, 0644, vmlogrdr_autorecording_show,
+		   vmlogrdr_autorecording_store);
+
+
+static ssize_t
+vmlogrdr_recording_store(struct device * dev, const char * buf, size_t count) {
+
+	struct vmlogrdr_priv_t *priv = dev->driver_data;
+	ssize_t ret;
+
+	switch (buf[0]) {
+	case '0':
+		ret = vmlogrdr_recording(priv,0,0);
+		break;
+	case '1':
+		ret = vmlogrdr_recording(priv,1,0);
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	if (ret)
+		return ret;
+	else
+		return count;
+
+}
+
+
+static DEVICE_ATTR(recording, 0200, NULL, vmlogrdr_recording_store);
+
+
+static ssize_t
+vmlogrdr_recording_status_show(struct device_driver *driver, char *buf) {
+
+	char cp_command[] = "QUERY RECORDING ";
+	int len;
+
+	cpcmd(cp_command, buf, 4096);
+	len = strlen(buf);
+	return len;
+}
+
+
+static DRIVER_ATTR(recording_status, 0444, vmlogrdr_recording_status_show,
+		   NULL);
+
+static struct attribute *vmlogrdr_attrs[] = {
+	&dev_attr_autopurge.attr,
+	&dev_attr_purge.attr,
+	&dev_attr_autorecording.attr,
+	&dev_attr_recording.attr,
+	NULL,
+};
+
+static struct attribute_group vmlogrdr_attr_group = {
+	.attrs = vmlogrdr_attrs,
+};
+
+static struct class_simple *vmlogrdr_class;
+static struct device_driver vmlogrdr_driver = {
+	.name = "vmlogrdr",
+	.bus  = &iucv_bus,
+};
+
+
+static int
+vmlogrdr_register_driver(void) {
+	int ret;
+
+	ret = driver_register(&vmlogrdr_driver);
+	if (ret) {
+		printk(KERN_ERR "vmlogrdr: failed to register driver.\n");
+		return ret;
+	}
+
+	ret = driver_create_file(&vmlogrdr_driver,
+				 &driver_attr_recording_status);
+	if (ret) {
+		printk(KERN_ERR "vmlogrdr: failed to add driver attribute.\n");
+		goto unregdriver;
+	}
+
+	vmlogrdr_class = class_simple_create(THIS_MODULE, "vmlogrdr");
+	if (IS_ERR(vmlogrdr_class)) {
+		printk(KERN_ERR "vmlogrdr: failed to create class.\n");
+		ret=PTR_ERR(vmlogrdr_class);
+		vmlogrdr_class=NULL;
+		goto unregattr;
+	}
+	return 0;
+
+unregattr:
+	driver_remove_file(&vmlogrdr_driver, &driver_attr_recording_status);
+unregdriver:
+	driver_unregister(&vmlogrdr_driver);
+	return ret;
+}
+
+
+static void
+vmlogrdr_unregister_driver(void) {
+	class_simple_destroy(vmlogrdr_class);
+	vmlogrdr_class = NULL;
+	driver_remove_file(&vmlogrdr_driver, &driver_attr_recording_status);
+	driver_unregister(&vmlogrdr_driver);
+	return;
+}
+
+
+static int
+vmlogrdr_register_device(struct vmlogrdr_priv_t *priv) {
+	struct device *dev;
+	int ret;
+
+	dev = kmalloc(sizeof(struct device), GFP_KERNEL);
+	if (dev) {
+		memset(dev, 0, sizeof(struct device));
+		snprintf(dev->bus_id, BUS_ID_SIZE, "%s",
+			 priv->internal_name);
+		dev->bus = &iucv_bus;
+		dev->parent = iucv_root;
+		dev->driver = &vmlogrdr_driver;
+		/*
+		 * The release function could be called after the
+		 * module has been unloaded. It's _only_ task is to
+		 * free the struct. Therefore, we specify kfree()
+		 * directly here. (Probably a little bit obfuscating
+		 * but legitime ...).
+		 */
+		dev->release = (void (*)(struct device *))kfree;
+	} else
+		return -ENOMEM;
+	ret = device_register(dev);
+	if (ret)
+		return ret;
+
+	ret = sysfs_create_group(&dev->kobj, &vmlogrdr_attr_group);
+	if (ret) {
+		device_unregister(dev);
+		return ret;
+	}
+	priv->class_device = class_simple_device_add(
+				vmlogrdr_class,
+				MKDEV(vmlogrdr_major, priv->minor_num),
+				dev,
+				"%s", dev->bus_id );
+	if (IS_ERR(priv->class_device)) {
+		ret = PTR_ERR(priv->class_device);
+		priv->class_device=NULL;
+		sysfs_remove_group(&dev->kobj, &vmlogrdr_attr_group);
+		device_unregister(dev);
+		return ret;
+	}
+	dev->driver_data = priv;
+	priv->device = dev;
+	return 0;
+}
+
+
+static int
+vmlogrdr_unregister_device(struct vmlogrdr_priv_t *priv ) {
+	class_simple_device_remove(MKDEV(vmlogrdr_major, priv->minor_num));
+	if (priv->device != NULL) {
+		sysfs_remove_group(&priv->device->kobj, &vmlogrdr_attr_group);
+		device_unregister(priv->device);
+		priv->device=NULL;
+	}
+	return 0;
+}
+
+
+static int
+vmlogrdr_register_cdev(dev_t dev) {
+	int rc = 0;
+	vmlogrdr_cdev = cdev_alloc();
+	if (!vmlogrdr_cdev) {
+		return -ENOMEM;
+	}
+	vmlogrdr_cdev->owner = THIS_MODULE;
+	vmlogrdr_cdev->ops = &vmlogrdr_fops;
+	vmlogrdr_cdev->dev = dev;
+	rc = cdev_add(vmlogrdr_cdev, vmlogrdr_cdev->dev, MAXMINOR);
+	if (!rc)
+		return 0;
+
+	// cleanup: cdev is not fully registered, no cdev_del here!
+	kobject_put(&vmlogrdr_cdev->kobj);
+	vmlogrdr_cdev=NULL;
+	return rc;
+}
+
+
+static void
+vmlogrdr_cleanup(void) {
+        int i;
+	if (vmlogrdr_cdev) {
+		cdev_del(vmlogrdr_cdev);
+		vmlogrdr_cdev=NULL;
+	}
+	for (i=0; i < MAXMINOR; ++i ) {
+		vmlogrdr_unregister_device(&sys_ser[i]);
+		free_page((unsigned long)sys_ser[i].buffer);
+	}
+	vmlogrdr_unregister_driver();
+	if (vmlogrdr_major) {
+		unregister_chrdev_region(MKDEV(vmlogrdr_major, 0), MAXMINOR);
+		vmlogrdr_major=0;
+	}
+}
+
+
+static int
+vmlogrdr_init(void)
+{
+	int rc;
+	int i;
+	dev_t dev;
+
+	if (! MACHINE_IS_VM) {
+		printk (KERN_ERR "vmlogrdr: not running under VM, "
+				"driver not loaded.\n");
+		return -ENODEV;
+	}
+
+        recording_class_AB = vmlogrdr_get_recording_class_AB();
+
+	rc = alloc_chrdev_region(&dev, 0, MAXMINOR, "vmlogrdr");
+	if (rc)
+		return rc;
+	vmlogrdr_major = MAJOR(dev);
+
+	rc=vmlogrdr_register_driver();
+	if (rc)
+		goto cleanup;
+
+	for (i=0; i < MAXMINOR; ++i ) {
+		sys_ser[i].buffer = (char *) get_zeroed_page(GFP_KERNEL);
+		if (!sys_ser[i].buffer) {
+			rc = ENOMEM;
+			break;
+		}
+		sys_ser[i].current_position = sys_ser[i].buffer;
+		rc=vmlogrdr_register_device(&sys_ser[i]);
+		if (rc)
+			break;
+	}
+	if (rc)
+		goto cleanup;
+
+	rc = vmlogrdr_register_cdev(dev);
+	if (rc)
+		goto cleanup;
+	printk (KERN_INFO "vmlogrdr: driver loaded\n");
+	return 0;
+
+cleanup:
+	vmlogrdr_cleanup();
+	printk (KERN_ERR "vmlogrdr: driver not loaded.\n");
+	return rc;
+}
+
+
+static void
+vmlogrdr_exit(void)
+{
+	vmlogrdr_cleanup();
+	printk (KERN_INFO "vmlogrdr: driver unloaded\n");
+	return;
+}
+
+
+module_init(vmlogrdr_init);
+module_exit(vmlogrdr_exit);

drivers/s390/char/vmwatchdog.c

@@ -0,0 +1,292 @@
+/*
+ * Watchdog implementation based on z/VM Watchdog Timer API
+ *
+ * The user space watchdog daemon can use this driver as
+ * /dev/vmwatchdog to have z/VM execute the specified CP
+ * command when the timeout expires. The default command is
+ * "IPL", which which cause an immediate reboot.
+ */
+#include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/kernel.h>
+#include <linux/miscdevice.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/watchdog.h>
+
+#include <asm/ebcdic.h>
+#include <asm/io.h>
+#include <asm/uaccess.h>
+
+#define MAX_CMDLEN 240
+#define MIN_INTERVAL 15
+static char vmwdt_cmd[MAX_CMDLEN] = "IPL";
+static int vmwdt_conceal;
+
+#ifdef CONFIG_WATCHDOG_NOWAYOUT
+static int vmwdt_nowayout = 1;
+#else
+static int vmwdt_nowayout = 0;
+#endif
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Arnd Bergmann <[email protected]>");
+MODULE_DESCRIPTION("z/VM Watchdog Timer");
+module_param_string(cmd, vmwdt_cmd, MAX_CMDLEN, 0644);
+MODULE_PARM_DESC(cmd, "CP command that is run when the watchdog triggers");
+module_param_named(conceal, vmwdt_conceal, bool, 0644);
+MODULE_PARM_DESC(conceal, "Enable the CONCEAL CP option while the watchdog "
+		" is active");
+module_param_named(nowayout, vmwdt_nowayout, bool, 0);
+MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started"
+		" (default=CONFIG_WATCHDOG_NOWAYOUT)");
+MODULE_ALIAS_MISCDEV(WATCHDOG_MINOR);
+
+static unsigned int vmwdt_interval = 60;
+static unsigned long vmwdt_is_open;
+static int vmwdt_expect_close;
+
+enum vmwdt_func {
+	/* function codes */
+	wdt_init   = 0,
+	wdt_change = 1,
+	wdt_cancel = 2,
+	/* flags */
+	wdt_conceal = 0x80000000,
+};
+
+static int __diag288(enum vmwdt_func func, unsigned int timeout,
+			    char *cmd, size_t len)
+{
+	register unsigned long __func asm("2");
+	register unsigned long __timeout asm("3");
+	register unsigned long __cmdp asm("4");
+	register unsigned long __cmdl asm("5");
+	int err;
+
+	__func = func;
+	__timeout = timeout;
+	__cmdp = virt_to_phys(cmd);
+	__cmdl = len;
+	err = 0;
+	asm volatile (
+#ifdef __s390x__
+		       "diag %2,%4,0x288\n"
+		"1:	\n"
+		".section .fixup,\"ax\"\n"
+		"2:	lghi %0,%1\n"
+		"	jg 1b\n"
+		".previous\n"
+		".section __ex_table,\"a\"\n"
+		"	.align 8\n"
+		"	.quad 1b,2b\n"
+		".previous\n"
+#else
+		       "diag %2,%4,0x288\n"
+		"1:	\n"
+		".section .fixup,\"ax\"\n"
+		"2:	lhi %0,%1\n"
+		"	bras 1,3f\n"
+		"	.long 1b\n"
+		"3:	l 1,0(1)\n"
+		"	br 1\n"
+		".previous\n"
+		".section __ex_table,\"a\"\n"
+		"	.align 4\n"
+		"	.long 1b,2b\n"
+		".previous\n"
+#endif
+		: "+&d"(err)
+		: "i"(-EINVAL), "d"(__func), "d"(__timeout),
+		  "d"(__cmdp), "d"(__cmdl)
+		: "1", "cc");
+	return err;
+}
+
+static int vmwdt_keepalive(void)
+{
+	/* we allocate new memory every time to avoid having
+	 * to track the state. static allocation is not an
+	 * option since that might not be contiguous in real
+	 * storage in case of a modular build */
+	static char *ebc_cmd;
+	size_t len;
+	int ret;
+	unsigned int func;
+
+	ebc_cmd = kmalloc(MAX_CMDLEN, GFP_KERNEL);
+	if (!ebc_cmd)
+		return -ENOMEM;
+
+	len = strlcpy(ebc_cmd, vmwdt_cmd, MAX_CMDLEN);
+	ASCEBC(ebc_cmd, MAX_CMDLEN);
+	EBC_TOUPPER(ebc_cmd, MAX_CMDLEN);
+
+	func = vmwdt_conceal ? (wdt_init | wdt_conceal) : wdt_init;
+	ret = __diag288(func, vmwdt_interval, ebc_cmd, len);
+	kfree(ebc_cmd);
+
+	if (ret) {
+		printk(KERN_WARNING "%s: problem setting interval %d, "
+			"cmd %s\n", __FUNCTION__, vmwdt_interval,
+			vmwdt_cmd);
+	}
+	return ret;
+}
+
+static int vmwdt_disable(void)
+{
+	int ret = __diag288(wdt_cancel, 0, "", 0);
+	if (ret) {
+		printk(KERN_WARNING "%s: problem disabling watchdog\n",
+			__FUNCTION__);
+	}
+	return ret;
+}
+
+static int __init vmwdt_probe(void)
+{
+	/* there is no real way to see if the watchdog is supported,
+	 * so we try initializing it with a NOP command ("BEGIN")
+	 * that won't cause any harm even if the following disable
+	 * fails for some reason */
+	static char __initdata ebc_begin[] = {
+		194, 197, 199, 201, 213
+	};
+	if (__diag288(wdt_init, 15, ebc_begin, sizeof(ebc_begin)) != 0) {
+		printk(KERN_INFO "z/VM watchdog not available\n");
+		return -EINVAL;
+	}
+	return vmwdt_disable();
+}
+
+static int vmwdt_open(struct inode *i, struct file *f)
+{
+	int ret;
+	if (test_and_set_bit(0, &vmwdt_is_open))
+		return -EBUSY;
+	ret = vmwdt_keepalive();
+	if (ret)
+		clear_bit(0, &vmwdt_is_open);
+	return ret ? ret : nonseekable_open(i, f);
+}
+
+static int vmwdt_close(struct inode *i, struct file *f)
+{
+	if (vmwdt_expect_close == 42)
+		vmwdt_disable();
+	vmwdt_expect_close = 0;
+	clear_bit(0, &vmwdt_is_open);
+	return 0;
+}
+
+static struct watchdog_info vmwdt_info = {
+	.options = WDIOF_SETTIMEOUT | WDIOF_KEEPALIVEPING | WDIOF_MAGICCLOSE,
+	.firmware_version = 0,
+	.identity = "z/VM Watchdog Timer",
+};
+
+static int vmwdt_ioctl(struct inode *i, struct file *f,
+			  unsigned int cmd, unsigned long arg)
+{
+	switch (cmd) {
+	case WDIOC_GETSUPPORT:
+		if (copy_to_user((void __user *)arg, &vmwdt_info,
+					sizeof(vmwdt_info)))
+			return -EFAULT;
+		return 0;
+	case WDIOC_GETSTATUS:
+	case WDIOC_GETBOOTSTATUS:
+		return put_user(0, (int *)arg);
+	case WDIOC_GETTEMP:
+		return -EINVAL;
+	case WDIOC_SETOPTIONS:
+		{
+			int options, ret;
+			if (get_user(options, (int __user *)arg))
+				return -EFAULT;
+			ret = -EINVAL;
+			if (options & WDIOS_DISABLECARD) {
+				ret = vmwdt_disable();
+				if (ret)
+					return ret;
+			}
+			if (options & WDIOS_ENABLECARD) {
+				ret = vmwdt_keepalive();
+			}
+			return ret;
+		}
+	case WDIOC_GETTIMEOUT:
+		return put_user(vmwdt_interval, (int __user *)arg);
+	case WDIOC_SETTIMEOUT:
+		{
+			int interval;
+			if (get_user(interval, (int __user *)arg))
+				return -EFAULT;
+			if (interval < MIN_INTERVAL)
+				return -EINVAL;
+			vmwdt_interval = interval;
+		}
+		return vmwdt_keepalive();
+	case WDIOC_KEEPALIVE:
+		return vmwdt_keepalive();
+	}
+
+	return -EINVAL;
+}
+
+static ssize_t vmwdt_write(struct file *f, const char __user *buf,
+				size_t count, loff_t *ppos)
+{
+	if(count) {
+		if (!vmwdt_nowayout) {
+			size_t i;
+
+			/* note: just in case someone wrote the magic character
+			 * five months ago... */
+			vmwdt_expect_close = 0;
+
+			for (i = 0; i != count; i++) {
+				char c;
+				if (get_user(c, buf+i))
+					return -EFAULT;
+				if (c == 'V')
+					vmwdt_expect_close = 42;
+			}
+		}
+		/* someone wrote to us, we should restart timer */
+		vmwdt_keepalive();
+	}
+	return count;
+}
+
+static struct file_operations vmwdt_fops = {
+	.open    = &vmwdt_open,
+	.release = &vmwdt_close,
+	.ioctl   = &vmwdt_ioctl,
+	.write   = &vmwdt_write,
+	.owner   = THIS_MODULE,
+};
+
+static struct miscdevice vmwdt_dev = {
+	.minor      = WATCHDOG_MINOR,
+	.name       = "watchdog",
+	.fops       = &vmwdt_fops,
+};
+
+static int __init vmwdt_init(void)
+{
+	int ret;
+
+	ret = vmwdt_probe();
+	if (ret)
+		return ret;
+	return misc_register(&vmwdt_dev);
+}
+module_init(vmwdt_init);
+
+static void __exit vmwdt_exit(void)
+{
+	WARN_ON(misc_deregister(&vmwdt_dev) != 0);
+}
+module_exit(vmwdt_exit);

drivers/s390/cio/Makefile

@@ -0,0 +1,10 @@
+#
+# Makefile for the S/390 common i/o drivers
+#
+
+obj-y += airq.o blacklist.o chsc.o cio.o css.o
+ccw_device-objs += device.o device_fsm.o device_ops.o
+ccw_device-objs += device_id.o device_pgid.o device_status.o
+obj-y += ccw_device.o cmf.o
+obj-$(CONFIG_CCWGROUP) += ccwgroup.o
+obj-$(CONFIG_QDIO) += qdio.o

drivers/s390/cio/airq.c

@@ -0,0 +1,87 @@
+/*
+ *  drivers/s390/cio/airq.c
+ *   S/390 common I/O routines -- support for adapter interruptions
+ *
+ *   $Revision: 1.12 $
+ *
+ *    Copyright (C) 1999-2002 IBM Deutschland Entwicklung GmbH,
+ *			      IBM Corporation
+ *    Author(s): Ingo Adlung (adlung@de.ibm.com)
+ *		 Cornelia Huck (cohuck@de.ibm.com)
+ *		 Arnd Bergmann (arndb@de.ibm.com)
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/rcupdate.h>
+
+#include "cio_debug.h"
+#include "airq.h"
+
+static adapter_int_handler_t adapter_handler;
+
+/*
+ * register for adapter interrupts
+ *
+ * With HiperSockets the zSeries architecture provides for
+ *  means of adapter interrups, pseudo I/O interrupts that are
+ *  not tied to an I/O subchannel, but to an adapter. However,
+ *  it doesn't disclose the info how to enable/disable them, but
+ *  to recognize them only. Perhaps we should consider them
+ *  being shared interrupts, and thus build a linked list
+ *  of adapter handlers ... to be evaluated ...
+ */
+int
+s390_register_adapter_interrupt (adapter_int_handler_t handler)
+{
+	int ret;
+	char dbf_txt[15];
+
+	CIO_TRACE_EVENT (4, "rgaint");
+
+	if (handler == NULL)
+		ret = -EINVAL;
+	else
+		ret = (cmpxchg(&adapter_handler, NULL, handler) ? -EBUSY : 0);
+	if (!ret)
+		synchronize_kernel();
+
+	sprintf (dbf_txt, "ret:%d", ret);
+	CIO_TRACE_EVENT (4, dbf_txt);
+
+	return ret;
+}
+
+int
+s390_unregister_adapter_interrupt (adapter_int_handler_t handler)
+{
+	int ret;
+	char dbf_txt[15];
+
+	CIO_TRACE_EVENT (4, "urgaint");
+
+	if (handler == NULL)
+		ret = -EINVAL;
+	else {
+		adapter_handler = NULL;
+		synchronize_kernel();
+		ret = 0;
+	}
+	sprintf (dbf_txt, "ret:%d", ret);
+	CIO_TRACE_EVENT (4, dbf_txt);
+
+	return ret;
+}
+
+void
+do_adapter_IO (void)
+{
+	CIO_TRACE_EVENT (6, "doaio");
+
+	if (adapter_handler)
+		(*adapter_handler) ();
+}
+
+EXPORT_SYMBOL (s390_register_adapter_interrupt);
+EXPORT_SYMBOL (s390_unregister_adapter_interrupt);

drivers/s390/cio/airq.h

@@ -0,0 +1,10 @@
+#ifndef S390_AINTERRUPT_H
+#define S390_AINTERRUPT_H
+
+typedef	int (*adapter_int_handler_t)(void);
+
+extern int s390_register_adapter_interrupt(adapter_int_handler_t handler);
+extern int s390_unregister_adapter_interrupt(adapter_int_handler_t handler);
+extern void do_adapter_IO (void);
+
+#endif

drivers/s390/cio/blacklist.c

@@ -0,0 +1,351 @@
+/*
+ *  drivers/s390/cio/blacklist.c
+ *   S/390 common I/O routines -- blacklisting of specific devices
+ *   $Revision: 1.33 $
+ *
+ *    Copyright (C) 1999-2002 IBM Deutschland Entwicklung GmbH,
+ *			      IBM Corporation
+ *    Author(s): Ingo Adlung (adlung@de.ibm.com)
+ *		 Cornelia Huck (cohuck@de.ibm.com)
+ *		 Arnd Bergmann (arndb@de.ibm.com)
+ */
+
+#include <linux/config.h>
+#include <linux/init.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+#include <linux/proc_fs.h>
+#include <linux/ctype.h>
+#include <linux/device.h>
+
+#include <asm/cio.h>
+#include <asm/uaccess.h>
+
+#include "blacklist.h"
+#include "cio.h"
+#include "cio_debug.h"
+#include "css.h"
+
+/*
+ * "Blacklisting" of certain devices:
+ * Device numbers given in the commandline as cio_ignore=... won't be known
+ * to Linux.
+ *
+ * These can be single devices or ranges of devices
+ */
+
+/* 65536 bits to indicate if a devno is blacklisted or not */
+#define __BL_DEV_WORDS (__MAX_SUBCHANNELS + (8*sizeof(long) - 1) / \
+			 (8*sizeof(long)))
+static unsigned long bl_dev[__BL_DEV_WORDS];
+typedef enum {add, free} range_action;
+
+/*
+ * Function: blacklist_range
+ * (Un-)blacklist the devices from-to
+ */
+static inline void
+blacklist_range (range_action action, unsigned int from, unsigned int to)
+{
+	if (!to)
+		to = from;
+
+	if (from > to || to > __MAX_SUBCHANNELS) {
+		printk (KERN_WARNING "Invalid blacklist range "
+			"0x%04x to 0x%04x, skipping\n", from, to);
+		return;
+	}
+	for (; from <= to; from++) {
+		if (action == add)
+			set_bit (from, bl_dev);
+		else
+			clear_bit (from, bl_dev);
+	}
+}
+
+/*
+ * Function: blacklist_busid
+ * Get devno/busid from given string.
+ * Shamelessly grabbed from dasd_devmap.c.
+ */
+static inline int
+blacklist_busid(char **str, int *id0, int *id1, int *devno)
+{
+	int val, old_style;
+	char *sav;
+
+	sav = *str;
+
+	/* check for leading '0x' */
+	old_style = 0;
+	if ((*str)[0] == '0' && (*str)[1] == 'x') {
+		*str += 2;
+		old_style = 1;
+	}
+	if (!isxdigit((*str)[0]))	/* We require at least one hex digit */
+		goto confused;
+	val = simple_strtoul(*str, str, 16);
+	if (old_style || (*str)[0] != '.') {
+		*id0 = *id1 = 0;
+		if (val < 0 || val > 0xffff)
+			goto confused;
+		*devno = val;
+		if ((*str)[0] != ',' && (*str)[0] != '-' &&
+		    (*str)[0] != '\n' && (*str)[0] != '\0')
+			goto confused;
+		return 0;
+	}
+	/* New style x.y.z busid */
+	if (val < 0 || val > 0xff)
+		goto confused;
+	*id0 = val;
+	(*str)++;
+	if (!isxdigit((*str)[0]))	/* We require at least one hex digit */
+		goto confused;
+	val = simple_strtoul(*str, str, 16);
+	if (val < 0 || val > 0xff || (*str)++[0] != '.')
+		goto confused;
+	*id1 = val;
+	if (!isxdigit((*str)[0]))	/* We require at least one hex digit */
+		goto confused;
+	val = simple_strtoul(*str, str, 16);
+	if (val < 0 || val > 0xffff)
+		goto confused;
+	*devno = val;
+	if ((*str)[0] != ',' && (*str)[0] != '-' &&
+	    (*str)[0] != '\n' && (*str)[0] != '\0')
+		goto confused;
+	return 0;
+confused:
+	strsep(str, ",\n");
+	printk(KERN_WARNING "Invalid cio_ignore parameter '%s'\n", sav);
+	return 1;
+}
+
+static inline int
+blacklist_parse_parameters (char *str, range_action action)
+{
+	unsigned int from, to, from_id0, to_id0, from_id1, to_id1;
+
+	while (*str != 0 && *str != '\n') {
+		range_action ra = action;
+		while(*str == ',')
+			str++;
+		if (*str == '!') {
+			ra = !action;
+			++str;
+		}
+
+		/*
+		 * Since we have to parse the proc commands and the
+		 * kernel arguments we have to check four cases
+		 */
+		if (strncmp(str,"all,",4) == 0 || strcmp(str,"all") == 0 ||
+		    strncmp(str,"all\n",4) == 0 || strncmp(str,"all ",4) == 0) {
+			from = 0;
+			to = __MAX_SUBCHANNELS;
+			str += 3;
+		} else {
+			int rc;
+
+			rc = blacklist_busid(&str, &from_id0,
+					     &from_id1, &from);
+			if (rc)
+				continue;
+			to = from;
+			to_id0 = from_id0;
+			to_id1 = from_id1;
+			if (*str == '-') {
+				str++;
+				rc = blacklist_busid(&str, &to_id0,
+						     &to_id1, &to);
+				if (rc)
+					continue;
+			}
+			if (*str == '-') {
+				printk(KERN_WARNING "invalid cio_ignore "
+					"parameter '%s'\n",
+					strsep(&str, ",\n"));
+				continue;
+			}
+			if ((from_id0 != to_id0) || (from_id1 != to_id1)) {
+				printk(KERN_WARNING "invalid cio_ignore range "
+					"%x.%x.%04x-%x.%x.%04x\n",
+					from_id0, from_id1, from,
+					to_id0, to_id1, to);
+				continue;
+			}
+		}
+		/* FIXME: ignoring id0 and id1 here. */
+		pr_debug("blacklist_setup: adding range "
+			 "from 0.0.%04x to 0.0.%04x\n", from, to);
+		blacklist_range (ra, from, to);
+	}
+	return 1;
+}
+
+/* Parsing the commandline for blacklist parameters, e.g. to blacklist
+ * bus ids 0.0.1234, 0.0.1235 and 0.0.1236, you could use any of:
+ * - cio_ignore=1234-1236
+ * - cio_ignore=0x1234-0x1235,1236
+ * - cio_ignore=0x1234,1235-1236
+ * - cio_ignore=1236 cio_ignore=1234-0x1236
+ * - cio_ignore=1234 cio_ignore=1236 cio_ignore=0x1235
+ * - cio_ignore=0.0.1234-0.0.1236
+ * - cio_ignore=0.0.1234,0x1235,1236
+ * - ...
+ */
+static int __init
+blacklist_setup (char *str)
+{
+	CIO_MSG_EVENT(6, "Reading blacklist parameters\n");
+	return blacklist_parse_parameters (str, add);
+}
+
+__setup ("cio_ignore=", blacklist_setup);
+
+/* Checking if devices are blacklisted */
+
+/*
+ * Function: is_blacklisted
+ * Returns 1 if the given devicenumber can be found in the blacklist,
+ * otherwise 0.
+ * Used by validate_subchannel()
+ */
+int
+is_blacklisted (int devno)
+{
+	return test_bit (devno, bl_dev);
+}
+
+#ifdef CONFIG_PROC_FS
+/*
+ * Function: s390_redo_validation
+ * Look for no longer blacklisted devices
+ * FIXME: there must be a better way to do this */
+static inline void
+s390_redo_validation (void)
+{
+	unsigned int irq;
+
+	CIO_TRACE_EVENT (0, "redoval");
+	for (irq = 0; irq < __MAX_SUBCHANNELS; irq++) {
+		int ret;
+		struct subchannel *sch;
+
+		sch = get_subchannel_by_schid(irq);
+		if (sch) {
+			/* Already known. */
+			put_device(&sch->dev);
+			continue;
+		}
+		ret = css_probe_device(irq);
+		if (ret == -ENXIO)
+			break; /* We're through. */
+		if (ret == -ENOMEM)
+			/*
+			 * Stop validation for now. Bad, but no need for a
+			 * panic.
+			 */
+			break;
+	}
+}
+
+/*
+ * Function: blacklist_parse_proc_parameters
+ * parse the stuff which is piped to /proc/cio_ignore
+ */
+static inline void
+blacklist_parse_proc_parameters (char *buf)
+{
+	if (strncmp (buf, "free ", 5) == 0) {
+		blacklist_parse_parameters (buf + 5, free);
+	} else if (strncmp (buf, "add ", 4) == 0) {
+		/* 
+		 * We don't need to check for known devices since
+		 * css_probe_device will handle this correctly. 
+		 */
+		blacklist_parse_parameters (buf + 4, add);
+	} else {
+		printk (KERN_WARNING "cio_ignore: Parse error; \n"
+			KERN_WARNING "try using 'free all|<devno-range>,"
+				     "<devno-range>,...'\n"
+			KERN_WARNING "or 'add <devno-range>,"
+				     "<devno-range>,...'\n");
+		return;
+	}
+
+	s390_redo_validation ();
+}
+
+/* FIXME: These should be real bus ids and not home-grown ones! */
+static int cio_ignore_read (char *page, char **start, off_t off,
+			    int count, int *eof, void *data)
+{
+	const unsigned int entry_size = 18; /* "0.0.ABCD-0.0.EFGH\n" */
+	long devno;
+	int len;
+
+	len = 0;
+	for (devno = off; /* abuse the page variable
+			   * as counter, see fs/proc/generic.c */
+	     devno <= __MAX_SUBCHANNELS && len + entry_size < count; devno++) {
+		if (!test_bit(devno, bl_dev))
+			continue;
+		len += sprintf(page + len, "0.0.%04lx", devno);
+		if (test_bit(devno + 1, bl_dev)) { /* print range */
+			while (++devno < __MAX_SUBCHANNELS)
+				if (!test_bit(devno, bl_dev))
+					break;
+			len += sprintf(page + len, "-0.0.%04lx", --devno);
+		}
+		len += sprintf(page + len, "\n");
+	}
+
+	if (devno <= __MAX_SUBCHANNELS)
+		*eof = 1;
+	*start = (char *) (devno - off); /* number of checked entries */
+	return len;
+}
+
+static int cio_ignore_write(struct file *file, const char __user *user_buf,
+			     unsigned long user_len, void *data)
+{
+	char *buf;
+
+	if (user_len > 65536)
+		user_len = 65536;
+	buf = vmalloc (user_len + 1); /* maybe better use the stack? */
+	if (buf == NULL)
+		return -ENOMEM;
+	if (strncpy_from_user (buf, user_buf, user_len) < 0) {
+		vfree (buf);
+		return -EFAULT;
+	}
+	buf[user_len] = '\0';
+
+	blacklist_parse_proc_parameters (buf);
+
+	vfree (buf);
+	return user_len;
+}
+
+static int
+cio_ignore_proc_init (void)
+{
+	struct proc_dir_entry *entry;
+
+	entry = create_proc_entry ("cio_ignore", S_IFREG | S_IRUGO | S_IWUSR,
+				   &proc_root);
+	if (!entry)
+		return 0;
+
+	entry->read_proc  = cio_ignore_read;
+	entry->write_proc = cio_ignore_write;
+
+	return 1;
+}
+
+__initcall (cio_ignore_proc_init);
+
+#endif /* CONFIG_PROC_FS */

drivers/s390/cio/blacklist.h

@@ -0,0 +1,6 @@
+#ifndef S390_BLACKLIST_H
+#define S390_BLACKLIST_H
+
+extern int is_blacklisted (int devno);
+
+#endif

drivers/s390/cio/ccwgroup.c

@@ -0,0 +1,482 @@
+/*
+ *  drivers/s390/cio/ccwgroup.c
+ *  bus driver for ccwgroup
+ *   $Revision: 1.29 $
+ *
+ *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
+ *                       IBM Corporation
+ *    Author(s): Arnd Bergmann (arndb@de.ibm.com)
+ *               Cornelia Huck (cohuck@de.ibm.com)
+ */
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/slab.h>
+#include <linux/list.h>
+#include <linux/device.h>
+#include <linux/init.h>
+#include <linux/ctype.h>
+#include <linux/dcache.h>
+
+#include <asm/semaphore.h>
+#include <asm/ccwdev.h>
+#include <asm/ccwgroup.h>
+
+/* In Linux 2.4, we had a channel device layer called "chandev"
+ * that did all sorts of obscure stuff for networking devices.
+ * This is another driver that serves as a replacement for just
+ * one of its functions, namely the translation of single subchannels
+ * to devices that use multiple subchannels.
+ */
+
+/* a device matches a driver if all its slave devices match the same
+ * entry of the driver */
+static int
+ccwgroup_bus_match (struct device * dev, struct device_driver * drv)
+{
+	struct ccwgroup_device *gdev;
+	struct ccwgroup_driver *gdrv;
+
+	gdev = container_of(dev, struct ccwgroup_device, dev);
+	gdrv = container_of(drv, struct ccwgroup_driver, driver);
+
+	if (gdev->creator_id == gdrv->driver_id)
+		return 1;
+
+	return 0;
+}
+static int
+ccwgroup_hotplug (struct device *dev, char **envp, int num_envp, char *buffer,
+		  int buffer_size)
+{
+	/* TODO */
+	return 0;
+}
+
+static struct bus_type ccwgroup_bus_type = {
+	.name    = "ccwgroup",
+	.match   = ccwgroup_bus_match,
+	.hotplug = ccwgroup_hotplug,
+};
+
+static inline void
+__ccwgroup_remove_symlinks(struct ccwgroup_device *gdev)
+{
+	int i;
+	char str[8];
+
+	for (i = 0; i < gdev->count; i++) {
+		sprintf(str, "cdev%d", i);
+		sysfs_remove_link(&gdev->dev.kobj, str);
+		sysfs_remove_link(&gdev->cdev[i]->dev.kobj, "group_device");
+	}
+	
+}
+
+/*
+ * Provide an 'ungroup' attribute so the user can remove group devices no
+ * longer needed or accidentially created. Saves memory :)
+ */
+static ssize_t
+ccwgroup_ungroup_store(struct device *dev, const char *buf, size_t count)
+{
+	struct ccwgroup_device *gdev;
+
+	gdev = to_ccwgroupdev(dev);
+
+	if (gdev->state != CCWGROUP_OFFLINE)
+		return -EINVAL;
+
+	__ccwgroup_remove_symlinks(gdev);
+	device_unregister(dev);
+
+	return count;
+}
+
+static DEVICE_ATTR(ungroup, 0200, NULL, ccwgroup_ungroup_store);
+
+static void
+ccwgroup_release (struct device *dev)
+{
+	struct ccwgroup_device *gdev;
+	int i;
+
+	gdev = to_ccwgroupdev(dev);
+
+	for (i = 0; i < gdev->count; i++) {
+		gdev->cdev[i]->dev.driver_data = NULL;
+		put_device(&gdev->cdev[i]->dev);
+	}
+	kfree(gdev);
+}
+
+static inline int
+__ccwgroup_create_symlinks(struct ccwgroup_device *gdev)
+{
+	char str[8];
+	int i, rc;
+
+	for (i = 0; i < gdev->count; i++) {
+		rc = sysfs_create_link(&gdev->cdev[i]->dev.kobj, &gdev->dev.kobj,
+				       "group_device");
+		if (rc) {
+			for (--i; i >= 0; i--)
+				sysfs_remove_link(&gdev->cdev[i]->dev.kobj,
+						  "group_device");
+			return rc;
+		}
+	}
+	for (i = 0; i < gdev->count; i++) {
+		sprintf(str, "cdev%d", i);
+		rc = sysfs_create_link(&gdev->dev.kobj, &gdev->cdev[i]->dev.kobj,
+				       str);
+		if (rc) {
+			for (--i; i >= 0; i--) {
+				sprintf(str, "cdev%d", i);
+				sysfs_remove_link(&gdev->dev.kobj, str);
+			}
+			for (i = 0; i < gdev->count; i++)
+				sysfs_remove_link(&gdev->cdev[i]->dev.kobj,
+						  "group_device");
+			return rc;
+		}
+	}
+	return 0;
+}
+
+/*
+ * try to add a new ccwgroup device for one driver
+ * argc and argv[] are a list of bus_id's of devices
+ * belonging to the driver.
+ */
+int
+ccwgroup_create(struct device *root,
+		unsigned int creator_id,
+		struct ccw_driver *cdrv,
+		int argc, char *argv[])
+{
+	struct ccwgroup_device *gdev;
+	int i;
+	int rc;
+	int del_drvdata;
+
+	if (argc > 256) /* disallow dumb users */
+		return -EINVAL;
+
+	gdev = kmalloc(sizeof(*gdev) + argc*sizeof(gdev->cdev[0]), GFP_KERNEL);
+	if (!gdev)
+		return -ENOMEM;
+
+	memset(gdev, 0, sizeof(*gdev) + argc*sizeof(gdev->cdev[0]));
+	atomic_set(&gdev->onoff, 0);
+
+	del_drvdata = 0;
+	for (i = 0; i < argc; i++) {
+		gdev->cdev[i] = get_ccwdev_by_busid(cdrv, argv[i]);
+
+		/* all devices have to be of the same type in
+		 * order to be grouped */
+		if (!gdev->cdev[i]
+		    || gdev->cdev[i]->id.driver_info !=
+		    gdev->cdev[0]->id.driver_info) {
+			rc = -EINVAL;
+			goto free_dev;
+		}
+		/* Don't allow a device to belong to more than one group. */
+		if (gdev->cdev[i]->dev.driver_data) {
+			rc = -EINVAL;
+			goto free_dev;
+		}
+	}
+	for (i = 0; i < argc; i++)
+		gdev->cdev[i]->dev.driver_data = gdev;
+	del_drvdata = 1;
+
+	gdev->creator_id = creator_id;
+	gdev->count = argc;
+	gdev->dev = (struct device ) {
+		.bus = &ccwgroup_bus_type,
+		.parent = root,
+		.release = ccwgroup_release,
+	};
+
+	snprintf (gdev->dev.bus_id, BUS_ID_SIZE, "%s",
+			gdev->cdev[0]->dev.bus_id);
+
+	rc = device_register(&gdev->dev);
+	
+	if (rc)
+		goto free_dev;
+	get_device(&gdev->dev);
+	rc = device_create_file(&gdev->dev, &dev_attr_ungroup);
+
+	if (rc) {
+		device_unregister(&gdev->dev);
+		goto error;
+	}
+
+	rc = __ccwgroup_create_symlinks(gdev);
+	if (!rc) {
+		put_device(&gdev->dev);
+		return 0;
+	}
+	device_remove_file(&gdev->dev, &dev_attr_ungroup);
+	device_unregister(&gdev->dev);
+error:
+	for (i = 0; i < argc; i++)
+		if (gdev->cdev[i]) {
+			put_device(&gdev->cdev[i]->dev);
+			gdev->cdev[i]->dev.driver_data = NULL;
+		}
+	put_device(&gdev->dev);
+	return rc;
+free_dev:
+	for (i = 0; i < argc; i++)
+		if (gdev->cdev[i]) {
+			put_device(&gdev->cdev[i]->dev);
+			if (del_drvdata)
+				gdev->cdev[i]->dev.driver_data = NULL;
+		}
+	kfree(gdev);
+	return rc;
+}
+
+static int __init
+init_ccwgroup (void)
+{
+	return bus_register (&ccwgroup_bus_type);
+}
+
+static void __exit
+cleanup_ccwgroup (void)
+{
+	bus_unregister (&ccwgroup_bus_type);
+}
+
+module_init(init_ccwgroup);
+module_exit(cleanup_ccwgroup);
+
+/************************** driver stuff ******************************/
+
+static int
+ccwgroup_set_online(struct ccwgroup_device *gdev)
+{
+	struct ccwgroup_driver *gdrv;
+	int ret;
+
+	if (atomic_compare_and_swap(0, 1, &gdev->onoff))
+		return -EAGAIN;
+	if (gdev->state == CCWGROUP_ONLINE) {
+		ret = 0;
+		goto out;
+	}
+	if (!gdev->dev.driver) {
+		ret = -EINVAL;
+		goto out;
+	}
+	gdrv = to_ccwgroupdrv (gdev->dev.driver);
+	if ((ret = gdrv->set_online(gdev)))
+		goto out;
+
+	gdev->state = CCWGROUP_ONLINE;
+ out:
+	atomic_set(&gdev->onoff, 0);
+	return ret;
+}
+
+static int
+ccwgroup_set_offline(struct ccwgroup_device *gdev)
+{
+	struct ccwgroup_driver *gdrv;
+	int ret;
+
+	if (atomic_compare_and_swap(0, 1, &gdev->onoff))
+		return -EAGAIN;
+	if (gdev->state == CCWGROUP_OFFLINE) {
+		ret = 0;
+		goto out;
+	}
+	if (!gdev->dev.driver) {
+		ret = -EINVAL;
+		goto out;
+	}
+	gdrv = to_ccwgroupdrv (gdev->dev.driver);
+	if ((ret = gdrv->set_offline(gdev)))
+		goto out;
+
+	gdev->state = CCWGROUP_OFFLINE;
+ out:
+	atomic_set(&gdev->onoff, 0);
+	return ret;
+}
+
+static ssize_t
+ccwgroup_online_store (struct device *dev, const char *buf, size_t count)
+{
+	struct ccwgroup_device *gdev;
+	struct ccwgroup_driver *gdrv;
+	unsigned int value;
+	int ret;
+
+	gdev = to_ccwgroupdev(dev);
+	if (!dev->driver)
+		return count;
+
+	gdrv = to_ccwgroupdrv (gdev->dev.driver);
+	if (!try_module_get(gdrv->owner))
+		return -EINVAL;
+
+	value = simple_strtoul(buf, 0, 0);
+	ret = count;
+	if (value == 1)
+		ccwgroup_set_online(gdev);
+	else if (value == 0)
+		ccwgroup_set_offline(gdev);
+	else
+		ret = -EINVAL;
+	module_put(gdrv->owner);
+	return ret;
+}
+
+static ssize_t
+ccwgroup_online_show (struct device *dev, char *buf)
+{
+	int online;
+
+	online = (to_ccwgroupdev(dev)->state == CCWGROUP_ONLINE);
+
+	return sprintf(buf, online ? "1\n" : "0\n");
+}
+
+static DEVICE_ATTR(online, 0644, ccwgroup_online_show, ccwgroup_online_store);
+
+static int
+ccwgroup_probe (struct device *dev)
+{
+	struct ccwgroup_device *gdev;
+	struct ccwgroup_driver *gdrv;
+
+	int ret;
+
+	gdev = to_ccwgroupdev(dev);
+	gdrv = to_ccwgroupdrv(dev->driver);
+
+	if ((ret = device_create_file(dev, &dev_attr_online)))
+		return ret;
+
+	pr_debug("%s: device %s\n", __func__, gdev->dev.bus_id);
+	ret = gdrv->probe ? gdrv->probe(gdev) : -ENODEV;
+	if (ret)
+		device_remove_file(dev, &dev_attr_online);
+
+	return ret;
+}
+
+static int
+ccwgroup_remove (struct device *dev)
+{
+	struct ccwgroup_device *gdev;
+	struct ccwgroup_driver *gdrv;
+
+	gdev = to_ccwgroupdev(dev);
+	gdrv = to_ccwgroupdrv(dev->driver);
+
+	pr_debug("%s: device %s\n", __func__, gdev->dev.bus_id);
+
+	device_remove_file(dev, &dev_attr_online);
+
+	if (gdrv && gdrv->remove)
+		gdrv->remove(gdev);
+	return 0;
+}
+
+int
+ccwgroup_driver_register (struct ccwgroup_driver *cdriver)
+{
+	/* register our new driver with the core */
+	cdriver->driver = (struct device_driver) {
+		.bus = &ccwgroup_bus_type,
+		.name = cdriver->name,
+		.probe = ccwgroup_probe,
+		.remove = ccwgroup_remove,
+	};
+
+	return driver_register(&cdriver->driver);
+}
+
+static inline struct device *
+__get_next_ccwgroup_device(struct device_driver *drv)
+{
+	struct device *dev, *d;
+
+	down_read(&drv->bus->subsys.rwsem);
+	dev = NULL;
+	list_for_each_entry(d, &drv->devices, driver_list) {
+		dev = get_device(d);
+		if (dev)
+			break;
+	}
+	up_read(&drv->bus->subsys.rwsem);
+	return dev;
+}
+
+void
+ccwgroup_driver_unregister (struct ccwgroup_driver *cdriver)
+{
+	struct device *dev;
+
+	/* We don't want ccwgroup devices to live longer than their driver. */
+	get_driver(&cdriver->driver);
+	while ((dev = __get_next_ccwgroup_device(&cdriver->driver))) {
+		__ccwgroup_remove_symlinks(to_ccwgroupdev(dev));
+		device_unregister(dev);
+		put_device(dev);
+	};
+	put_driver(&cdriver->driver);
+	driver_unregister(&cdriver->driver);
+}
+
+int
+ccwgroup_probe_ccwdev(struct ccw_device *cdev)
+{
+	return 0;
+}
+
+static inline struct ccwgroup_device *
+__ccwgroup_get_gdev_by_cdev(struct ccw_device *cdev)
+{
+	struct ccwgroup_device *gdev;
+
+	if (cdev->dev.driver_data) {
+		gdev = (struct ccwgroup_device *)cdev->dev.driver_data;
+		if (get_device(&gdev->dev)) {
+			if (!list_empty(&gdev->dev.node))
+				return gdev;
+			put_device(&gdev->dev);
+		}
+		return NULL;
+	}
+	return NULL;
+}
+
+void
+ccwgroup_remove_ccwdev(struct ccw_device *cdev)
+{
+	struct ccwgroup_device *gdev;
+
+	/* Ignore offlining errors, device is gone anyway. */
+	ccw_device_set_offline(cdev);
+	/* If one of its devices is gone, the whole group is done for. */
+	gdev = __ccwgroup_get_gdev_by_cdev(cdev);
+	if (gdev) {
+		__ccwgroup_remove_symlinks(gdev);
+		device_unregister(&gdev->dev);
+		put_device(&gdev->dev);
+	}
+}
+
+MODULE_LICENSE("GPL");
+EXPORT_SYMBOL(ccwgroup_driver_register);
+EXPORT_SYMBOL(ccwgroup_driver_unregister);
+EXPORT_SYMBOL(ccwgroup_create);
+EXPORT_SYMBOL(ccwgroup_probe_ccwdev);
+EXPORT_SYMBOL(ccwgroup_remove_ccwdev);

drivers/s390/cio/chsc.h

@@ -0,0 +1,66 @@
+#ifndef S390_CHSC_H
+#define S390_CHSC_H
+
+#define NR_CHPIDS 256
+
+#define CHSC_SEI_ACC_CHPID        1
+#define CHSC_SEI_ACC_LINKADDR     2
+#define CHSC_SEI_ACC_FULLLINKADDR 3
+
+struct chsc_header {
+	u16 length;
+	u16 code;
+};
+
+struct channel_path_desc {
+	u8 flags;
+	u8 lsn;
+	u8 desc;
+	u8 chpid;
+	u8 swla;
+	u8 zeroes;
+	u8 chla;
+	u8 chpp;
+};
+
+struct channel_path {
+	int id;
+	int state;
+	struct channel_path_desc desc;
+	struct device dev;
+};
+
+extern void s390_process_css( void );
+extern void chsc_validate_chpids(struct subchannel *);
+extern void chpid_is_actually_online(int);
+
+struct css_general_char {
+	u64 : 41;
+	u32 aif : 1;     /* bit 41 */
+	u32 : 3;
+	u32 mcss : 1;    /* bit 45 */
+	u32 : 2;
+	u32 ext_mb : 1;  /* bit 48 */
+	u32 : 7;
+	u32 aif_tdd : 1; /* bit 56 */
+	u32 : 10;
+	u32 aif_osa : 1; /* bit 67 */
+	u32 : 28;
+}__attribute__((packed));
+
+struct css_chsc_char {
+	u64 res;
+	u64 : 43;
+	u32 scssc : 1;  /* bit 107 */
+	u32 scsscf : 1; /* bit 108 */
+	u32 : 19;
+}__attribute__((packed));
+
+extern struct css_general_char css_general_characteristics;
+extern struct css_chsc_char css_chsc_characteristics;
+
+extern int chsc_determine_css_characteristics(void);
+extern int css_characteristics_avail;
+
+extern void *chsc_get_chp_desc(struct subchannel*, int);
+#endif

drivers/s390/cio/cio.c

@@ -0,0 +1,860 @@
+/*
+ *  drivers/s390/cio/cio.c
+ *   S/390 common I/O routines -- low level i/o calls
+ *   $Revision: 1.131 $
+ *
+ *    Copyright (C) 1999-2002 IBM Deutschland Entwicklung GmbH,
+ *			      IBM Corporation
+ *    Author(s): Ingo Adlung (adlung@de.ibm.com)
+ *		 Cornelia Huck (cohuck@de.ibm.com)
+ *		 Arnd Bergmann (arndb@de.ibm.com)
+ *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
+ */
+
+#include <linux/module.h>
+#include <linux/config.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/kernel_stat.h>
+#include <linux/interrupt.h>
+
+#include <asm/cio.h>
+#include <asm/delay.h>
+#include <asm/irq.h>
+
+#include "airq.h"
+#include "cio.h"
+#include "css.h"
+#include "chsc.h"
+#include "ioasm.h"
+#include "blacklist.h"
+#include "cio_debug.h"
+
+debug_info_t *cio_debug_msg_id;
+debug_info_t *cio_debug_trace_id;
+debug_info_t *cio_debug_crw_id;
+
+int cio_show_msg;
+
+static int __init
+cio_setup (char *parm)
+{
+	if (!strcmp (parm, "yes"))
+		cio_show_msg = 1;
+	else if (!strcmp (parm, "no"))
+		cio_show_msg = 0;
+	else
+		printk (KERN_ERR "cio_setup : invalid cio_msg parameter '%s'",
+			parm);
+	return 1;
+}
+
+__setup ("cio_msg=", cio_setup);
+
+/*
+ * Function: cio_debug_init
+ * Initializes three debug logs (under /proc/s390dbf) for common I/O:
+ * - cio_msg logs the messages which are printk'ed when CONFIG_DEBUG_IO is on
+ * - cio_trace logs the calling of different functions
+ * - cio_crw logs the messages which are printk'ed when CONFIG_DEBUG_CRW is on
+ * debug levels depend on CONFIG_DEBUG_IO resp. CONFIG_DEBUG_CRW
+ */
+static int __init
+cio_debug_init (void)
+{
+	cio_debug_msg_id = debug_register ("cio_msg", 4, 4, 16*sizeof (long));
+	if (!cio_debug_msg_id)
+		goto out_unregister;
+	debug_register_view (cio_debug_msg_id, &debug_sprintf_view);
+	debug_set_level (cio_debug_msg_id, 2);
+	cio_debug_trace_id = debug_register ("cio_trace", 4, 4, 8);
+	if (!cio_debug_trace_id)
+		goto out_unregister;
+	debug_register_view (cio_debug_trace_id, &debug_hex_ascii_view);
+	debug_set_level (cio_debug_trace_id, 2);
+	cio_debug_crw_id = debug_register ("cio_crw", 2, 4, 16*sizeof (long));
+	if (!cio_debug_crw_id)
+		goto out_unregister;
+	debug_register_view (cio_debug_crw_id, &debug_sprintf_view);
+	debug_set_level (cio_debug_crw_id, 2);
+	pr_debug("debugging initialized\n");
+	return 0;
+
+out_unregister:
+	if (cio_debug_msg_id)
+		debug_unregister (cio_debug_msg_id);
+	if (cio_debug_trace_id)
+		debug_unregister (cio_debug_trace_id);
+	if (cio_debug_crw_id)
+		debug_unregister (cio_debug_crw_id);
+	pr_debug("could not initialize debugging\n");
+	return -1;
+}
+
+arch_initcall (cio_debug_init);
+
+int
+cio_set_options (struct subchannel *sch, int flags)
+{
+       sch->options.suspend = (flags & DOIO_ALLOW_SUSPEND) != 0;
+       sch->options.prefetch = (flags & DOIO_DENY_PREFETCH) != 0;
+       sch->options.inter = (flags & DOIO_SUPPRESS_INTER) != 0;
+       return 0;
+}
+
+/* FIXME: who wants to use this? */
+int
+cio_get_options (struct subchannel *sch)
+{
+       int flags;
+
+       flags = 0;
+       if (sch->options.suspend)
+		flags |= DOIO_ALLOW_SUSPEND;
+       if (sch->options.prefetch)
+		flags |= DOIO_DENY_PREFETCH;
+       if (sch->options.inter)
+		flags |= DOIO_SUPPRESS_INTER;
+       return flags;
+}
+
+/*
+ * Use tpi to get a pending interrupt, call the interrupt handler and
+ * return a pointer to the subchannel structure.
+ */
+static inline int
+cio_tpi(void)
+{
+	struct tpi_info *tpi_info;
+	struct subchannel *sch;
+	struct irb *irb;
+
+	tpi_info = (struct tpi_info *) __LC_SUBCHANNEL_ID;
+	if (tpi (NULL) != 1)
+		return 0;
+	irb = (struct irb *) __LC_IRB;
+	/* Store interrupt response block to lowcore. */
+	if (tsch (tpi_info->irq, irb) != 0)
+		/* Not status pending or not operational. */
+		return 1;
+	sch = (struct subchannel *)(unsigned long)tpi_info->intparm;
+	if (!sch)
+		return 1;
+	local_bh_disable();
+	irq_enter ();
+	spin_lock(&sch->lock);
+	memcpy (&sch->schib.scsw, &irb->scsw, sizeof (struct scsw));
+	if (sch->driver && sch->driver->irq)
+		sch->driver->irq(&sch->dev);
+	spin_unlock(&sch->lock);
+	irq_exit ();
+	__local_bh_enable();
+	return 1;
+}
+
+static inline int
+cio_start_handle_notoper(struct subchannel *sch, __u8 lpm)
+{
+	char dbf_text[15];
+
+	if (lpm != 0)
+		sch->lpm &= ~lpm;
+	else
+		sch->lpm = 0;
+
+	stsch (sch->irq, &sch->schib);
+
+	CIO_MSG_EVENT(0, "cio_start: 'not oper' status for "
+		      "subchannel %04x!\n", sch->irq);
+	sprintf(dbf_text, "no%s", sch->dev.bus_id);
+	CIO_TRACE_EVENT(0, dbf_text);
+	CIO_HEX_EVENT(0, &sch->schib, sizeof (struct schib));
+
+	return (sch->lpm ? -EACCES : -ENODEV);
+}
+
+int
+cio_start_key (struct subchannel *sch,	/* subchannel structure */
+	       struct ccw1 * cpa,	/* logical channel prog addr */
+	       __u8 lpm,		/* logical path mask */
+	       __u8 key)                /* storage key */
+{
+	char dbf_txt[15];
+	int ccode;
+
+	CIO_TRACE_EVENT (4, "stIO");
+	CIO_TRACE_EVENT (4, sch->dev.bus_id);
+
+	/* sch is always under 2G. */
+	sch->orb.intparm = (__u32)(unsigned long)sch;
+	sch->orb.fmt = 1;
+
+	sch->orb.pfch = sch->options.prefetch == 0;
+	sch->orb.spnd = sch->options.suspend;
+	sch->orb.ssic = sch->options.suspend && sch->options.inter;
+	sch->orb.lpm = (lpm != 0) ? (lpm & sch->opm) : sch->lpm;
+#ifdef CONFIG_ARCH_S390X
+	/*
+	 * for 64 bit we always support 64 bit IDAWs with 4k page size only
+	 */
+	sch->orb.c64 = 1;
+	sch->orb.i2k = 0;
+#endif
+	sch->orb.key = key >> 4;
+	/* issue "Start Subchannel" */
+	sch->orb.cpa = (__u32) __pa (cpa);
+	ccode = ssch (sch->irq, &sch->orb);
+
+	/* process condition code */
+	sprintf (dbf_txt, "ccode:%d", ccode);
+	CIO_TRACE_EVENT (4, dbf_txt);
+
+	switch (ccode) {
+	case 0:
+		/*
+		 * initialize device status information
+		 */
+		sch->schib.scsw.actl |= SCSW_ACTL_START_PEND;
+		return 0;
+	case 1:		/* status pending */
+	case 2:		/* busy */
+		return -EBUSY;
+	default:		/* device/path not operational */
+		return cio_start_handle_notoper(sch, lpm);
+	}
+}
+
+int
+cio_start (struct subchannel *sch, struct ccw1 *cpa, __u8 lpm)
+{
+	return cio_start_key(sch, cpa, lpm, default_storage_key);
+}
+
+/*
+ * resume suspended I/O operation
+ */
+int
+cio_resume (struct subchannel *sch)
+{
+	char dbf_txt[15];
+	int ccode;
+
+	CIO_TRACE_EVENT (4, "resIO");
+	CIO_TRACE_EVENT (4, sch->dev.bus_id);
+
+	ccode = rsch (sch->irq);
+
+	sprintf (dbf_txt, "ccode:%d", ccode);
+	CIO_TRACE_EVENT (4, dbf_txt);
+
+	switch (ccode) {
+	case 0:
+		sch->schib.scsw.actl |= SCSW_ACTL_RESUME_PEND;
+		return 0;
+	case 1:
+		return -EBUSY;
+	case 2:
+		return -EINVAL;
+	default:
+		/*
+		 * useless to wait for request completion
+		 *  as device is no longer operational !
+		 */
+		return -ENODEV;
+	}
+}
+
+/*
+ * halt I/O operation
+ */
+int
+cio_halt(struct subchannel *sch)
+{
+	char dbf_txt[15];
+	int ccode;
+
+	if (!sch)
+		return -ENODEV;
+
+	CIO_TRACE_EVENT (2, "haltIO");
+	CIO_TRACE_EVENT (2, sch->dev.bus_id);
+
+	/*
+	 * Issue "Halt subchannel" and process condition code
+	 */
+	ccode = hsch (sch->irq);
+
+	sprintf (dbf_txt, "ccode:%d", ccode);
+	CIO_TRACE_EVENT (2, dbf_txt);
+
+	switch (ccode) {
+	case 0:
+		sch->schib.scsw.actl |= SCSW_ACTL_HALT_PEND;
+		return 0;
+	case 1:		/* status pending */
+	case 2:		/* busy */
+		return -EBUSY;
+	default:		/* device not operational */
+		return -ENODEV;
+	}
+}
+
+/*
+ * Clear I/O operation
+ */
+int
+cio_clear(struct subchannel *sch)
+{
+	char dbf_txt[15];
+	int ccode;
+
+	if (!sch)
+		return -ENODEV;
+
+	CIO_TRACE_EVENT (2, "clearIO");
+	CIO_TRACE_EVENT (2, sch->dev.bus_id);
+
+	/*
+	 * Issue "Clear subchannel" and process condition code
+	 */
+	ccode = csch (sch->irq);
+
+	sprintf (dbf_txt, "ccode:%d", ccode);
+	CIO_TRACE_EVENT (2, dbf_txt);
+
+	switch (ccode) {
+	case 0:
+		sch->schib.scsw.actl |= SCSW_ACTL_CLEAR_PEND;
+		return 0;
+	default:		/* device not operational */
+		return -ENODEV;
+	}
+}
+
+/*
+ * Function: cio_cancel
+ * Issues a "Cancel Subchannel" on the specified subchannel
+ * Note: We don't need any fancy intparms and flags here
+ *	 since xsch is executed synchronously.
+ * Only for common I/O internal use as for now.
+ */
+int
+cio_cancel (struct subchannel *sch)
+{
+	char dbf_txt[15];
+	int ccode;
+
+	if (!sch)
+		return -ENODEV;
+
+	CIO_TRACE_EVENT (2, "cancelIO");
+	CIO_TRACE_EVENT (2, sch->dev.bus_id);
+
+	ccode = xsch (sch->irq);
+
+	sprintf (dbf_txt, "ccode:%d", ccode);
+	CIO_TRACE_EVENT (2, dbf_txt);
+
+	switch (ccode) {
+	case 0:		/* success */
+		/* Update information in scsw. */
+		stsch (sch->irq, &sch->schib);
+		return 0;
+	case 1:		/* status pending */
+		return -EBUSY;
+	case 2:		/* not applicable */
+		return -EINVAL;
+	default:	/* not oper */
+		return -ENODEV;
+	}
+}
+
+/*
+ * Function: cio_modify
+ * Issues a "Modify Subchannel" on the specified subchannel
+ */
+int
+cio_modify (struct subchannel *sch)
+{
+	int ccode, retry, ret;
+
+	ret = 0;
+	for (retry = 0; retry < 5; retry++) {
+		ccode = msch_err (sch->irq, &sch->schib);
+		if (ccode < 0)	/* -EIO if msch gets a program check. */
+			return ccode;
+		switch (ccode) {
+		case 0: /* successfull */
+			return 0;
+		case 1:	/* status pending */
+			return -EBUSY;
+		case 2:	/* busy */
+			udelay (100);	/* allow for recovery */
+			ret = -EBUSY;
+			break;
+		case 3:	/* not operational */
+			return -ENODEV;
+		}
+	}
+	return ret;
+}
+
+/*
+ * Enable subchannel.
+ */
+int
+cio_enable_subchannel (struct subchannel *sch, unsigned int isc)
+{
+	char dbf_txt[15];
+	int ccode;
+	int retry;
+	int ret;
+
+	CIO_TRACE_EVENT (2, "ensch");
+	CIO_TRACE_EVENT (2, sch->dev.bus_id);
+
+	ccode = stsch (sch->irq, &sch->schib);
+	if (ccode)
+		return -ENODEV;
+
+	for (retry = 5, ret = 0; retry > 0; retry--) {
+		sch->schib.pmcw.ena = 1;
+		sch->schib.pmcw.isc = isc;
+		sch->schib.pmcw.intparm = (__u32)(unsigned long)sch;
+		ret = cio_modify(sch);
+		if (ret == -ENODEV)
+			break;
+		if (ret == -EIO)
+			/*
+			 * Got a program check in cio_modify. Try without
+			 * the concurrent sense bit the next time.
+			 */
+			sch->schib.pmcw.csense = 0;
+		if (ret == 0) {
+			stsch (sch->irq, &sch->schib);
+			if (sch->schib.pmcw.ena)
+				break;
+		}
+		if (ret == -EBUSY) {
+			struct irb irb;
+			if (tsch(sch->irq, &irb) != 0)
+				break;
+		}
+	}
+	sprintf (dbf_txt, "ret:%d", ret);
+	CIO_TRACE_EVENT (2, dbf_txt);
+	return ret;
+}
+
+/*
+ * Disable subchannel.
+ */
+int
+cio_disable_subchannel (struct subchannel *sch)
+{
+	char dbf_txt[15];
+	int ccode;
+	int retry;
+	int ret;
+
+	CIO_TRACE_EVENT (2, "dissch");
+	CIO_TRACE_EVENT (2, sch->dev.bus_id);
+
+	ccode = stsch (sch->irq, &sch->schib);
+	if (ccode == 3)		/* Not operational. */
+		return -ENODEV;
+
+	if (sch->schib.scsw.actl != 0)
+		/*
+		 * the disable function must not be called while there are
+		 *  requests pending for completion !
+		 */
+		return -EBUSY;
+
+	for (retry = 5, ret = 0; retry > 0; retry--) {
+		sch->schib.pmcw.ena = 0;
+		ret = cio_modify(sch);
+		if (ret == -ENODEV)
+			break;
+		if (ret == -EBUSY)
+			/*
+			 * The subchannel is busy or status pending.
+			 * We'll disable when the next interrupt was delivered
+			 * via the state machine.
+			 */
+			break;
+		if (ret == 0) {
+			stsch (sch->irq, &sch->schib);
+			if (!sch->schib.pmcw.ena)
+				break;
+		}
+	}
+	sprintf (dbf_txt, "ret:%d", ret);
+	CIO_TRACE_EVENT (2, dbf_txt);
+	return ret;
+}
+
+/*
+ * cio_validate_subchannel()
+ *
+ * Find out subchannel type and initialize struct subchannel.
+ * Return codes:
+ *   SUBCHANNEL_TYPE_IO for a normal io subchannel
+ *   SUBCHANNEL_TYPE_CHSC for a chsc subchannel
+ *   SUBCHANNEL_TYPE_MESSAGE for a messaging subchannel
+ *   SUBCHANNEL_TYPE_ADM for a adm(?) subchannel
+ *   -ENXIO for non-defined subchannels
+ *   -ENODEV for subchannels with invalid device number or blacklisted devices
+ */
+int
+cio_validate_subchannel (struct subchannel *sch, unsigned int irq)
+{
+	char dbf_txt[15];
+	int ccode;
+
+	sprintf (dbf_txt, "valsch%x", irq);
+	CIO_TRACE_EVENT (4, dbf_txt);
+
+	/* Nuke all fields. */
+	memset(sch, 0, sizeof(struct subchannel));
+
+	spin_lock_init(&sch->lock);
+
+	/* Set a name for the subchannel */
+	snprintf (sch->dev.bus_id, BUS_ID_SIZE, "0.0.%04x", irq);
+
+	/*
+	 * The first subchannel that is not-operational (ccode==3)
+	 *  indicates that there aren't any more devices available.
+	 */
+	sch->irq = irq;
+	ccode = stsch (irq, &sch->schib);
+	if (ccode)
+		return -ENXIO;
+
+	/* Copy subchannel type from path management control word. */
+	sch->st = sch->schib.pmcw.st;
+
+	/*
+	 * ... just being curious we check for non I/O subchannels
+	 */
+	if (sch->st != 0) {
+		CIO_DEBUG(KERN_INFO, 0,
+			  "Subchannel %04X reports "
+			  "non-I/O subchannel type %04X\n",
+			  sch->irq, sch->st);
+		/* We stop here for non-io subchannels. */
+		return sch->st;
+	}
+
+	/* Initialization for io subchannels. */
+	if (!sch->schib.pmcw.dnv)
+		/* io subchannel but device number is invalid. */
+		return -ENODEV;
+
+	/* Devno is valid. */
+	if (is_blacklisted (sch->schib.pmcw.dev)) {
+		/*
+		 * This device must not be known to Linux. So we simply
+		 * say that there is no device and return ENODEV.
+		 */
+		CIO_MSG_EVENT(0, "Blacklisted device detected "
+			      "at devno %04X\n", sch->schib.pmcw.dev);
+		return -ENODEV;
+	}
+	sch->opm = 0xff;
+	chsc_validate_chpids(sch);
+	sch->lpm = sch->schib.pmcw.pim &
+		sch->schib.pmcw.pam &
+		sch->schib.pmcw.pom &
+		sch->opm;
+
+	CIO_DEBUG(KERN_INFO, 0,
+		  "Detected device %04X on subchannel %04X"
+		  " - PIM = %02X, PAM = %02X, POM = %02X\n",
+		  sch->schib.pmcw.dev, sch->irq, sch->schib.pmcw.pim,
+		  sch->schib.pmcw.pam, sch->schib.pmcw.pom);
+
+	/*
+	 * We now have to initially ...
+	 *  ... set "interruption subclass"
+	 *  ... enable "concurrent sense"
+	 *  ... enable "multipath mode" if more than one
+	 *	  CHPID is available. This is done regardless
+	 *	  whether multiple paths are available for us.
+	 */
+	sch->schib.pmcw.isc = 3;	/* could be smth. else */
+	sch->schib.pmcw.csense = 1;	/* concurrent sense */
+	sch->schib.pmcw.ena = 0;
+	if ((sch->lpm & (sch->lpm - 1)) != 0)
+		sch->schib.pmcw.mp = 1;	/* multipath mode */
+	return 0;
+}
+
+/*
+ * do_IRQ() handles all normal I/O device IRQ's (the special
+ *	    SMP cross-CPU interrupts have their own specific
+ *	    handlers).
+ *
+ */
+void
+do_IRQ (struct pt_regs *regs)
+{
+	struct tpi_info *tpi_info;
+	struct subchannel *sch;
+	struct irb *irb;
+
+	irq_enter ();
+	asm volatile ("mc 0,0");
+	if (S390_lowcore.int_clock >= S390_lowcore.jiffy_timer)
+		/**
+		 * Make sure that the i/o interrupt did not "overtake"
+		 * the last HZ timer interrupt.
+		 */
+		account_ticks(regs);
+	/*
+	 * Get interrupt information from lowcore
+	 */
+	tpi_info = (struct tpi_info *) __LC_SUBCHANNEL_ID;
+	irb = (struct irb *) __LC_IRB;
+	do {
+		kstat_cpu(smp_processor_id()).irqs[IO_INTERRUPT]++;
+		/*
+		 * Non I/O-subchannel thin interrupts are processed differently
+		 */
+		if (tpi_info->adapter_IO == 1 &&
+		    tpi_info->int_type == IO_INTERRUPT_TYPE) {
+			do_adapter_IO();
+			continue;
+		}
+		sch = (struct subchannel *)(unsigned long)tpi_info->intparm;
+		if (sch)
+			spin_lock(&sch->lock);
+		/* Store interrupt response block to lowcore. */
+		if (tsch (tpi_info->irq, irb) == 0 && sch) {
+			/* Keep subchannel information word up to date. */
+			memcpy (&sch->schib.scsw, &irb->scsw,
+				sizeof (irb->scsw));
+			/* Call interrupt handler if there is one. */
+			if (sch->driver && sch->driver->irq)
+				sch->driver->irq(&sch->dev);
+		}
+		if (sch)
+			spin_unlock(&sch->lock);
+		/*
+		 * Are more interrupts pending?
+		 * If so, the tpi instruction will update the lowcore
+		 * to hold the info for the next interrupt.
+		 * We don't do this for VM because a tpi drops the cpu
+		 * out of the sie which costs more cycles than it saves.
+		 */
+	} while (!MACHINE_IS_VM && tpi (NULL) != 0);
+	irq_exit ();
+}
+
+#ifdef CONFIG_CCW_CONSOLE
+static struct subchannel console_subchannel;
+static int console_subchannel_in_use;
+
+/*
+ * busy wait for the next interrupt on the console
+ */
+void
+wait_cons_dev (void)
+{
+	unsigned long cr6      __attribute__ ((aligned (8)));
+	unsigned long save_cr6 __attribute__ ((aligned (8)));
+
+	/* 
+	 * before entering the spinlock we may already have
+	 * processed the interrupt on a different CPU...
+	 */
+	if (!console_subchannel_in_use)
+		return;
+
+	/* disable all but isc 7 (console device) */
+	__ctl_store (save_cr6, 6, 6);
+	cr6 = 0x01000000;
+	__ctl_load (cr6, 6, 6);
+
+	do {
+		spin_unlock(&console_subchannel.lock);
+		if (!cio_tpi())
+			cpu_relax();
+		spin_lock(&console_subchannel.lock);
+	} while (console_subchannel.schib.scsw.actl != 0);
+	/*
+	 * restore previous isc value
+	 */
+	__ctl_load (save_cr6, 6, 6);
+}
+
+static int
+cio_console_irq(void)
+{
+	int irq;
+	
+	if (console_irq != -1) {
+		/* VM provided us with the irq number of the console. */
+		if (stsch(console_irq, &console_subchannel.schib) != 0 ||
+		    !console_subchannel.schib.pmcw.dnv)
+			return -1;
+		console_devno = console_subchannel.schib.pmcw.dev;
+	} else if (console_devno != -1) {
+		/* At least the console device number is known. */
+		for (irq = 0; irq < __MAX_SUBCHANNELS; irq++) {
+			if (stsch(irq, &console_subchannel.schib) != 0)
+				break;
+			if (console_subchannel.schib.pmcw.dnv &&
+			    console_subchannel.schib.pmcw.dev ==
+			    console_devno) {
+				console_irq = irq;
+				break;
+			}
+		}
+		if (console_irq == -1)
+			return -1;
+	} else {
+		/* unlike in 2.4, we cannot autoprobe here, since
+		 * the channel subsystem is not fully initialized.
+		 * With some luck, the HWC console can take over */
+		printk(KERN_WARNING "No ccw console found!\n");
+		return -1;
+	}
+	return console_irq;
+}
+
+struct subchannel *
+cio_probe_console(void)
+{
+	int irq, ret;
+
+	if (xchg(&console_subchannel_in_use, 1) != 0)
+		return ERR_PTR(-EBUSY);
+	irq = cio_console_irq();
+	if (irq == -1) {
+		console_subchannel_in_use = 0;
+		return ERR_PTR(-ENODEV);
+	}
+	memset(&console_subchannel, 0, sizeof(struct subchannel));
+	ret = cio_validate_subchannel(&console_subchannel, irq);
+	if (ret) {
+		console_subchannel_in_use = 0;
+		return ERR_PTR(-ENODEV);
+	}
+
+	/*
+	 * enable console I/O-interrupt subclass 7
+	 */
+	ctl_set_bit(6, 24);
+	console_subchannel.schib.pmcw.isc = 7;
+	console_subchannel.schib.pmcw.intparm =
+		(__u32)(unsigned long)&console_subchannel;
+	ret = cio_modify(&console_subchannel);
+	if (ret) {
+		console_subchannel_in_use = 0;
+		return ERR_PTR(ret);
+	}
+	return &console_subchannel;
+}
+
+void
+cio_release_console(void)
+{
+	console_subchannel.schib.pmcw.intparm = 0;
+	cio_modify(&console_subchannel);
+	ctl_clear_bit(6, 24);
+	console_subchannel_in_use = 0;
+}
+
+/* Bah... hack to catch console special sausages. */
+int
+cio_is_console(int irq)
+{
+	if (!console_subchannel_in_use)
+		return 0;
+	return (irq == console_subchannel.irq);
+}
+
+struct subchannel *
+cio_get_console_subchannel(void)
+{
+	if (!console_subchannel_in_use)
+		return 0;
+	return &console_subchannel;
+}
+
+#endif
+static inline int
+__disable_subchannel_easy(unsigned int schid, struct schib *schib)
+{
+	int retry, cc;
+
+	cc = 0;
+	for (retry=0;retry<3;retry++) {
+		schib->pmcw.ena = 0;
+		cc = msch(schid, schib);
+		if (cc)
+			return (cc==3?-ENODEV:-EBUSY);
+		stsch(schid, schib);
+		if (!schib->pmcw.ena)
+			return 0;
+	}
+	return -EBUSY; /* uhm... */
+}
+
+static inline int
+__clear_subchannel_easy(unsigned int schid)
+{
+	int retry;
+
+	if (csch(schid))
+		return -ENODEV;
+	for (retry=0;retry<20;retry++) {
+		struct tpi_info ti;
+
+		if (tpi(&ti)) {
+			tsch(schid, (struct irb *)__LC_IRB);
+			return 0;
+		}
+		udelay(100);
+	}
+	return -EBUSY;
+}
+
+extern void do_reipl(unsigned long devno);
+
+/* Clear all subchannels. */
+void
+clear_all_subchannels(void)
+{
+	unsigned int schid;
+
+	local_irq_disable();
+	for (schid=0;schid<=highest_subchannel;schid++) {
+		struct schib schib;
+		if (stsch(schid, &schib))
+			break; /* break out of the loop */
+		if (!schib.pmcw.ena)
+			continue;
+		switch(__disable_subchannel_easy(schid, &schib)) {
+		case 0:
+		case -ENODEV:
+			break;
+		default: /* -EBUSY */
+			if (__clear_subchannel_easy(schid))
+				break; /* give up... jump out of switch */
+			stsch(schid, &schib);
+			__disable_subchannel_easy(schid, &schib);
+		}
+	}
+}
+
+/* Make sure all subchannels are quiet before we re-ipl an lpar. */
+void
+reipl(unsigned long devno)
+{
+	clear_all_subchannels();
+	do_reipl(devno);
+}

drivers/s390/cio/cio.h

@@ -0,0 +1,143 @@
+#ifndef S390_CIO_H
+#define S390_CIO_H
+
+/*
+ * where we put the ssd info
+ */
+struct ssd_info {
+	__u8  valid:1;
+	__u8  type:7;		/* subchannel type */
+	__u8  chpid[8];		/* chpids */
+	__u16 fla[8];		/* full link addresses */
+} __attribute__ ((packed));
+
+/*
+ * path management control word
+ */
+struct pmcw {
+	__u32 intparm;		/* interruption parameter */
+	__u32 qf   : 1;		/* qdio facility */
+	__u32 res0 : 1;		/* reserved zeros */
+	__u32 isc  : 3;		/* interruption sublass */
+	__u32 res5 : 3;		/* reserved zeros */
+	__u32 ena  : 1;		/* enabled */
+	__u32 lm   : 2;		/* limit mode */
+	__u32 mme  : 2;		/* measurement-mode enable */
+	__u32 mp   : 1;		/* multipath mode */
+	__u32 tf   : 1;		/* timing facility */
+	__u32 dnv  : 1;		/* device number valid */
+	__u32 dev  : 16;	/* device number */
+	__u8  lpm;		/* logical path mask */
+	__u8  pnom;		/* path not operational mask */
+	__u8  lpum;		/* last path used mask */
+	__u8  pim;		/* path installed mask */
+	__u16 mbi;		/* measurement-block index */
+	__u8  pom;		/* path operational mask */
+	__u8  pam;		/* path available mask */
+	__u8  chpid[8];		/* CHPID 0-7 (if available) */
+	__u32 unused1 : 8;	/* reserved zeros */
+	__u32 st      : 3;	/* subchannel type */
+	__u32 unused2 : 18;	/* reserved zeros */
+	__u32 mbfc    : 1;      /* measurement block format control */
+	__u32 xmwme   : 1;      /* extended measurement word mode enable */
+	__u32 csense  : 1;	/* concurrent sense; can be enabled ...*/
+				/*  ... per MSCH, however, if facility */
+				/*  ... is not installed, this results */
+				/*  ... in an operand exception.       */
+} __attribute__ ((packed));
+
+/*
+ * subchannel information block
+ */
+struct schib {
+	struct pmcw pmcw;	 /* path management control word */
+	struct scsw scsw;	 /* subchannel status word */
+	__u64 mba;               /* measurement block address */
+	__u8 mda[4];		 /* model dependent area */
+} __attribute__ ((packed,aligned(4)));
+
+/*
+ * operation request block
+ */
+struct orb {
+	__u32 intparm;		/* interruption parameter */
+	__u32 key  : 4; 	/* flags, like key, suspend control, etc. */
+	__u32 spnd : 1; 	/* suspend control */
+	__u32 res1 : 1; 	/* reserved */
+	__u32 mod  : 1; 	/* modification control */
+	__u32 sync : 1; 	/* synchronize control */
+	__u32 fmt  : 1; 	/* format control */
+	__u32 pfch : 1; 	/* prefetch control */
+	__u32 isic : 1; 	/* initial-status-interruption control */
+	__u32 alcc : 1; 	/* address-limit-checking control */
+	__u32 ssic : 1; 	/* suppress-suspended-interr. control */
+	__u32 res2 : 1; 	/* reserved */
+	__u32 c64  : 1; 	/* IDAW/QDIO 64 bit control  */
+	__u32 i2k  : 1; 	/* IDAW 2/4kB block size control */
+	__u32 lpm  : 8; 	/* logical path mask */
+	__u32 ils  : 1; 	/* incorrect length */
+	__u32 zero : 6; 	/* reserved zeros */
+	__u32 orbx : 1; 	/* ORB extension control */
+	__u32 cpa;		/* channel program address */
+}  __attribute__ ((packed,aligned(4)));
+
+/* subchannel data structure used by I/O subroutines */
+struct subchannel {
+	unsigned int irq;	/* aka. subchannel number */
+	spinlock_t lock;	/* subchannel lock */
+
+	enum {
+		SUBCHANNEL_TYPE_IO = 0,
+		SUBCHANNEL_TYPE_CHSC = 1,
+		SUBCHANNEL_TYPE_MESSAGE = 2,
+		SUBCHANNEL_TYPE_ADM = 3,
+	} st;			/* subchannel type */
+
+	struct {
+		unsigned int suspend:1; /* allow suspend */
+		unsigned int prefetch:1;/* deny prefetch */
+		unsigned int inter:1;   /* suppress intermediate interrupts */
+	} __attribute__ ((packed)) options;
+
+	__u8 vpm;		/* verified path mask */
+	__u8 lpm;		/* logical path mask */
+	__u8 opm;               /* operational path mask */
+	struct schib schib;	/* subchannel information block */
+	struct orb orb;		/* operation request block */
+	struct ccw1 sense_ccw;	/* static ccw for sense command */
+	struct ssd_info ssd_info;	/* subchannel description */
+	struct device dev;	/* entry in device tree */
+	struct css_driver *driver;
+} __attribute__ ((aligned(8)));
+
+#define IO_INTERRUPT_TYPE	   0 /* I/O interrupt type */
+
+#define to_subchannel(n) container_of(n, struct subchannel, dev)
+
+extern int cio_validate_subchannel (struct subchannel *, unsigned int);
+extern int cio_enable_subchannel (struct subchannel *, unsigned int);
+extern int cio_disable_subchannel (struct subchannel *);
+extern int cio_cancel (struct subchannel *);
+extern int cio_clear (struct subchannel *);
+extern int cio_resume (struct subchannel *);
+extern int cio_halt (struct subchannel *);
+extern int cio_start (struct subchannel *, struct ccw1 *, __u8);
+extern int cio_start_key (struct subchannel *, struct ccw1 *, __u8, __u8);
+extern int cio_cancel (struct subchannel *);
+extern int cio_set_options (struct subchannel *, int);
+extern int cio_get_options (struct subchannel *);
+extern int cio_modify (struct subchannel *);
+/* Use with care. */
+#ifdef CONFIG_CCW_CONSOLE
+extern struct subchannel *cio_probe_console(void);
+extern void cio_release_console(void);
+extern int cio_is_console(int irq);
+extern struct subchannel *cio_get_console_subchannel(void);
+#else
+#define cio_is_console(irq) 0
+#define cio_get_console_subchannel() NULL
+#endif
+
+extern int cio_show_msg;
+
+#endif

drivers/s390/cio/cio_debug.h

@@ -0,0 +1,32 @@
+#ifndef CIO_DEBUG_H
+#define CIO_DEBUG_H
+
+#include <asm/debug.h>
+
+#define CIO_TRACE_EVENT(imp, txt) do { \
+		debug_text_event(cio_debug_trace_id, imp, txt); \
+	} while (0)
+
+#define CIO_MSG_EVENT(imp, args...) do { \
+		debug_sprintf_event(cio_debug_msg_id, imp , ##args); \
+	} while (0)
+
+#define CIO_CRW_EVENT(imp, args...) do { \
+		debug_sprintf_event(cio_debug_crw_id, imp , ##args); \
+	} while (0)
+
+#define CIO_HEX_EVENT(imp, args...) do { \
+                debug_event(cio_debug_trace_id, imp, ##args); \
+        } while (0)
+
+#define CIO_DEBUG(printk_level,event_level,msg...) ({ \
+	if (cio_show_msg) printk(printk_level msg); \
+	CIO_MSG_EVENT (event_level, msg); \
+})
+
+/* for use of debug feature */
+extern debug_info_t *cio_debug_msg_id;
+extern debug_info_t *cio_debug_trace_id;
+extern debug_info_t *cio_debug_crw_id;
+
+#endif

drivers/s390/cio/css.c

@@ -0,0 +1,575 @@
+/*
+ *  drivers/s390/cio/css.c
+ *  driver for channel subsystem
+ *   $Revision: 1.85 $
+ *
+ *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
+ *			 IBM Corporation
+ *    Author(s): Arnd Bergmann (arndb@de.ibm.com)
+ *		 Cornelia Huck (cohuck@de.ibm.com)
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/device.h>
+#include <linux/slab.h>
+#include <linux/errno.h>
+#include <linux/list.h>
+
+#include "css.h"
+#include "cio.h"
+#include "cio_debug.h"
+#include "ioasm.h"
+#include "chsc.h"
+
+unsigned int highest_subchannel;
+int need_rescan = 0;
+int css_init_done = 0;
+
+struct pgid global_pgid;
+int css_characteristics_avail = 0;
+
+struct device css_bus_device = {
+	.bus_id = "css0",
+};
+
+static struct subchannel *
+css_alloc_subchannel(int irq)
+{
+	struct subchannel *sch;
+	int ret;
+
+	sch = kmalloc (sizeof (*sch), GFP_KERNEL | GFP_DMA);
+	if (sch == NULL)
+		return ERR_PTR(-ENOMEM);
+	ret = cio_validate_subchannel (sch, irq);
+	if (ret < 0) {
+		kfree(sch);
+		return ERR_PTR(ret);
+	}
+	if (irq > highest_subchannel)
+		highest_subchannel = irq;
+
+	if (sch->st != SUBCHANNEL_TYPE_IO) {
+		/* For now we ignore all non-io subchannels. */
+		kfree(sch);
+		return ERR_PTR(-EINVAL);
+	}
+
+	/* 
+	 * Set intparm to subchannel address.
+	 * This is fine even on 64bit since the subchannel is always located
+	 * under 2G.
+	 */
+	sch->schib.pmcw.intparm = (__u32)(unsigned long)sch;
+	ret = cio_modify(sch);
+	if (ret) {
+		kfree(sch);
+		return ERR_PTR(ret);
+	}
+	return sch;
+}
+
+static void
+css_free_subchannel(struct subchannel *sch)
+{
+	if (sch) {
+		/* Reset intparm to zeroes. */
+		sch->schib.pmcw.intparm = 0;
+		cio_modify(sch);
+		kfree(sch);
+	}
+	
+}
+
+static void
+css_subchannel_release(struct device *dev)
+{
+	struct subchannel *sch;
+
+	sch = to_subchannel(dev);
+	if (!cio_is_console(sch->irq))
+		kfree(sch);
+}
+
+extern int css_get_ssd_info(struct subchannel *sch);
+
+static int
+css_register_subchannel(struct subchannel *sch)
+{
+	int ret;
+
+	/* Initialize the subchannel structure */
+	sch->dev.parent = &css_bus_device;
+	sch->dev.bus = &css_bus_type;
+	sch->dev.release = &css_subchannel_release;
+	
+	/* make it known to the system */
+	ret = device_register(&sch->dev);
+	if (ret)
+		printk (KERN_WARNING "%s: could not register %s\n",
+			__func__, sch->dev.bus_id);
+	else
+		css_get_ssd_info(sch);
+	return ret;
+}
+
+int
+css_probe_device(int irq)
+{
+	int ret;
+	struct subchannel *sch;
+
+	sch = css_alloc_subchannel(irq);
+	if (IS_ERR(sch))
+		return PTR_ERR(sch);
+	ret = css_register_subchannel(sch);
+	if (ret)
+		css_free_subchannel(sch);
+	return ret;
+}
+
+struct subchannel *
+get_subchannel_by_schid(int irq)
+{
+	struct subchannel *sch;
+	struct list_head *entry;
+	struct device *dev;
+
+	if (!get_bus(&css_bus_type))
+		return NULL;
+	down_read(&css_bus_type.subsys.rwsem);
+	sch = NULL;
+	list_for_each(entry, &css_bus_type.devices.list) {
+		dev = get_device(container_of(entry,
+					      struct device, bus_list));
+		if (!dev)
+			continue;
+		sch = to_subchannel(dev);
+		if (sch->irq == irq)
+			break;
+		put_device(dev);
+		sch = NULL;
+	}
+	up_read(&css_bus_type.subsys.rwsem);
+	put_bus(&css_bus_type);
+
+	return sch;
+}
+
+static inline int
+css_get_subchannel_status(struct subchannel *sch, int schid)
+{
+	struct schib schib;
+	int cc;
+
+	cc = stsch(schid, &schib);
+	if (cc)
+		return CIO_GONE;
+	if (!schib.pmcw.dnv)
+		return CIO_GONE;
+	if (sch && sch->schib.pmcw.dnv &&
+	    (schib.pmcw.dev != sch->schib.pmcw.dev))
+		return CIO_REVALIDATE;
+	if (sch && !sch->lpm)
+		return CIO_NO_PATH;
+	return CIO_OPER;
+}
+	
+static int
+css_evaluate_subchannel(int irq, int slow)
+{
+	int event, ret, disc;
+	struct subchannel *sch;
+	unsigned long flags;
+
+	sch = get_subchannel_by_schid(irq);
+	disc = sch ? device_is_disconnected(sch) : 0;
+	if (disc && slow) {
+		if (sch)
+			put_device(&sch->dev);
+		return 0; /* Already processed. */
+	}
+	/*
+	 * We've got a machine check, so running I/O won't get an interrupt.
+	 * Kill any pending timers.
+	 */
+	if (sch)
+		device_kill_pending_timer(sch);
+	if (!disc && !slow) {
+		if (sch)
+			put_device(&sch->dev);
+		return -EAGAIN; /* Will be done on the slow path. */
+	}
+	event = css_get_subchannel_status(sch, irq);
+	CIO_MSG_EVENT(4, "Evaluating schid %04x, event %d, %s, %s path.\n",
+		      irq, event, sch?(disc?"disconnected":"normal"):"unknown",
+		      slow?"slow":"fast");
+	switch (event) {
+	case CIO_NO_PATH:
+	case CIO_GONE:
+		if (!sch) {
+			/* Never used this subchannel. Ignore. */
+			ret = 0;
+			break;
+		}
+		if (disc && (event == CIO_NO_PATH)) {
+			/*
+			 * Uargh, hack again. Because we don't get a machine
+			 * check on configure on, our path bookkeeping can
+			 * be out of date here (it's fine while we only do
+			 * logical varying or get chsc machine checks). We
+			 * need to force reprobing or we might miss devices
+			 * coming operational again. It won't do harm in real
+			 * no path situations.
+			 */
+			spin_lock_irqsave(&sch->lock, flags);
+			device_trigger_reprobe(sch);
+			spin_unlock_irqrestore(&sch->lock, flags);
+			ret = 0;
+			break;
+		}
+		if (sch->driver && sch->driver->notify &&
+		    sch->driver->notify(&sch->dev, event)) {
+			cio_disable_subchannel(sch);
+			device_set_disconnected(sch);
+			ret = 0;
+			break;
+		}
+		/*
+		 * Unregister subchannel.
+		 * The device will be killed automatically.
+		 */
+		cio_disable_subchannel(sch);
+		device_unregister(&sch->dev);
+		/* Reset intparm to zeroes. */
+		sch->schib.pmcw.intparm = 0;
+		cio_modify(sch);
+		put_device(&sch->dev);
+		ret = 0;
+		break;
+	case CIO_REVALIDATE:
+		/* 
+		 * Revalidation machine check. Sick.
+		 * We don't notify the driver since we have to throw the device
+		 * away in any case.
+		 */
+		if (!disc) {
+			device_unregister(&sch->dev);
+			/* Reset intparm to zeroes. */
+			sch->schib.pmcw.intparm = 0;
+			cio_modify(sch);
+			put_device(&sch->dev);
+			ret = css_probe_device(irq);
+		} else {
+			/*
+			 * We can't immediately deregister the disconnected
+			 * device since it might block.
+			 */
+			spin_lock_irqsave(&sch->lock, flags);
+			device_trigger_reprobe(sch);
+			spin_unlock_irqrestore(&sch->lock, flags);
+			ret = 0;
+		}
+		break;
+	case CIO_OPER:
+		if (disc) {
+			spin_lock_irqsave(&sch->lock, flags);
+			/* Get device operational again. */
+			device_trigger_reprobe(sch);
+			spin_unlock_irqrestore(&sch->lock, flags);
+		}
+		ret = sch ? 0 : css_probe_device(irq);
+		break;
+	default:
+		BUG();
+		ret = 0;
+	}
+	return ret;
+}
+
+static void
+css_rescan_devices(void)
+{
+	int irq, ret;
+
+	for (irq = 0; irq < __MAX_SUBCHANNELS; irq++) {
+		ret = css_evaluate_subchannel(irq, 1);
+		/* No more memory. It doesn't make sense to continue. No
+		 * panic because this can happen in midflight and just
+		 * because we can't use a new device is no reason to crash
+		 * the system. */
+		if (ret == -ENOMEM)
+			break;
+		/* -ENXIO indicates that there are no more subchannels. */
+		if (ret == -ENXIO)
+			break;
+	}
+}
+
+struct slow_subchannel {
+	struct list_head slow_list;
+	unsigned long schid;
+};
+
+static LIST_HEAD(slow_subchannels_head);
+static DEFINE_SPINLOCK(slow_subchannel_lock);
+
+static void
+css_trigger_slow_path(void)
+{
+	CIO_TRACE_EVENT(4, "slowpath");
+
+	if (need_rescan) {
+		need_rescan = 0;
+		css_rescan_devices();
+		return;
+	}
+
+	spin_lock_irq(&slow_subchannel_lock);
+	while (!list_empty(&slow_subchannels_head)) {
+		struct slow_subchannel *slow_sch =
+			list_entry(slow_subchannels_head.next,
+				   struct slow_subchannel, slow_list);
+
+		list_del_init(slow_subchannels_head.next);
+		spin_unlock_irq(&slow_subchannel_lock);
+		css_evaluate_subchannel(slow_sch->schid, 1);
+		spin_lock_irq(&slow_subchannel_lock);
+		kfree(slow_sch);
+	}
+	spin_unlock_irq(&slow_subchannel_lock);
+}
+
+typedef void (*workfunc)(void *);
+DECLARE_WORK(slow_path_work, (workfunc)css_trigger_slow_path, NULL);
+struct workqueue_struct *slow_path_wq;
+
+/*
+ * Rescan for new devices. FIXME: This is slow.
+ * This function is called when we have lost CRWs due to overflows and we have
+ * to do subchannel housekeeping.
+ */
+void
+css_reiterate_subchannels(void)
+{
+	css_clear_subchannel_slow_list();
+	need_rescan = 1;
+}
+
+/*
+ * Called from the machine check handler for subchannel report words.
+ */
+int
+css_process_crw(int irq)
+{
+	int ret;
+
+	CIO_CRW_EVENT(2, "source is subchannel %04X\n", irq);
+
+	if (need_rescan)
+		/* We need to iterate all subchannels anyway. */
+		return -EAGAIN;
+	/* 
+	 * Since we are always presented with IPI in the CRW, we have to
+	 * use stsch() to find out if the subchannel in question has come
+	 * or gone.
+	 */
+	ret = css_evaluate_subchannel(irq, 0);
+	if (ret == -EAGAIN) {
+		if (css_enqueue_subchannel_slow(irq)) {
+			css_clear_subchannel_slow_list();
+			need_rescan = 1;
+		}
+	}
+	return ret;
+}
+
+static void __init
+css_generate_pgid(void)
+{
+	/* Let's build our path group ID here. */
+	if (css_characteristics_avail && css_general_characteristics.mcss)
+		global_pgid.cpu_addr = 0x8000;
+	else {
+#ifdef CONFIG_SMP
+		global_pgid.cpu_addr = hard_smp_processor_id();
+#else
+		global_pgid.cpu_addr = 0;
+#endif
+	}
+	global_pgid.cpu_id = ((cpuid_t *) __LC_CPUID)->ident;
+	global_pgid.cpu_model = ((cpuid_t *) __LC_CPUID)->machine;
+	global_pgid.tod_high = (__u32) (get_clock() >> 32);
+}
+
+/*
+ * Now that the driver core is running, we can setup our channel subsystem.
+ * The struct subchannel's are created during probing (except for the
+ * static console subchannel).
+ */
+static int __init
+init_channel_subsystem (void)
+{
+	int ret, irq;
+
+	if (chsc_determine_css_characteristics() == 0)
+		css_characteristics_avail = 1;
+
+	css_generate_pgid();
+
+	if ((ret = bus_register(&css_bus_type)))
+		goto out;
+	if ((ret = device_register (&css_bus_device)))
+		goto out_bus;
+
+	css_init_done = 1;
+
+	ctl_set_bit(6, 28);
+
+	for (irq = 0; irq < __MAX_SUBCHANNELS; irq++) {
+		struct subchannel *sch;
+
+		if (cio_is_console(irq))
+			sch = cio_get_console_subchannel();
+		else {
+			sch = css_alloc_subchannel(irq);
+			if (IS_ERR(sch))
+				ret = PTR_ERR(sch);
+			else
+				ret = 0;
+			if (ret == -ENOMEM)
+				panic("Out of memory in "
+				      "init_channel_subsystem\n");
+			/* -ENXIO: no more subchannels. */
+			if (ret == -ENXIO)
+				break;
+			if (ret)
+				continue;
+		}
+		/*
+		 * We register ALL valid subchannels in ioinfo, even those
+		 * that have been present before init_channel_subsystem.
+		 * These subchannels can't have been registered yet (kmalloc
+		 * not working) so we do it now. This is true e.g. for the
+		 * console subchannel.
+		 */
+		css_register_subchannel(sch);
+	}
+	return 0;
+
+out_bus:
+	bus_unregister(&css_bus_type);
+out:
+	return ret;
+}
+
+/*
+ * find a driver for a subchannel. They identify by the subchannel
+ * type with the exception that the console subchannel driver has its own
+ * subchannel type although the device is an i/o subchannel
+ */
+static int
+css_bus_match (struct device *dev, struct device_driver *drv)
+{
+	struct subchannel *sch = container_of (dev, struct subchannel, dev);
+	struct css_driver *driver = container_of (drv, struct css_driver, drv);
+
+	if (sch->st == driver->subchannel_type)
+		return 1;
+
+	return 0;
+}
+
+struct bus_type css_bus_type = {
+	.name  = "css",
+	.match = &css_bus_match,
+};
+
+subsys_initcall(init_channel_subsystem);
+
+/*
+ * Register root devices for some drivers. The release function must not be
+ * in the device drivers, so we do it here.
+ */
+static void
+s390_root_dev_release(struct device *dev)
+{
+	kfree(dev);
+}
+
+struct device *
+s390_root_dev_register(const char *name)
+{
+	struct device *dev;
+	int ret;
+
+	if (!strlen(name))
+		return ERR_PTR(-EINVAL);
+	dev = kmalloc(sizeof(struct device), GFP_KERNEL);
+	if (!dev)
+		return ERR_PTR(-ENOMEM);
+	memset(dev, 0, sizeof(struct device));
+	strncpy(dev->bus_id, name, min(strlen(name), (size_t)BUS_ID_SIZE));
+	dev->release = s390_root_dev_release;
+	ret = device_register(dev);
+	if (ret) {
+		kfree(dev);
+		return ERR_PTR(ret);
+	}
+	return dev;
+}
+
+void
+s390_root_dev_unregister(struct device *dev)
+{
+	if (dev)
+		device_unregister(dev);
+}
+
+int
+css_enqueue_subchannel_slow(unsigned long schid)
+{
+	struct slow_subchannel *new_slow_sch;
+	unsigned long flags;
+
+	new_slow_sch = kmalloc(sizeof(struct slow_subchannel), GFP_ATOMIC);
+	if (!new_slow_sch)
+		return -ENOMEM;
+	memset(new_slow_sch, 0, sizeof(struct slow_subchannel));
+	new_slow_sch->schid = schid;
+	spin_lock_irqsave(&slow_subchannel_lock, flags);
+	list_add_tail(&new_slow_sch->slow_list, &slow_subchannels_head);
+	spin_unlock_irqrestore(&slow_subchannel_lock, flags);
+	return 0;
+}
+
+void
+css_clear_subchannel_slow_list(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&slow_subchannel_lock, flags);
+	while (!list_empty(&slow_subchannels_head)) {
+		struct slow_subchannel *slow_sch =
+			list_entry(slow_subchannels_head.next,
+				   struct slow_subchannel, slow_list);
+
+		list_del_init(slow_subchannels_head.next);
+		kfree(slow_sch);
+	}
+	spin_unlock_irqrestore(&slow_subchannel_lock, flags);
+}
+
+
+
+int
+css_slow_subchannels_exist(void)
+{
+	return (!list_empty(&slow_subchannels_head));
+}
+
+MODULE_LICENSE("GPL");
+EXPORT_SYMBOL(css_bus_type);
+EXPORT_SYMBOL(s390_root_dev_register);
+EXPORT_SYMBOL(s390_root_dev_unregister);
+EXPORT_SYMBOL_GPL(css_characteristics_avail);

drivers/s390/cio/css.h

@@ -0,0 +1,155 @@
+#ifndef _CSS_H
+#define _CSS_H
+
+#include <linux/wait.h>
+#include <linux/workqueue.h>
+
+#include <asm/cio.h>
+
+/*
+ * path grouping stuff
+ */
+#define SPID_FUNC_SINGLE_PATH	   0x00
+#define SPID_FUNC_MULTI_PATH	   0x80
+#define SPID_FUNC_ESTABLISH	   0x00
+#define SPID_FUNC_RESIGN	   0x40
+#define SPID_FUNC_DISBAND	   0x20
+
+#define SNID_STATE1_RESET	   0
+#define SNID_STATE1_UNGROUPED	   2
+#define SNID_STATE1_GROUPED	   3
+
+#define SNID_STATE2_NOT_RESVD	   0
+#define SNID_STATE2_RESVD_ELSE	   2
+#define SNID_STATE2_RESVD_SELF	   3
+
+#define SNID_STATE3_MULTI_PATH	   1
+#define SNID_STATE3_SINGLE_PATH	   0
+
+struct path_state {
+	__u8  state1 : 2;	/* path state value 1 */
+	__u8  state2 : 2;	/* path state value 2 */
+	__u8  state3 : 1;	/* path state value 3 */
+	__u8  resvd  : 3;	/* reserved */
+} __attribute__ ((packed));
+
+struct pgid {
+	union {
+		__u8 fc;   	/* SPID function code */
+		struct path_state ps;	/* SNID path state */
+	} inf;
+	__u32 cpu_addr	: 16;	/* CPU address */
+	__u32 cpu_id	: 24;	/* CPU identification */
+	__u32 cpu_model : 16;	/* CPU model */
+	__u32 tod_high;		/* high word TOD clock */
+} __attribute__ ((packed));
+
+extern struct pgid global_pgid;
+
+#define MAX_CIWS 8
+
+/*
+ * sense-id response buffer layout
+ */
+struct senseid {
+	/* common part */
+	__u8  reserved;     	/* always 0x'FF' */
+	__u16 cu_type;	     	/* control unit type */
+	__u8  cu_model;     	/* control unit model */
+	__u16 dev_type;     	/* device type */
+	__u8  dev_model;    	/* device model */
+	__u8  unused;	     	/* padding byte */
+	/* extended part */
+	struct ciw ciw[MAX_CIWS];	/* variable # of CIWs */
+}  __attribute__ ((packed,aligned(4)));
+
+struct ccw_device_private {
+	int state;		/* device state */
+	atomic_t onoff;
+	unsigned long registered;
+	__u16 devno;		/* device number */
+	__u16 irq;		/* subchannel number */
+	__u8 imask;		/* lpm mask for SNID/SID/SPGID */
+	int iretry;		/* retry counter SNID/SID/SPGID */
+	struct {
+		unsigned int fast:1;	/* post with "channel end" */
+		unsigned int repall:1;	/* report every interrupt status */
+		unsigned int pgroup:1;  /* do path grouping */
+		unsigned int force:1;   /* allow forced online */
+	} __attribute__ ((packed)) options;
+	struct {
+		unsigned int pgid_single:1; /* use single path for Set PGID */
+		unsigned int esid:1;        /* Ext. SenseID supported by HW */
+		unsigned int dosense:1;	    /* delayed SENSE required */
+		unsigned int doverify:1;    /* delayed path verification */
+		unsigned int donotify:1;    /* call notify function */
+		unsigned int recog_done:1;  /* dev. recog. complete */
+		unsigned int fake_irb:1;    /* deliver faked irb */
+	} __attribute__((packed)) flags;
+	unsigned long intparm;	/* user interruption parameter */
+	struct qdio_irq *qdio_data;
+	struct irb irb;		/* device status */
+	struct senseid senseid;	/* SenseID info */
+	struct pgid pgid;	/* path group ID */
+	struct ccw1 iccws[2];	/* ccws for SNID/SID/SPGID commands */
+	struct work_struct kick_work;
+	wait_queue_head_t wait_q;
+	struct timer_list timer;
+	void *cmb;			/* measurement information */
+	struct list_head cmb_list;	/* list of measured devices */
+	u64 cmb_start_time;		/* clock value of cmb reset */
+	void *cmb_wait;			/* deferred cmb enable/disable */
+};
+
+/*
+ * A css driver handles all subchannels of one type.
+ * Currently, we only care about I/O subchannels (type 0), these
+ * have a ccw_device connected to them.
+ */
+struct css_driver {
+	unsigned int subchannel_type;
+	struct device_driver drv;
+	void (*irq)(struct device *);
+	int (*notify)(struct device *, int);
+	void (*verify)(struct device *);
+	void (*termination)(struct device *);
+};
+
+/*
+ * all css_drivers have the css_bus_type
+ */
+extern struct bus_type css_bus_type;
+extern struct css_driver io_subchannel_driver;
+
+int css_probe_device(int irq);
+extern struct subchannel * get_subchannel_by_schid(int irq);
+extern unsigned int highest_subchannel;
+extern int css_init_done;
+
+#define __MAX_SUBCHANNELS 65536
+
+extern struct bus_type css_bus_type;
+extern struct device css_bus_device;
+
+/* Some helper functions for disconnected state. */
+int device_is_disconnected(struct subchannel *);
+void device_set_disconnected(struct subchannel *);
+void device_trigger_reprobe(struct subchannel *);
+
+/* Helper functions for vary on/off. */
+int device_is_online(struct subchannel *);
+void device_set_waiting(struct subchannel *);
+
+/* Machine check helper function. */
+void device_kill_pending_timer(struct subchannel *);
+
+/* Helper functions to build lists for the slow path. */
+int css_enqueue_subchannel_slow(unsigned long schid);
+void css_walk_subchannel_slow_list(void (*fn)(unsigned long));
+void css_clear_subchannel_slow_list(void);
+int css_slow_subchannels_exist(void);
+extern int need_rescan;
+
+extern struct workqueue_struct *slow_path_wq;
+extern struct work_struct slow_path_work;
+#endif

drivers/s390/cio/device.h

@@ -0,0 +1,115 @@
+#ifndef S390_DEVICE_H
+#define S390_DEVICE_H
+
+/*
+ * states of the device statemachine
+ */
+enum dev_state {
+	DEV_STATE_NOT_OPER,
+	DEV_STATE_SENSE_PGID,
+	DEV_STATE_SENSE_ID,
+	DEV_STATE_OFFLINE,
+	DEV_STATE_VERIFY,
+	DEV_STATE_ONLINE,
+	DEV_STATE_W4SENSE,
+	DEV_STATE_DISBAND_PGID,
+	DEV_STATE_BOXED,
+	/* states to wait for i/o completion before doing something */
+	DEV_STATE_CLEAR_VERIFY,
+	DEV_STATE_TIMEOUT_KILL,
+	DEV_STATE_WAIT4IO,
+	DEV_STATE_QUIESCE,
+	/* special states for devices gone not operational */
+	DEV_STATE_DISCONNECTED,
+	DEV_STATE_DISCONNECTED_SENSE_ID,
+	DEV_STATE_CMFCHANGE,
+	/* last element! */
+	NR_DEV_STATES
+};
+
+/*
+ * asynchronous events of the device statemachine
+ */
+enum dev_event {
+	DEV_EVENT_NOTOPER,
+	DEV_EVENT_INTERRUPT,
+	DEV_EVENT_TIMEOUT,
+	DEV_EVENT_VERIFY,
+	/* last element! */
+	NR_DEV_EVENTS
+};
+
+struct ccw_device;
+
+/*
+ * action called through jumptable
+ */
+typedef void (fsm_func_t)(struct ccw_device *, enum dev_event);
+extern fsm_func_t *dev_jumptable[NR_DEV_STATES][NR_DEV_EVENTS];
+
+static inline void
+dev_fsm_event(struct ccw_device *cdev, enum dev_event dev_event)
+{
+	dev_jumptable[cdev->private->state][dev_event](cdev, dev_event);
+}
+
+/*
+ * Delivers 1 if the device state is final.
+ */
+static inline int
+dev_fsm_final_state(struct ccw_device *cdev)
+{
+	return (cdev->private->state == DEV_STATE_NOT_OPER ||
+		cdev->private->state == DEV_STATE_OFFLINE ||
+		cdev->private->state == DEV_STATE_ONLINE ||
+		cdev->private->state == DEV_STATE_BOXED);
+}
+
+extern struct workqueue_struct *ccw_device_work;
+extern struct workqueue_struct *ccw_device_notify_work;
+
+void io_subchannel_recog_done(struct ccw_device *cdev);
+
+int ccw_device_cancel_halt_clear(struct ccw_device *);
+
+int ccw_device_register(struct ccw_device *);
+void ccw_device_do_unreg_rereg(void *);
+void ccw_device_call_sch_unregister(void *);
+
+int ccw_device_recognition(struct ccw_device *);
+int ccw_device_online(struct ccw_device *);
+int ccw_device_offline(struct ccw_device *);
+
+/* Function prototypes for device status and basic sense stuff. */
+void ccw_device_accumulate_irb(struct ccw_device *, struct irb *);
+void ccw_device_accumulate_basic_sense(struct ccw_device *, struct irb *);
+int ccw_device_accumulate_and_sense(struct ccw_device *, struct irb *);
+int ccw_device_do_sense(struct ccw_device *, struct irb *);
+
+/* Function prototypes for sense id stuff. */
+void ccw_device_sense_id_start(struct ccw_device *);
+void ccw_device_sense_id_irq(struct ccw_device *, enum dev_event);
+void ccw_device_sense_id_done(struct ccw_device *, int);
+
+/* Function prototypes for path grouping stuff. */
+void ccw_device_sense_pgid_start(struct ccw_device *);
+void ccw_device_sense_pgid_irq(struct ccw_device *, enum dev_event);
+void ccw_device_sense_pgid_done(struct ccw_device *, int);
+
+void ccw_device_verify_start(struct ccw_device *);
+void ccw_device_verify_irq(struct ccw_device *, enum dev_event);
+void ccw_device_verify_done(struct ccw_device *, int);
+
+void ccw_device_disband_start(struct ccw_device *);
+void ccw_device_disband_irq(struct ccw_device *, enum dev_event);
+void ccw_device_disband_done(struct ccw_device *, int);
+
+int ccw_device_call_handler(struct ccw_device *);
+
+int ccw_device_stlck(struct ccw_device *);
+
+/* qdio needs this. */
+void ccw_device_set_timeout(struct ccw_device *, int);
+
+void retry_set_schib(struct ccw_device *cdev);
+#endif

drivers/s390/cio/device_id.c

@@ -0,0 +1,355 @@
+/*
+ * drivers/s390/cio/device_id.c
+ *
+ *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
+ *			 IBM Corporation
+ *    Author(s): Cornelia Huck(cohuck@de.ibm.com)
+ *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *
+ * Sense ID functions.
+ */
+
+#include <linux/module.h>
+#include <linux/config.h>
+#include <linux/init.h>
+
+#include <asm/ccwdev.h>
+#include <asm/delay.h>
+#include <asm/cio.h>
+#include <asm/lowcore.h>
+
+#include "cio.h"
+#include "cio_debug.h"
+#include "css.h"
+#include "device.h"
+#include "ioasm.h"
+
+/*
+ * diag210 is used under VM to get information about a virtual device
+ */
+#ifdef CONFIG_ARCH_S390X
+int
+diag210(struct diag210 * addr)
+{
+	/*
+	 * diag 210 needs its data below the 2GB border, so we
+	 * use a static data area to be sure
+	 */
+	static struct diag210 diag210_tmp;
+	static DEFINE_SPINLOCK(diag210_lock);
+	unsigned long flags;
+	int ccode;
+
+	spin_lock_irqsave(&diag210_lock, flags);
+	diag210_tmp = *addr;
+
+	asm volatile (
+		"   lhi	  %0,-1\n"
+		"   sam31\n"
+		"   diag  %1,0,0x210\n"
+		"0: ipm	  %0\n"
+		"   srl	  %0,28\n"
+		"1: sam64\n"
+		".section __ex_table,\"a\"\n"
+		"    .align 8\n"
+		"    .quad 0b,1b\n"
+		".previous"
+		: "=&d" (ccode) : "a" (__pa(&diag210_tmp)) : "cc", "memory" );
+
+	*addr = diag210_tmp;
+	spin_unlock_irqrestore(&diag210_lock, flags);
+
+	return ccode;
+}
+#else
+int
+diag210(struct diag210 * addr)
+{
+	int ccode;
+
+	asm volatile (
+		"   lhi	  %0,-1\n"
+		"   diag  %1,0,0x210\n"
+		"0: ipm	  %0\n"
+		"   srl	  %0,28\n"
+		"1:\n"
+		".section __ex_table,\"a\"\n"
+		"    .align 4\n"
+		"    .long 0b,1b\n"
+		".previous"
+		: "=&d" (ccode) : "a" (__pa(addr)) : "cc", "memory" );
+
+	return ccode;
+}
+#endif
+
+/*
+ * Input :
+ *   devno - device number
+ *   ps	   - pointer to sense ID data area
+ * Output : none
+ */
+static void
+VM_virtual_device_info (__u16 devno, struct senseid *ps)
+{
+	static struct {
+		int vrdcvcla, vrdcvtyp, cu_type;
+	} vm_devices[] = {
+		{ 0x08, 0x01, 0x3480 },
+		{ 0x08, 0x02, 0x3430 },
+		{ 0x08, 0x10, 0x3420 },
+		{ 0x08, 0x42, 0x3424 },
+		{ 0x08, 0x44, 0x9348 },
+		{ 0x08, 0x81, 0x3490 },
+		{ 0x08, 0x82, 0x3422 },
+		{ 0x10, 0x41, 0x1403 },
+		{ 0x10, 0x42, 0x3211 },
+		{ 0x10, 0x43, 0x3203 },
+		{ 0x10, 0x45, 0x3800 },
+		{ 0x10, 0x47, 0x3262 },
+		{ 0x10, 0x48, 0x3820 },
+		{ 0x10, 0x49, 0x3800 },
+		{ 0x10, 0x4a, 0x4245 },
+		{ 0x10, 0x4b, 0x4248 },
+		{ 0x10, 0x4d, 0x3800 },
+		{ 0x10, 0x4e, 0x3820 },
+		{ 0x10, 0x4f, 0x3820 },
+		{ 0x10, 0x82, 0x2540 },
+		{ 0x10, 0x84, 0x3525 },
+		{ 0x20, 0x81, 0x2501 },
+		{ 0x20, 0x82, 0x2540 },
+		{ 0x20, 0x84, 0x3505 },
+		{ 0x40, 0x01, 0x3278 },
+		{ 0x40, 0x04, 0x3277 },
+		{ 0x40, 0x80, 0x2250 },
+		{ 0x40, 0xc0, 0x5080 },
+		{ 0x80, 0x00, 0x3215 },
+	};
+	struct diag210 diag_data;
+	int ccode, i;
+
+	CIO_TRACE_EVENT (4, "VMvdinf");
+
+	diag_data = (struct diag210) {
+		.vrdcdvno = devno,
+		.vrdclen = sizeof (diag_data),
+	};
+
+	ccode = diag210 (&diag_data);
+	ps->reserved = 0xff;
+
+	/* Special case for bloody osa devices. */
+	if (diag_data.vrdcvcla == 0x02 &&
+	    diag_data.vrdcvtyp == 0x20) {
+		ps->cu_type = 0x3088;
+		ps->cu_model = 0x60;
+		return;
+	}
+	for (i = 0; i < sizeof(vm_devices) / sizeof(vm_devices[0]); i++)
+		if (diag_data.vrdcvcla == vm_devices[i].vrdcvcla &&
+		    diag_data.vrdcvtyp == vm_devices[i].vrdcvtyp) {
+			ps->cu_type = vm_devices[i].cu_type;
+			return;
+		}
+	CIO_MSG_EVENT(0, "DIAG X'210' for device %04X returned (cc = %d):"
+		      "vdev class : %02X, vdev type : %04X \n ...  "
+		      "rdev class : %02X, rdev type : %04X, "
+		      "rdev model: %02X\n",
+		      devno, ccode,
+		      diag_data.vrdcvcla, diag_data.vrdcvtyp,
+		      diag_data.vrdcrccl, diag_data.vrdccrty,
+		      diag_data.vrdccrmd);
+}
+
+/*
+ * Start Sense ID helper function.
+ * Try to obtain the 'control unit'/'device type' information
+ *  associated with the subchannel.
+ */
+static int
+__ccw_device_sense_id_start(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	struct ccw1 *ccw;
+	int ret;
+
+	sch = to_subchannel(cdev->dev.parent);
+	/* Setup sense channel program. */
+	ccw = cdev->private->iccws;
+	if (sch->schib.pmcw.pim != 0x80) {
+		/* more than one path installed. */
+		ccw->cmd_code = CCW_CMD_SUSPEND_RECONN;
+		ccw->cda = 0;
+		ccw->count = 0;
+		ccw->flags = CCW_FLAG_SLI | CCW_FLAG_CC;
+		ccw++;
+	}
+	ccw->cmd_code = CCW_CMD_SENSE_ID;
+	ccw->cda = (__u32) __pa (&cdev->private->senseid);
+	ccw->count = sizeof (struct senseid);
+	ccw->flags = CCW_FLAG_SLI;
+
+	/* Reset device status. */
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+
+	/* Try on every path. */
+	ret = -ENODEV;
+	while (cdev->private->imask != 0) {
+		if ((sch->opm & cdev->private->imask) != 0 &&
+		    cdev->private->iretry > 0) {
+			cdev->private->iretry--;
+			ret = cio_start (sch, cdev->private->iccws,
+					 cdev->private->imask);
+			/* ret is 0, -EBUSY, -EACCES or -ENODEV */
+			if (ret != -EACCES)
+				return ret;
+		}
+		cdev->private->imask >>= 1;
+		cdev->private->iretry = 5;
+	}
+	return ret;
+}
+
+void
+ccw_device_sense_id_start(struct ccw_device *cdev)
+{
+	int ret;
+
+	memset (&cdev->private->senseid, 0, sizeof (struct senseid));
+	cdev->private->senseid.cu_type = 0xFFFF;
+	cdev->private->imask = 0x80;
+	cdev->private->iretry = 5;
+	ret = __ccw_device_sense_id_start(cdev);
+	if (ret && ret != -EBUSY)
+		ccw_device_sense_id_done(cdev, ret);
+}
+
+/*
+ * Called from interrupt context to check if a valid answer
+ * to Sense ID was received.
+ */
+static int
+ccw_device_check_sense_id(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+
+	sch = to_subchannel(cdev->dev.parent);
+	irb = &cdev->private->irb;
+	/* Did we get a proper answer ? */
+	if (cdev->private->senseid.cu_type != 0xFFFF && 
+	    cdev->private->senseid.reserved == 0xFF) {
+		if (irb->scsw.count < sizeof (struct senseid) - 8)
+			cdev->private->flags.esid = 1;
+		return 0; /* Success */
+	}
+	/* Check the error cases. */
+	if (irb->scsw.fctl & (SCSW_FCTL_HALT_FUNC | SCSW_FCTL_CLEAR_FUNC))
+		return -ETIME;
+	if (irb->esw.esw0.erw.cons && (irb->ecw[0] & SNS0_CMD_REJECT)) {
+		/*
+		 * if the device doesn't support the SenseID
+		 *  command further retries wouldn't help ...
+		 * NB: We don't check here for intervention required like we
+		 *     did before, because tape devices with no tape inserted
+		 *     may present this status *in conjunction with* the
+		 *     sense id information. So, for intervention required,
+		 *     we use the "whack it until it talks" strategy...
+		 */
+		CIO_MSG_EVENT(2, "SenseID : device %04x on Subchannel %04x "
+			      "reports cmd reject\n",
+			      cdev->private->devno, sch->irq);
+		return -EOPNOTSUPP;
+	}
+	if (irb->esw.esw0.erw.cons) {
+		CIO_MSG_EVENT(2, "SenseID : UC on dev %04x, "
+			      "lpum %02X, cnt %02d, sns :"
+			      " %02X%02X%02X%02X %02X%02X%02X%02X ...\n",
+			      cdev->private->devno,
+			      irb->esw.esw0.sublog.lpum,
+			      irb->esw.esw0.erw.scnt,
+			      irb->ecw[0], irb->ecw[1],
+			      irb->ecw[2], irb->ecw[3],
+			      irb->ecw[4], irb->ecw[5],
+			      irb->ecw[6], irb->ecw[7]);
+		return -EAGAIN;
+	}
+	if (irb->scsw.cc == 3) {
+		if ((sch->orb.lpm &
+		     sch->schib.pmcw.pim & sch->schib.pmcw.pam) != 0)
+			CIO_MSG_EVENT(2, "SenseID : path %02X for device %04x on"
+				      " subchannel %04x is 'not operational'\n",
+				      sch->orb.lpm, cdev->private->devno,
+				      sch->irq);
+		return -EACCES;
+	}
+	/* Hmm, whatever happened, try again. */
+	CIO_MSG_EVENT(2, "SenseID : start_IO() for device %04x on "
+		      "subchannel %04x returns status %02X%02X\n",
+		      cdev->private->devno, sch->irq,
+		      irb->scsw.dstat, irb->scsw.cstat);
+	return -EAGAIN;
+}
+
+/*
+ * Got interrupt for Sense ID.
+ */
+void
+ccw_device_sense_id_irq(struct ccw_device *cdev, enum dev_event dev_event)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+	int ret;
+
+	sch = to_subchannel(cdev->dev.parent);
+	irb = (struct irb *) __LC_IRB;
+	/* Retry sense id, if needed. */
+	if (irb->scsw.stctl ==
+	    (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) {
+		if ((irb->scsw.cc == 1) || !irb->scsw.actl) {
+			ret = __ccw_device_sense_id_start(cdev);
+			if (ret && ret != -EBUSY)
+				ccw_device_sense_id_done(cdev, ret);
+		}
+		return;
+	}
+	if (ccw_device_accumulate_and_sense(cdev, irb) != 0)
+		return;
+	ret = ccw_device_check_sense_id(cdev);
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	switch (ret) {
+	/* 0, -ETIME, -EOPNOTSUPP, -EAGAIN or -EACCES */
+	case 0:			/* Sense id succeeded. */
+	case -ETIME:		/* Sense id stopped by timeout. */
+		ccw_device_sense_id_done(cdev, ret);
+		break;
+	case -EACCES:		/* channel is not operational. */
+		sch->lpm &= ~cdev->private->imask;
+		cdev->private->imask >>= 1;
+		cdev->private->iretry = 5;
+		/* fall through. */
+	case -EAGAIN:		/* try again. */
+		ret = __ccw_device_sense_id_start(cdev);
+		if (ret == 0 || ret == -EBUSY)
+			break;
+		/* fall through. */
+	default:		/* Sense ID failed. Try asking VM. */
+		if (MACHINE_IS_VM) {
+			VM_virtual_device_info (cdev->private->devno,
+						&cdev->private->senseid);
+			if (cdev->private->senseid.cu_type != 0xFFFF) {
+				/* Got the device information from VM. */
+				ccw_device_sense_id_done(cdev, 0);
+				return;
+			}
+		}
+		/*
+		 * If we can't couldn't identify the device type we
+		 *  consider the device "not operational".
+		 */
+		ccw_device_sense_id_done(cdev, -ENODEV);
+		break;
+	}
+}
+
+EXPORT_SYMBOL(diag210);

drivers/s390/cio/device_ops.c

@@ -0,0 +1,603 @@
+/*
+ *  drivers/s390/cio/device_ops.c
+ *
+ *   $Revision: 1.55 $
+ *
+ *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
+ *			 IBM Corporation
+ *    Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *               Cornelia Huck (cohuck@de.ibm.com)
+ */
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/slab.h>
+#include <linux/list.h>
+#include <linux/device.h>
+#include <linux/delay.h>
+
+#include <asm/ccwdev.h>
+#include <asm/idals.h>
+#include <asm/qdio.h>
+
+#include "cio.h"
+#include "cio_debug.h"
+#include "css.h"
+#include "chsc.h"
+#include "device.h"
+#include "qdio.h"
+
+int
+ccw_device_set_options(struct ccw_device *cdev, unsigned long flags)
+{
+       /*
+	* The flag usage is mutal exclusive ...
+	*/
+	if ((flags & CCWDEV_EARLY_NOTIFICATION) &&
+	    (flags & CCWDEV_REPORT_ALL))
+		return -EINVAL;
+	cdev->private->options.fast = (flags & CCWDEV_EARLY_NOTIFICATION) != 0;
+	cdev->private->options.repall = (flags & CCWDEV_REPORT_ALL) != 0;
+	cdev->private->options.pgroup = (flags & CCWDEV_DO_PATHGROUP) != 0;
+	cdev->private->options.force = (flags & CCWDEV_ALLOW_FORCE) != 0;
+	return 0;
+}
+
+int
+ccw_device_clear(struct ccw_device *cdev, unsigned long intparm)
+{
+	struct subchannel *sch;
+	int ret;
+
+	if (!cdev)
+		return -ENODEV;
+	if (cdev->private->state == DEV_STATE_NOT_OPER)
+		return -ENODEV;
+	if (cdev->private->state != DEV_STATE_ONLINE &&
+	    cdev->private->state != DEV_STATE_WAIT4IO &&
+	    cdev->private->state != DEV_STATE_W4SENSE)
+		return -EINVAL;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch)
+		return -ENODEV;
+	ret = cio_clear(sch);
+	if (ret == 0)
+		cdev->private->intparm = intparm;
+	return ret;
+}
+
+int
+ccw_device_start_key(struct ccw_device *cdev, struct ccw1 *cpa,
+		     unsigned long intparm, __u8 lpm, __u8 key,
+		     unsigned long flags)
+{
+	struct subchannel *sch;
+	int ret;
+
+	if (!cdev)
+		return -ENODEV;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch)
+		return -ENODEV;
+	if (cdev->private->state == DEV_STATE_NOT_OPER)
+		return -ENODEV;
+	if (cdev->private->state == DEV_STATE_VERIFY) {
+		/* Remember to fake irb when finished. */
+		if (!cdev->private->flags.fake_irb) {
+			cdev->private->flags.fake_irb = 1;
+			cdev->private->intparm = intparm;
+			return 0;
+		} else
+			/* There's already a fake I/O around. */
+			return -EBUSY;
+	}
+	if (cdev->private->state != DEV_STATE_ONLINE ||
+	    ((sch->schib.scsw.stctl & SCSW_STCTL_PRIM_STATUS) &&
+	     !(sch->schib.scsw.stctl & SCSW_STCTL_SEC_STATUS)) ||
+	    cdev->private->flags.doverify)
+		return -EBUSY;
+	ret = cio_set_options (sch, flags);
+	if (ret)
+		return ret;
+	ret = cio_start_key (sch, cpa, lpm, key);
+	if (ret == 0)
+		cdev->private->intparm = intparm;
+	return ret;
+}
+
+
+int
+ccw_device_start_timeout_key(struct ccw_device *cdev, struct ccw1 *cpa,
+			     unsigned long intparm, __u8 lpm, __u8 key,
+			     unsigned long flags, int expires)
+{
+	int ret;
+
+	if (!cdev)
+		return -ENODEV;
+	ccw_device_set_timeout(cdev, expires);
+	ret = ccw_device_start_key(cdev, cpa, intparm, lpm, key, flags);
+	if (ret != 0)
+		ccw_device_set_timeout(cdev, 0);
+	return ret;
+}
+
+int
+ccw_device_start(struct ccw_device *cdev, struct ccw1 *cpa,
+		 unsigned long intparm, __u8 lpm, unsigned long flags)
+{
+	return ccw_device_start_key(cdev, cpa, intparm, lpm,
+				    default_storage_key, flags);
+}
+
+int
+ccw_device_start_timeout(struct ccw_device *cdev, struct ccw1 *cpa,
+			 unsigned long intparm, __u8 lpm, unsigned long flags,
+			 int expires)
+{
+	return ccw_device_start_timeout_key(cdev, cpa, intparm, lpm,
+					    default_storage_key, flags,
+					    expires);
+}
+
+
+int
+ccw_device_halt(struct ccw_device *cdev, unsigned long intparm)
+{
+	struct subchannel *sch;
+	int ret;
+
+	if (!cdev)
+		return -ENODEV;
+	if (cdev->private->state == DEV_STATE_NOT_OPER)
+		return -ENODEV;
+	if (cdev->private->state != DEV_STATE_ONLINE &&
+	    cdev->private->state != DEV_STATE_WAIT4IO &&
+	    cdev->private->state != DEV_STATE_W4SENSE)
+		return -EINVAL;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch)
+		return -ENODEV;
+	ret = cio_halt(sch);
+	if (ret == 0)
+		cdev->private->intparm = intparm;
+	return ret;
+}
+
+int
+ccw_device_resume(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+
+	if (!cdev)
+		return -ENODEV;
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch)
+		return -ENODEV;
+	if (cdev->private->state == DEV_STATE_NOT_OPER)
+		return -ENODEV;
+	if (cdev->private->state != DEV_STATE_ONLINE ||
+	    !(sch->schib.scsw.actl & SCSW_ACTL_SUSPENDED))
+		return -EINVAL;
+	return cio_resume(sch);
+}
+
+/*
+ * Pass interrupt to device driver.
+ */
+int
+ccw_device_call_handler(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	unsigned int stctl;
+	int ending_status;
+
+	sch = to_subchannel(cdev->dev.parent);
+
+	/*
+	 * we allow for the device action handler if .
+	 *  - we received ending status
+	 *  - the action handler requested to see all interrupts
+	 *  - we received an intermediate status
+	 *  - fast notification was requested (primary status)
+	 *  - unsolicited interrupts
+	 */
+	stctl = cdev->private->irb.scsw.stctl;
+	ending_status = (stctl & SCSW_STCTL_SEC_STATUS) ||
+		(stctl == (SCSW_STCTL_ALERT_STATUS | SCSW_STCTL_STATUS_PEND)) ||
+		(stctl == SCSW_STCTL_STATUS_PEND);
+	if (!ending_status &&
+	    !cdev->private->options.repall &&
+	    !(stctl & SCSW_STCTL_INTER_STATUS) &&
+	    !(cdev->private->options.fast &&
+	      (stctl & SCSW_STCTL_PRIM_STATUS)))
+		return 0;
+
+	/*
+	 * Now we are ready to call the device driver interrupt handler.
+	 */
+	if (cdev->handler)
+		cdev->handler(cdev, cdev->private->intparm,
+			      &cdev->private->irb);
+
+	/*
+	 * Clear the old and now useless interrupt response block.
+	 */
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+
+	return 1;
+}
+
+/*
+ * Search for CIW command in extended sense data.
+ */
+struct ciw *
+ccw_device_get_ciw(struct ccw_device *cdev, __u32 ct)
+{
+	int ciw_cnt;
+
+	if (cdev->private->flags.esid == 0)
+		return NULL;
+	for (ciw_cnt = 0; ciw_cnt < MAX_CIWS; ciw_cnt++)
+		if (cdev->private->senseid.ciw[ciw_cnt].ct == ct)
+			return cdev->private->senseid.ciw + ciw_cnt;
+	return NULL;
+}
+
+__u8
+ccw_device_get_path_mask(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+
+	sch = to_subchannel(cdev->dev.parent);
+	if (!sch)
+		return 0;
+	else
+		return sch->vpm;
+}
+
+static void
+ccw_device_wake_up(struct ccw_device *cdev, unsigned long ip, struct irb *irb)
+{
+	if (!ip)
+		/* unsolicited interrupt */
+		return;
+
+	/* Abuse intparm for error reporting. */
+	if (IS_ERR(irb))
+		cdev->private->intparm = -EIO;
+	else if ((irb->scsw.dstat !=
+		  (DEV_STAT_CHN_END|DEV_STAT_DEV_END)) ||
+		 (irb->scsw.cstat != 0)) {
+		/*
+		 * We didn't get channel end / device end. Check if path
+		 * verification has been started; we can retry after it has
+		 * finished. We also retry unit checks except for command reject
+		 * or intervention required.
+		 */
+		 if (cdev->private->flags.doverify ||
+			 cdev->private->state == DEV_STATE_VERIFY)
+			 cdev->private->intparm = -EAGAIN;
+		 if ((irb->scsw.dstat & DEV_STAT_UNIT_CHECK) &&
+		     !(irb->ecw[0] &
+		       (SNS0_CMD_REJECT | SNS0_INTERVENTION_REQ)))
+			 cdev->private->intparm = -EAGAIN;
+		 else
+			 cdev->private->intparm = -EIO;
+			 
+	} else
+		cdev->private->intparm = 0;
+	wake_up(&cdev->private->wait_q);
+}
+
+static inline int
+__ccw_device_retry_loop(struct ccw_device *cdev, struct ccw1 *ccw, long magic, __u8 lpm)
+{
+	int ret;
+	struct subchannel *sch;
+
+	sch = to_subchannel(cdev->dev.parent);
+	do {
+		ret = cio_start (sch, ccw, lpm);
+		if ((ret == -EBUSY) || (ret == -EACCES)) {
+			/* Try again later. */
+			spin_unlock_irq(&sch->lock);
+			msleep(10);
+			spin_lock_irq(&sch->lock);
+			continue;
+		}
+		if (ret != 0)
+			/* Non-retryable error. */
+			break;
+		/* Wait for end of request. */
+		cdev->private->intparm = magic;
+		spin_unlock_irq(&sch->lock);
+		wait_event(cdev->private->wait_q,
+			   (cdev->private->intparm == -EIO) ||
+			   (cdev->private->intparm == -EAGAIN) ||
+			   (cdev->private->intparm == 0));
+		spin_lock_irq(&sch->lock);
+		/* Check at least for channel end / device end */
+		if (cdev->private->intparm == -EIO) {
+			/* Non-retryable error. */
+			ret = -EIO;
+			break;
+		}
+		if (cdev->private->intparm == 0)
+			/* Success. */
+			break;
+		/* Try again later. */
+		spin_unlock_irq(&sch->lock);
+		msleep(10);
+		spin_lock_irq(&sch->lock);
+	} while (1);
+
+	return ret;
+}
+
+/**
+ * read_dev_chars() - read device characteristics
+ * @param cdev   target ccw device
+ * @param buffer pointer to buffer for rdc data
+ * @param length size of rdc data
+ * @returns 0 for success, negative error value on failure
+ *
+ * Context:
+ *   called for online device, lock not held
+ **/
+int
+read_dev_chars (struct ccw_device *cdev, void **buffer, int length)
+{
+	void (*handler)(struct ccw_device *, unsigned long, struct irb *);
+	struct subchannel *sch;
+	int ret;
+	struct ccw1 *rdc_ccw;
+
+	if (!cdev)
+		return -ENODEV;
+	if (!buffer || !length)
+		return -EINVAL;
+	sch = to_subchannel(cdev->dev.parent);
+
+	CIO_TRACE_EVENT (4, "rddevch");
+	CIO_TRACE_EVENT (4, sch->dev.bus_id);
+
+	rdc_ccw = kmalloc(sizeof(struct ccw1), GFP_KERNEL | GFP_DMA);
+	if (!rdc_ccw)
+		return -ENOMEM;
+	memset(rdc_ccw, 0, sizeof(struct ccw1));
+	rdc_ccw->cmd_code = CCW_CMD_RDC;
+	rdc_ccw->count = length;
+	rdc_ccw->flags = CCW_FLAG_SLI;
+	ret = set_normalized_cda (rdc_ccw, (*buffer));
+	if (ret != 0) {
+		kfree(rdc_ccw);
+		return ret;
+	}
+
+	spin_lock_irq(&sch->lock);
+	/* Save interrupt handler. */
+	handler = cdev->handler;
+	/* Temporarily install own handler. */
+	cdev->handler = ccw_device_wake_up;
+	if (cdev->private->state != DEV_STATE_ONLINE)
+		ret = -ENODEV;
+	else if (((sch->schib.scsw.stctl & SCSW_STCTL_PRIM_STATUS) &&
+		  !(sch->schib.scsw.stctl & SCSW_STCTL_SEC_STATUS)) ||
+		 cdev->private->flags.doverify)
+		ret = -EBUSY;
+	else
+		/* 0x00D9C4C3 == ebcdic "RDC" */
+		ret = __ccw_device_retry_loop(cdev, rdc_ccw, 0x00D9C4C3, 0);
+
+	/* Restore interrupt handler. */
+	cdev->handler = handler;
+	spin_unlock_irq(&sch->lock);
+
+	clear_normalized_cda (rdc_ccw);
+	kfree(rdc_ccw);
+
+	return ret;
+}
+
+/*
+ *  Read Configuration data using path mask
+ */
+int
+read_conf_data_lpm (struct ccw_device *cdev, void **buffer, int *length, __u8 lpm)
+{
+	void (*handler)(struct ccw_device *, unsigned long, struct irb *);
+	struct subchannel *sch;
+	struct ciw *ciw;
+	char *rcd_buf;
+	int ret;
+	struct ccw1 *rcd_ccw;
+
+	if (!cdev)
+		return -ENODEV;
+	if (!buffer || !length)
+		return -EINVAL;
+	sch = to_subchannel(cdev->dev.parent);
+
+	CIO_TRACE_EVENT (4, "rdconf");
+	CIO_TRACE_EVENT (4, sch->dev.bus_id);
+
+	/*
+	 * scan for RCD command in extended SenseID data
+	 */
+	ciw = ccw_device_get_ciw(cdev, CIW_TYPE_RCD);
+	if (!ciw || ciw->cmd == 0)
+		return -EOPNOTSUPP;
+
+	rcd_ccw = kmalloc(sizeof(struct ccw1), GFP_KERNEL | GFP_DMA);
+	if (!rcd_ccw)
+		return -ENOMEM;
+	memset(rcd_ccw, 0, sizeof(struct ccw1));
+	rcd_buf = kmalloc(ciw->count, GFP_KERNEL | GFP_DMA);
+ 	if (!rcd_buf) {
+		kfree(rcd_ccw);
+		return -ENOMEM;
+	}
+ 	memset (rcd_buf, 0, ciw->count);
+	rcd_ccw->cmd_code = ciw->cmd;
+	rcd_ccw->cda = (__u32) __pa (rcd_buf);
+	rcd_ccw->count = ciw->count;
+	rcd_ccw->flags = CCW_FLAG_SLI;
+
+	spin_lock_irq(&sch->lock);
+	/* Save interrupt handler. */
+	handler = cdev->handler;
+	/* Temporarily install own handler. */
+	cdev->handler = ccw_device_wake_up;
+	if (cdev->private->state != DEV_STATE_ONLINE)
+		ret = -ENODEV;
+	else if (((sch->schib.scsw.stctl & SCSW_STCTL_PRIM_STATUS) &&
+		  !(sch->schib.scsw.stctl & SCSW_STCTL_SEC_STATUS)) ||
+		 cdev->private->flags.doverify)
+		ret = -EBUSY;
+	else
+		/* 0x00D9C3C4 == ebcdic "RCD" */
+		ret = __ccw_device_retry_loop(cdev, rcd_ccw, 0x00D9C3C4, lpm);
+
+	/* Restore interrupt handler. */
+	cdev->handler = handler;
+	spin_unlock_irq(&sch->lock);
+
+ 	/*
+ 	 * on success we update the user input parms
+ 	 */
+ 	if (ret) {
+ 		kfree (rcd_buf);
+ 		*buffer = NULL;
+ 		*length = 0;
+ 	} else {
+		*length = ciw->count;
+		*buffer = rcd_buf;
+	}
+	kfree(rcd_ccw);
+
+	return ret;
+}
+
+/*
+ *  Read Configuration data
+ */
+int
+read_conf_data (struct ccw_device *cdev, void **buffer, int *length)
+{
+	return read_conf_data_lpm (cdev, buffer, length, 0);
+}
+
+/*
+ * Try to break the lock on a boxed device.
+ */
+int
+ccw_device_stlck(struct ccw_device *cdev)
+{
+	void *buf, *buf2;
+	unsigned long flags;
+	struct subchannel *sch;
+	int ret;
+
+	if (!cdev)
+		return -ENODEV;
+
+	if (cdev->drv && !cdev->private->options.force)
+		return -EINVAL;
+
+	sch = to_subchannel(cdev->dev.parent);
+	
+	CIO_TRACE_EVENT(2, "stl lock");
+	CIO_TRACE_EVENT(2, cdev->dev.bus_id);
+
+	buf = kmalloc(32*sizeof(char), GFP_DMA|GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+	buf2 = kmalloc(32*sizeof(char), GFP_DMA|GFP_KERNEL);
+	if (!buf2) {
+		kfree(buf);
+		return -ENOMEM;
+	}
+	spin_lock_irqsave(&sch->lock, flags);
+	ret = cio_enable_subchannel(sch, 3);
+	if (ret)
+		goto out_unlock;
+	/*
+	 * Setup ccw. We chain an unconditional reserve and a release so we
+	 * only break the lock.
+	 */
+	cdev->private->iccws[0].cmd_code = CCW_CMD_STLCK;
+	cdev->private->iccws[0].cda = (__u32) __pa(buf);
+	cdev->private->iccws[0].count = 32;
+	cdev->private->iccws[0].flags = CCW_FLAG_CC;
+	cdev->private->iccws[1].cmd_code = CCW_CMD_RELEASE;
+	cdev->private->iccws[1].cda = (__u32) __pa(buf2);
+	cdev->private->iccws[1].count = 32;
+	cdev->private->iccws[1].flags = 0;
+	ret = cio_start(sch, cdev->private->iccws, 0);
+	if (ret) {
+		cio_disable_subchannel(sch); //FIXME: return code?
+		goto out_unlock;
+	}
+	cdev->private->irb.scsw.actl |= SCSW_ACTL_START_PEND;
+	spin_unlock_irqrestore(&sch->lock, flags);
+	wait_event(cdev->private->wait_q, cdev->private->irb.scsw.actl == 0);
+	spin_lock_irqsave(&sch->lock, flags);
+	cio_disable_subchannel(sch); //FIXME: return code?
+	if ((cdev->private->irb.scsw.dstat !=
+	     (DEV_STAT_CHN_END|DEV_STAT_DEV_END)) ||
+	    (cdev->private->irb.scsw.cstat != 0))
+		ret = -EIO;
+	/* Clear irb. */
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+out_unlock:
+	if (buf)
+		kfree(buf);
+	if (buf2)
+		kfree(buf2);
+	spin_unlock_irqrestore(&sch->lock, flags);
+	return ret;
+}
+
+void *
+ccw_device_get_chp_desc(struct ccw_device *cdev, int chp_no)
+{
+	struct subchannel *sch;
+
+	sch = to_subchannel(cdev->dev.parent);
+	return chsc_get_chp_desc(sch, chp_no);
+}
+
+// FIXME: these have to go:
+
+int
+_ccw_device_get_subchannel_number(struct ccw_device *cdev)
+{
+	return cdev->private->irq;
+}
+
+int
+_ccw_device_get_device_number(struct ccw_device *cdev)
+{
+	return cdev->private->devno;
+}
+
+
+MODULE_LICENSE("GPL");
+EXPORT_SYMBOL(ccw_device_set_options);
+EXPORT_SYMBOL(ccw_device_clear);
+EXPORT_SYMBOL(ccw_device_halt);
+EXPORT_SYMBOL(ccw_device_resume);
+EXPORT_SYMBOL(ccw_device_start_timeout);
+EXPORT_SYMBOL(ccw_device_start);
+EXPORT_SYMBOL(ccw_device_start_timeout_key);
+EXPORT_SYMBOL(ccw_device_start_key);
+EXPORT_SYMBOL(ccw_device_get_ciw);
+EXPORT_SYMBOL(ccw_device_get_path_mask);
+EXPORT_SYMBOL(read_conf_data);
+EXPORT_SYMBOL(read_dev_chars);
+EXPORT_SYMBOL(_ccw_device_get_subchannel_number);
+EXPORT_SYMBOL(_ccw_device_get_device_number);
+EXPORT_SYMBOL_GPL(ccw_device_get_chp_desc);
+EXPORT_SYMBOL_GPL(read_conf_data_lpm);

drivers/s390/cio/device_pgid.c

@@ -0,0 +1,448 @@
+/*
+ * drivers/s390/cio/device_pgid.c
+ *
+ *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
+ *			 IBM Corporation
+ *    Author(s): Cornelia Huck(cohuck@de.ibm.com)
+ *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *
+ * Path Group ID functions.
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/init.h>
+
+#include <asm/ccwdev.h>
+#include <asm/cio.h>
+#include <asm/delay.h>
+#include <asm/lowcore.h>
+
+#include "cio.h"
+#include "cio_debug.h"
+#include "css.h"
+#include "device.h"
+
+/*
+ * Start Sense Path Group ID helper function. Used in ccw_device_recog
+ * and ccw_device_sense_pgid.
+ */
+static int
+__ccw_device_sense_pgid_start(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	struct ccw1 *ccw;
+	int ret;
+
+	sch = to_subchannel(cdev->dev.parent);
+	/* Setup sense path group id channel program. */
+	ccw = cdev->private->iccws;
+	ccw->cmd_code = CCW_CMD_SENSE_PGID;
+	ccw->cda = (__u32) __pa (&cdev->private->pgid);
+	ccw->count = sizeof (struct pgid);
+	ccw->flags = CCW_FLAG_SLI;
+
+	/* Reset device status. */
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	/* Try on every path. */
+	ret = -ENODEV;
+	while (cdev->private->imask != 0) {
+		/* Try every path multiple times. */
+		if (cdev->private->iretry > 0) {
+			cdev->private->iretry--;
+			ret = cio_start (sch, cdev->private->iccws, 
+					 cdev->private->imask);
+			/* ret is 0, -EBUSY, -EACCES or -ENODEV */
+			if (ret != -EACCES)
+				return ret;
+			CIO_MSG_EVENT(2, "SNID - Device %04x on Subchannel "
+				      "%04x, lpm %02X, became 'not "
+				      "operational'\n",
+				      cdev->private->devno, sch->irq,
+				      cdev->private->imask);
+
+		}
+		cdev->private->imask >>= 1;
+		cdev->private->iretry = 5;
+	}
+	return ret;
+}
+
+void
+ccw_device_sense_pgid_start(struct ccw_device *cdev)
+{
+	int ret;
+
+	cdev->private->state = DEV_STATE_SENSE_PGID;
+	cdev->private->imask = 0x80;
+	cdev->private->iretry = 5;
+	memset (&cdev->private->pgid, 0, sizeof (struct pgid));
+	ret = __ccw_device_sense_pgid_start(cdev);
+	if (ret && ret != -EBUSY)
+		ccw_device_sense_pgid_done(cdev, ret);
+}
+
+/*
+ * Called from interrupt context to check if a valid answer
+ * to Sense Path Group ID was received.
+ */
+static int
+__ccw_device_check_sense_pgid(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+
+	sch = to_subchannel(cdev->dev.parent);
+	irb = &cdev->private->irb;
+	if (irb->scsw.fctl & (SCSW_FCTL_HALT_FUNC | SCSW_FCTL_CLEAR_FUNC))
+		return -ETIME;
+	if (irb->esw.esw0.erw.cons &&
+	    (irb->ecw[0]&(SNS0_CMD_REJECT|SNS0_INTERVENTION_REQ))) {
+		/*
+		 * If the device doesn't support the Sense Path Group ID
+		 *  command further retries wouldn't help ...
+		 */
+		return -EOPNOTSUPP;
+	}
+	if (irb->esw.esw0.erw.cons) {
+		CIO_MSG_EVENT(2, "SNID - device %04x, unit check, "
+			      "lpum %02X, cnt %02d, sns : "
+			      "%02X%02X%02X%02X %02X%02X%02X%02X ...\n",
+			      cdev->private->devno,
+			      irb->esw.esw0.sublog.lpum,
+			      irb->esw.esw0.erw.scnt,
+			      irb->ecw[0], irb->ecw[1],
+			      irb->ecw[2], irb->ecw[3],
+			      irb->ecw[4], irb->ecw[5],
+			      irb->ecw[6], irb->ecw[7]);
+		return -EAGAIN;
+	}
+	if (irb->scsw.cc == 3) {
+		CIO_MSG_EVENT(2, "SNID - Device %04x on Subchannel "
+			      "%04x, lpm %02X, became 'not operational'\n",
+			      cdev->private->devno, sch->irq, sch->orb.lpm);
+		return -EACCES;
+	}
+	if (cdev->private->pgid.inf.ps.state2 == SNID_STATE2_RESVD_ELSE) {
+		CIO_MSG_EVENT(2, "SNID - Device %04x on Subchannel %04x "
+			      "is reserved by someone else\n",
+			      cdev->private->devno, sch->irq);
+		return -EUSERS;
+	}
+	return 0;
+}
+
+/*
+ * Got interrupt for Sense Path Group ID.
+ */
+void
+ccw_device_sense_pgid_irq(struct ccw_device *cdev, enum dev_event dev_event)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+	int ret;
+
+	irb = (struct irb *) __LC_IRB;
+	/* Retry sense pgid for cc=1. */
+	if (irb->scsw.stctl ==
+	    (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) {
+		if (irb->scsw.cc == 1) {
+			ret = __ccw_device_sense_pgid_start(cdev);
+			if (ret && ret != -EBUSY)
+				ccw_device_sense_pgid_done(cdev, ret);
+		}
+		return;
+	}
+	if (ccw_device_accumulate_and_sense(cdev, irb) != 0)
+		return;
+	sch = to_subchannel(cdev->dev.parent);
+	ret = __ccw_device_check_sense_pgid(cdev);
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	switch (ret) {
+	/* 0, -ETIME, -EOPNOTSUPP, -EAGAIN, -EACCES or -EUSERS */
+	case 0:			/* Sense Path Group ID successful. */
+		if (cdev->private->pgid.inf.ps.state1 == SNID_STATE1_RESET)
+			memcpy(&cdev->private->pgid, &global_pgid,
+			       sizeof(struct pgid));
+		ccw_device_sense_pgid_done(cdev, 0);
+		break;
+	case -EOPNOTSUPP:	/* Sense Path Group ID not supported */
+		ccw_device_sense_pgid_done(cdev, -EOPNOTSUPP);
+		break;
+	case -ETIME:		/* Sense path group id stopped by timeout. */
+		ccw_device_sense_pgid_done(cdev, -ETIME);
+		break;
+	case -EACCES:		/* channel is not operational. */
+		sch->lpm &= ~cdev->private->imask;
+		cdev->private->imask >>= 1;
+		cdev->private->iretry = 5;
+		/* Fall through. */
+	case -EAGAIN:		/* Try again. */
+		ret = __ccw_device_sense_pgid_start(cdev);
+		if (ret != 0 && ret != -EBUSY)
+			ccw_device_sense_pgid_done(cdev, -ENODEV);
+		break;
+	case -EUSERS:		/* device is reserved for someone else. */
+		ccw_device_sense_pgid_done(cdev, -EUSERS);
+		break;
+	}
+}
+
+/*
+ * Path Group ID helper function.
+ */
+static int
+__ccw_device_do_pgid(struct ccw_device *cdev, __u8 func)
+{
+	struct subchannel *sch;
+	struct ccw1 *ccw;
+	int ret;
+
+	sch = to_subchannel(cdev->dev.parent);
+
+	/* Setup sense path group id channel program. */
+	cdev->private->pgid.inf.fc = func;
+	ccw = cdev->private->iccws;
+	if (!cdev->private->flags.pgid_single) {
+		cdev->private->pgid.inf.fc |= SPID_FUNC_MULTI_PATH;
+		ccw->cmd_code = CCW_CMD_SUSPEND_RECONN;
+		ccw->cda = 0;
+		ccw->count = 0;
+		ccw->flags = CCW_FLAG_SLI | CCW_FLAG_CC;
+		ccw++;
+	} else
+		cdev->private->pgid.inf.fc |= SPID_FUNC_SINGLE_PATH;
+
+	ccw->cmd_code = CCW_CMD_SET_PGID;
+	ccw->cda = (__u32) __pa (&cdev->private->pgid);
+	ccw->count = sizeof (struct pgid);
+	ccw->flags = CCW_FLAG_SLI;
+
+	/* Reset device status. */
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+
+	/* Try multiple times. */
+	ret = -ENODEV;
+	if (cdev->private->iretry > 0) {
+		cdev->private->iretry--;
+		ret = cio_start (sch, cdev->private->iccws,
+				 cdev->private->imask);
+		/* ret is 0, -EBUSY, -EACCES or -ENODEV */
+		if ((ret != -EACCES) && (ret != -ENODEV))
+			return ret;
+	}
+	/* PGID command failed on this path. Switch it off. */
+	sch->lpm &= ~cdev->private->imask;
+	sch->vpm &= ~cdev->private->imask;
+	CIO_MSG_EVENT(2, "SPID - Device %04x on Subchannel "
+		      "%04x, lpm %02X, became 'not operational'\n",
+		      cdev->private->devno, sch->irq, cdev->private->imask);
+	return ret;
+}
+
+/*
+ * Called from interrupt context to check if a valid answer
+ * to Set Path Group ID was received.
+ */
+static int
+__ccw_device_check_pgid(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+
+	sch = to_subchannel(cdev->dev.parent);
+	irb = &cdev->private->irb;
+	if (irb->scsw.fctl & (SCSW_FCTL_HALT_FUNC | SCSW_FCTL_CLEAR_FUNC))
+		return -ETIME;
+	if (irb->esw.esw0.erw.cons) {
+		if (irb->ecw[0] & SNS0_CMD_REJECT)
+			return -EOPNOTSUPP;
+		/* Hmm, whatever happened, try again. */
+		CIO_MSG_EVENT(2, "SPID - device %04x, unit check, cnt %02d, "
+			      "sns : %02X%02X%02X%02X %02X%02X%02X%02X ...\n",
+			      cdev->private->devno, irb->esw.esw0.erw.scnt,
+			      irb->ecw[0], irb->ecw[1],
+			      irb->ecw[2], irb->ecw[3],
+			      irb->ecw[4], irb->ecw[5],
+			      irb->ecw[6], irb->ecw[7]);
+		return -EAGAIN;
+	}
+	if (irb->scsw.cc == 3) {
+		CIO_MSG_EVENT(2, "SPID - Device %04x on Subchannel "
+			      "%04x, lpm %02X, became 'not operational'\n",
+			      cdev->private->devno, sch->irq,
+			      cdev->private->imask);
+		return -EACCES;
+	}
+	return 0;
+}
+
+static void
+__ccw_device_verify_start(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	__u8 imask, func;
+	int ret;
+
+	sch = to_subchannel(cdev->dev.parent);
+	while (sch->vpm != sch->lpm) {
+		/* Find first unequal bit in vpm vs. lpm */
+		for (imask = 0x80; imask != 0; imask >>= 1)
+			if ((sch->vpm & imask) != (sch->lpm & imask))
+				break;
+		cdev->private->imask = imask;
+		func = (sch->vpm & imask) ?
+			SPID_FUNC_RESIGN : SPID_FUNC_ESTABLISH;
+		ret = __ccw_device_do_pgid(cdev, func);
+		if (ret == 0 || ret == -EBUSY)
+			return;
+		cdev->private->iretry = 5;
+	}
+	ccw_device_verify_done(cdev, (sch->lpm != 0) ? 0 : -ENODEV);
+}
+		
+/*
+ * Got interrupt for Set Path Group ID.
+ */
+void
+ccw_device_verify_irq(struct ccw_device *cdev, enum dev_event dev_event)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+	int ret;
+
+	irb = (struct irb *) __LC_IRB;
+	/* Retry set pgid for cc=1. */
+	if (irb->scsw.stctl ==
+	    (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) {
+		if (irb->scsw.cc == 1)
+			__ccw_device_verify_start(cdev);
+		return;
+	}
+	if (ccw_device_accumulate_and_sense(cdev, irb) != 0)
+		return;
+	sch = to_subchannel(cdev->dev.parent);
+	ret = __ccw_device_check_pgid(cdev);
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	switch (ret) {
+	/* 0, -ETIME, -EAGAIN, -EOPNOTSUPP or -EACCES */
+	case 0:
+		/* Establish or Resign Path Group done. Update vpm. */
+		if ((sch->lpm & cdev->private->imask) != 0)
+			sch->vpm |= cdev->private->imask;
+		else
+			sch->vpm &= ~cdev->private->imask;
+		cdev->private->iretry = 5;
+		__ccw_device_verify_start(cdev);
+		break;
+	case -EOPNOTSUPP:
+		/*
+		 * One of those strange devices which claim to be able
+		 * to do multipathing but not for Set Path Group ID.
+		 */
+		if (cdev->private->flags.pgid_single) {
+			ccw_device_verify_done(cdev, -EOPNOTSUPP);
+			break;
+		}
+		cdev->private->flags.pgid_single = 1;
+		/* fall through. */
+	case -EAGAIN:		/* Try again. */
+		__ccw_device_verify_start(cdev);
+		break;
+	case -ETIME:		/* Set path group id stopped by timeout. */
+		ccw_device_verify_done(cdev, -ETIME);
+		break;
+	case -EACCES:		/* channel is not operational. */
+		sch->lpm &= ~cdev->private->imask;
+		sch->vpm &= ~cdev->private->imask;
+		cdev->private->iretry = 5;
+		__ccw_device_verify_start(cdev);
+		break;
+	}
+}
+
+void
+ccw_device_verify_start(struct ccw_device *cdev)
+{
+	cdev->private->flags.pgid_single = 0;
+	cdev->private->iretry = 5;
+	__ccw_device_verify_start(cdev);
+}
+
+static void
+__ccw_device_disband_start(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+	int ret;
+
+	sch = to_subchannel(cdev->dev.parent);
+	while (cdev->private->imask != 0) {
+		if (sch->lpm & cdev->private->imask) {
+			ret = __ccw_device_do_pgid(cdev, SPID_FUNC_DISBAND);
+			if (ret == 0)
+				return;
+		}
+		cdev->private->iretry = 5;
+		cdev->private->imask >>= 1;
+	}
+	ccw_device_verify_done(cdev, (sch->lpm != 0) ? 0 : -ENODEV);
+}
+
+/*
+ * Got interrupt for Unset Path Group ID.
+ */
+void
+ccw_device_disband_irq(struct ccw_device *cdev, enum dev_event dev_event)
+{
+	struct subchannel *sch;
+	struct irb *irb;
+	int ret;
+
+	irb = (struct irb *) __LC_IRB;
+	/* Retry set pgid for cc=1. */
+	if (irb->scsw.stctl ==
+	    (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) {
+		if (irb->scsw.cc == 1)
+			__ccw_device_disband_start(cdev);
+		return;
+	}
+	if (ccw_device_accumulate_and_sense(cdev, irb) != 0)
+		return;
+	sch = to_subchannel(cdev->dev.parent);
+	ret = __ccw_device_check_pgid(cdev);
+	memset(&cdev->private->irb, 0, sizeof(struct irb));
+	switch (ret) {
+	/* 0, -ETIME, -EAGAIN, -EOPNOTSUPP or -EACCES */
+	case 0:			/* disband successful. */
+		sch->vpm = 0;
+		ccw_device_disband_done(cdev, ret);
+		break;
+	case -EOPNOTSUPP:
+		/*
+		 * One of those strange devices which claim to be able
+		 * to do multipathing but not for Unset Path Group ID.
+		 */
+		cdev->private->flags.pgid_single = 1;
+		/* fall through. */
+	case -EAGAIN:		/* Try again. */
+		__ccw_device_disband_start(cdev);
+		break;
+	case -ETIME:		/* Set path group id stopped by timeout. */
+		ccw_device_disband_done(cdev, -ETIME);
+		break;
+	case -EACCES:		/* channel is not operational. */
+		cdev->private->imask >>= 1;
+		cdev->private->iretry = 5;
+		__ccw_device_disband_start(cdev);
+		break;
+	}
+}
+
+void
+ccw_device_disband_start(struct ccw_device *cdev)
+{
+	cdev->private->flags.pgid_single = 0;
+	cdev->private->iretry = 5;
+	cdev->private->imask = 0x80;
+	__ccw_device_disband_start(cdev);
+}

drivers/s390/cio/device_status.c

@@ -0,0 +1,385 @@
+/*
+ * drivers/s390/cio/device_status.c
+ *
+ *    Copyright (C) 2002 IBM Deutschland Entwicklung GmbH,
+ *			 IBM Corporation
+ *    Author(s): Cornelia Huck(cohuck@de.ibm.com)
+ *		 Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *
+ * Status accumulation and basic sense functions.
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/init.h>
+
+#include <asm/ccwdev.h>
+#include <asm/cio.h>
+
+#include "cio.h"
+#include "cio_debug.h"
+#include "css.h"
+#include "device.h"
+#include "ioasm.h"
+
+/*
+ * Check for any kind of channel or interface control check but don't
+ * issue the message for the console device
+ */
+static inline void
+ccw_device_msg_control_check(struct ccw_device *cdev, struct irb *irb)
+{
+	if (!(irb->scsw.cstat & (SCHN_STAT_CHN_DATA_CHK |
+				 SCHN_STAT_CHN_CTRL_CHK |
+				 SCHN_STAT_INTF_CTRL_CHK)))
+		return;
+		
+	CIO_MSG_EVENT(0, "Channel-Check or Interface-Control-Check "
+		      "received"
+		      " ... device %04X on subchannel %04X, dev_stat "
+		      ": %02X sch_stat : %02X\n",
+		      cdev->private->devno, cdev->private->irq,
+		      cdev->private->irb.scsw.dstat,
+		      cdev->private->irb.scsw.cstat);
+
+	if (irb->scsw.cc != 3) {
+		char dbf_text[15];
+
+		sprintf(dbf_text, "chk%x", cdev->private->irq);
+		CIO_TRACE_EVENT(0, dbf_text);
+		CIO_HEX_EVENT(0, &cdev->private->irb, sizeof (struct irb));
+	}
+}
+
+/*
+ * Some paths became not operational (pno bit in scsw is set).
+ */
+static void
+ccw_device_path_notoper(struct ccw_device *cdev)
+{
+	struct subchannel *sch;
+
+	sch = to_subchannel(cdev->dev.parent);
+	stsch (sch->irq, &sch->schib);
+
+	CIO_MSG_EVENT(0, "%s(%04x) - path(s) %02x are "
+		      "not operational \n", __FUNCTION__, sch->irq,
+		      sch->schib.pmcw.pnom);
+
+	sch->lpm &= ~sch->schib.pmcw.pnom;
+	if (cdev->private->options.pgroup)
+		cdev->private->flags.doverify = 1;
+}
+
+/*
+ * Copy valid bits from the extended control word to device irb.
+ */
+static inline void
+ccw_device_accumulate_ecw(struct ccw_device *cdev, struct irb *irb)
+{
+	/*
+	 * Copy extended control bit if it is valid... yes there
+	 * are condition that have to be met for the extended control
+	 * bit to have meaning. Sick.
+	 */
+	cdev->private->irb.scsw.ectl = 0;
+	if ((irb->scsw.stctl & SCSW_STCTL_ALERT_STATUS) &&
+	    !(irb->scsw.stctl & SCSW_STCTL_INTER_STATUS))
+		cdev->private->irb.scsw.ectl = irb->scsw.ectl;
+	/* Check if extended control word is valid. */
+	if (!cdev->private->irb.scsw.ectl)
+		return;
+	/* Copy concurrent sense / model dependent information. */
+	memcpy (&cdev->private->irb.ecw, irb->ecw, sizeof (irb->ecw));
+}
+
+/*
+ * Check if extended status word is valid.
+ */
+static inline int
+ccw_device_accumulate_esw_valid(struct irb *irb)
+{
+	if (!irb->scsw.eswf && irb->scsw.stctl == SCSW_STCTL_STATUS_PEND)
+		return 0;
+	if (irb->scsw.stctl == 
+	    		(SCSW_STCTL_INTER_STATUS|SCSW_STCTL_STATUS_PEND) &&
+	    !(irb->scsw.actl & SCSW_ACTL_SUSPENDED))
+		return 0;
+	return 1;
+}
+
+/*
+ * Copy valid bits from the extended status word to device irb.
+ */
+static inline void
+ccw_device_accumulate_esw(struct ccw_device *cdev, struct irb *irb)
+{
+	struct irb *cdev_irb;
+	struct sublog *cdev_sublog, *sublog;
+
+	if (!ccw_device_accumulate_esw_valid(irb))
+		return;
+
+	cdev_irb = &cdev->private->irb;
+
+	/* Copy last path used mask. */
+	cdev_irb->esw.esw1.lpum = irb->esw.esw1.lpum;
+
+	/* Copy subchannel logout information if esw is of format 0. */
+	if (irb->scsw.eswf) {
+		cdev_sublog = &cdev_irb->esw.esw0.sublog;
+		sublog = &irb->esw.esw0.sublog;
+		/* Copy extended status flags. */
+		cdev_sublog->esf = sublog->esf;
+		/*
+		 * Copy fields that have a meaning for channel data check
+		 * channel control check and interface control check.
+		 */
+		if (irb->scsw.cstat & (SCHN_STAT_CHN_DATA_CHK |
+				       SCHN_STAT_CHN_CTRL_CHK |
+				       SCHN_STAT_INTF_CTRL_CHK)) {
+			/* Copy ancillary report bit. */
+			cdev_sublog->arep = sublog->arep;
+			/* Copy field-validity-flags. */
+			cdev_sublog->fvf = sublog->fvf;
+			/* Copy storage access code. */
+			cdev_sublog->sacc = sublog->sacc;
+			/* Copy termination code. */
+			cdev_sublog->termc = sublog->termc;
+			/* Copy sequence code. */
+			cdev_sublog->seqc = sublog->seqc;
+		}
+		/* Copy device status check. */
+		cdev_sublog->devsc = sublog->devsc;
+		/* Copy secondary error. */
+		cdev_sublog->serr = sublog->serr;
+		/* Copy i/o-error alert. */
+		cdev_sublog->ioerr = sublog->ioerr;
+		/* Copy channel path timeout bit. */
+		if (irb->scsw.cstat & SCHN_STAT_INTF_CTRL_CHK)
+			cdev_irb->esw.esw0.erw.cpt = irb->esw.esw0.erw.cpt;
+		/* Copy failing storage address validity flag. */
+		cdev_irb->esw.esw0.erw.fsavf = irb->esw.esw0.erw.fsavf;
+		if (cdev_irb->esw.esw0.erw.fsavf) {
+			/* ... and copy the failing storage address. */
+			memcpy(cdev_irb->esw.esw0.faddr, irb->esw.esw0.faddr,
+			       sizeof (irb->esw.esw0.faddr));
+			/* ... and copy the failing storage address format. */
+			cdev_irb->esw.esw0.erw.fsaf = irb->esw.esw0.erw.fsaf;
+		}
+		/* Copy secondary ccw address validity bit. */
+		cdev_irb->esw.esw0.erw.scavf = irb->esw.esw0.erw.scavf;
+		if (irb->esw.esw0.erw.scavf)
+			/* ... and copy the secondary ccw address. */
+			cdev_irb->esw.esw0.saddr = irb->esw.esw0.saddr;
+		
+	}
+	/* FIXME: DCTI for format 2? */
+
+	/* Copy authorization bit. */
+	cdev_irb->esw.esw0.erw.auth = irb->esw.esw0.erw.auth;
+	/* Copy path verification required flag. */
+	cdev_irb->esw.esw0.erw.pvrf = irb->esw.esw0.erw.pvrf;
+	if (irb->esw.esw0.erw.pvrf && cdev->private->options.pgroup)
+		cdev->private->flags.doverify = 1;
+	/* Copy concurrent sense bit. */
+	cdev_irb->esw.esw0.erw.cons = irb->esw.esw0.erw.cons;
+	if (irb->esw.esw0.erw.cons)
+		cdev_irb->esw.esw0.erw.scnt = irb->esw.esw0.erw.scnt;
+}
+
+/*
+ * Accumulate status from irb to devstat.
+ */
+void
+ccw_device_accumulate_irb(struct ccw_device *cdev, struct irb *irb)
+{
+	struct irb *cdev_irb;
+
+	/*
+	 * Check if the status pending bit is set in stctl.
+	 * If not, the remaining bit have no meaning and we must ignore them.
+	 * The esw is not meaningful as well...
+	 */
+	if (!(irb->scsw.stctl & SCSW_STCTL_STATUS_PEND))
+		return;
+
+	/* Check for channel checks and interface control checks. */
+	ccw_device_msg_control_check(cdev, irb);
+
+	/* Check for path not operational. */
+	if (irb->scsw.pno && irb->scsw.fctl != 0 &&
+	    (!(irb->scsw.stctl & SCSW_STCTL_INTER_STATUS) ||
+	     (irb->scsw.actl & SCSW_ACTL_SUSPENDED)))
+		ccw_device_path_notoper(cdev);
+
+	/*
+	 * Don't accumulate unsolicited interrupts.
+	 */
+	if ((irb->scsw.stctl ==
+	     (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) &&
+	    (!irb->scsw.cc))
+		return;
+
+	cdev_irb = &cdev->private->irb;
+
+	/* Copy bits which are valid only for the start function. */
+	if (irb->scsw.fctl & SCSW_FCTL_START_FUNC) {
+		/* Copy key. */
+		cdev_irb->scsw.key = irb->scsw.key;
+		/* Copy suspend control bit. */
+		cdev_irb->scsw.sctl = irb->scsw.sctl;
+		/* Accumulate deferred condition code. */
+		cdev_irb->scsw.cc |= irb->scsw.cc;
+		/* Copy ccw format bit. */
+		cdev_irb->scsw.fmt = irb->scsw.fmt;
+		/* Copy prefetch bit. */
+		cdev_irb->scsw.pfch = irb->scsw.pfch;
+		/* Copy initial-status-interruption-control. */
+		cdev_irb->scsw.isic = irb->scsw.isic;
+		/* Copy address limit checking control. */
+		cdev_irb->scsw.alcc = irb->scsw.alcc;
+		/* Copy suppress suspend bit. */
+		cdev_irb->scsw.ssi = irb->scsw.ssi;
+	}
+
+	/* Take care of the extended control bit and extended control word. */
+	ccw_device_accumulate_ecw(cdev, irb);
+	    
+	/* Accumulate function control. */
+	cdev_irb->scsw.fctl |= irb->scsw.fctl;
+	/* Copy activity control. */
+	cdev_irb->scsw.actl= irb->scsw.actl;
+	/* Accumulate status control. */
+	cdev_irb->scsw.stctl |= irb->scsw.stctl;
+	/*
+	 * Copy ccw address if it is valid. This is a bit simplified
+	 * but should be close enough for all practical purposes.
+	 */
+	if ((irb->scsw.stctl & SCSW_STCTL_PRIM_STATUS) ||
+	    ((irb->scsw.stctl == 
+	      (SCSW_STCTL_INTER_STATUS|SCSW_STCTL_STATUS_PEND)) &&
+	     (irb->scsw.actl & SCSW_ACTL_DEVACT) &&
+	     (irb->scsw.actl & SCSW_ACTL_SCHACT)) ||
+	    (irb->scsw.actl & SCSW_ACTL_SUSPENDED))
+		cdev_irb->scsw.cpa = irb->scsw.cpa;
+	/* Accumulate device status, but not the device busy flag. */
+	cdev_irb->scsw.dstat &= ~DEV_STAT_BUSY;
+	cdev_irb->scsw.dstat |= irb->scsw.dstat;
+	/* Accumulate subchannel status. */
+	cdev_irb->scsw.cstat |= irb->scsw.cstat;
+	/* Copy residual count if it is valid. */
+	if ((irb->scsw.stctl & SCSW_STCTL_PRIM_STATUS) &&
+	    (irb->scsw.cstat & ~(SCHN_STAT_PCI | SCHN_STAT_INCORR_LEN)) == 0)
+		cdev_irb->scsw.count = irb->scsw.count;
+
+	/* Take care of bits in the extended status word. */
+	ccw_device_accumulate_esw(cdev, irb);
+
+	/*
+	 * Check whether we must issue a SENSE CCW ourselves if there is no
+	 * concurrent sense facility installed for the subchannel.
+	 * No sense is required if no delayed sense is pending
+	 * and we did not get a unit check without sense information.
+	 *
+	 * Note: We should check for ioinfo[irq]->flags.consns but VM
+	 *	 violates the ESA/390 architecture and doesn't present an
+	 *	 operand exception for virtual devices without concurrent
+	 *	 sense facility available/supported when enabling the
+	 *	 concurrent sense facility.
+	 */
+	if ((cdev_irb->scsw.dstat & DEV_STAT_UNIT_CHECK) &&
+	    !(cdev_irb->esw.esw0.erw.cons))
+		cdev->private->flags.dosense = 1;
+}
+
+/*
+ * Do a basic sense.
+ */
+int
+ccw_device_do_sense(struct ccw_device *cdev, struct irb *irb)
+{
+	struct subchannel *sch;
+
+	sch = to_subchannel(cdev->dev.parent);
+
+	/* A sense is required, can we do it now ? */
+	if ((irb->scsw.actl  & (SCSW_ACTL_DEVACT | SCSW_ACTL_SCHACT)) != 0)
+		/*
+		 * we received an Unit Check but we have no final
+		 *  status yet, therefore we must delay the SENSE
+		 *  processing. We must not report this intermediate
+		 *  status to the device interrupt handler.
+		 */
+		return -EBUSY;
+
+	/*
+	 * We have ending status but no sense information. Do a basic sense.
+	 */
+	sch = to_subchannel(cdev->dev.parent);
+	sch->sense_ccw.cmd_code = CCW_CMD_BASIC_SENSE;
+	sch->sense_ccw.cda = (__u32) __pa(cdev->private->irb.ecw);
+	sch->sense_ccw.count = SENSE_MAX_COUNT;
+	sch->sense_ccw.flags = CCW_FLAG_SLI;
+
+	return cio_start (sch, &sch->sense_ccw, 0xff);
+}
+
+/*
+ * Add information from basic sense to devstat.
+ */
+void
+ccw_device_accumulate_basic_sense(struct ccw_device *cdev, struct irb *irb)
+{
+	/*
+	 * Check if the status pending bit is set in stctl.
+	 * If not, the remaining bit have no meaning and we must ignore them.
+	 * The esw is not meaningful as well...
+	 */
+	if (!(irb->scsw.stctl & SCSW_STCTL_STATUS_PEND))
+		return;
+
+	/* Check for channel checks and interface control checks. */
+	ccw_device_msg_control_check(cdev, irb);
+
+	/* Check for path not operational. */
+	if (irb->scsw.pno && irb->scsw.fctl != 0 &&
+	    (!(irb->scsw.stctl & SCSW_STCTL_INTER_STATUS) ||
+	     (irb->scsw.actl & SCSW_ACTL_SUSPENDED)))
+		ccw_device_path_notoper(cdev);
+
+	if (!(irb->scsw.dstat & DEV_STAT_UNIT_CHECK) &&
+	    (irb->scsw.dstat & DEV_STAT_CHN_END)) {
+		cdev->private->irb.esw.esw0.erw.cons = 1;
+		cdev->private->flags.dosense = 0;
+	}
+	/* Check if path verification is required. */
+	if (ccw_device_accumulate_esw_valid(irb) &&
+	    irb->esw.esw0.erw.pvrf && cdev->private->options.pgroup) 
+		cdev->private->flags.doverify = 1;
+}
+
+/*
+ * This function accumulates the status into the private devstat and
+ * starts a basic sense if one is needed.
+ */
+int
+ccw_device_accumulate_and_sense(struct ccw_device *cdev, struct irb *irb)
+{
+	ccw_device_accumulate_irb(cdev, irb);
+	if ((irb->scsw.actl  & (SCSW_ACTL_DEVACT | SCSW_ACTL_SCHACT)) != 0)
+		return -EBUSY;
+	/* Check for basic sense. */
+	if (cdev->private->flags.dosense &&
+	    !(irb->scsw.dstat & DEV_STAT_UNIT_CHECK)) {
+		cdev->private->irb.esw.esw0.erw.cons = 1;
+		cdev->private->flags.dosense = 0;
+		return 0;
+	}
+	if (cdev->private->flags.dosense) {
+		ccw_device_do_sense(cdev, irb);
+		return -EBUSY;
+	}
+	return 0;
+}
+

drivers/s390/cio/ioasm.h

@@ -0,0 +1,228 @@
+#ifndef S390_CIO_IOASM_H
+#define S390_CIO_IOASM_H
+
+/*
+ * TPI info structure
+ */
+struct tpi_info {
+	__u32 reserved1	 : 16;	 /* reserved 0x00000001 */
+	__u32 irq	 : 16;	 /* aka. subchannel number */
+	__u32 intparm;		 /* interruption parameter */
+	__u32 adapter_IO : 1;
+	__u32 reserved2	 : 1;
+	__u32 isc	 : 3;
+	__u32 reserved3	 : 12;
+	__u32 int_type	 : 3;
+	__u32 reserved4	 : 12;
+} __attribute__ ((packed));
+
+
+/*
+ * Some S390 specific IO instructions as inline
+ */
+
+extern __inline__ int stsch(int irq, volatile struct schib *addr)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   stsch 0(%2)\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000), "a" (addr)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int msch(int irq, volatile struct schib *addr)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   msch  0(%2)\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L), "a" (addr)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int msch_err(int irq, volatile struct schib *addr)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"    lhi  %0,%3\n"
+		"    lr	  1,%1\n"
+		"    msch 0(%2)\n"
+		"0:  ipm  %0\n"
+		"    srl  %0,28\n"
+		"1:\n"
+#ifdef CONFIG_ARCH_S390X
+		".section __ex_table,\"a\"\n"
+		"   .align 8\n"
+		"   .quad 0b,1b\n"
+		".previous"
+#else
+		".section __ex_table,\"a\"\n"
+		"   .align 4\n"
+		"   .long 0b,1b\n"
+		".previous"
+#endif
+		: "=&d" (ccode)
+		: "d" (irq | 0x10000L), "a" (addr), "K" (-EIO)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int tsch(int irq, volatile struct irb *addr)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   tsch  0(%2)\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L), "a" (addr)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int tpi( volatile struct tpi_info *addr)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   tpi	  0(%1)\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "a" (addr)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int ssch(int irq, volatile struct orb *addr)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   ssch  0(%2)\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L), "a" (addr)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int rsch(int irq)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   rsch\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int csch(int irq)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   csch\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int hsch(int irq)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   hsch\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int xsch(int irq)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   .insn rre,0xb2760000,%1,0\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (irq | 0x10000L)
+		: "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int chsc(void *chsc_area)
+{
+	int cc;
+
+	__asm__ __volatile__ (
+		".insn	rre,0xb25f0000,%1,0	\n\t"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		: "=d" (cc)
+		: "d" (chsc_area)
+		: "cc" );
+
+	return cc;
+}
+
+extern __inline__ int iac( void)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   iac	  1\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode) : : "cc", "1" );
+	return ccode;
+}
+
+extern __inline__ int rchp(int chpid)
+{
+	int ccode;
+
+	__asm__ __volatile__(
+		"   lr	  1,%1\n"
+		"   rchp\n"
+		"   ipm	  %0\n"
+		"   srl	  %0,28"
+		: "=d" (ccode)
+		: "d" (chpid)
+		: "cc", "1" );
+	return ccode;
+}
+
+#endif

drivers/s390/cio/qdio.h

@@ -0,0 +1,648 @@
+#ifndef _CIO_QDIO_H
+#define _CIO_QDIO_H
+
+#define VERSION_CIO_QDIO_H "$Revision: 1.26 $"
+
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_VERBOSE_LEVEL 9
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_VERBOSE_LEVEL 5
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_USE_PROCESSING_STATE
+
+#ifdef CONFIG_QDIO_PERF_STATS
+#define QDIO_PERFORMANCE_STATS
+#endif /* CONFIG_QDIO_PERF_STATS */
+
+#define QDIO_MINIMAL_BH_RELIEF_TIME 16
+#define QDIO_TIMER_POLL_VALUE 1
+#define IQDIO_TIMER_POLL_VALUE 1
+
+/*
+ * unfortunately this can't be (QDIO_MAX_BUFFERS_PER_Q*4/3) or so -- as
+ * we never know, whether we'll get initiative again, e.g. to give the
+ * transmit skb's back to the stack, however the stack may be waiting for
+ * them... therefore we define 4 as threshold to start polling (which
+ * will stop as soon as the asynchronous queue catches up)
+ * btw, this only applies to the asynchronous HiperSockets queue
+ */
+#define IQDIO_FILL_LEVEL_TO_POLL 4
+
+#define TIQDIO_THININT_ISC 3
+#define TIQDIO_DELAY_TARGET 0
+#define QDIO_BUSY_BIT_PATIENCE 100 /* in microsecs */
+#define QDIO_BUSY_BIT_GIVE_UP 10000000 /* 10 seconds */
+#define IQDIO_GLOBAL_LAPS 2 /* GLOBAL_LAPS are not used as we */
+#define IQDIO_GLOBAL_LAPS_INT 1 /* don't global summary */
+#define IQDIO_LOCAL_LAPS 4
+#define IQDIO_LOCAL_LAPS_INT 1
+#define IQDIO_GLOBAL_SUMMARY_CC_MASK 2
+/*#define IQDIO_IQDC_INT_PARM 0x1234*/
+
+#define QDIO_Q_LAPS 5
+
+#define QDIO_STORAGE_KEY 0
+
+#define L2_CACHELINE_SIZE 256
+#define INDICATORS_PER_CACHELINE (L2_CACHELINE_SIZE/sizeof(__u32))
+
+#define QDIO_PERF "qdio_perf"
+
+/* must be a power of 2 */
+/*#define QDIO_STATS_NUMBER 4
+
+#define QDIO_STATS_CLASSES 2
+#define QDIO_STATS_COUNT_NEEDED 2*/
+
+#define QDIO_NO_USE_COUNT_TIMEOUT (1*HZ) /* wait for 1 sec on each q before
+					    exiting without having use_count
+					    of the queue to 0 */
+
+#define QDIO_ESTABLISH_TIMEOUT (1*HZ)
+#define QDIO_ACTIVATE_TIMEOUT ((5*HZ)>>10)
+#define QDIO_CLEANUP_CLEAR_TIMEOUT (20*HZ)
+#define QDIO_CLEANUP_HALT_TIMEOUT (10*HZ)
+
+enum qdio_irq_states {
+	QDIO_IRQ_STATE_INACTIVE,
+	QDIO_IRQ_STATE_ESTABLISHED,
+	QDIO_IRQ_STATE_ACTIVE,
+	QDIO_IRQ_STATE_STOPPED,
+	QDIO_IRQ_STATE_CLEANUP,
+	QDIO_IRQ_STATE_ERR,
+	NR_QDIO_IRQ_STATES,
+};
+
+/* used as intparm in do_IO: */
+#define QDIO_DOING_SENSEID 0
+#define QDIO_DOING_ESTABLISH 1
+#define QDIO_DOING_ACTIVATE 2
+#define QDIO_DOING_CLEANUP 3
+
+/************************* DEBUG FACILITY STUFF *********************/
+
+#define QDIO_DBF_HEX(ex,name,level,addr,len) \
+	do { \
+	if (ex) \
+		debug_exception(qdio_dbf_##name,level,(void*)(addr),len); \
+	else \
+		debug_event(qdio_dbf_##name,level,(void*)(addr),len); \
+	} while (0)
+#define QDIO_DBF_TEXT(ex,name,level,text) \
+	do { \
+	if (ex) \
+		debug_text_exception(qdio_dbf_##name,level,text); \
+	else \
+		debug_text_event(qdio_dbf_##name,level,text); \
+	} while (0)
+
+
+#define QDIO_DBF_HEX0(ex,name,addr,len) QDIO_DBF_HEX(ex,name,0,addr,len)
+#define QDIO_DBF_HEX1(ex,name,addr,len) QDIO_DBF_HEX(ex,name,1,addr,len)
+#define QDIO_DBF_HEX2(ex,name,addr,len) QDIO_DBF_HEX(ex,name,2,addr,len)
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_DBF_HEX3(ex,name,addr,len) QDIO_DBF_HEX(ex,name,3,addr,len)
+#define QDIO_DBF_HEX4(ex,name,addr,len) QDIO_DBF_HEX(ex,name,4,addr,len)
+#define QDIO_DBF_HEX5(ex,name,addr,len) QDIO_DBF_HEX(ex,name,5,addr,len)
+#define QDIO_DBF_HEX6(ex,name,addr,len) QDIO_DBF_HEX(ex,name,6,addr,len)
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_DBF_HEX3(ex,name,addr,len) do {} while (0)
+#define QDIO_DBF_HEX4(ex,name,addr,len) do {} while (0)
+#define QDIO_DBF_HEX5(ex,name,addr,len) do {} while (0)
+#define QDIO_DBF_HEX6(ex,name,addr,len) do {} while (0)
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_DBF_TEXT0(ex,name,text) QDIO_DBF_TEXT(ex,name,0,text)
+#define QDIO_DBF_TEXT1(ex,name,text) QDIO_DBF_TEXT(ex,name,1,text)
+#define QDIO_DBF_TEXT2(ex,name,text) QDIO_DBF_TEXT(ex,name,2,text)
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_DBF_TEXT3(ex,name,text) QDIO_DBF_TEXT(ex,name,3,text)
+#define QDIO_DBF_TEXT4(ex,name,text) QDIO_DBF_TEXT(ex,name,4,text)
+#define QDIO_DBF_TEXT5(ex,name,text) QDIO_DBF_TEXT(ex,name,5,text)
+#define QDIO_DBF_TEXT6(ex,name,text) QDIO_DBF_TEXT(ex,name,6,text)
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_DBF_TEXT3(ex,name,text) do {} while (0)
+#define QDIO_DBF_TEXT4(ex,name,text) do {} while (0)
+#define QDIO_DBF_TEXT5(ex,name,text) do {} while (0)
+#define QDIO_DBF_TEXT6(ex,name,text) do {} while (0)
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_DBF_SETUP_NAME "qdio_setup"
+#define QDIO_DBF_SETUP_LEN 8
+#define QDIO_DBF_SETUP_INDEX 2
+#define QDIO_DBF_SETUP_NR_AREAS 1
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_DBF_SETUP_LEVEL 6
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_DBF_SETUP_LEVEL 2
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_DBF_SBAL_NAME "qdio_labs" /* sbal */
+#define QDIO_DBF_SBAL_LEN 256
+#define QDIO_DBF_SBAL_INDEX 2
+#define QDIO_DBF_SBAL_NR_AREAS 2
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_DBF_SBAL_LEVEL 6
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_DBF_SBAL_LEVEL 2
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_DBF_TRACE_NAME "qdio_trace"
+#define QDIO_DBF_TRACE_LEN 8
+#define QDIO_DBF_TRACE_NR_AREAS 2
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_DBF_TRACE_INDEX 4
+#define QDIO_DBF_TRACE_LEVEL 4 /* -------- could be even more verbose here */
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_DBF_TRACE_INDEX 2
+#define QDIO_DBF_TRACE_LEVEL 2
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_DBF_SENSE_NAME "qdio_sense"
+#define QDIO_DBF_SENSE_LEN 64
+#define QDIO_DBF_SENSE_INDEX 1
+#define QDIO_DBF_SENSE_NR_AREAS 1
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_DBF_SENSE_LEVEL 6
+#else /* CONFIG_QDIO_DEBUG */
+#define QDIO_DBF_SENSE_LEVEL 2
+#endif /* CONFIG_QDIO_DEBUG */
+
+#ifdef CONFIG_QDIO_DEBUG
+#define QDIO_TRACE_QTYPE QDIO_ZFCP_QFMT
+
+#define QDIO_DBF_SLSB_OUT_NAME "qdio_slsb_out"
+#define QDIO_DBF_SLSB_OUT_LEN QDIO_MAX_BUFFERS_PER_Q
+#define QDIO_DBF_SLSB_OUT_INDEX 8
+#define QDIO_DBF_SLSB_OUT_NR_AREAS 1
+#define QDIO_DBF_SLSB_OUT_LEVEL 6
+
+#define QDIO_DBF_SLSB_IN_NAME "qdio_slsb_in"
+#define QDIO_DBF_SLSB_IN_LEN QDIO_MAX_BUFFERS_PER_Q
+#define QDIO_DBF_SLSB_IN_INDEX 8
+#define QDIO_DBF_SLSB_IN_NR_AREAS 1
+#define QDIO_DBF_SLSB_IN_LEVEL 6
+#endif /* CONFIG_QDIO_DEBUG */
+
+#define QDIO_PRINTK_HEADER QDIO_NAME ": "
+
+#if QDIO_VERBOSE_LEVEL>8
+#define QDIO_PRINT_STUPID(x...) printk( KERN_DEBUG QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_STUPID(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>7
+#define QDIO_PRINT_ALL(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_ALL(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>6
+#define QDIO_PRINT_INFO(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_INFO(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>5
+#define QDIO_PRINT_WARN(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_WARN(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>4
+#define QDIO_PRINT_ERR(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_ERR(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>3
+#define QDIO_PRINT_CRIT(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_CRIT(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>2
+#define QDIO_PRINT_ALERT(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_ALERT(x...)
+#endif
+
+#if QDIO_VERBOSE_LEVEL>1
+#define QDIO_PRINT_EMERG(x...) printk( QDIO_PRINTK_HEADER x)
+#else
+#define QDIO_PRINT_EMERG(x...)
+#endif
+
+#define HEXDUMP16(importance,header,ptr) \
+QDIO_PRINT_##importance(header "%02x %02x %02x %02x  " \
+			"%02x %02x %02x %02x  %02x %02x %02x %02x  " \
+			"%02x %02x %02x %02x\n",*(((char*)ptr)), \
+			*(((char*)ptr)+1),*(((char*)ptr)+2), \
+			*(((char*)ptr)+3),*(((char*)ptr)+4), \
+			*(((char*)ptr)+5),*(((char*)ptr)+6), \
+			*(((char*)ptr)+7),*(((char*)ptr)+8), \
+			*(((char*)ptr)+9),*(((char*)ptr)+10), \
+			*(((char*)ptr)+11),*(((char*)ptr)+12), \
+			*(((char*)ptr)+13),*(((char*)ptr)+14), \
+			*(((char*)ptr)+15)); \
+QDIO_PRINT_##importance(header "%02x %02x %02x %02x  %02x %02x %02x %02x  " \
+			"%02x %02x %02x %02x  %02x %02x %02x %02x\n", \
+			*(((char*)ptr)+16),*(((char*)ptr)+17), \
+			*(((char*)ptr)+18),*(((char*)ptr)+19), \
+			*(((char*)ptr)+20),*(((char*)ptr)+21), \
+			*(((char*)ptr)+22),*(((char*)ptr)+23), \
+			*(((char*)ptr)+24),*(((char*)ptr)+25), \
+			*(((char*)ptr)+26),*(((char*)ptr)+27), \
+			*(((char*)ptr)+28),*(((char*)ptr)+29), \
+			*(((char*)ptr)+30),*(((char*)ptr)+31));
+
+/****************** END OF DEBUG FACILITY STUFF *********************/
+
+/*
+ * Some instructions as assembly
+ */
+extern __inline__ int 
+do_siga_sync(unsigned int irq, unsigned int mask1, unsigned int mask2)
+{
+	int cc;
+
+#ifndef CONFIG_ARCH_S390X
+	asm volatile (
+		"lhi	0,2	\n\t"
+		"lr	1,%1	\n\t"
+		"lr	2,%2	\n\t"
+		"lr	3,%3	\n\t"
+		"siga   0	\n\t"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		: "=d" (cc)
+		: "d" (0x10000|irq), "d" (mask1), "d" (mask2)
+		: "cc", "0", "1", "2", "3"
+		);
+#else /* CONFIG_ARCH_S390X */
+	asm volatile (
+		"lghi	0,2	\n\t"
+		"llgfr	1,%1	\n\t"
+		"llgfr	2,%2	\n\t"
+		"llgfr	3,%3	\n\t"
+		"siga   0	\n\t"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		: "=d" (cc)
+		: "d" (0x10000|irq), "d" (mask1), "d" (mask2)
+		: "cc", "0", "1", "2", "3"
+		);
+#endif /* CONFIG_ARCH_S390X */
+	return cc;
+}
+
+extern __inline__ int
+do_siga_input(unsigned int irq, unsigned int mask)
+{
+	int cc;
+
+#ifndef CONFIG_ARCH_S390X
+	asm volatile (
+		"lhi	0,1	\n\t"
+		"lr	1,%1	\n\t"
+		"lr	2,%2	\n\t"
+		"siga   0	\n\t"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		: "=d" (cc)
+		: "d" (0x10000|irq), "d" (mask)
+		: "cc", "0", "1", "2", "memory"
+		);
+#else /* CONFIG_ARCH_S390X */
+	asm volatile (
+		"lghi	0,1	\n\t"
+		"llgfr	1,%1	\n\t"
+		"llgfr	2,%2	\n\t"
+		"siga   0	\n\t"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		: "=d" (cc)
+		: "d" (0x10000|irq), "d" (mask)
+		: "cc", "0", "1", "2", "memory"
+		);
+#endif /* CONFIG_ARCH_S390X */
+	
+	return cc;
+}
+
+extern __inline__ int
+do_siga_output(unsigned long irq, unsigned long mask, __u32 *bb)
+{
+	int cc;
+	__u32 busy_bit;
+
+#ifndef CONFIG_ARCH_S390X
+	asm volatile (
+		"lhi	0,0	\n\t"
+		"lr	1,%2	\n\t"
+		"lr	2,%3	\n\t"
+		"siga	0	\n\t"
+		"0:"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		"srl	0,31	\n\t"
+		"lr	%1,0	\n\t"
+		"1:	\n\t"
+		".section .fixup,\"ax\"\n\t"
+		"2:	\n\t"
+		"lhi	%0,%4	\n\t"
+		"bras	1,3f	\n\t"
+		".long 1b	\n\t"
+		"3:	\n\t"
+		"l	1,0(1)	\n\t"
+		"br	1	\n\t"
+		".previous	\n\t"
+		".section __ex_table,\"a\"\n\t"
+		".align 4	\n\t"
+		".long	0b,2b	\n\t"
+		".previous	\n\t"
+		: "=d" (cc), "=d" (busy_bit)
+		: "d" (0x10000|irq), "d" (mask),
+		"i" (QDIO_SIGA_ERROR_ACCESS_EXCEPTION)
+		: "cc", "0", "1", "2", "memory"
+		);
+#else /* CONFIG_ARCH_S390X */
+	asm volatile (
+		"lghi	0,0	\n\t"
+		"llgfr	1,%2	\n\t"
+		"llgfr	2,%3	\n\t"
+		"siga	0	\n\t"
+		"0:"
+		"ipm	%0	\n\t"
+		"srl	%0,28	\n\t"
+		"srl	0,31	\n\t"
+		"llgfr	%1,0	\n\t"
+		"1:	\n\t"
+		".section .fixup,\"ax\"\n\t"
+		"lghi	%0,%4	\n\t"
+		"jg	1b	\n\t"
+		".previous\n\t"
+		".section __ex_table,\"a\"\n\t"
+		".align 8	\n\t"
+		".quad	0b,1b	\n\t"
+		".previous	\n\t"
+		: "=d" (cc), "=d" (busy_bit)
+		: "d" (0x10000|irq), "d" (mask),
+		"i" (QDIO_SIGA_ERROR_ACCESS_EXCEPTION)
+		: "cc", "0", "1", "2", "memory"
+		);
+#endif /* CONFIG_ARCH_S390X */
+	
+	(*bb) = busy_bit;
+	return cc;
+}
+
+extern __inline__ unsigned long
+do_clear_global_summary(void)
+{
+
+	unsigned long time;
+
+#ifndef CONFIG_ARCH_S390X
+	asm volatile (
+		"lhi	1,3	\n\t"
+		".insn	rre,0xb2650000,2,0	\n\t"
+		"lr	%0,3	\n\t"
+		: "=d" (time) : : "cc", "1", "2", "3"
+		);
+#else /* CONFIG_ARCH_S390X */
+	asm volatile (
+		"lghi	1,3	\n\t"
+		".insn	rre,0xb2650000,2,0	\n\t"
+		"lgr	%0,3	\n\t"
+		: "=d" (time) : : "cc", "1", "2", "3"
+		);
+#endif /* CONFIG_ARCH_S390X */
+	
+	return time;
+}
+	
+/*
+ * QDIO device commands returned by extended Sense-ID
+ */
+#define DEFAULT_ESTABLISH_QS_CMD 0x1b
+#define DEFAULT_ESTABLISH_QS_COUNT 0x1000
+#define DEFAULT_ACTIVATE_QS_CMD 0x1f
+#define DEFAULT_ACTIVATE_QS_COUNT 0
+
+/*
+ * additional CIWs returned by extended Sense-ID
+ */
+#define CIW_TYPE_EQUEUE 0x3       /* establish QDIO queues */
+#define CIW_TYPE_AQUEUE 0x4       /* activate QDIO queues */
+
+#define QDIO_CHSC_RESPONSE_CODE_OK 1
+/* flags for st qdio sch data */
+#define CHSC_FLAG_QDIO_CAPABILITY 0x80
+#define CHSC_FLAG_VALIDITY 0x40
+
+#define CHSC_FLAG_SIGA_INPUT_NECESSARY 0x40
+#define CHSC_FLAG_SIGA_OUTPUT_NECESSARY 0x20
+#define CHSC_FLAG_SIGA_SYNC_NECESSARY 0x10
+#define CHSC_FLAG_SIGA_SYNC_DONE_ON_THININTS 0x08
+#define CHSC_FLAG_SIGA_SYNC_DONE_ON_OUTB_PCIS 0x04
+
+#ifdef QDIO_PERFORMANCE_STATS
+struct qdio_perf_stats {
+	unsigned int tl_runs;
+
+	unsigned int siga_outs;
+	unsigned int siga_ins;
+	unsigned int siga_syncs;
+	unsigned int pcis;
+	unsigned int thinints;
+	unsigned int fast_reqs;
+
+	__u64 start_time_outbound;
+	unsigned int outbound_cnt;
+	unsigned int outbound_time;
+	__u64 start_time_inbound;
+	unsigned int inbound_cnt;
+	unsigned int inbound_time;
+};
+#endif /* QDIO_PERFORMANCE_STATS */
+
+#define atomic_swap(a,b) xchg((int*)a.counter,b)
+
+/* unlikely as the later the better */
+#define SYNC_MEMORY if (unlikely(q->siga_sync)) qdio_siga_sync_q(q)
+#define SYNC_MEMORY_ALL if (unlikely(q->siga_sync)) \
+	qdio_siga_sync(q,~0U,~0U)
+#define SYNC_MEMORY_ALL_OUTB if (unlikely(q->siga_sync)) \
+	qdio_siga_sync(q,~0U,0)
+
+#define NOW qdio_get_micros()
+#define SAVE_TIMESTAMP(q) q->timing.last_transfer_time=NOW
+#define GET_SAVED_TIMESTAMP(q) (q->timing.last_transfer_time)
+#define SAVE_FRONTIER(q,val) q->last_move_ftc=val
+#define GET_SAVED_FRONTIER(q) (q->last_move_ftc)
+
+#define MY_MODULE_STRING(x) #x
+
+#ifdef CONFIG_ARCH_S390X
+#define QDIO_GET_ADDR(x) ((__u32)(unsigned long)x)
+#else /* CONFIG_ARCH_S390X */
+#define QDIO_GET_ADDR(x) ((__u32)(long)x)
+#endif /* CONFIG_ARCH_S390X */
+
+#ifdef CONFIG_QDIO_DEBUG
+#define set_slsb(x,y) \
+  if(q->queue_type==QDIO_TRACE_QTYPE) { \
+        if(q->is_input_q) { \
+            QDIO_DBF_HEX2(0,slsb_in,&q->slsb,QDIO_MAX_BUFFERS_PER_Q); \
+        } else { \
+            QDIO_DBF_HEX2(0,slsb_out,&q->slsb,QDIO_MAX_BUFFERS_PER_Q); \
+        } \
+  } \
+  qdio_set_slsb(x,y); \
+  if(q->queue_type==QDIO_TRACE_QTYPE) { \
+        if(q->is_input_q) { \
+            QDIO_DBF_HEX2(0,slsb_in,&q->slsb,QDIO_MAX_BUFFERS_PER_Q); \
+        } else { \
+            QDIO_DBF_HEX2(0,slsb_out,&q->slsb,QDIO_MAX_BUFFERS_PER_Q); \
+        } \
+  }
+#else /* CONFIG_QDIO_DEBUG */
+#define set_slsb(x,y) qdio_set_slsb(x,y)
+#endif /* CONFIG_QDIO_DEBUG */
+
+struct qdio_q {
+	volatile struct slsb slsb;
+
+	char unused[QDIO_MAX_BUFFERS_PER_Q];
+
+	__u32 * volatile dev_st_chg_ind;
+
+	int is_input_q;
+	int irq;
+	struct ccw_device *cdev;
+
+	unsigned int is_iqdio_q;
+	unsigned int is_thinint_q;
+
+	/* bit 0 means queue 0, bit 1 means queue 1, ... */
+	unsigned int mask;
+	unsigned int q_no;
+
+	qdio_handler_t (*handler);
+
+	/* points to the next buffer to be checked for having
+	 * been processed by the card (outbound)
+	 * or to the next buffer the program should check for (inbound) */
+	volatile int first_to_check;
+	/* and the last time it was: */
+	volatile int last_move_ftc;
+
+	atomic_t number_of_buffers_used;
+	atomic_t polling;
+
+	unsigned int siga_in;
+	unsigned int siga_out;
+	unsigned int siga_sync;
+	unsigned int siga_sync_done_on_thinints;
+	unsigned int siga_sync_done_on_outb_tis;
+	unsigned int hydra_gives_outbound_pcis;
+
+	/* used to save beginning position when calling dd_handlers */
+	int first_element_to_kick;
+
+	atomic_t use_count;
+	atomic_t is_in_shutdown;
+
+	void *irq_ptr;
+
+#ifdef QDIO_USE_TIMERS_FOR_POLLING
+	struct timer_list timer;
+	atomic_t timer_already_set;
+	spinlock_t timer_lock;
+#else /* QDIO_USE_TIMERS_FOR_POLLING */
+	struct tasklet_struct tasklet;
+#endif /* QDIO_USE_TIMERS_FOR_POLLING */
+
+	enum qdio_irq_states state;
+
+	/* used to store the error condition during a data transfer */
+	unsigned int qdio_error;
+	unsigned int siga_error;
+	unsigned int error_status_flags;
+
+	/* list of interesting queues */
+	volatile struct qdio_q *list_next;
+	volatile struct qdio_q *list_prev;
+
+	struct sl *sl;
+	volatile struct sbal *sbal[QDIO_MAX_BUFFERS_PER_Q];
+
+	struct qdio_buffer *qdio_buffers[QDIO_MAX_BUFFERS_PER_Q];
+
+	unsigned long int_parm;
+
+	/*struct {
+		int in_bh_check_limit;
+		int threshold;
+	} threshold_classes[QDIO_STATS_CLASSES];*/
+
+	struct {
+		/* inbound: the time to stop polling
+		   outbound: the time to kick peer */
+		int threshold; /* the real value */
+
+		/* outbound: last time of do_QDIO
+		   inbound: last time of noticing incoming data */
+		/*__u64 last_transfer_times[QDIO_STATS_NUMBER];
+		int last_transfer_index; */
+
+		__u64 last_transfer_time;
+		__u64 busy_start;
+	} timing;
+	atomic_t busy_siga_counter;
+        unsigned int queue_type;
+
+	/* leave this member at the end. won't be cleared in qdio_fill_qs */
+	struct slib *slib; /* a page is allocated under this pointer,
+			      sl points into this page, offset PAGE_SIZE/2
+			      (after slib) */
+} __attribute__ ((aligned(256)));
+
+struct qdio_irq {
+	__u32 * volatile dev_st_chg_ind;
+
+	unsigned long int_parm;
+	int irq;
+
+	unsigned int is_iqdio_irq;
+	unsigned int is_thinint_irq;
+	unsigned int hydra_gives_outbound_pcis;
+	unsigned int sync_done_on_outb_pcis;
+
+	enum qdio_irq_states state;
+
+	unsigned int no_input_qs;
+	unsigned int no_output_qs;
+
+	unsigned char qdioac;
+
+	struct ccw1 ccw;
+
+	struct ciw equeue;
+	struct ciw aqueue;
+
+	struct qib qib;
+	
+ 	void (*original_int_handler) (struct ccw_device *,
+ 				      unsigned long, struct irb *);
+
+	/* leave these four members together at the end. won't be cleared in qdio_fill_irq */
+	struct qdr *qdr;
+	struct qdio_q *input_qs[QDIO_MAX_QUEUES_PER_IRQ];
+	struct qdio_q *output_qs[QDIO_MAX_QUEUES_PER_IRQ];
+	struct semaphore setting_up_sema;
+};
+#endif

drivers/s390/crypto/Makefile

@@ -0,0 +1,6 @@
+#
+# S/390 crypto devices
+#
+
+z90crypt-objs := z90main.o z90hardware.o
+obj-$(CONFIG_Z90CRYPT) += z90crypt.o

drivers/s390/crypto/z90common.h

@@ -0,0 +1,168 @@
+/*
+ *  linux/drivers/s390/crypto/z90common.h
+ *
+ *  z90crypt 1.3.2
+ *
+ *  Copyright (C)  2001, 2004 IBM Corporation
+ *  Author(s): Robert Burroughs (burrough@us.ibm.com)
+ *             Eric Rossman (edrossma@us.ibm.com)
+ *
+ *  Hotplug & misc device support: Jochen Roehrig (roehrig@de.ibm.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _Z90COMMON_H_
+#define _Z90COMMON_H_
+
+#define VERSION_Z90COMMON_H "$Revision: 1.16 $"
+
+
+#define RESPBUFFSIZE 256
+#define PCI_FUNC_KEY_DECRYPT 0x5044
+#define PCI_FUNC_KEY_ENCRYPT 0x504B
+extern int ext_bitlens;
+
+enum devstat {
+	DEV_GONE,
+	DEV_ONLINE,
+	DEV_QUEUE_FULL,
+	DEV_EMPTY,
+	DEV_NO_WORK,
+	DEV_BAD_MESSAGE,
+	DEV_TSQ_EXCEPTION,
+	DEV_RSQ_EXCEPTION,
+	DEV_SEN_EXCEPTION,
+	DEV_REC_EXCEPTION
+};
+
+enum hdstat {
+	HD_NOT_THERE,
+	HD_BUSY,
+	HD_DECONFIGURED,
+	HD_CHECKSTOPPED,
+	HD_ONLINE,
+	HD_TSQ_EXCEPTION
+};
+
+#define Z90C_NO_DEVICES		1
+#define Z90C_AMBIGUOUS_DOMAIN	2
+#define Z90C_INCORRECT_DOMAIN	3
+#define ENOTINIT		4
+
+#define SEN_BUSY	 7
+#define SEN_USER_ERROR	 8
+#define SEN_QUEUE_FULL	11
+#define SEN_NOT_AVAIL	16
+#define SEN_PAD_ERROR	17
+#define SEN_RETRY	18
+#define SEN_RELEASED	24
+
+#define REC_EMPTY	 4
+#define REC_BUSY	 6
+#define REC_OPERAND_INV	 8
+#define REC_OPERAND_SIZE 9
+#define REC_EVEN_MOD	10
+#define REC_NO_WORK	11
+#define REC_HARDWAR_ERR	12
+#define REC_NO_RESPONSE	13
+#define REC_RETRY_DEV	14
+#define REC_USER_GONE	15
+#define REC_BAD_MESSAGE	16
+#define REC_INVALID_PAD	17
+#define REC_USE_PCICA	18
+
+#define WRONG_DEVICE_TYPE 20
+
+#define REC_FATAL_ERROR 32
+#define SEN_FATAL_ERROR 33
+#define TSQ_FATAL_ERROR 34
+#define RSQ_FATAL_ERROR 35
+
+#define Z90CRYPT_NUM_TYPES	5
+#define PCICA		0
+#define PCICC		1
+#define PCIXCC_MCL2	2
+#define PCIXCC_MCL3	3
+#define CEX2C		4
+#define NILDEV		-1
+#define ANYDEV		-1
+#define PCIXCC_UNK	-2
+
+enum hdevice_type {
+	PCICC_HW  = 3,
+	PCICA_HW  = 4,
+	PCIXCC_HW = 5,
+	OTHER_HW  = 6,
+	CEX2C_HW  = 7
+};
+
+struct CPRBX {
+	unsigned short cprb_len;
+	unsigned char  cprb_ver_id;
+	unsigned char  pad_000[3];
+	unsigned char  func_id[2];
+	unsigned char  cprb_flags[4];
+	unsigned int   req_parml;
+	unsigned int   req_datal;
+	unsigned int   rpl_msgbl;
+	unsigned int   rpld_parml;
+	unsigned int   rpl_datal;
+	unsigned int   rpld_datal;
+	unsigned int   req_extbl;
+	unsigned char  pad_001[4];
+	unsigned int   rpld_extbl;
+	unsigned char  req_parmb[16];
+	unsigned char  req_datab[16];
+	unsigned char  rpl_parmb[16];
+	unsigned char  rpl_datab[16];
+	unsigned char  req_extb[16];
+	unsigned char  rpl_extb[16];
+	unsigned short ccp_rtcode;
+	unsigned short ccp_rscode;
+	unsigned int   mac_data_len;
+	unsigned char  logon_id[8];
+	unsigned char  mac_value[8];
+	unsigned char  mac_content_flgs;
+	unsigned char  pad_002;
+	unsigned short domain;
+	unsigned char  pad_003[12];
+	unsigned char  pad_004[36];
+};
+
+#ifndef DEV_NAME
+#define DEV_NAME	"z90crypt"
+#endif
+#define PRINTK(fmt, args...) \
+	printk(KERN_DEBUG DEV_NAME ": %s -> " fmt, __FUNCTION__ , ## args)
+#define PRINTKN(fmt, args...) \
+	printk(KERN_DEBUG DEV_NAME ": " fmt, ## args)
+#define PRINTKW(fmt, args...) \
+	printk(KERN_WARNING DEV_NAME ": %s -> " fmt, __FUNCTION__ , ## args)
+#define PRINTKC(fmt, args...) \
+	printk(KERN_CRIT DEV_NAME ": %s -> " fmt, __FUNCTION__ , ## args)
+
+#ifdef Z90CRYPT_DEBUG
+#define PDEBUG(fmt, args...) \
+	printk(KERN_DEBUG DEV_NAME ": %s -> " fmt, __FUNCTION__ , ## args)
+#else
+#define PDEBUG(fmt, args...) do {} while (0)
+#endif
+
+#define UMIN(a,b) ((a) < (b) ? (a) : (b))
+#define IS_EVEN(x) ((x) == (2 * ((x) / 2)))
+
+
+#endif

drivers/s390/crypto/z90crypt.h

@@ -0,0 +1,258 @@
+/*
+ *  linux/drivers/s390/crypto/z90crypt.h
+ *
+ *  z90crypt 1.3.2
+ *
+ *  Copyright (C)  2001, 2004 IBM Corporation
+ *  Author(s): Robert Burroughs (burrough@us.ibm.com)
+ *             Eric Rossman (edrossma@us.ibm.com)
+ *
+ *  Hotplug & misc device support: Jochen Roehrig (roehrig@de.ibm.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _Z90CRYPT_H_
+#define _Z90CRYPT_H_
+
+#include <linux/ioctl.h>
+
+#define VERSION_Z90CRYPT_H "$Revision: 1.11 $"
+
+#define z90crypt_VERSION 1
+#define z90crypt_RELEASE 3	// 2 = PCIXCC, 3 = rewrite for coding standards
+#define z90crypt_VARIANT 2	// 2 = added PCIXCC MCL3 and CEX2C support
+
+/**
+ * If we are not using the sparse checker, __user has no use.
+ */
+#ifdef __CHECKER__
+# define __user		__attribute__((noderef, address_space(1)))
+#else
+# define __user
+#endif
+
+/**
+ * struct ica_rsa_modexpo
+ *
+ * Requirements:
+ * - outputdatalength is at least as large as inputdatalength.
+ * - All key parts are right justified in their fields, padded on
+ *   the left with zeroes.
+ * - length(b_key) = inputdatalength
+ * - length(n_modulus) = inputdatalength
+ */
+struct ica_rsa_modexpo {
+	char __user *	inputdata;
+	unsigned int	inputdatalength;
+	char __user *	outputdata;
+	unsigned int	outputdatalength;
+	char __user *	b_key;
+	char __user *	n_modulus;
+};
+
+/**
+ * struct ica_rsa_modexpo_crt
+ *
+ * Requirements:
+ * - inputdatalength is even.
+ * - outputdatalength is at least as large as inputdatalength.
+ * - All key parts are right justified in their fields, padded on
+ *   the left with zeroes.
+ * - length(bp_key)	= inputdatalength/2 + 8
+ * - length(bq_key)	= inputdatalength/2
+ * - length(np_key)	= inputdatalength/2 + 8
+ * - length(nq_key)	= inputdatalength/2
+ * - length(u_mult_inv) = inputdatalength/2 + 8
+ */
+struct ica_rsa_modexpo_crt {
+	char __user *	inputdata;
+	unsigned int	inputdatalength;
+	char __user *	outputdata;
+	unsigned int	outputdatalength;
+	char __user *	bp_key;
+	char __user *	bq_key;
+	char __user *	np_prime;
+	char __user *	nq_prime;
+	char __user *	u_mult_inv;
+};
+
+#define Z90_IOCTL_MAGIC 'z'  // NOTE:  Need to allocate from linux folks
+
+/**
+ * Interface notes:
+ *
+ * The ioctl()s which are implemented (along with relevant details)
+ * are:
+ *
+ *   ICARSAMODEXPO
+ *     Perform an RSA operation using a Modulus-Exponent pair
+ *     This takes an ica_rsa_modexpo struct as its arg.
+ *
+ *     NOTE: please refer to the comments preceding this structure
+ *           for the implementation details for the contents of the
+ *           block
+ *
+ *   ICARSACRT
+ *     Perform an RSA operation using a Chinese-Remainder Theorem key
+ *     This takes an ica_rsa_modexpo_crt struct as its arg.
+ *
+ *     NOTE: please refer to the comments preceding this structure
+ *           for the implementation details for the contents of the
+ *           block
+ *
+ *   Z90STAT_TOTALCOUNT
+ *     Return an integer count of all device types together.
+ *
+ *   Z90STAT_PCICACOUNT
+ *     Return an integer count of all PCICAs.
+ *
+ *   Z90STAT_PCICCCOUNT
+ *     Return an integer count of all PCICCs.
+ *
+ *   Z90STAT_PCIXCCMCL2COUNT
+ *     Return an integer count of all MCL2 PCIXCCs.
+ *
+ *   Z90STAT_PCIXCCMCL3COUNT
+ *     Return an integer count of all MCL3 PCIXCCs.
+ *
+ *   Z90STAT_CEX2CCOUNT
+ *     Return an integer count of all CEX2Cs.
+ *
+ *   Z90STAT_REQUESTQ_COUNT
+ *     Return an integer count of the number of entries waiting to be
+ *     sent to a device.
+ *
+ *   Z90STAT_PENDINGQ_COUNT
+ *     Return an integer count of the number of entries sent to a
+ *     device awaiting the reply.
+ *
+ *   Z90STAT_TOTALOPEN_COUNT
+ *     Return an integer count of the number of open file handles.
+ *
+ *   Z90STAT_DOMAIN_INDEX
+ *     Return the integer value of the Cryptographic Domain.
+ *
+ *   Z90STAT_STATUS_MASK
+ *     Return an 64 element array of unsigned chars for the status of
+ *     all devices.
+ *       0x01: PCICA
+ *       0x02: PCICC
+ *       0x03: PCIXCC_MCL2
+ *       0x04: PCIXCC_MCL3
+ *       0x05: CEX2C
+ *       0x0d: device is disabled via the proc filesystem
+ *
+ *   Z90STAT_QDEPTH_MASK
+ *     Return an 64 element array of unsigned chars for the queue
+ *     depth of all devices.
+ *
+ *   Z90STAT_PERDEV_REQCNT
+ *     Return an 64 element array of unsigned integers for the number
+ *     of successfully completed requests per device since the device
+ *     was detected and made available.
+ *
+ *   ICAZ90STATUS (deprecated)
+ *     Return some device driver status in a ica_z90_status struct
+ *     This takes an ica_z90_status struct as its arg.
+ *
+ *     NOTE: this ioctl() is deprecated, and has been replaced with
+ *           single ioctl()s for each type of status being requested
+ *
+ *   Z90STAT_PCIXCCCOUNT (deprecated)
+ *     Return an integer count of all PCIXCCs (MCL2 + MCL3).
+ *     This is DEPRECATED now that MCL3 PCIXCCs are treated differently from
+ *     MCL2 PCIXCCs.
+ *
+ *   Z90QUIESCE (not recommended)
+ *     Quiesce the driver.  This is intended to stop all new
+ *     requests from being processed.  Its use is NOT recommended,
+ *     except in circumstances where there is no other way to stop
+ *     callers from accessing the driver.  Its original use was to
+ *     allow the driver to be "drained" of work in preparation for
+ *     a system shutdown.
+ *
+ *     NOTE: once issued, this ban on new work cannot be undone
+ *           except by unloading and reloading the driver.
+ */
+
+/**
+ * Supported ioctl calls
+ */
+#define ICARSAMODEXPO	_IOC(_IOC_READ|_IOC_WRITE, Z90_IOCTL_MAGIC, 0x05, 0)
+#define ICARSACRT	_IOC(_IOC_READ|_IOC_WRITE, Z90_IOCTL_MAGIC, 0x06, 0)
+
+/* DEPRECATED status calls (bound for removal at some point) */
+#define ICAZ90STATUS	_IOR(Z90_IOCTL_MAGIC, 0x10, struct ica_z90_status)
+#define Z90STAT_PCIXCCCOUNT	_IOR(Z90_IOCTL_MAGIC, 0x43, int)
+
+/* unrelated to ICA callers */
+#define Z90QUIESCE	_IO(Z90_IOCTL_MAGIC, 0x11)
+
+/* New status calls */
+#define Z90STAT_TOTALCOUNT	_IOR(Z90_IOCTL_MAGIC, 0x40, int)
+#define Z90STAT_PCICACOUNT	_IOR(Z90_IOCTL_MAGIC, 0x41, int)
+#define Z90STAT_PCICCCOUNT	_IOR(Z90_IOCTL_MAGIC, 0x42, int)
+#define Z90STAT_PCIXCCMCL2COUNT	_IOR(Z90_IOCTL_MAGIC, 0x4b, int)
+#define Z90STAT_PCIXCCMCL3COUNT	_IOR(Z90_IOCTL_MAGIC, 0x4c, int)
+#define Z90STAT_CEX2CCOUNT	_IOR(Z90_IOCTL_MAGIC, 0x4d, int)
+#define Z90STAT_REQUESTQ_COUNT	_IOR(Z90_IOCTL_MAGIC, 0x44, int)
+#define Z90STAT_PENDINGQ_COUNT	_IOR(Z90_IOCTL_MAGIC, 0x45, int)
+#define Z90STAT_TOTALOPEN_COUNT _IOR(Z90_IOCTL_MAGIC, 0x46, int)
+#define Z90STAT_DOMAIN_INDEX	_IOR(Z90_IOCTL_MAGIC, 0x47, int)
+#define Z90STAT_STATUS_MASK	_IOR(Z90_IOCTL_MAGIC, 0x48, char[64])
+#define Z90STAT_QDEPTH_MASK	_IOR(Z90_IOCTL_MAGIC, 0x49, char[64])
+#define Z90STAT_PERDEV_REQCNT	_IOR(Z90_IOCTL_MAGIC, 0x4a, int[64])
+
+/**
+ * local errno definitions
+ */
+#define ENOBUFF	  129	// filp->private_data->...>work_elem_p->buffer is NULL
+#define EWORKPEND 130	// user issues ioctl while another pending
+#define ERELEASED 131	// user released while ioctl pending
+#define EQUIESCE  132	// z90crypt quiescing (no more work allowed)
+#define ETIMEOUT  133	// request timed out
+#define EUNKNOWN  134	// some unrecognized error occured (retry may succeed)
+#define EGETBUFF  135	// Error getting buffer or hardware lacks capability
+			// (retry in software)
+
+/**
+ * DEPRECATED STRUCTURES
+ */
+
+/**
+ * This structure is DEPRECATED and the corresponding ioctl() has been
+ * replaced with individual ioctl()s for each piece of data!
+ * This structure will NOT survive past version 1.3.1, so switch to the
+ * new ioctl()s.
+ */
+#define MASK_LENGTH 64 // mask length
+struct ica_z90_status {
+	int totalcount;
+	int leedslitecount; // PCICA
+	int leeds2count;    // PCICC
+	// int PCIXCCCount; is not in struct for backward compatibility
+	int requestqWaitCount;
+	int pendingqWaitCount;
+	int totalOpenCount;
+	int cryptoDomain;
+	// status: 0=not there, 1=PCICA, 2=PCICC, 3=PCIXCC_MCL2, 4=PCIXCC_MCL3,
+	//         5=CEX2C
+	unsigned char status[MASK_LENGTH];
+	// qdepth: # work elements waiting for each device
+	unsigned char qdepth[MASK_LENGTH];
+};
+
+#endif /* _Z90CRYPT_H_ */

drivers/s390/ebcdic.c

@@ -0,0 +1,246 @@
+/*
+ *  arch/s390/kernel/ebcdic.c
+ *    ECBDIC -> ASCII, ASCII -> ECBDIC conversion tables.
+ *
+ *  S390 version
+ *    Copyright (C) 1998 IBM Corporation
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <asm/types.h>
+
+/*
+ * ASCII -> EBCDIC
+ */
+__u8 _ascebc[256] =
+{
+ /*00  NL    SH    SX    EX    ET    NQ    AK    BL */
+     0x00, 0x01, 0x02, 0x03, 0x37, 0x2D, 0x2E, 0x2F,
+ /*08  BS    HT    LF    VT    FF    CR    SO    SI */
+     0x16, 0x05, 0x15, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
+ /*10  DL    D1    D2    D3    D4    NK    SN    EB */
+     0x10, 0x11, 0x12, 0x13, 0x3C, 0x15, 0x32, 0x26,
+ /*18  CN    EM    SB    EC    FS    GS    RS    US */
+     0x18, 0x19, 0x3F, 0x27, 0x1C, 0x1D, 0x1E, 0x1F,
+ /*20  SP     !     "     #     $     %     &     ' */
+     0x40, 0x5A, 0x7F, 0x7B, 0x5B, 0x6C, 0x50, 0x7D,
+ /*28   (     )     *     +     ,     -    .      / */
+     0x4D, 0x5D, 0x5C, 0x4E, 0x6B, 0x60, 0x4B, 0x61,
+ /*30   0     1     2     3     4     5     6     7 */
+     0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7,
+ /*38   8     9     :     ;     <     =     >     ? */
+     0xF8, 0xF9, 0x7A, 0x5E, 0x4C, 0x7E, 0x6E, 0x6F,
+ /*40   @     A     B     C     D     E     F     G */
+     0x7C, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7,
+ /*48   H     I     J     K     L     M     N     O */
+     0xC8, 0xC9, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6,
+ /*50   P     Q     R     S     T     U     V     W */
+     0xD7, 0xD8, 0xD9, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6,
+ /*58   X     Y     Z     [     \     ]     ^     _ */
+     0xE7, 0xE8, 0xE9, 0xAD, 0xE0, 0xBD, 0x5F, 0x6D,
+ /*60   `     a     b     c     d     e     f     g */
+     0x79, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87,
+ /*68   h     i     j     k     l     m     n     o */
+     0x88, 0x89, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96,
+ /*70   p     q     r     s     t     u     v     w */
+     0x97, 0x98, 0x99, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6,
+ /*78   x     y     z     {     |     }     ~    DL */
+     0xA7, 0xA8, 0xA9, 0xC0, 0x4F, 0xD0, 0xA1, 0x07,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F,
+     0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0x3F, 0xFF
+};
+
+/*
+ * EBCDIC -> ASCII
+ */
+__u8 _ebcasc[256] =
+{
+ /* 0x00   NUL   SOH   STX   ETX  *SEL    HT  *RNL   DEL */
+          0x00, 0x01, 0x02, 0x03, 0x07, 0x09, 0x07, 0x7F,
+ /* 0x08   -GE  -SPS  -RPT    VT    FF    CR    SO    SI */
+          0x07, 0x07, 0x07, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
+ /* 0x10   DLE   DC1   DC2   DC3  -RES   -NL    BS  -POC
+                                  -ENP  ->LF             */
+          0x10, 0x11, 0x12, 0x13, 0x07, 0x0A, 0x08, 0x07,
+ /* 0x18   CAN    EM  -UBS  -CU1  -IFS  -IGS  -IRS  -ITB
+                                                    -IUS */
+          0x18, 0x19, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
+ /* 0x20   -DS  -SOS    FS  -WUS  -BYP    LF   ETB   ESC
+                                  -INP                   */
+          0x07, 0x07, 0x1C, 0x07, 0x07, 0x0A, 0x17, 0x1B,
+ /* 0x28   -SA  -SFE   -SM  -CSP  -MFA   ENQ   ACK   BEL
+                       -SW                               */ 
+          0x07, 0x07, 0x07, 0x07, 0x07, 0x05, 0x06, 0x07,
+ /* 0x30  ----  ----   SYN   -IR   -PP  -TRN  -NBS   EOT */
+          0x07, 0x07, 0x16, 0x07, 0x07, 0x07, 0x07, 0x04,
+ /* 0x38  -SBS   -IT  -RFF  -CU3   DC4   NAK  ----   SUB */
+          0x07, 0x07, 0x07, 0x07, 0x14, 0x15, 0x07, 0x1A,
+ /* 0x40    SP   RSP           ∽              ----       */
+          0x20, 0xFF, 0x83, 0x84, 0x85, 0xA0, 0x07, 0x86,
+ /* 0x48                       .     <     (     +     | */
+          0x87, 0xA4, 0x9B, 0x2E, 0x3C, 0x28, 0x2B, 0x7C,
+ /* 0x50     &                                      ---- */
+          0x26, 0x82, 0x88, 0x89, 0x8A, 0xA1, 0x8C, 0x07,
+ /* 0x58           ≡     !     $     *     )     ;       */
+          0x8D, 0xE1, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0xAA,
+ /* 0x60     -     /  ----     <20>  ----  ----  ----       */
+          0x2D, 0x2F, 0x07, 0x8E, 0x07, 0x07, 0x07, 0x8F,
+ /* 0x68              ----     ,     %     _     >     ? */ 
+          0x80, 0xA5, 0x07, 0x2C, 0x25, 0x5F, 0x3E, 0x3F,
+ /* 0x70  ----        ----  ----  ----  ----  ----  ---- */
+          0x07, 0x90, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
+ /* 0x78     *     `     :     #     @     '     =     " */
+          0x70, 0x60, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22,
+ /* 0x80     *     a     b     c     d     e     f     g */
+          0x07, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
+ /* 0x88     h     i              ----  ----  ----       */
+          0x68, 0x69, 0xAE, 0xAF, 0x07, 0x07, 0x07, 0xF1,
+ /* 0x90     <20>     j     k     l     m     n     o     p */
+          0xF8, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70,
+ /* 0x98     q     r                    ----        ---- */
+          0x71, 0x72, 0xA6, 0xA7, 0x91, 0x07, 0x92, 0x07,
+ /* 0xA0           ~     s     t     u     v     w     x */
+          0xE6, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78,
+ /* 0xA8     y     z              ----  ----  ----  ---- */
+          0x79, 0x7A, 0xAD, 0xAB, 0x07, 0x07, 0x07, 0x07,
+ /* 0xB0     ^                    ----     〒  ----       */
+          0x5E, 0x9C, 0x9D, 0xFA, 0x07, 0x07, 0x07, 0xAC,
+ /* 0xB8        ----     [     ]  ----  ----  ----  ---- */
+          0xAB, 0x07, 0x5B, 0x5D, 0x07, 0x07, 0x07, 0x07,
+ /* 0xC0     {     A     B     C     D     E     F     G */
+          0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,
+ /* 0xC8     H     I  ----           ‡              ---- */
+          0x48, 0x49, 0x07, 0x93, 0x94, 0x95, 0xA2, 0x07,
+ /* 0xD0     }     J     K     L     M     N     O     P */
+          0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50,
+ /* 0xD8     Q     R  ----           ◯                   */
+          0x51, 0x52, 0x07, 0x96, 0x81, 0x97, 0xA3, 0x98,
+ /* 0xE0     \           S     T     U     V     W     X */
+          0x5C, 0xF6, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58,
+ /* 0xE8     Y     Z        ----     <20>  ----  ----  ---- */
+          0x59, 0x5A, 0xFD, 0x07, 0x99, 0x07, 0x07, 0x07,
+ /* 0xF0     0     1     2     3     4     5     6     7 */
+          0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
+ /* 0xF8     8     9  ----  ----     ⌒  ----  ----  ---- */
+          0x38, 0x39, 0x07, 0x07, 0x9A, 0x07, 0x07, 0x07
+};
+
+/*
+ * EBCDIC (capitals) -> ASCII (small case)
+ */
+__u8 _ebcasc_reduce_case[256] =
+{
+ /* 0x00   NUL   SOH   STX   ETX  *SEL    HT  *RNL   DEL */
+          0x00, 0x01, 0x02, 0x03, 0x07, 0x09, 0x07, 0x7F,
+
+ /* 0x08   -GE  -SPS  -RPT    VT    FF    CR    SO    SI */
+          0x07, 0x07, 0x07, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
+
+ /* 0x10   DLE   DC1   DC2   DC3  -RES   -NL    BS  -POC
+                                  -ENP  ->LF             */
+          0x10, 0x11, 0x12, 0x13, 0x07, 0x0A, 0x08, 0x07,
+
+ /* 0x18   CAN    EM  -UBS  -CU1  -IFS  -IGS  -IRS  -ITB
+                                                    -IUS */
+          0x18, 0x19, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
+
+ /* 0x20   -DS  -SOS    FS  -WUS  -BYP    LF   ETB   ESC
+                                  -INP                   */
+          0x07, 0x07, 0x1C, 0x07, 0x07, 0x0A, 0x17, 0x1B,
+
+ /* 0x28   -SA  -SFE   -SM  -CSP  -MFA   ENQ   ACK   BEL
+                       -SW                               */ 
+          0x07, 0x07, 0x07, 0x07, 0x07, 0x05, 0x06, 0x07,
+
+ /* 0x30  ----  ----   SYN   -IR   -PP  -TRN  -NBS   EOT */
+          0x07, 0x07, 0x16, 0x07, 0x07, 0x07, 0x07, 0x04,
+
+ /* 0x38  -SBS   -IT  -RFF  -CU3   DC4   NAK  ----   SUB */
+          0x07, 0x07, 0x07, 0x07, 0x14, 0x15, 0x07, 0x1A,
+
+ /* 0x40    SP   RSP           ∽              ----       */
+          0x20, 0xFF, 0x83, 0x84, 0x85, 0xA0, 0x07, 0x86,
+
+ /* 0x48                       .     <     (     +     | */
+          0x87, 0xA4, 0x9B, 0x2E, 0x3C, 0x28, 0x2B, 0x7C,
+
+ /* 0x50     &                                      ---- */
+          0x26, 0x82, 0x88, 0x89, 0x8A, 0xA1, 0x8C, 0x07,
+
+ /* 0x58           ≡     !     $     *     )     ;       */
+          0x8D, 0xE1, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0xAA,
+
+ /* 0x60     -     /  ----     <20>  ----  ----  ----       */
+          0x2D, 0x2F, 0x07, 0x84, 0x07, 0x07, 0x07, 0x8F,
+
+ /* 0x68              ----     ,     %     _     >     ? */ 
+          0x80, 0xA5, 0x07, 0x2C, 0x25, 0x5F, 0x3E, 0x3F,
+
+ /* 0x70  ----        ----  ----  ----  ----  ----  ---- */
+          0x07, 0x90, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
+
+ /* 0x78     *     `     :     #     @     '     =     " */
+          0x70, 0x60, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22,
+
+ /* 0x80     *     a     b     c     d     e     f     g */
+          0x07, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
+
+ /* 0x88     h     i              ----  ----  ----       */
+          0x68, 0x69, 0xAE, 0xAF, 0x07, 0x07, 0x07, 0xF1,
+
+ /* 0x90     <20>     j     k     l     m     n     o     p */
+          0xF8, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70,
+
+ /* 0x98     q     r                    ----        ---- */
+          0x71, 0x72, 0xA6, 0xA7, 0x91, 0x07, 0x92, 0x07,
+
+ /* 0xA0           ~     s     t     u     v     w     x */
+          0xE6, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78,
+
+ /* 0xA8     y     z              ----  ----  ----  ---- */
+          0x79, 0x7A, 0xAD, 0xAB, 0x07, 0x07, 0x07, 0x07,
+
+ /* 0xB0     ^                    ----     〒  ----       */
+          0x5E, 0x9C, 0x9D, 0xFA, 0x07, 0x07, 0x07, 0xAC,
+
+ /* 0xB8        ----     [     ]  ----  ----  ----  ---- */
+          0xAB, 0x07, 0x5B, 0x5D, 0x07, 0x07, 0x07, 0x07,
+
+ /* 0xC0     {     A     B     C     D     E     F     G */
+          0x7B, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
+
+ /* 0xC8     H     I  ----           ‡              ---- */
+          0x68, 0x69, 0x07, 0x93, 0x94, 0x95, 0xA2, 0x07,
+
+ /* 0xD0     }     J     K     L     M     N     O     P */
+          0x7D, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70,
+
+ /* 0xD8     Q     R  ----           ◯                   */
+          0x71, 0x72, 0x07, 0x96, 0x81, 0x97, 0xA3, 0x98,
+
+ /* 0xE0     \           S     T     U     V     W     X */
+          0x5C, 0xF6, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78,
+
+ /* 0xE8     Y     Z        ----     <20>  ----  ----  ---- */
+          0x79, 0x7A, 0xFD, 0x07, 0x94, 0x07, 0x07, 0x07,
+
+ /* 0xF0     0     1     2     3     4     5     6     7 */
+          0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
+
+ /* 0xF8     8     9  ----  ----     ⌒  ----  ----  ---- */
+          0x38, 0x39, 0x07, 0x07, 0x81, 0x07, 0x07, 0x07
+};

drivers/s390/net/Kconfig

@@ -0,0 +1,108 @@
+menu "S/390 network device drivers"
+	depends on NETDEVICES && ARCH_S390
+
+config LCS
+	tristate "Lan Channel Station Interface"
+	depends on NETDEVICES && (NET_ETHERNET || TR || FDDI)
+	help
+	   Select this option if you want to use LCS networking  on IBM S/390
+  	   or zSeries. This device driver supports Token Ring (IEEE 802.5),
+  	   FDDI (IEEE 802.7) and Ethernet. 
+	   This option is also available as a module which will be
+	   called lcs.ko. If you do not know what it is, it's safe to say "Y".
+
+config CTC
+	tristate "CTC device support"
+	depends on NETDEVICES
+	help
+	  Select this option if you want to use channel-to-channel networking
+	  on IBM S/390 or zSeries. This device driver supports real CTC
+	  coupling using ESCON. It also supports virtual CTCs when running
+	  under VM. It will use the channel device configuration if this is
+	  available.  This option is also available as a module which will be
+	  called ctc.ko.  If you do not know what it is, it's safe to say "Y".
+
+config IUCV
+	tristate "IUCV support (VM only)"
+	help
+	  Select this option if you want to use inter-user communication
+	  under VM or VIF. If unsure, say "Y" to enable a fast communication
+	  link between VM guests. At boot time the user ID of the guest needs
+	  to be passed to the kernel. Note that both kernels need to be
+	  compiled with this option and both need to be booted with the user ID
+	  of the other VM guest.
+
+config NETIUCV
+	tristate "IUCV network device support (VM only)"
+	depends on IUCV && NETDEVICES
+	help
+	  Select this option if you want to use inter-user communication
+	  vehicle networking under VM or VIF. It enables a fast communication
+	  link between VM guests. Using ifconfig a point-to-point connection
+	  can be established to the Linux for zSeries and S7390 system
+	  running on the other VM guest. This option is also available
+	  as a module which will be called netiucv.ko. If unsure, say "Y".
+
+config SMSGIUCV
+	tristate "IUCV special message support (VM only)"
+	depends on IUCV
+	help
+	  Select this option if you want to be able to receive SMSG messages
+	  from other VM guest systems.
+
+config CLAW
+	tristate "CLAW device support"
+	depends on NETDEVICES
+	help
+	  This driver supports channel attached CLAW devices.
+	  CLAW is Common Link Access for Workstation.  Common devices
+          that use CLAW are RS/6000s, Cisco Routers (CIP) and 3172 devices.
+	  To compile as a module choose M here:  The module will be called
+	  claw.ko to compile into the kernel choose Y
+
+config QETH
+	tristate "Gigabit Ethernet device support"
+	depends on NETDEVICES && IP_MULTICAST && QDIO
+	help
+	  This driver supports the IBM S/390 and zSeries OSA Express adapters
+	  in QDIO mode (all media types), HiperSockets interfaces and VM GuestLAN
+	  interfaces in QDIO and HIPER mode.
+	
+	  For details please refer to the documentation provided by IBM at   
+	  <http://www10.software.ibm.com/developerworks/opensource/linux390>
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called qeth.ko.
+
+
+comment "Gigabit Ethernet default settings"
+	depends on QETH
+
+config QETH_IPV6
+	bool "IPv6 support for gigabit ethernet"
+	depends on (QETH = IPV6) || (QETH && IPV6 = 'y')
+	help
+	  If CONFIG_QETH is switched on, this option will include IPv6
+	  support in the qeth device driver.
+	
+config QETH_VLAN
+	bool "VLAN support for gigabit ethernet"
+	depends on (QETH = VLAN_8021Q) || (QETH && VLAN_8021Q = 'y')
+	help
+	  If CONFIG_QETH is switched on, this option will include IEEE
+	  802.1q VLAN support in the qeth device driver.
+
+config QETH_PERF_STATS
+	bool "Performance statistics in /proc"
+	depends on QETH
+	help
+	  When switched on, this option will add a file in the proc-fs
+	  (/proc/qeth_perf_stats) containing performance statistics. It
+	  may slightly impact performance, so this is only recommended for
+	  internal tuning of the device driver.
+
+config CCWGROUP
+ 	tristate
+	default (LCS || CTC || QETH)
+
+endmenu

drivers/s390/net/Makefile

@@ -0,0 +1,14 @@
+#
+# S/390 network devices
+#
+
+ctc-objs := ctcmain.o ctctty.o ctcdbug.o
+
+obj-$(CONFIG_IUCV) += iucv.o
+obj-$(CONFIG_NETIUCV) += netiucv.o fsm.o
+obj-$(CONFIG_SMSGIUCV) += smsgiucv.o
+obj-$(CONFIG_CTC) += ctc.o fsm.o cu3088.o
+obj-$(CONFIG_LCS) += lcs.o cu3088.o
+qeth-y := qeth_main.o qeth_mpc.o qeth_sys.o qeth_eddp.o qeth_tso.o
+qeth-$(CONFIG_PROC_FS) += qeth_proc.o
+obj-$(CONFIG_QETH) += qeth.o

drivers/s390/net/claw.h

@@ -0,0 +1,335 @@
+/*******************************************************
+*  Define constants                                    *
+*                                                      *
+********************************************************/
+#define VERSION_CLAW_H "$Revision: 1.6 $"
+/*-----------------------------------------------------*
+*     CCW command codes for CLAW protocol              *
+*------------------------------------------------------*/
+
+#define CCW_CLAW_CMD_WRITE           0x01      /* write - not including link */
+#define CCW_CLAW_CMD_READ            0x02      /* read */
+#define CCW_CLAW_CMD_NOP             0x03      /* NOP */
+#define CCW_CLAW_CMD_SENSE           0x04      /* Sense */
+#define CCW_CLAW_CMD_SIGNAL_SMOD     0x05      /* Signal Status Modifier */
+#define CCW_CLAW_CMD_TIC             0x08      /* TIC */
+#define CCW_CLAW_CMD_READHEADER      0x12      /* read header data */
+#define CCW_CLAW_CMD_READFF          0x22      /* read an FF */
+#define CCW_CLAW_CMD_SENSEID         0xe4      /* Sense ID */
+
+
+/*-----------------------------------------------------*
+*    CLAW Unique constants                             *
+*------------------------------------------------------*/
+
+#define MORE_to_COME_FLAG       0x04   /* OR with write CCW in case of m-t-c */
+#define CLAW_IDLE               0x00   /* flag to indicate CLAW is idle */
+#define CLAW_BUSY               0xff   /* flag to indicate CLAW is busy */
+#define CLAW_PENDING            0x00   /* flag to indicate i/o is pending */
+#define CLAW_COMPLETE           0xff   /* flag to indicate i/o completed */
+
+/*-----------------------------------------------------*
+*     CLAW control comand code                         *
+*------------------------------------------------------*/
+
+#define SYSTEM_VALIDATE_REQUEST   0x01  /* System Validate request */
+#define SYSTEM_VALIDATE_RESPONSE  0x02  /* System Validate response */
+#define CONNECTION_REQUEST        0x21  /* Connection request */
+#define CONNECTION_RESPONSE       0x22  /* Connection response */
+#define CONNECTION_CONFIRM        0x23  /* Connection confirm */
+#define DISCONNECT                0x24  /* Disconnect */
+#define CLAW_ERROR                0x41  /* CLAW error message */
+#define CLAW_VERSION_ID           2     /* CLAW version ID */
+
+/*-----------------------------------------------------*
+*  CLAW adater sense bytes                             *
+*------------------------------------------------------*/
+
+#define CLAW_ADAPTER_SENSE_BYTE 0x41   /* Stop command issued to adapter */
+
+/*-----------------------------------------------------*
+*      CLAW control command return codes               *
+*------------------------------------------------------*/
+
+#define CLAW_RC_NAME_MISMATCH       166  /*  names do not match */
+#define CLAW_RC_WRONG_VERSION       167  /*  wrong CLAW version number */
+#define CLAW_RC_HOST_RCV_TOO_SMALL  180  /*  Host maximum receive is   */
+					 /*  less than Linux on zSeries*/
+                                         /*  transmit size             */
+
+/*-----------------------------------------------------*
+*      CLAW Constants application name                 *
+*------------------------------------------------------*/
+
+#define HOST_APPL_NAME          "TCPIP   "
+#define WS_APPL_NAME_IP_LINK    "TCPIP   "
+#define WS_APPL_NAME_IP_NAME	"IP      "
+#define WS_APPL_NAME_API_LINK   "API     "
+#define WS_APPL_NAME_PACKED     "PACKED  "
+#define WS_NAME_NOT_DEF         "NOT_DEF "
+#define PACKING_ASK		1
+#define PACK_SEND		2
+#define DO_PACKED		3
+
+#define MAX_ENVELOPE_SIZE       65536
+#define CLAW_DEFAULT_MTU_SIZE   4096
+#define DEF_PACK_BUFSIZE	32768
+#define READ                    0
+#define WRITE                   1
+
+#define TB_TX                   0          /* sk buffer handling in process  */
+#define TB_STOP                 1          /* network device stop in process */
+#define TB_RETRY                2          /* retry in process               */
+#define TB_NOBUFFER             3          /* no buffer on free queue        */
+#define CLAW_MAX_LINK_ID        1
+#define CLAW_MAX_DEV            256        /*      max claw devices          */
+#define MAX_NAME_LEN            8          /* host name, adapter name length */
+#define CLAW_FRAME_SIZE         4096
+#define CLAW_ID_SIZE            BUS_ID_SIZE+3
+
+/* state machine codes used in claw_irq_handler */
+
+#define CLAW_STOP                0
+#define CLAW_START_HALT_IO       1
+#define CLAW_START_SENSEID       2
+#define CLAW_START_READ          3
+#define CLAW_START_WRITE         4
+
+/*-----------------------------------------------------*
+*    Lock flag                                         *
+*------------------------------------------------------*/
+#define LOCK_YES             0
+#define LOCK_NO              1
+
+/*-----------------------------------------------------*
+*    DBF Debug macros                                  *
+*------------------------------------------------------*/
+#define CLAW_DBF_TEXT(level, name, text) \
+	do { \
+		debug_text_event(claw_dbf_##name, level, text); \
+	} while (0)
+
+#define CLAW_DBF_HEX(level,name,addr,len) \
+do { \
+	debug_event(claw_dbf_##name,level,(void*)(addr),len); \
+} while (0)
+
+#define CLAW_DBF_TEXT_(level,name,text...) \
+do {                                       \
+	sprintf(debug_buffer, text);  \
+		debug_text_event(claw_dbf_##name,level, debug_buffer);\
+} while (0)
+
+/*******************************************************
+*  Define Control Blocks                               *
+*                                                      *
+********************************************************/
+
+/*------------------------------------------------------*/
+/*     CLAW header                                      */
+/*------------------------------------------------------*/
+
+struct clawh {
+        __u16  length;     /* length of data read by preceding read CCW */
+        __u8   opcode;     /* equivalent read CCW */
+        __u8   flag;       /* flag of FF to indicate read was completed */
+};
+
+/*------------------------------------------------------*/
+/*     CLAW Packing header   4 bytes                    */
+/*------------------------------------------------------*/
+struct clawph {
+       __u16 len;  	/* Length of Packed Data Area   */
+       __u8  flag;  	/* Reserved not used            */
+       __u8  link_num;	/* Link ID                      */
+};
+
+/*------------------------------------------------------*/
+/*     CLAW Ending struct ccwbk                         */
+/*------------------------------------------------------*/
+struct endccw {
+	__u32     real;            /* real address of this block */
+       __u8      write1;          /* write 1 is active */
+        __u8      read1;           /* read 1 is active  */
+        __u16     reserved;        /* reserved for future use */
+        struct ccw1    write1_nop1;
+        struct ccw1    write1_nop2;
+        struct ccw1    write2_nop1;
+        struct ccw1    write2_nop2;
+        struct ccw1    read1_nop1;
+        struct ccw1    read1_nop2;
+        struct ccw1    read2_nop1;
+        struct ccw1    read2_nop2;
+};
+
+/*------------------------------------------------------*/
+/*     CLAW struct ccwbk                                       */
+/*------------------------------------------------------*/
+struct ccwbk {
+        void   *next;        /* pointer to next ccw block */
+        __u32     real;         /* real address of this ccw */
+        void      *p_buffer;    /* virtual address of data */
+        struct clawh     header;       /* claw header */
+        struct ccw1    write;   /* write CCW    */
+        struct ccw1    w_read_FF; /* read FF */
+        struct ccw1    w_TIC_1;        /* TIC */
+        struct ccw1    read;         /* read CCW  */
+        struct ccw1    read_h;        /* read header */
+        struct ccw1    signal;       /* signal SMOD  */
+        struct ccw1    r_TIC_1;        /* TIC1 */
+        struct ccw1    r_read_FF;      /* read FF  */
+        struct ccw1    r_TIC_2;        /* TIC2 */
+};
+
+/*------------------------------------------------------*/
+/*     CLAW control block                               */
+/*------------------------------------------------------*/
+struct clawctl {
+        __u8    command;      /* control command */
+        __u8    version;      /* CLAW protocol version */
+        __u8    linkid;       /* link ID   */
+        __u8    correlator;   /* correlator */
+        __u8    rc;           /* return code */
+        __u8    reserved1;    /* reserved */
+        __u8    reserved2;    /* reserved */
+        __u8    reserved3;    /* reserved */
+        __u8    data[24];     /* command specific fields */
+};
+
+/*------------------------------------------------------*/
+/*     Data for SYSTEMVALIDATE command                  */
+/*------------------------------------------------------*/
+struct sysval  {
+        char    WS_name[8];        /* Workstation System name  */
+        char    host_name[8];      /* Host system name     */
+        __u16   read_frame_size;   /* read frame size */
+        __u16   write_frame_size;  /* write frame size */
+        __u8    reserved[4];       /* reserved */
+};
+
+/*------------------------------------------------------*/
+/*     Data for Connect command                         */
+/*------------------------------------------------------*/
+struct conncmd  {
+        char     WS_name[8];       /* Workstation application name  */
+        char     host_name[8];     /* Host application name      */
+        __u16    reserved1[2];     /* read frame size */
+        __u8     reserved2[4];     /* reserved  */
+};
+
+/*------------------------------------------------------*/
+/*     Data for CLAW error                              */
+/*------------------------------------------------------*/
+struct clawwerror  {
+        char      reserved1[8];   /* reserved */
+        char      reserved2[8];   /* reserved  */
+        char      reserved3[8];   /* reserved  */
+};
+
+/*------------------------------------------------------*/
+/*     Data buffer for CLAW                             */
+/*------------------------------------------------------*/
+struct clawbuf  {
+       char      buffer[MAX_ENVELOPE_SIZE];   /* data buffer */
+};
+
+/*------------------------------------------------------*/
+/*     Channel control block for read and write channel */
+/*------------------------------------------------------*/
+
+struct chbk {
+        unsigned int        devno;
+        int                 irq;
+	char 		    id[CLAW_ID_SIZE];
+       __u32               IO_active;
+        __u8                claw_state;
+        struct irb          *irb;
+       	struct ccw_device   *cdev;  /* pointer to the channel device */
+	struct net_device   *ndev;
+        wait_queue_head_t   wait;
+        struct tasklet_struct    tasklet;
+        struct timer_list   timer;
+        unsigned long       flag_a;    /* atomic flags */
+#define CLAW_BH_ACTIVE      0
+        unsigned long       flag_b;    /* atomic flags */
+#define CLAW_WRITE_ACTIVE   0
+        __u8                last_dstat;
+        __u8                flag;
+	struct sk_buff_head collect_queue;
+	spinlock_t collect_lock;
+#define CLAW_WRITE      0x02      /* - Set if this is a write channel */
+#define CLAW_READ	0x01      /* - Set if this is a read channel  */
+#define CLAW_TIMER      0x80      /* - Set if timer made the wake_up  */
+};
+
+/*--------------------------------------------------------------*
+*           CLAW  environment block                             *
+*---------------------------------------------------------------*/
+
+struct claw_env {
+        unsigned int            devno[2];       /* device number */
+        char                    host_name[9];   /* Host name */
+        char                    adapter_name [9]; /* adapter name */
+        char                    api_type[9];    /* TCPIP, API or PACKED */
+        void                    *p_priv;        /* privptr */
+        __u16                   read_buffers;   /* read buffer number */
+        __u16                   write_buffers;  /* write buffer number */
+        __u16                   read_size;      /* read buffer size */
+        __u16                   write_size;     /* write buffer size */
+        __u16                   dev_id;         /* device ident */
+	__u8			packing;	/* are we packing? */
+	volatile __u8		queme_switch;   /* gate for imed packing  */
+	volatile unsigned long	pk_delay;	/* Delay for adaptive packing */
+        __u8                    in_use;         /* device active flag */
+        struct net_device       *ndev;    	/* backward ptr to the net dev*/
+};
+
+/*--------------------------------------------------------------*
+*           CLAW  main control block                            *
+*---------------------------------------------------------------*/
+
+struct claw_privbk {
+        void *p_buff_ccw;
+        __u32      p_buff_ccw_num;
+        void  *p_buff_read;
+        __u32      p_buff_read_num;
+        __u32      p_buff_pages_perread;
+        void  *p_buff_write;
+        __u32      p_buff_write_num;
+        __u32      p_buff_pages_perwrite;
+        long       active_link_ID;           /* Active logical link ID */
+        struct ccwbk *p_write_free_chain;     /* pointer to free ccw chain */
+        struct ccwbk *p_write_active_first;   /* ptr to the first write ccw */
+        struct ccwbk *p_write_active_last;    /* ptr to the last write ccw */
+        struct ccwbk *p_read_active_first;    /* ptr to the first read ccw */
+        struct ccwbk *p_read_active_last;     /* ptr to the last read ccw */
+        struct endccw *p_end_ccw;              /*ptr to ending ccw */
+        struct ccwbk *p_claw_signal_blk;      /* ptr to signal block */
+        __u32      write_free_count;       /* number of free bufs for write */
+	struct     net_device_stats  stats; /* 	 device status    */
+        struct chbk channel[2];            /* Channel control blocks */
+        __u8       mtc_skipping;
+        int        mtc_offset;
+        int        mtc_logical_link;
+        void       *p_mtc_envelope;
+	struct	   sk_buff	*pk_skb;	/* packing buffer    */
+	int	   pk_cnt;
+        struct clawctl ctl_bk;
+        struct claw_env *p_env;
+        __u8       system_validate_comp;
+        __u8       release_pend;
+        __u8      checksum_received_ip_pkts;
+	__u8      buffs_alloc;
+        struct endccw  end_ccw;
+        unsigned long  tbusy;
+
+};
+
+
+/************************************************************/
+/* define global constants                                  */
+/************************************************************/
+
+#define CCWBK_SIZE sizeof(struct ccwbk)
+
+

drivers/s390/net/ctcdbug.c

@@ -0,0 +1,83 @@
+/*
+ *
+ * linux/drivers/s390/net/ctcdbug.c ($Revision: 1.4 $)
+ *
+ * CTC / ESCON network driver - s390 dbf exploit.
+ *
+ * Copyright 2000,2003 IBM Corporation
+ *
+ *    Author(s): Original Code written by
+ *			  Peter Tiedemann (ptiedem@de.ibm.com)
+ *
+ *    $Revision: 1.4 $	 $Date: 2004/08/04 10:11:59 $
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.	 See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#include "ctcdbug.h"
+
+/**
+ * Debug Facility Stuff
+ */
+debug_info_t *ctc_dbf_setup = NULL;
+debug_info_t *ctc_dbf_data = NULL;
+debug_info_t *ctc_dbf_trace = NULL;
+
+DEFINE_PER_CPU(char[256], ctc_dbf_txt_buf);
+
+void
+ctc_unregister_dbf_views(void)
+{
+	if (ctc_dbf_setup)
+		debug_unregister(ctc_dbf_setup);
+	if (ctc_dbf_data)
+		debug_unregister(ctc_dbf_data);
+	if (ctc_dbf_trace)
+		debug_unregister(ctc_dbf_trace);
+}
+int
+ctc_register_dbf_views(void)
+{
+	ctc_dbf_setup = debug_register(CTC_DBF_SETUP_NAME,
+					CTC_DBF_SETUP_INDEX,
+					CTC_DBF_SETUP_NR_AREAS,
+					CTC_DBF_SETUP_LEN);
+	ctc_dbf_data = debug_register(CTC_DBF_DATA_NAME,
+				       CTC_DBF_DATA_INDEX,
+				       CTC_DBF_DATA_NR_AREAS,
+				       CTC_DBF_DATA_LEN);
+	ctc_dbf_trace = debug_register(CTC_DBF_TRACE_NAME,
+					CTC_DBF_TRACE_INDEX,
+					CTC_DBF_TRACE_NR_AREAS,
+					CTC_DBF_TRACE_LEN);
+
+	if ((ctc_dbf_setup == NULL) || (ctc_dbf_data == NULL) ||
+	    (ctc_dbf_trace == NULL)) {
+		ctc_unregister_dbf_views();
+		return -ENOMEM;
+	}
+	debug_register_view(ctc_dbf_setup, &debug_hex_ascii_view);
+	debug_set_level(ctc_dbf_setup, CTC_DBF_SETUP_LEVEL);
+
+	debug_register_view(ctc_dbf_data, &debug_hex_ascii_view);
+	debug_set_level(ctc_dbf_data, CTC_DBF_DATA_LEVEL);
+
+	debug_register_view(ctc_dbf_trace, &debug_hex_ascii_view);
+	debug_set_level(ctc_dbf_trace, CTC_DBF_TRACE_LEVEL);
+
+	return 0;
+}
+
+

drivers/s390/net/ctcdbug.h

@@ -0,0 +1,123 @@
+/*
+ *
+ * linux/drivers/s390/net/ctcdbug.h ($Revision: 1.4 $)
+ *
+ * CTC / ESCON network driver - s390 dbf exploit.
+ *
+ * Copyright 2000,2003 IBM Corporation
+ *
+ *    Author(s): Original Code written by
+ *			  Peter Tiedemann (ptiedem@de.ibm.com)
+ *
+ *    $Revision: 1.4 $	 $Date: 2004/10/15 09:26:58 $
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.	 See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+
+#include <asm/debug.h>
+/**
+ * Debug Facility stuff
+ */
+#define CTC_DBF_SETUP_NAME "ctc_setup"
+#define CTC_DBF_SETUP_LEN 16
+#define CTC_DBF_SETUP_INDEX 3
+#define CTC_DBF_SETUP_NR_AREAS 1
+#define CTC_DBF_SETUP_LEVEL 3
+
+#define CTC_DBF_DATA_NAME "ctc_data"
+#define CTC_DBF_DATA_LEN 128
+#define CTC_DBF_DATA_INDEX 3
+#define CTC_DBF_DATA_NR_AREAS 1
+#define CTC_DBF_DATA_LEVEL 2
+
+#define CTC_DBF_TRACE_NAME "ctc_trace"
+#define CTC_DBF_TRACE_LEN 16
+#define CTC_DBF_TRACE_INDEX 2
+#define CTC_DBF_TRACE_NR_AREAS 2
+#define CTC_DBF_TRACE_LEVEL 3
+
+#define DBF_TEXT(name,level,text) \
+	do { \
+		debug_text_event(ctc_dbf_##name,level,text); \
+	} while (0)
+
+#define DBF_HEX(name,level,addr,len) \
+	do { \
+		debug_event(ctc_dbf_##name,level,(void*)(addr),len); \
+	} while (0)
+
+DECLARE_PER_CPU(char[256], ctc_dbf_txt_buf);
+extern debug_info_t *ctc_dbf_setup;
+extern debug_info_t *ctc_dbf_data;
+extern debug_info_t *ctc_dbf_trace;
+
+
+#define DBF_TEXT_(name,level,text...)				\
+	do {								\
+		char* ctc_dbf_txt_buf = get_cpu_var(ctc_dbf_txt_buf);	\
+		sprintf(ctc_dbf_txt_buf, text);			  	\
+		debug_text_event(ctc_dbf_##name,level,ctc_dbf_txt_buf);	\
+		put_cpu_var(ctc_dbf_txt_buf);				\
+	} while (0)
+
+#define DBF_SPRINTF(name,level,text...) \
+	do { \
+		debug_sprintf_event(ctc_dbf_trace, level, ##text ); \
+		debug_sprintf_event(ctc_dbf_trace, level, text ); \
+	} while (0)
+
+
+int ctc_register_dbf_views(void);
+
+void ctc_unregister_dbf_views(void);
+
+/**
+ * some more debug stuff
+ */
+
+#define HEXDUMP16(importance,header,ptr) \
+PRINT_##importance(header "%02x %02x %02x %02x  %02x %02x %02x %02x  " \
+		   "%02x %02x %02x %02x  %02x %02x %02x %02x\n", \
+		   *(((char*)ptr)),*(((char*)ptr)+1),*(((char*)ptr)+2), \
+		   *(((char*)ptr)+3),*(((char*)ptr)+4),*(((char*)ptr)+5), \
+		   *(((char*)ptr)+6),*(((char*)ptr)+7),*(((char*)ptr)+8), \
+		   *(((char*)ptr)+9),*(((char*)ptr)+10),*(((char*)ptr)+11), \
+		   *(((char*)ptr)+12),*(((char*)ptr)+13), \
+		   *(((char*)ptr)+14),*(((char*)ptr)+15)); \
+PRINT_##importance(header "%02x %02x %02x %02x  %02x %02x %02x %02x  " \
+		   "%02x %02x %02x %02x  %02x %02x %02x %02x\n", \
+		   *(((char*)ptr)+16),*(((char*)ptr)+17), \
+		   *(((char*)ptr)+18),*(((char*)ptr)+19), \
+		   *(((char*)ptr)+20),*(((char*)ptr)+21), \
+		   *(((char*)ptr)+22),*(((char*)ptr)+23), \
+		   *(((char*)ptr)+24),*(((char*)ptr)+25), \
+		   *(((char*)ptr)+26),*(((char*)ptr)+27), \
+		   *(((char*)ptr)+28),*(((char*)ptr)+29), \
+		   *(((char*)ptr)+30),*(((char*)ptr)+31));
+
+static inline void
+hex_dump(unsigned char *buf, size_t len)
+{
+	size_t i;
+
+	for (i = 0; i < len; i++) {
+		if (i && !(i % 16))
+			printk("\n");
+		printk("%02x ", *(buf + i));
+	}
+	printk("\n");
+}
+

drivers/s390/net/ctctty.h

@@ -0,0 +1,37 @@
+/*
+ * $Id: ctctty.h,v 1.4 2003/09/18 08:01:10 mschwide Exp $
+ *
+ * CTC / ESCON network driver, tty interface.
+ *
+ * Copyright (C) 2001 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s): Fritz Elfert (elfert@de.ibm.com, felfert@millenux.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _CTCTTY_H_
+#define _CTCTTY_H_
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+
+extern int  ctc_tty_register_netdev(struct net_device *);
+extern void ctc_tty_unregister_netdev(struct net_device *);
+extern void ctc_tty_netif_rx(struct sk_buff *);
+extern int  ctc_tty_init(void);
+extern void ctc_tty_cleanup(void);
+extern void ctc_tty_setcarrier(struct net_device *, int);
+
+#endif

drivers/s390/net/cu3088.c

@@ -0,0 +1,166 @@
+/*
+ * $Id: cu3088.c,v 1.34 2004/06/15 13:16:27 pavlic Exp $
+ *
+ * CTC / LCS ccw_device driver
+ *
+ * Copyright (C) 2002 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Author(s): Arnd Bergmann <arndb@de.ibm.com>
+ *            Cornelia Huck <cohuck@de.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/err.h>
+
+#include <asm/ccwdev.h>
+#include <asm/ccwgroup.h>
+
+#include "cu3088.h"
+
+const char *cu3088_type[] = {
+	"not a channel",
+	"CTC/A",
+	"ESCON channel",
+	"FICON channel",
+	"P390 LCS card",
+	"OSA LCS card",
+	"unknown channel type",
+	"unsupported channel type",
+};
+
+/* static definitions */
+
+static struct ccw_device_id cu3088_ids[] = {
+	{ CCW_DEVICE(0x3088, 0x08), .driver_info = channel_type_parallel },
+	{ CCW_DEVICE(0x3088, 0x1f), .driver_info = channel_type_escon },
+	{ CCW_DEVICE(0x3088, 0x1e), .driver_info = channel_type_ficon },
+	{ CCW_DEVICE(0x3088, 0x01), .driver_info = channel_type_p390 },
+	{ CCW_DEVICE(0x3088, 0x60), .driver_info = channel_type_osa2 },
+	{ /* end of list */ }
+};
+
+static struct ccw_driver cu3088_driver;
+
+struct device *cu3088_root_dev;
+
+static ssize_t
+group_write(struct device_driver *drv, const char *buf, size_t count)
+{
+	const char *start, *end;
+	char bus_ids[2][BUS_ID_SIZE], *argv[2];
+	int i;
+	int ret;
+	struct ccwgroup_driver *cdrv;
+
+	cdrv = to_ccwgroupdrv(drv);
+	if (!cdrv)
+		return -EINVAL;
+	start = buf;
+	for (i=0; i<2; i++) {
+		static const char delim[] = {',', '\n'};
+		int len;
+
+		if (!(end = strchr(start, delim[i])))
+			return count;
+		len = min_t(ptrdiff_t, BUS_ID_SIZE, end - start + 1);
+		strlcpy (bus_ids[i], start, len);
+		argv[i] = bus_ids[i];
+		start = end + 1;
+	}
+
+	ret = ccwgroup_create(cu3088_root_dev, cdrv->driver_id,
+			      &cu3088_driver, 2, argv);
+
+	return (ret == 0) ? count : ret;
+}
+
+static DRIVER_ATTR(group, 0200, NULL, group_write);
+
+/* Register-unregister for ctc&lcs */
+int
+register_cu3088_discipline(struct ccwgroup_driver *dcp) 
+{
+	int rc;
+
+	if (!dcp)
+		return -EINVAL;
+
+	/* Register discipline.*/
+	rc = ccwgroup_driver_register(dcp);
+	if (rc)
+		return rc;
+
+	rc = driver_create_file(&dcp->driver, &driver_attr_group);
+	if (rc)
+		ccwgroup_driver_unregister(dcp);
+		
+	return rc;
+
+}
+
+void
+unregister_cu3088_discipline(struct ccwgroup_driver *dcp)
+{
+	if (!dcp)
+		return;
+
+	driver_remove_file(&dcp->driver, &driver_attr_group);
+	ccwgroup_driver_unregister(dcp);
+}
+
+static struct ccw_driver cu3088_driver = {
+	.owner	     = THIS_MODULE,
+	.ids	     = cu3088_ids,
+	.name        = "cu3088",
+	.probe	     = ccwgroup_probe_ccwdev,
+	.remove	     = ccwgroup_remove_ccwdev,
+};
+
+/* module setup */
+static int __init
+cu3088_init (void)
+{
+	int rc;
+	
+	cu3088_root_dev = s390_root_dev_register("cu3088");
+	if (IS_ERR(cu3088_root_dev))
+		return PTR_ERR(cu3088_root_dev);
+	rc = ccw_driver_register(&cu3088_driver);
+	if (rc)
+		s390_root_dev_unregister(cu3088_root_dev);
+
+	return rc;
+}
+
+static void __exit
+cu3088_exit (void)
+{
+	ccw_driver_unregister(&cu3088_driver);
+	s390_root_dev_unregister(cu3088_root_dev);
+}
+
+MODULE_DEVICE_TABLE(ccw,cu3088_ids);
+MODULE_AUTHOR("Arnd Bergmann <[email protected]>");
+MODULE_LICENSE("GPL");
+
+module_init(cu3088_init);
+module_exit(cu3088_exit);
+
+EXPORT_SYMBOL_GPL(cu3088_type);
+EXPORT_SYMBOL_GPL(register_cu3088_discipline);
+EXPORT_SYMBOL_GPL(unregister_cu3088_discipline);