* use 8 clk setting for active clocks == 7 (was 12 clk)
* use 12 clk setting for active clocks > 12 (was 8 clk)
* do 66MHz bus fixup before mapping active clocks
* fix setup of PIO command timings
Acked-by: Jung-Ik (John) Lee <[email protected]>
Signed-off-by: Bartlomiej Zolnierkiewicz <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
MWDMA modes are not supported by this driver currently.
Acked-by: Jung-Ik (John) Lee <[email protected]>
Signed-off-by: Bartlomiej Zolnierkiewicz <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
While trying to work around spurious detection retries for
non-existent devices on slave links, commit
816ab89782 incorrectly added link
offline check logic before ata_eh_thaw() was called. This means that
if an occupied link goes down briefly at the time that offline check
was performed, device class will be cleared to ATA_DEV_NONE and libata
wouldn't retry thus failing detection of the device.
The offline check should be done after the port is thawed together
with online check so that such link glitches can be detected by the
interrupt handler and handled properly.
Signed-off-by: Tejun Heo <[email protected]>
Reported-by: Tim Blechmann <[email protected]>
Cc: [email protected]
Signed-off-by: Jeff Garzik <[email protected]>
This is a fix for a bug which was a result of wrong use of checksum offload flag.
The status of tx-checksumming was not changed from on to off
after a 'ethtool -K <ifname> tx off' operation.
Use the proper checksum offload flag NETIF_F_HW_CSUM instead of
NETIF_F_IP_CSUM and NETIF_F_IPV6_CSUM.
Patch is against net-2.6 tree.
Signed-off-by: Ajit Khaparde <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
MCC_STATUS_NOT_SUPPORTED should be decimal 66 not hex 66.
This patch fixes this typo. Patch against net-2.6 tree.
Signed-off-by: Ajit Khaparde <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
While updating the statistics to be passed via the get_stats,
tx multicast frames were being accounted instead of rx multicast frames.
This patch fixes the bug. This patch is against the net-2.6 tree.
Signed-off-by: Ajit Khaparde <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Get rid of spinlock and private mutex usage for exclusive access to the
HW semaphore register. rtnl_lock already creates exclusive access to
this register in all driver API.
Add rtnl to firmware worker threads that also use the HW semaphore register.
Signed-off-by: Ron Mercer <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Check that we are not already polling firmware events before we queue the
firmware event worker, then disable firmware interrupts.
Otherwise we can queue the same event multiple times.
Signed-off-by: Ron Mercer <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This fixes a bug introduced by this commit ID:
commit 537a1bf059
Author: Krzysztof Helt <[email protected]>
Date: Tue Jun 30 11:41:29 2009 -0700
fbdev: add mutex for fb_mmap locking
In which a mutex was added when changing smem_start and smem_len fields,
so the mutex inside the fb_mmap() call is actually used.
The problem was that set_fb_fix, which modifies the above 2 fields,
was called before and after registering the framebuffer,
which when used before registration, lead to a failed attempt to
use an uninitialized mutex.
Solution: Don't use mutex before framebuffer registration.
Signed-off-by: Sergio Aguirre <[email protected]>
Acked-by: Tomi Valkeinen <[email protected]>
Acked-by: Imre Deak <[email protected]>
Tested-by: Cory Maccarrone <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
I was using Coccinelle with the mutex_unlock semantic patch, and it
unconvered this problem. It appears to be a valid missing unlock issue.
This change should correct it by moving the unlock below the label.
This patch is against the mainline kernel.
Cc: Julia Lawall <[email protected]>
Cc: Hiroshi DOYU <[email protected]>
Signed-off-by: Daniel Walker <[email protected]>
Signed-off-by: Kevin Hilman <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
the original flush operation is to flush the function address which is
copied from.
But we do not change the function code and it is not necessary to flush it.
Signed-off-by: janboe <[email protected]>
Acked-by: Paul Walmsley <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
exit_pi_state() is called from do_exit() but not from do_execve().
Move it to release_mm() so it gets called from do_execve() as well.
Signed-off-by: Thomas Gleixner <[email protected]>
LKML-Reference: <new-submission>
Cc: [email protected]
Cc: Anirban Sinha <[email protected]>
Cc: Peter Zijlstra <[email protected]>
The robust list pointers of user space held futexes are kept intact
over an exec() call. When the exec'ed task exits exit_robust_list() is
called with the stale pointer. The risk of corruption is minimal, but
still it is incorrect to keep the pointers valid. Actually glibc
should uninstall the robust list before calling exec() but we have to
deal with it anyway.
Nullify the pointers after [compat_]exit_robust_list() has been
called.
Reported-by: Anirban Sinha <[email protected]>
Signed-off-by: Peter Zijlstra <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
LKML-Reference: <new-submission>
Cc: [email protected]
The state char variable S should be reassigned, if S == 0.
We are missing the state of the task that is going to sleep for the
context switch events (in the raw mode).
Fortunately the problem arises with the sched_switch/wake_up
tracers, not the sched trace events.
The formers are legacy now. But still, that was buggy.
Signed-off-by: Hiroshi Shimamoto <[email protected]>
Cc: Steven Rostedt <[email protected]>
Acked-by: Frederic Weisbecker <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
Some architectures such as Sparc, ARM and MIPS (basically
everything with flush_dcache_page()) need to deal with dcache
aliases by carefully placing pages in both kernel and user maps.
These architectures typically have to use vmalloc_user() for this.
However, on other architectures, vmalloc() is not needed and has
the downsides of being more restricted and slower than regular
allocations.
Signed-off-by: Peter Zijlstra <[email protected]>
Acked-by: David Miller <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Jens Axboe <[email protected]>
Cc: Paul Mackerras <[email protected]>
LKML-Reference: <1254830228.21044.272.camel@laptop>
Signed-off-by: Ingo Molnar <[email protected]>
alc_subsystem_id() tries to pick up a headphone pin if not configured,
but this caused side-effects as the problem in commit
15870f05e9.
This patch fixes the driver behavior to pick up invalid HP pins; at least,
the pins that are listed as the primary outputs aren't taken any more.
Signed-off-by: Takashi Iwai <[email protected]>
Asm routines that end up having size equal to zero are not really
zero sized, and as now we do kernel_maps__fixup_sym_end, at least
for kernel routines this gets fixed.
A similar fixup needs to be done for the userspace bits as well,
but as this fixup started only because in /proc/kallsyms we don't
have the end address nor the function size, it appeared here first.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
Cc: Frédéric Weisbecker <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Mike Galbraith <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
Commit 7069331 (connector: Provide the sender's credentials to the
callback, 2009-10-02) changed callbacks to take two arguments but missed
this one.
drivers/connector/cn_proc.c: In function ‘cn_proc_init’:
drivers/connector/cn_proc.c:263: warning: passing argument 3 of
‘cn_add_callback’ from incompatible pointer type
Signed-off-by: Stephen Boyd <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
the global refcount wasn't being increased after the first reference.
this caused an oops on unload on a multi-gpu card.
Signed-off-by: Dave Airlie <[email protected]>
Curiously, Aspire 3810T issues many SATA feature enable commands via
_GTF, of which one is invalid and another is not supported by the
drive. In the process, it also enables FPDMA non-zero offset.
However, the feature also needs to be supported and enabled from the
controller and it's wrong to enable it from _GTF unless the controller
can do it by default.
Currently, this ends up enabling FPDMA non-zero offset only on the
drive side leading to NCQ command failures and eventual disabling of
NCQ. This patch makes libata filter out FPDMA non-zero offset enable
for the machine.
This was reported by Marcus Meissner in bnc#522790.
https://bugzilla.novell.com/show_bug.cgi?id=522790
Reported-by: Marcus Meissner <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
Add ->gtf_filter to ata_device and set it to ata_acpi_gtf_filter when
initializing ata_link. This is to allow quirks which apply different
gtf filters.
Signed-off-by: Tejun Heo <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
Currently libata-acpi can only filter DIPM among SATA feature enables
via _GTF. This patch adds the capability to filter out FPDMA non-zero
offset, in-order guarantee and auto-activation.
Signed-off-by: Tejun Heo <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
We're about to add more SATA_* and ATA_ACPI_FILTER_* constants.
Reformat them in preparation.
Signed-off-by: Tejun Heo <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
Update the AHCI driver to display all of the HBA capabilities defined in the
AHCI 1.3 specification. Some of these are in a new CAP2 (HBA Capabilities
Extended) register which is only defined on AHCI 1.2 or later. The spec says
that undefined registers should always return 0 on read, but to be safe we
assume a value of 0 unless the controller reports AHCI version 1.2 or later.
The value can also be retrieved through sysfs as with the existing capability
field.
For example, on an Intel Ibex Peak (PCH) controller:
ahci 0000:00:1f.2: flags: 64bit ncq sntf stag pm led clo pmp pio slum part ems
sxs apst
We don't do anything special with the new flags yet.
Also, change the code that displays the flags to use the same bit enumerations
that are used to control actual operation.
Signed-off-by: Robert Hancock <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
something-bility is spelled as something-blity
so a grep for 'blit' would find these lines
I broke this one out from the rest as it actually changes
the output of a kernel message - so it could in theory
change the behavior of tools that parse that ouput
Signed-off-by: Dirk Hohndel <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
Till now only one board, ASUS M2A-VM, can do 64bit dma with recent
BIOSen. Enabling 64bit DMA by default already broke three boards.
Enabling 64bit DMA isn't worth these regressions. Disable 64bit DMA
by default and enable it only on boards which are known to work.
Signed-off-by: Tejun Heo <[email protected]>
Reported-by: Gabriele Balducci <[email protected]>
Reported-by: [email protected]
Cc: Shane Huang <[email protected]>
Cc: [email protected]
Signed-off-by: Jeff Garzik <[email protected]>
The previous patches had some unwanted side effects, I've fixed
the lack of 32bpp working, and fixed up 16bpp so it should also work.
this also adds the interface to allow the driver to set a preferred
console depth so for example low memory rn50 can set it to 8bpp.
It also catches 24bpp on cards that can't do it and forces 32bpp.
Tested on r100/r600/i945.
Signed-off-by: Dave Airlie <[email protected]>
Lock DPLL5 at 120MHz at boot. The USBHOST 120MHz f-clock and
USBTLL f-clock are the only users of this DPLL, and 120MHz is
is the only recommended rate for these clocks.
With this patch, the 60 MHz ULPI clock is generated correctly.
Tested on an OMAP3430 SDP.
Signed-off-by: Rajendra Nayak <[email protected]>
Signed-off-by: Anand Gadiyar <[email protected]>
Signed-off-by: Paul Walmsley <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
Commit cd92204924 added
support for omap850. However, the patch accidentally
removed the wrong ifdef:
# define cpu_is_omap730() 1
# endif
#endif
+#else
+# if defined(CONFIG_ARCH_OMAP850)
+# undef cpu_is_omap850
+# define cpu_is_omap850() 1
+# endif
+#endif
...
void omap2_check_revision(void);
#endif /* defined(CONFIG_ARCH_OMAP2) || defined(CONFIG_ARCH_OMAP3) */
-
-#endif
Instead of removing removing the #endif at the end of the file,
the #endif before #else should have been removed.
But we cannot have multiple #else statements as pointed out by
Alistair Buxton <[email protected]>. So the fix is to:
- remove the non-multi-omap special handling, as we need to
detect between omap730 and omap850 anyways.
- add the missing #endif back to the end of the file
Reported-by: Sanjeev Premi <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
Rich reported a lock imbalance in the futex code:
http://bugzilla.kernel.org/show_bug.cgi?id=14288
It's caused by the displacement of the retry_private label in
futex_wake_op(). The code unlocks the hash bucket locks in the
error handling path and retries without locking them again which
makes the next unlock fail.
Move retry_private so we lock the hash bucket locks when we retry.
Reported-by: Rich Ercolany <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Darren Hart <[email protected]>
Cc: stable-2.6.31 <[email protected]>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <[email protected]>
Commit ffd71da4e3 ("panic: decrease oops_in_progress only after
having done the panic") moved bust_spinlocks(0) to the end of the
function, which in practice is never reached.
As a result console_unblank() is not called, and on some systems
the user may not see the panic message.
Move it back up to before the unblanking.
Signed-off-by: Aaro Koskinen <[email protected]>
Reviewed-by: Frederic Weisbecker <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
