Miao Xie
708c1bbc9d
mempolicy: restructure rebinding-mempolicy functions ...
Nick Piggin reported that the allocator may see an empty nodemask when
changing cpuset's mems[1]. It happens only on the kernel that do not do
atomic nodemask_t stores. (MAX_NUMNODES > BITS_PER_LONG)
But I found that there is also a problem on the kernel that can do atomic
nodemask_t stores. The problem is that the allocator can't find a node to
alloc page when changing cpuset's mems though there is a lot of free
memory. The reason is like this:
(mpol: mempolicy)
task1 task1's mpol task2
alloc page 1
alloc on node0? NO 1
1 change mems from 1 to 0
1 rebind task1's mpol
0-1 set new bits
0 clear disallowed bits
alloc on node1? NO 0
...
can't alloc page
goto oom
I can use the attached program reproduce it by the following step:
# mkdir /dev/cpuset
# mount -t cpuset cpuset /dev/cpuset
# mkdir /dev/cpuset/1
# echo `cat /dev/cpuset/cpus` > /dev/cpuset/1/cpus
# echo `cat /dev/cpuset/mems` > /dev/cpuset/1/mems
# echo $$ > /dev/cpuset/1/tasks
# numactl --membind=`cat /dev/cpuset/mems` ./cpuset_mem_hog <nr_tasks> &
<nr_tasks> = max(nr_cpus - 1, 1)
# killall -s SIGUSR1 cpuset_mem_hog
# ./change_mems.sh
several hours later, oom will happen though there is a lot of free memory.
This patchset fixes this problem by expanding the nodes range first(set
newly allowed bits) and shrink it lazily(clear newly disallowed bits). So
we use a variable to tell the write-side task that read-side task is
reading nodemask, and the write-side task clears newly disallowed nodes
after read-side task ends the current memory allocation.
This patch:
In order to fix no node to alloc memory, when we want to update mempolicy
and mems_allowed, we expand the set of nodes first (set all the newly
nodes) and shrink the set of nodes lazily(clean disallowed nodes), But the
mempolicy's rebind functions may breaks the expanding.
So we restructure the mempolicy's rebind functions and split the rebind
work to two steps, just like the update of cpuset's mems: The 1st step:
expand the set of the mempolicy's nodes. The 2nd step: shrink the set of
the mempolicy's nodes. It is used when there is no real lock to protect
the mempolicy in the read-side. Otherwise we can do rebind work at once.
In order to implement it, we define
enum mpol_rebind_step {
MPOL_REBIND_ONCE,
MPOL_REBIND_STEP1,
MPOL_REBIND_STEP2,
MPOL_REBIND_NSTEP,
};
If the mempolicy needn't be updated by two steps, we can pass
MPOL_REBIND_ONCE to the rebind functions. Or we can pass
MPOL_REBIND_STEP1 to do the first step of the rebind work and pass
MPOL_REBIND_STEP2 to do the second step work.
Besides that, it maybe long time between these two step and we have to
release the lock that protects mempolicy and mems_allowed. If we hold the
lock once again, we must check whether the current mempolicy is under the
rebinding (the first step has been done) or not, because the task may
alloc a new mempolicy when we don't hold the lock. So we defined the
following flag to identify it:
#define MPOL_F_REBINDING (1 << 2)
The new functions will be used in the next patch.
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Paul Menage <menage@google.com>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Cc: Ravikiran Thirumalai <kiran@scalex86.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-05-25 08:06:57 -07:00
Merge branch 'devel-stable' into devel
2010-05-17 17:24:04 +01:00
caif: Rewritten socket implementation
2010-04-28 12:55:14 -07:00
can: sja1000 platform data fixes
2010-05-17 22:39:48 -07:00
Revert "V4L/DVB: Add FE_CAN_PSK_8 to allow apps to identify PSK_8 capable DVB devices"
2010-05-19 12:57:48 -03:00
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
2010-05-20 09:41:44 -07:00
sh: allow platforms to specify SD-card supported voltages
2010-05-22 17:05:22 +09:00
IB/mlx4: Add support for masked atomic operations
2010-04-21 16:37:49 -07:00
mtd: nand: support alternate BB marker locations on MLC
2010-05-14 01:56:12 +01:00
netfilter: xtables: change hotdrop pointer to direct modification
2010-05-11 18:35:27 +02:00
nfsd: further comment typos
2010-05-03 08:33:00 -04:00
Merge branch 'master' into for-davem
2010-04-23 14:43:45 -04:00
ssb: Fix order of definitions and some text space indents
2010-04-26 13:51:09 -04:00
SUNRPC: Don't spam gssd with upcall requests when the kerberos key expired
2010-05-14 15:09:37 -04:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
2010-05-20 21:26:12 -07:00
x86, acpi/irq: Teach acpi_get_override_irq to take a gsi not an isa_irq
2010-05-04 13:34:27 -07:00
ahci: Add platform driver
2010-05-14 17:08:01 -04:00
serial: Add driver for the Altera JTAG UART
2010-05-21 09:34:30 -07:00
serial: Add driver for the Altera UART
2010-05-21 09:34:30 -07:00
libata-sff: prd is BMDMA specific
2010-05-19 13:38:54 -04:00
writeback: fixups for !dirty_writeback_centisecs
2010-05-21 20:00:35 +02:00
Merge branch 'core-hweight-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2010-05-18 09:17:01 -07:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6
2010-05-21 15:26:46 -07:00
cgroup: Check task_lock in task_subsys_state()
2010-05-04 09:25:02 -07:00
clocksource: Add clocksource_register_hz/khz interface
2010-05-10 14:24:26 +02:00
coda: move backing-dev.h kernel include inside __KERNEL__
2010-04-28 09:20:33 +02:00
crypto: Use ARCH_KMALLOC_MINALIGN for CRYPTO_MINALIGN now that it's exposed
2010-05-19 22:03:14 +03:00
Fix the regression created by "set S_DEAD on unlink()..." commit
2010-05-15 07:16:33 -04:00
remove DCB_PROTO_VERSION as we don't do netlink versioning
2010-04-22 18:32:12 -07:00
Debugobjects transition check
2010-05-10 16:08:01 -07:00
Merge remote branch 'origin' into secretlab/next-devicetree
2010-05-22 00:36:56 -06:00
Merge branch 'ioat' into dmaengine
2010-05-17 16:30:58 -07:00
quota: unify ->set_dqblk
2010-05-21 19:30:44 +02:00
drbd: This is now equivalent to drbd release 8.3.8rc1
2010-05-21 21:12:01 +02:00
drbd: Four new configuration settings for resync speed control
2010-05-18 01:25:00 +02:00
drbd: Four new configuration settings for resync speed control
2010-05-18 01:25:00 +02:00
ds2782_battery: Add support for ds2786 battery gas gauge
2010-04-26 22:03:42 +04:00
[S390] add breaking event address for user space
2010-05-17 10:00:15 +02:00
linux/elfcore.h: hide kernel functions
2010-05-21 20:29:10 -07:00
ext2: Add ext2_sb_info s_lock spinlock
2010-05-21 19:30:39 +02:00
vga16fb, drm: vga16fb->drm handoff
2010-05-18 16:19:30 +10:00
pipe: add support for shrinking and growing pipes
2010-05-21 21:12:40 +02:00
net: rtnetlink: decouple rtnetlink address families from real address families
2010-04-26 16:13:54 +02:00
net: Socket filter ancilliary data access for skb->dev->type
2010-04-22 16:05:44 -07:00
firmware_class: fix memory leak - free allocated pages
2010-05-21 09:37:28 -07:00
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
2010-05-21 19:37:45 -07:00
Merge branch 'tracing-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2010-05-18 08:35:04 -07:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
2010-05-21 07:19:18 -07:00
fs: xattr_handler table should be const
2010-05-21 18:31:18 -04:00
tty: n_gsm line discipline
2010-05-21 09:34:29 -07:00
Merge branches 'upstream-fixes', 'bkl-removal', 'debugfs-fixes' and 'hid-suspend' into for-linus
2010-05-19 14:05:06 +02:00
hw-breakpoints: Get the number of available registers on boot dynamically
2010-05-01 04:32:14 +02:00
i2c-omap: add mpu wake up latency constraint in i2c
2010-05-20 00:18:59 +01:00
Merge remote branch 'origin' into secretlab/next-devicetree
2010-05-22 00:36:56 -06:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide-2.6
2010-05-24 08:05:29 -07:00
netpoll: add generic support for bridge and bonding devices
2010-05-06 00:47:21 -07:00
net: Add netlink support for virtual port management (was iovnl)
2010-05-17 22:49:55 -07:00
net: adjust handle_macvlan to pass port struct to hook
2010-05-15 23:48:02 -07:00
tun: add ioctl to modify vnet header size
2010-05-03 12:33:13 +03:00
X25: Add if_x25.h and x25 to device identifiers
2010-04-22 16:12:36 -07:00
IPv6: data structure changes for new socket options
2010-04-23 23:35:28 -07:00
rcu: remove all rcu head initializations, except on_stack initializations
2010-05-11 16:10:47 -07:00
interrupt.h: fix fatal kernel-doc error
2010-05-21 10:48:12 -07:00
resource: shared I/O region support
2010-05-11 12:01:10 -07:00
ipv6: ip6mr: support multiple tables
2010-05-11 14:40:55 +02:00
genirq: Add CPU mask affinity hint
2010-05-03 11:50:57 +02:00
isapnp: move definitions to mod_devicetable.h so file2alias can reach them.
2010-05-19 17:33:38 +09:30
jbd: Provide function to check whether transaction will issue data barrier
2010-05-21 19:30:40 +02:00
X25: Add if_x25.h and x25 to device identifiers
2010-04-22 16:12:36 -07:00
printk,kdb: capture printk() when in kdb shell
2010-05-20 21:04:27 -05:00
Merge git://git.infradead.org/iommu-2.6
2010-05-21 17:25:01 -07:00
Merge branch 'master' into for-next
2010-04-23 02:08:44 +02:00
x86, kgdb, init: Add early and late debug states
2010-05-20 21:04:29 -05:00
netns: Teach network device kobjects which namespace they are in.
2010-05-21 09:37:32 -07:00
kref: remove kref_set
2010-05-21 09:37:29 -07:00
ks8842: Add platform data for setting mac address
2010-04-21 16:33:29 -07:00
ktime: introduce ktime_to_ms()
2010-05-14 15:09:32 -04:00
KVM: PPC: Add OSI hypercall interface
2010-05-17 12:17:10 +03:00
KVM: Let vcpu structure alignment be determined at runtime
2010-05-19 11:36:29 +03:00
libata-sff: separate out BMDMA qc_issue
2010-05-19 13:38:55 -04:00
lockdep: Add novalidate class for dev->mutex conversion
2010-05-21 09:37:30 -07:00
LSM Audit: rename LSM_AUDIT_NO_AUDIT to LSM_AUDIT_DATA_NONE
2010-04-28 08:51:12 +10:00
mempolicy: restructure rebinding-mempolicy functions
2010-05-25 08:06:57 -07:00
V4L/DVB: meye: remove last V4L1 remnants from the code and add v4l2_device
2010-05-18 00:52:36 -03:00
mm: remove return value of putback_lru_pages()
2010-05-25 08:06:57 -07:00
Merge branch 'modules' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus
2010-05-21 17:15:44 -07:00
ipv6: ip6mr: support multiple tables
2010-05-11 14:40:55 +02:00
drivers: video: msm: add include msm_mdp.h
2010-04-28 15:16:48 -07:00
ncpfs: BKL ioctl pushdown
2010-05-17 05:27:42 +02:00
ncpfs: add bdi backing to mount session
2010-04-22 12:31:11 +02:00
net: sock_def_readable() and friends RCU conversion
2010-05-01 15:00:15 -07:00
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next-2.6
2010-05-20 21:04:44 -07:00
netlink: Implment netlink_broadcast_filtered
2010-05-21 09:37:32 -07:00
netpoll: Use 'bool' for netpoll_rx() return type.
2010-05-06 01:31:27 -07:00
NFS: Add helper functions for allocating filehandles and fattr structs
2010-05-14 15:09:21 -04:00
NFSv4: Clean up the NFSv4 setclientid operation
2010-05-14 15:09:30 -04:00
NFSv4: Clean up the NFSv4 setclientid operation
2010-05-14 15:09:30 -04:00
nilfs2: enlarge s_volume_name member in nilfs_super_block
2010-05-10 11:32:33 +09:00
cfg80211/mac80211: better channel handling
2010-05-07 14:55:50 -04:00
netpoll: add generic support for bridge and bonding devices
2010-05-06 00:47:21 -07:00
of: change of_match_device to work with struct device
2010-05-22 00:10:41 -06:00
of/flattree: Make unflatten_device_tree() safe to call from any arch
2010-04-28 18:20:29 -06:00
of: Remove duplicate fields from of_platform_driver
2010-05-22 00:10:40 -06:00
padata: Add some code comments
2010-05-19 13:44:27 +10:00
x86/PCI: irq and pci_ids patch for additional Intel Cougar Point DeviceIDs
2010-05-11 12:01:40 -07:00
Merge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6
2010-05-21 18:58:52 -07:00
Revert "perf: Fix exit() vs PERF_FORMAT_GROUP"
2010-05-11 08:31:49 +02:00
pipe: set lower and upper limit on max pages in the pipe page array
2010-05-21 21:12:52 +02:00
platform: Make platform resource input parameters const
2010-05-17 21:37:40 +02:00
PM QOS update
2010-05-10 23:08:19 +02:00
i2c: Fix bus-level power management callbacks
2010-05-10 23:09:30 +02:00
PM: pm_wakeup - switch to using bool
2010-05-10 23:08:15 +02:00
hugetlb: fix infinite loop in get_futex_key() when backed by huge pages
2010-04-24 11:31:25 -07:00
power_supply: Add support for writeable properties
2010-05-19 12:14:42 +04:00
quota: unify ->set_dqblk
2010-05-21 19:30:44 +02:00
quota: Refactor dquot_transfer code so that OCFS2 can pass in its references
2010-05-21 19:30:45 +02:00
ramfs: replace inode uid,gid,mode initialization with helper function
2010-05-21 18:31:26 -04:00
Merge branch 'x86-pat-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2010-05-18 09:28:04 -07:00
net: rcu fixes
2010-05-03 15:53:54 -07:00
rcu head introduce rcu head init on stack
2010-05-10 16:53:55 -07:00
Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2010-05-18 08:27:54 -07:00
Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2010-05-18 08:27:54 -07:00
reiserfs: constify xattr_handler
2010-05-21 18:31:19 -04:00
reiserfs: constify xattr_handler
2010-05-21 18:31:19 -04:00
ring-buffer: Make non-consuming read less expensive with lots of cpus.
2010-04-27 13:06:35 -04:00
rtc-v3020: make bitfield unsigned
2010-05-11 10:09:47 +02:00
ipv6: ip6mr: support multiple tables
2010-05-11 14:40:55 +02:00
Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2010-05-18 08:27:54 -07:00
Merge branch 'master' into next
2010-05-06 10:56:07 +10:00
Merge branch 'kdb-merge' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/linux-2.6-kgdb
2010-05-21 11:08:05 -07:00
sh: fixup the docbook paths for clock framework shuffling.
2010-05-13 18:42:25 +09:00
SH: constify multiple DMA related objects and references to them
2010-04-26 15:50:50 +09:00
net: add a noref bit on skb dst
2010-05-17 17:18:50 -07:00
mm: Move ARCH_SLAB_MINALIGN and ARCH_KMALLOC_MINALIGN to <linux/slab_def.h>
2010-05-19 22:03:13 +03:00
mm: Move ARCH_SLAB_MINALIGN and ARCH_KMALLOC_MINALIGN to <linux/slob_def.h>
2010-05-19 22:03:13 +03:00
mm: Move ARCH_SLAB_MINALIGN and ARCH_KMALLOC_MINALIGN to <linux/slub_def.h>
2010-05-19 22:03:13 +03:00
smbfs: add bdi backing to mount session
2010-04-22 12:37:07 +02:00
pipe: add support for shrinking and growing pipes
2010-05-21 21:12:40 +02:00
rcu: make SRCU usable in modules
2010-05-10 11:08:35 -07:00
cpu_stop: add dummy implementation for UP
2010-05-08 17:12:33 +02:00
tmpfs: insert tmpfs cache pages to inactive list at first
2010-05-25 08:06:56 -07:00
sysctl: add proc_do_large_bitmap
2010-05-15 23:28:39 -07:00
sysfs: add struct file* to bin_attr callbacks
2010-05-21 09:37:31 -07:00
KVM: VMX: enable VMXON check with SMX enabled (Intel TXT)
2010-05-19 11:36:34 +03:00
Input: add keypad driver for keys interfaced to TCA6416
2010-05-03 23:50:42 -07:00
sched: Intoduce get_cpu_iowait_time_us()
2010-05-09 19:35:27 +02:00
tipc: Update commenting in TIPC API
2010-05-12 23:02:23 -07:00
tipc: Add support for "-s" configuration option
2010-05-12 23:02:23 -07:00
tracing: Fix tracepoint.h DECLARE_TRACE() to allow more than one header
2010-05-05 11:46:17 -04:00
tty: n_gsm line discipline
2010-05-21 09:34:29 -07:00
atomic_t: Remove volatile from atomic_t definition
2010-05-17 07:57:27 -07:00
USB: remove unused usb_buffer_alloc and usb_buffer_free macros
2010-05-20 13:21:50 -07:00
viafb: move some include files to include/linux
2010-05-11 16:07:59 -06:00
viafb: move some include files to include/linux
2010-05-11 16:07:59 -06:00
viafb: move some include files to include/linux
2010-05-11 16:07:59 -06:00
V4L/DVB: Add a new ERROR flag for DQBUF after recoverable streaming errors
2010-05-19 12:58:37 -03:00
virtio: add_buf_gfp
2010-05-19 22:15:46 +09:30
Add virtio disk identification support
2010-05-19 22:15:40 +09:30
virtio: console: Use a control message to add ports
2010-05-19 22:15:49 +09:30
Merge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
2010-05-20 21:26:12 -07:00
writeback: fix problem with !CONFIG_BLOCK compilation
2010-05-21 20:01:03 +02:00
fs: xattr_handler table should be const
2010-05-21 18:31:18 -04:00
m68k: amiga - Zorro host bridge platform device conversion
2010-05-17 21:37:42 +02:00
