GitHub/moto-9609/android_kernel_motorola_exynos9610.git
11 years agoMerge back earlier 'pm-cpufreq' material.
Rafael J. Wysocki [Thu, 22 Aug 2013 22:55:13 +0000 (00:55 +0200)]
Merge back earlier 'pm-cpufreq' material.

11 years agocpufreq: fix bad unlock balance on !CONFIG_SMP
Li Zhong [Tue, 20 Aug 2013 23:31:08 +0000 (01:31 +0200)]
cpufreq: fix bad unlock balance on !CONFIG_SMP

This patch tries to fix lockdep complaint attached below.

It seems that we should always read acquire the cpufreq_rwsem,
whether CONFIG_SMP is enabled or not.  And CONFIG_HOTPLUG_CPU
depends on CONFIG_SMP, so it seems we don't need CONFIG_SMP for the
code enabled by CONFIG_HOTPLUG_CPU.

[    0.504191] =====================================
[    0.504627] [ BUG: bad unlock balance detected! ]
[    0.504627] 3.11.0-rc6-next-20130819 #1 Not tainted
[    0.504627] -------------------------------------
[    0.504627] swapper/1 is trying to release lock (cpufreq_rwsem) at:
[    0.504627] [<ffffffff813d927a>] cpufreq_add_dev+0x13a/0x3e0
[    0.504627] but there are no more locks to release!
[    0.504627]
[    0.504627] other info that might help us debug this:
[    0.504627] 1 lock held by swapper/1:
[    0.504627]  #0:  (subsys mutex#4){+.+.+.}, at: [<ffffffff8134a7bf>] subsys_interface_register+0x4f/0xe0
[    0.504627]
[    0.504627] stack backtrace:
[    0.504627] CPU: 0 PID: 1 Comm: swapper Not tainted 3.11.0-rc6-next-20130819 #1
[    0.504627] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2007
[    0.504627]  ffffffff813d927a ffff88007f847c98 ffffffff814c062b ffff88007f847cc8
[    0.504627]  ffffffff81098bce ffff88007f847cf8 ffffffff81aadc30 ffffffff813d927a
[    0.504627]  00000000ffffffff ffff88007f847d68 ffffffff8109d0be 0000000000000006
[    0.504627] Call Trace:
[    0.504627]  [<ffffffff813d927a>] ? cpufreq_add_dev+0x13a/0x3e0
[    0.504627]  [<ffffffff814c062b>] dump_stack+0x19/0x1b
[    0.504627]  [<ffffffff81098bce>] print_unlock_imbalance_bug+0xfe/0x110
[    0.504627]  [<ffffffff813d927a>] ? cpufreq_add_dev+0x13a/0x3e0
[    0.504627]  [<ffffffff8109d0be>] lock_release_non_nested+0x1ee/0x310
[    0.504627]  [<ffffffff81099d0e>] ? mark_held_locks+0xae/0x120
[    0.504627]  [<ffffffff811510cb>] ? kfree+0xcb/0x1d0
[    0.504627]  [<ffffffff813d77ea>] ? cpufreq_policy_free+0x4a/0x60
[    0.504627]  [<ffffffff813d927a>] ? cpufreq_add_dev+0x13a/0x3e0
[    0.504627]  [<ffffffff8109d2a4>] lock_release+0xc4/0x250
[    0.504627]  [<ffffffff8106c9f3>] up_read+0x23/0x40
[    0.504627]  [<ffffffff813d927a>] cpufreq_add_dev+0x13a/0x3e0
[    0.504627]  [<ffffffff8134a809>] subsys_interface_register+0x99/0xe0
[    0.504627]  [<ffffffff81b19f3b>] ? cpufreq_gov_dbs_init+0x12/0x12
[    0.504627]  [<ffffffff813d7f0d>] cpufreq_register_driver+0x9d/0x1d0
[    0.504627]  [<ffffffff81b19f3b>] ? cpufreq_gov_dbs_init+0x12/0x12
[    0.504627]  [<ffffffff81b1a039>] acpi_cpufreq_init+0xfe/0x1f8
[    0.504627]  [<ffffffff810002ba>] do_one_initcall+0xda/0x180
[    0.504627]  [<ffffffff81ae301e>] kernel_init_freeable+0x12c/0x1bb
[    0.504627]  [<ffffffff81ae2841>] ? do_early_param+0x8c/0x8c
[    0.504627]  [<ffffffff814b4dd0>] ? rest_init+0x140/0x140
[    0.504627]  [<ffffffff814b4dde>] kernel_init+0xe/0xf0
[    0.504627]  [<ffffffff814d029a>] ret_from_fork+0x7a/0xb0
[    0.504627]  [<ffffffff814b4dd0>] ? rest_init+0x140/0x140

Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com>
Acked-and-tested-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq: Use cpufreq_policy_list for iterating over policies
Viresh Kumar [Tue, 20 Aug 2013 06:38:26 +0000 (12:08 +0530)]
cpufreq: Use cpufreq_policy_list for iterating over policies

To iterate over all policies we currently iterate over all online
CPUs and then get the policy for each of them which is suboptimal.
Use the newly created cpufreq_policy_list for this purpose instead.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq: remove cpufreq_policy_cpu per-cpu variable
Viresh Kumar [Tue, 20 Aug 2013 06:38:25 +0000 (12:08 +0530)]
cpufreq: remove cpufreq_policy_cpu per-cpu variable

cpufreq_policy_cpu per-cpu variables are used for storing the ID of
the CPU that manages the given CPU's policy.  However, we also store
a policy pointer for each cpu in cpufreq_cpu_data, so the
cpufreq_policy_cpu information is simply redundant.

It is better to use cpufreq_cpu_data to retrieve a policy and get
policy->cpu from there, so make that happen everywhere and drop the
cpufreq_policy_cpu per-cpu variables which aren't necessary any more.

[rjw: Changelog]
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq: remove unnecessary check in __cpufreq_governor()
Viresh Kumar [Tue, 20 Aug 2013 06:38:24 +0000 (12:08 +0530)]
cpufreq: remove unnecessary check in __cpufreq_governor()

We don't need to check if event is CPUFREQ_GOV_POLICY_INIT and put
governor module as we are sure event can only be START/STOP here.

Remove the useless check.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq: remove policy from cpufreq_policy_list during suspend
Viresh Kumar [Tue, 20 Aug 2013 06:38:23 +0000 (12:08 +0530)]
cpufreq: remove policy from cpufreq_policy_list during suspend

cpufreq_policy_list is a list of active policies.  We do remove
policies from this list when all CPUs belonging to that policy are
removed.  But during system suspend we don't really free a policy
struct as it will be used again during resume, so we didn't remove
it from cpufreq_policy_list as well..

However, this is incorrect.  We are saying this policy isn't valid
anymore and must not be referenced (though we haven't freed it), but
it can still be used by code that iterates over cpufreq_policy_list.

Remove policy from this list during system suspend as well.
Of course, we must add it back whenever the first CPU belonging to
that policy shows up.

[rjw: Changelog]
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq: Fix white space in __cpufreq_remove_dev()
Viresh Kumar [Tue, 20 Aug 2013 06:38:22 +0000 (12:08 +0530)]
cpufreq: Fix white space in __cpufreq_remove_dev()

Align closing brace '}' of an if block.

[rjw: Subject and changelog]
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agoLinux 3.11-rc6
Linus Torvalds [Sun, 18 Aug 2013 21:36:53 +0000 (14:36 -0700)]
Linux 3.11-rc6

11 years agoMerge branch 'for-3.11-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj...
Linus Torvalds [Sun, 18 Aug 2013 15:51:28 +0000 (08:51 -0700)]
Merge branch 'for-3.11-fixes' of git://git./linux/kernel/git/tj/cgroup

Pull cgroup fix from Tejun Heo:
 "This contains one patch to fix the return value of cpuset's cgroups
  interface function, which used to always return -ENODEV for the writes
  on the 'memory_pressure_enabled' file"

* 'for-3.11-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cpuset: fix the return value of cpuset_write_u64()

11 years agoRevert "cpufreq: Use cpufreq_policy_list for iterating over policies"
Rafael J. Wysocki [Sun, 18 Aug 2013 13:35:59 +0000 (15:35 +0200)]
Revert "cpufreq: Use cpufreq_policy_list for iterating over policies"

Revert commit eb60852 (cpufreq: Use cpufreq_policy_list for iterating
over policies), because it breaks system suspend/resume on multiple
machines.

It either causes resume to block indefinitely or causes the BUG_ON()
in lock_policy_rwsem_##mode() to trigger on sysfs accesses to cpufreq
attributes.

Conflicts:
drivers/cpufreq/cpufreq.c

11 years agoMerge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Sat, 17 Aug 2013 17:43:19 +0000 (10:43 -0700)]
Merge tag 'ext4_for_linus' of git://git./linux/kernel/git/tytso/ext4

Pull jbd2 bug fixes from Ted Ts'o:
 "Two jbd2 bug fixes, one of which is a regression fix"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  jbd2: Fix oops in jbd2_journal_file_inode()
  jbd2: Fix use after free after error in jbd2_journal_dirty_metadata()

11 years agos390: Fix broken build
Guenter Roeck [Sat, 17 Aug 2013 03:50:55 +0000 (20:50 -0700)]
s390: Fix broken build

Fix this build error:

  In file included from fs/exec.c:61:0:
  arch/s390/include/asm/tlb.h:35:23: error: expected identifier or '(' before 'unsigned'
  arch/s390/include/asm/tlb.h:36:1: warning: no semicolon at end of struct or union [enabled by default]
  arch/s390/include/asm/tlb.h: In function 'tlb_gather_mmu':
  arch/s390/include/asm/tlb.h:57:5: error: 'struct mmu_gather' has no member named 'end'

Broken due to commit 2b047252d0 ("Fix TLB gather virtual address range
invalidation corner cases").

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: stable@vger.kernel.org
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
[ Oh well. We had build testing for ppc amd um, but no s390  - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMAINTAINERS: Change ownership for SGI specific modules.
Robin Holt [Fri, 16 Aug 2013 23:01:42 +0000 (18:01 -0500)]
MAINTAINERS: Change ownership for SGI specific modules.

I have taken a different job.  I am removing myself as maintainer of
GRU.  Dimitri will continue to maintain the SGI GRU driver, changing the
XP/XPC/XPNET maintainer to Cliff Whickman, but leaving behind my
personal email address to answer any questions about the design or
operation of the XP family of drivers.

Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agojbd2: Fix oops in jbd2_journal_file_inode()
Jan Kara [Sat, 17 Aug 2013 01:19:41 +0000 (21:19 -0400)]
jbd2: Fix oops in jbd2_journal_file_inode()

Commit 0713ed0cde76438d05849f1537d3aab46e099475 added
jbd2_journal_file_inode() call into ext4_block_zero_page_range().
However that function gets called from truncate path and thus inode
needn't have jinode attached - that happens in ext4_file_open() but
the file needn't be ever open since mount. Calling
jbd2_journal_file_inode() without jinode attached results in the oops.

We fix the problem by attaching jinode to inode also in ext4_truncate()
and ext4_punch_hole() when we are going to zero out partial blocks.

Reported-by: majianpeng <majianpeng@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
11 years agoMerge branch 'fixes' of git://git.linaro.org/people/rmk/linux-arm
Linus Torvalds [Fri, 16 Aug 2013 23:52:29 +0000 (16:52 -0700)]
Merge branch 'fixes' of git://git.linaro.org/people/rmk/linux-arm

Pull ARM fixes from Russell King:
 "The usual collection of random fixes.  Also some further fixes to the
  last set of security fixes, and some more from Will (which you may
  already have in a slightly different form)"

* 'fixes' of git://git.linaro.org/people/rmk/linux-arm:
  ARM: 7807/1: kexec: validate CPU hotplug support
  ARM: 7812/1: rwlocks: retry trylock operation if strex fails on free lock
  ARM: 7811/1: locks: use early clobber in arch_spin_trylock
  ARM: 7810/1: perf: Fix array out of bounds access in armpmu_map_hw_event()
  ARM: 7809/1: perf: fix event validation for software group leaders
  ARM: Fix FIQ code on VIVT CPUs
  ARM: Fix !kuser helpers case
  ARM: Fix the world famous typo with is_gate_vma()

11 years agoMerge branch 'for-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux...
Linus Torvalds [Fri, 16 Aug 2013 23:49:06 +0000 (16:49 -0700)]
Merge branch 'for-3.11' of git://git./linux/kernel/git/geert/linux-m68k

Pull m68k fixes from Geert Uytterhoeven:
 "These are two critical fixes, needed by distro kernels, and thus also
  destined for stable:

   - The do_div() commit fixes a crash in mounting btrfs volumes, which
     was a regression from 3.2,

   - The ARAnyM fix allows to have NatFeat drivers as loadable modules,
     which is needed for initrds"

* 'for-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k: Truncate base in do_div()
  m68k/atari: ARAnyM - Fix NatFeat module support

11 years agoMerge tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mturquette/linux
Linus Torvalds [Fri, 16 Aug 2013 17:00:18 +0000 (10:00 -0700)]
Merge tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mturquette/linux

Pull clock controller fixes from Michael Turquette:
 "Two small fixes for the Zynq clock controller introduced in 3.11-rc1
  and another Exynos clock patch which fixes a regression that prevents
  the video pipeline from functioning on that platform"

* tag 'clk-fixes-for-linus' of git://git.linaro.org/people/mturquette/linux:
  clk: exynos4: Add CLK_GET_RATE_NOCACHE flag for the Exynos4x12 ISP clocks
  clk/zynq/clkc: Add CLK_SET_RATE_PARENT flag to ethernet muxes
  clk/zynq/clkc: Add dedicated spinlock for the SWDT

11 years agoMerge tag 'pm-3.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Linus Torvalds [Fri, 16 Aug 2013 16:59:00 +0000 (09:59 -0700)]
Merge tag 'pm-3.11-rc6' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fix from Rafael Wysocki:
 "The removal of delayed_work_pending() checks from kernel/power/qos.c
  done in 3.9 introduced a deadlock in pm_qos_work_fn().

  Fix from Stephen Boyd"

* tag 'pm-3.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  PM / QoS: Fix workqueue deadlock when using pm_qos_update_request_timeout()

11 years agoMerge tag 'sound-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Linus Torvalds [Fri, 16 Aug 2013 16:58:21 +0000 (09:58 -0700)]
Merge tag 'sound-3.11' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "This batch contains a few USB audio fixes, a couple of HD-audio
  quirks, various small ASoC driver fixes in addition to an ASoC core
  fix that may lead to memory corruption.

  Unfortunately slightly more volume than the previous pull request, but
  all are reasonable regression fixes"

* tag 'sound-3.11' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - Add a fixup for Gateway LT27
  ASoC: tegra: fix Tegra30 I2S capture parameter setup
  ALSA: usb-audio: Fix invalid volume resolution for Logitech HD Webcam C525
  ALSA: hda - Fix missing mute controls for CX5051
  ALSA: usb-audio: fix automatic Roland/Yamaha MIDI detection
  ALSA: 6fire: make buffers DMA-able (midi)
  ALSA: 6fire: make buffers DMA-able (pcm)
  ALSA: hda - Add pinfix for LG LW25 laptop
  ASoC: cs42l52: Add new TLV for Beep Volume
  ASoC: cs42l52: Reorder Min/Max and update to SX_TLV for Beep Volume
  ASoC: dapm: Fix empty list check in dapm_new_mux()
  ASoC: sgtl5000: fix buggy 'Capture Attenuate Switch' control
  ASoC: sgtl5000: prevent playback to be muted when terminating concurrent capture

11 years agoMerge tag 'usb-3.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 16 Aug 2013 16:57:38 +0000 (09:57 -0700)]
Merge tag 'usb-3.11-rc6' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some small USB fixes for 3.11-rc6 that have accumulated.

  Nothing huge, a EHCI fix that solves a much-reported audio USB
  problem, some usb-serial driver endian fixes and other minor fixes, a
  wireless USB oops fix, and two new quirks"

* tag 'usb-3.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  USB: keyspan: fix null-deref at disconnect and release
  USB: mos7720: fix broken control requests
  usb: add two quirky touchscreen
  USB: ti_usb_3410_5052: fix big-endian firmware handling
  USB: adutux: fix big-endian device-type reporting
  USB: usbtmc: fix big-endian probe of Rigol devices
  USB: mos7840: fix big-endian probe
  USB-Serial: Fix error handling of usb_wwan
  wusbcore: fix kernel panic when disconnecting a wireless USB->serial device
  USB: EHCI: accept very late isochronous URBs

11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 16 Aug 2013 16:35:29 +0000 (09:35 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Fix SKB leak in 8139cp, from Dave Jones.

 2) Fix use of *_PAGES interfaces with mlx5 firmware, from Moshe Lazar.

 3) RCU conversion of macvtap introduced two races, fixes by Eric
    Dumazet

 4) Synchronize statistic flows in bnx2x driver to prevent corruption,
    from Dmitry Kravkov

 5) Undo optimization in IP tunneling, we were using the inner IP header
    in some cases to inherit the IP ID, but that isn't correct in some
    circumstances.  From Pravin B Shelar

 6) Use correct struct size when parsing netlink attributes in
    rtnl_bridge_getlink().  From Asbjoern Sloth Toennesen

 7) Length verifications in tun_get_user() are bogus, from Weiping Pan
    and Dan Carpenter

 8) Fix bad merge resolution during 3.11 networking development in
    openvswitch, albeit a harmless one which added some unreachable
    code.  From Jesse Gross

 9) Wrong size used in flexible array allocation in openvswitch, from
    Pravin B Shelar

10) Clear out firmware capability flags the be2net driver isn't ready to
    handle yet, from Sarveshwar Bandi

11) Revert DMA mapping error checking addition to cxgb3 driver, it's
    buggy.  From Alexey Kardashevskiy

12) Fix regression in packet scheduler rate limiting when working with a
    link layer of ATM.  From Jesper Dangaard Brouer

13) Fix several errors in TCP Cubic congestion control, in particular
    overflow errors in timestamp calculations.  From Eric Dumazet and
    Van Jacobson

14) In ipv6 routing lookups, we need to backtrack if subtree traversal
    don't result in a match.  From Hannes Frederic Sowa

15) ipgre_header() returns incorrect packet offset.  Fix from Timo Teräs

16) Get "low latency" out of the new MIB counter names.  From Eliezer
    Tamir

17) State check in ndo_dflt_fdb_del() is inverted, from Sridhar
    Samudrala

18) Handle TCP Fast Open properly in netfilter conntrack, from Yuchung
    Cheng

19) Wrong memcpy length in pcan_usb driver, from Stephane Grosjean

20) Fix dealock in TIPC, from Wang Weidong and Ding Tianhong

21) call_rcu() call to destroy SCTP transport is done too early and
    might result in an oops.  From Daniel Borkmann

22) Fix races in genetlink family dumps, from Johannes Berg

23) Flags passed into macvlan by the user need to be validated properly,
    from Michael S Tsirkin

24) Fix skge build on 32-bit, from Stephen Hemminger

25) Handle malformed TCP headers properly in xt_TCPMSS, from Pablo Neira
    Ayuso

26) Fix handling of stacked vlans in vlan_dev_real_dev(), from Nikolay
    Aleksandrov

27) Eliminate MTU calculation overflows in esp{4,6}, from Daniel
    Borkmann

28) neigh_parms need to be setup before calling the ->ndo_neigh_setup()
    method.  From Veaceslav Falico

29) Kill out-of-bounds prefetch in fib_trie, from Eric Dumazet

30) Don't dereference MLD query message if the length isn't value in the
    bridge multicast code, from Linus LĂ¼ssing

31) Fix VXLAN IGMP join regression due to an inverted check, from Cong
    Wang

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (70 commits)
  net/mlx5_core: Support MANAGE_PAGES and QUERY_PAGES firmware command changes
  tun: signedness bug in tun_get_user()
  qlcnic: Fix diagnostic interrupt test for 83xx adapters
  qlcnic: Fix beacon state return status handling
  qlcnic: Fix set driver version command
  net: tg3: fix NULL pointer dereference in tg3_io_error_detected and tg3_io_slot_reset
  net_sched: restore "linklayer atm" handling
  drivers/net/ethernet/via/via-velocity.c: update napi implementation
  Revert "cxgb3: Check and handle the dma mapping errors"
  be2net: Clear any capability flags that driver is not interested in.
  openvswitch: Reset tunnel key between input and output.
  openvswitch: Use correct type while allocating flex array.
  openvswitch: Fix bad merge resolution.
  tun: compare with 0 instead of total_len
  rtnetlink: rtnl_bridge_getlink: Call nlmsg_find_attr() with ifinfomsg header
  ethernet/arc/arc_emac - fix NAPI "work > weight" warning
  ip_tunnel: Do not use inner ip-header-id for tunnel ip-header-id.
  bnx2x: prevent crash in shutdown flow with CNIC
  bnx2x: fix PTE write access error
  bnx2x: fix memory leak in VF
  ...

11 years agoFix TLB gather virtual address range invalidation corner cases
Linus Torvalds [Thu, 15 Aug 2013 18:42:25 +0000 (11:42 -0700)]
Fix TLB gather virtual address range invalidation corner cases

Ben Tebulin reported:

 "Since v3.7.2 on two independent machines a very specific Git
  repository fails in 9/10 cases on git-fsck due to an SHA1/memory
  failures.  This only occurs on a very specific repository and can be
  reproduced stably on two independent laptops.  Git mailing list ran
  out of ideas and for me this looks like some very exotic kernel issue"

and bisected the failure to the backport of commit 53a59fc67f97 ("mm:
limit mmu_gather batching to fix soft lockups on !CONFIG_PREEMPT").

That commit itself is not actually buggy, but what it does is to make it
much more likely to hit the partial TLB invalidation case, since it
introduces a new case in tlb_next_batch() that previously only ever
happened when running out of memory.

The real bug is that the TLB gather virtual memory range setup is subtly
buggered.  It was introduced in commit 597e1c3580b7 ("mm/mmu_gather:
enable tlb flush range in generic mmu_gather"), and the range handling
was already fixed at least once in commit e6c495a96ce0 ("mm: fix the TLB
range flushed when __tlb_remove_page() runs out of slots"), but that fix
was not complete.

The problem with the TLB gather virtual address range is that it isn't
set up by the initial tlb_gather_mmu() initialization (which didn't get
the TLB range information), but it is set up ad-hoc later by the
functions that actually flush the TLB.  And so any such case that forgot
to update the TLB range entries would potentially miss TLB invalidates.

Rather than try to figure out exactly which particular ad-hoc range
setup was missing (I personally suspect it's the hugetlb case in
zap_huge_pmd(), which didn't have the same logic as zap_pte_range()
did), this patch just gets rid of the problem at the source: make the
TLB range information available to tlb_gather_mmu(), and initialize it
when initializing all the other tlb gather fields.

This makes the patch larger, but conceptually much simpler.  And the end
result is much more understandable; even if you want to play games with
partial ranges when invalidating the TLB contents in chunks, now the
range information is always there, and anybody who doesn't want to
bother with it won't introduce subtle bugs.

Ben verified that this fixes his problem.

Reported-bisected-and-tested-by: Ben Tebulin <tebulin@googlemail.com>
Build-testing-by: Stephen Rothwell <sfr@canb.auug.org.au>
Build-testing-by: Richard Weinberger <richard.weinberger@gmail.com>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoALSA: hda - Add a fixup for Gateway LT27
Takashi Iwai [Fri, 16 Aug 2013 06:17:05 +0000 (08:17 +0200)]
ALSA: hda - Add a fixup for Gateway LT27

Gateway LT27 needs a fixup for the inverted digital mic.

Reported-by: "Nathanael D. Noblet" <nathanael@gnat.ca>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
11 years agonet/mlx5_core: Support MANAGE_PAGES and QUERY_PAGES firmware command changes
Moshe Lazer [Wed, 14 Aug 2013 14:46:48 +0000 (17:46 +0300)]
net/mlx5_core: Support MANAGE_PAGES and QUERY_PAGES firmware command changes

In the previous QUERY_PAGES command version we used one command to get the
required amount of boot, init and post init pages.  The new version uses the
op_mod field to specify whether the query is for the required amount of boot,
init or post init pages. In addition the output field size for the required
amount of pages increased from 16 to 32 bits.

In MANAGE_PAGES command the input_num_entries and output_num_entries fields
sizes changed from 16 to 32 bits and the PAS tables offset changed to 0x10.

In the pages request event the num_pages field also changed to 32 bits.

In the HCA-capabilities-layout the size and location of max_qp_mcg field has
been changed to support 24 bits.

This patch isn't compatible with firmware versions < 5; however, it  turns out that the
first GA firmware we will publish will not support previous versions so this should be OK.

Signed-off-by: Moshe Lazer <moshel@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotun: signedness bug in tun_get_user()
Dan Carpenter [Thu, 15 Aug 2013 12:52:57 +0000 (15:52 +0300)]
tun: signedness bug in tun_get_user()

The recent fix d9bf5f1309 "tun: compare with 0 instead of total_len" is
not totally correct.  Because "len" and "sizeof()" are size_t type, that
means they are never less than zero.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Fix diagnostic interrupt test for 83xx adapters
Manish Chopra [Thu, 15 Aug 2013 12:29:29 +0000 (08:29 -0400)]
qlcnic: Fix diagnostic interrupt test for 83xx adapters

o Do not allow interrupt test when adapter is resetting.

Signed-off-by: Manish Chopra <manish.chopra@qlogic.com>
Signed-off-by: Sucheta Chakraborty <sucheta.chakraborty@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Fix beacon state return status handling
Sucheta Chakraborty [Thu, 15 Aug 2013 12:29:28 +0000 (08:29 -0400)]
qlcnic: Fix beacon state return status handling

o Driver was misinterpreting the return status for beacon
  state query leading to incorrect interpretation of beacon
  state and logging an error message for successful status.
  Fixed the driver to properly interpret the return status.

Signed-off-by: Sucheta Chakraborty <sucheta.chakraborty@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Fix set driver version command
Himanshu Madhani [Thu, 15 Aug 2013 12:29:27 +0000 (08:29 -0400)]
qlcnic: Fix set driver version command

Driver was issuing set driver version command through all
functions in the adapter. Fix the driver to issue set driver
version once per adapter, through function 0.

Signed-off-by: Himanshu Madhani <himanshu.madhani@qlogic.com>
Signed-off-by: Sucheta Chakraborty <sucheta.chakraborty@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: tg3: fix NULL pointer dereference in tg3_io_error_detected and tg3_io_slot_reset
Daniel Borkmann [Tue, 13 Aug 2013 18:45:13 +0000 (11:45 -0700)]
net: tg3: fix NULL pointer dereference in tg3_io_error_detected and tg3_io_slot_reset

Commit d8af4dfd8 ("net/tg3: Fix kernel crash") introduced a possible
NULL pointer dereference in tg3 driver when !netdev || !netif_running(netdev)
condition is met and netdev is NULL. Then, the jump to the 'done' label
calls dev_close() with a netdevice that is NULL. Therefore, only call
dev_close() when we have a netdevice, but one that is not running.

[ Add the same checks in tg3_io_slot_reset() per Gavin Shan - by Nithin
Nayak Sujir ]

Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Gavin Shan <shangw@linux.vnet.ibm.com>
Cc: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge tag 'asoc-v3.11-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie...
Takashi Iwai [Thu, 15 Aug 2013 18:43:46 +0000 (20:43 +0200)]
Merge tag 'asoc-v3.11-rc5' of git://git./linux/kernel/git/broonie/sound into for-linus

ASoC: Fixes for v3.11

A few driver specific fixes here plus one core fix for a memory
corruption issue in DAPM initialisation which could lead to crashes.

11 years agoMerge remote-tracking branch 'asoc/fix/tegra' into asoc-linus
Mark Brown [Thu, 15 Aug 2013 10:37:54 +0000 (11:37 +0100)]
Merge remote-tracking branch 'asoc/fix/tegra' into asoc-linus

11 years agoMerge remote-tracking branch 'asoc/fix/sgtl5000' into asoc-linus
Mark Brown [Thu, 15 Aug 2013 10:37:53 +0000 (11:37 +0100)]
Merge remote-tracking branch 'asoc/fix/sgtl5000' into asoc-linus

11 years agoMerge remote-tracking branch 'asoc/fix/dapm' into asoc-linus
Mark Brown [Thu, 15 Aug 2013 10:37:53 +0000 (11:37 +0100)]
Merge remote-tracking branch 'asoc/fix/dapm' into asoc-linus

11 years agoMerge remote-tracking branch 'asoc/fix/cs42l52' into asoc-linus
Mark Brown [Thu, 15 Aug 2013 10:37:52 +0000 (11:37 +0100)]
Merge remote-tracking branch 'asoc/fix/cs42l52' into asoc-linus

11 years agoASoC: tegra: fix Tegra30 I2S capture parameter setup
Stephen Warren [Wed, 14 Aug 2013 20:24:16 +0000 (14:24 -0600)]
ASoC: tegra: fix Tegra30 I2S capture parameter setup

The Tegra30 I2S driver was writing the AHUB interface parameters to the
playback path register rather than the capture path register. This
caused the capture parameters not to be configured at all, so if
capturing using non-HW-default parameters (e.g. 16-bit stereo rather
than 8-bit mono) the audio would be corrupted.

With this fixed, audio capture from an analog microphone works correctly
on the Cardhu board.

Cc: stable@vger.kernel.org
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
11 years agonet_sched: restore "linklayer atm" handling
Jesper Dangaard Brouer [Wed, 14 Aug 2013 21:47:11 +0000 (23:47 +0200)]
net_sched: restore "linklayer atm" handling

commit 56b765b79 ("htb: improved accuracy at high rates")
broke the "linklayer atm" handling.

 tc class add ... htb rate X ceil Y linklayer atm

The linklayer setting is implemented by modifying the rate table
which is send to the kernel.  No direct parameter were
transferred to the kernel indicating the linklayer setting.

The commit 56b765b79 ("htb: improved accuracy at high rates")
removed the use of the rate table system.

To keep compatible with older iproute2 utils, this patch detects
the linklayer by parsing the rate table.  It also supports future
versions of iproute2 to send this linklayer parameter to the
kernel directly. This is done by using the __reserved field in
struct tc_ratespec, to convey the choosen linklayer option, but
only using the lower 4 bits of this field.

Linklayer detection is limited to speeds below 100Mbit/s, because
at high rates the rtab is gets too inaccurate, so bad that
several fields contain the same values, this resembling the ATM
detect.  Fields even start to contain "0" time to send, e.g. at
1000Mbit/s sending a 96 bytes packet cost "0", thus the rtab have
been more broken than we first realized.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jesse/openvswitch
David S. Miller [Thu, 15 Aug 2013 08:41:10 +0000 (01:41 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/jesse/openvswitch

Jesse Gross says:

====================
Three bug fixes that are fairly small either way but resolve obviously
incorrect code. For net/3.11.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agodrivers/net/ethernet/via/via-velocity.c: update napi implementation
Julia Lawall [Wed, 14 Aug 2013 14:26:53 +0000 (16:26 +0200)]
drivers/net/ethernet/via/via-velocity.c: update napi implementation

Drivers supporting NAPI should use a NAPI-specific function for receiving
packets.  Hence netif_rx is changed to netif_receive_skb.

Furthermore netif_napi_del should be used in the probe and remove function
to clean up the NAPI resource information.

Thanks to Francois Romieu, David Shwatrz and Rami Rosen for their help on
this patch.

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoRevert "cxgb3: Check and handle the dma mapping errors"
Alexey Kardashevskiy [Wed, 14 Aug 2013 09:19:01 +0000 (19:19 +1000)]
Revert "cxgb3: Check and handle the dma mapping errors"

This reverts commit f83331bab149e29fa2c49cf102c0cd8c3f1ce9f9.

As the tests PPC64 (powernv platform) show, IOMMU pages are leaking
when transferring big amount of small packets (<=64 bytes),
"ping -f" and waiting for 15 seconds is the simplest way to confirm the bug.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Santosh Rastapur <santosh@chelsio.com>
Cc: Jay Fenlason <fenlason@redhat.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Divy Le ray <divy@chelsio.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Acked-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobe2net: Clear any capability flags that driver is not interested in.
Sarveshwar Bandi [Wed, 14 Aug 2013 07:51:47 +0000 (13:21 +0530)]
be2net: Clear any capability flags that driver is not interested in.

It is possible for some versions of firmware to advertise capabilities that driver
is not ready to handle. This may lead to controller stall. Since the driver is
interested only in subset of flags, clearing the rest.

Signed-off-by: Sarveshwar Bandi <sarveshwar.bandi@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoopenvswitch: Reset tunnel key between input and output.
Jesse Gross [Wed, 14 Aug 2013 22:50:36 +0000 (15:50 -0700)]
openvswitch: Reset tunnel key between input and output.

It doesn't make sense to output a tunnel packet using the same
parameters that it was received with since that will generally
just result in the packet going back to us. As a result, userspace
assumes that the tunnel key is cleared when transitioning through
the switch. In the majority of cases this doesn't matter since a
packet is either going to a tunnel port (in which the key is
overwritten with new values) or to a non-tunnel port (in which
case the key is ignored). However, it's theoreticaly possible that
userspace could rely on the documented behavior, so this corrects
it.

Signed-off-by: Jesse Gross <jesse@nicira.com>
11 years agoopenvswitch: Use correct type while allocating flex array.
Pravin B Shelar [Tue, 30 Jul 2013 22:44:14 +0000 (15:44 -0700)]
openvswitch: Use correct type while allocating flex array.

Flex array is used to allocate hash buckets which is type struct
hlist_head, but we use `struct hlist_head *` to calculate
array size.  Since hlist_head is of size pointer it works fine.

Following patch use correct type.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
11 years agoopenvswitch: Fix bad merge resolution.
Jesse Gross [Mon, 13 May 2013 15:41:06 +0000 (08:41 -0700)]
openvswitch: Fix bad merge resolution.

git silently included an extra hunk in vport_cmd_set() during
automatic merging. This code is unreachable so it does not actually
introduce a problem but it is clearly incorrect.

Signed-off-by: Jesse Gross <jesse@nicira.com>
11 years agocpufreq: unicore2: Staticize local symbol
Jingoo Han [Fri, 9 Aug 2013 07:14:51 +0000 (16:14 +0900)]
cpufreq: unicore2: Staticize local symbol

This local symbol is used only in this file.
Fix the following sparse warnings:

drivers/cpufreq/unicore2-cpufreq.c:27:5: warning: symbol 'ucv2_verify_speed' was not declared. Should it be static?

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq / s3c24xx: Fix s3c_cpufreq_initclks() __init attribute location
Hanjun Guo [Tue, 13 Aug 2013 10:20:10 +0000 (18:20 +0800)]
cpufreq / s3c24xx: Fix s3c_cpufreq_initclks() __init attribute location

__init belongs after the return type on functions, not before it.

Signed-off-by: Hanjun Guo <hanjun.guo@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq / pxa2xx: Fix pxa_cpufreq_init_voltages() __init attribute location
Hanjun Guo [Tue, 13 Aug 2013 10:20:09 +0000 (18:20 +0800)]
cpufreq / pxa2xx: Fix pxa_cpufreq_init_voltages() __init attribute location

__init belongs after the return type on functions, not before it.

Signed-off-by: Hanjun Guo <hanjun.guo@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq / gx: Fix gx_detect_chipset() __init attribute location
Hanjun Guo [Tue, 13 Aug 2013 10:20:08 +0000 (18:20 +0800)]
cpufreq / gx: Fix gx_detect_chipset() __init attribute location

__init belongs after the return type on functions, not before it.

Signed-off-by: Hanjun Guo <hanjun.guo@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agopxa3xx-cpufreq.c: Avoid using ARRAY_AND_SIZE(e) as a function argument
Julia Lawall [Tue, 13 Aug 2013 13:02:25 +0000 (15:02 +0200)]
pxa3xx-cpufreq.c: Avoid using ARRAY_AND_SIZE(e) as a function argument

Replace ARRAY_AND_SIZE(e) in function argument position to avoid
hiding the arity of the called function.

The semantic match that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e,f;
@@

f(...,
- ARRAY_AND_SIZE(e)
+ e,ARRAY_SIZE(e)
  ,...)
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agocpufreq: cpufreq-cpu0: NULL is a valid regulator
Mark Brown [Tue, 13 Aug 2013 12:58:24 +0000 (14:58 +0200)]
cpufreq: cpufreq-cpu0: NULL is a valid regulator

Since NULL could in theory be a valid regulator we ought to check for
IS_ERR() rather than for NULL. In practice this is unlikely to be an
issue but it's better for neatness.

Signed-off-by: Mark Brown <broonie@linaro.org>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agoacpi-cpufreq: Use cpufreq_freq_attr_rw to define the cpb attribute
Lan Tianyu [Tue, 13 Aug 2013 02:05:53 +0000 (10:05 +0800)]
acpi-cpufreq: Use cpufreq_freq_attr_rw to define the cpb attribute

Standardise the defintion of the cpb (Core Performance Boost)
attribute in the acpi-cpufreq driver via the cpufreq_freq_attr_rw
macro.

[rjw: Subject and changelog]
Signed-off-by: Lan Tianyu <tianyu.lan@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agoMerge branch 'cpufreq-fixes' of git://git.linaro.org/people/vireshk/linux into pm...
Rafael J. Wysocki [Wed, 14 Aug 2013 20:22:57 +0000 (22:22 +0200)]
Merge branch 'cpufreq-fixes' of git://git.linaro.org/people/vireshk/linux into pm-cpufreq

Pull ARM cpufreq fixes from Viresh Kumar.

* 'cpufreq-fixes' of git://git.linaro.org/people/vireshk/linux:
  cpufreq: fix EXYNOS drivers selection
  cpufreq: exynos5440: Fix to skip when new frequency same as current

11 years agoMerge back earlier 'pm-cpufreq' material
Rafael J. Wysocki [Wed, 14 Aug 2013 20:21:16 +0000 (22:21 +0200)]
Merge back earlier 'pm-cpufreq' material

11 years agoUSB: keyspan: fix null-deref at disconnect and release
Johan Hovold [Tue, 13 Aug 2013 11:27:35 +0000 (13:27 +0200)]
USB: keyspan: fix null-deref at disconnect and release

Make sure to fail properly if the device is not accepted during attach
in order to avoid null-pointer derefs (of missing interface private
data) at disconnect or release.

Cc: stable@vger.kernel.org
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoUSB: mos7720: fix broken control requests
Johan Hovold [Tue, 13 Aug 2013 11:27:34 +0000 (13:27 +0200)]
USB: mos7720: fix broken control requests

The parallel-port code of the drivers used a stack allocated
control-request buffer for asynchronous (and possibly deferred) control
requests. This not only violates the no-DMA-from-stack requirement but
could also lead to corrupt control requests being submitted.

Cc: stable@vger.kernel.org
Signed-off-by: Johan Hovold <jhovold@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agousb: add two quirky touchscreen
Oliver Neukum [Wed, 14 Aug 2013 09:01:46 +0000 (11:01 +0200)]
usb: add two quirky touchscreen

These devices tend to become unresponsive after S3

Signed-off-by: Oliver Neukum <oneukum@suse.de>
CC: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
11 years agoMerge branch 'akpm' (patches from Andrew Morton)
Linus Torvalds [Wed, 14 Aug 2013 17:04:43 +0000 (10:04 -0700)]
Merge branch 'akpm' (patches from Andrew Morton)

Merge a bunch of fixes from Andrew Morton.

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  fs/proc/task_mmu.c: fix buffer overflow in add_page_map()
  arch: *: Kconfig: add "kernel/Kconfig.freezer" to "arch/*/Kconfig"
  ocfs2: fix null pointer dereference in ocfs2_dir_foreach_blk_id()
  x86 get_unmapped_area(): use proper mmap base for bottom-up direction
  ocfs2: fix NULL pointer dereference in ocfs2_duplicate_clusters_by_page
  ocfs2: Revert 40bd62e to avoid regression in extended allocation
  drivers/rtc/rtc-stmp3xxx.c: provide timeout for potentially endless loop polling a HW bit
  hugetlb: fix lockdep splat caused by pmd sharing
  aoe: adjust ref of head for compound page tails
  microblaze: fix clone syscall
  mm: save soft-dirty bits on file pages
  mm: save soft-dirty bits on swapped pages
  memcg: don't initialize kmem-cache destroying work for root caches

11 years agom68k: Truncate base in do_div()
Andreas Schwab [Fri, 9 Aug 2013 13:14:08 +0000 (15:14 +0200)]
m68k: Truncate base in do_div()

Explicitly truncate the second operand of do_div() to 32 bits to guard
against bogus code calling it with a 64-bit divisor.

[Thorsten]

After upgrading from 3.2 to 3.10, mounting a btrfs volume fails with:

btrfs: setting nodatacow, compression disabled
btrfs: enabling auto recovery
btrfs: disk space caching is enabled
*** ZERO DIVIDE ***   FORMAT=2
Current process id is 722
BAD KERNEL TRAP: 00000000
Modules linked in: evdev mac_hid ext4 crc16 jbd2 mbcache btrfs xor lzo_compress zlib_deflate raid6_pq crc32c libcrc32c
PC: [<319535b2>] __btrfs_map_block+0x11c/0x119a [btrfs]
SR: 2000  SP: 30c1fab4  a2: 30f0faf0
d0: 00000000    d1: 00001000    d2: 00000000    d3: 00000000
d4: 00010000    d5: 00000000    a0: 3085c72c    a1: 3085c72c
Process mount (pid: 722, task=30f0faf0)
Frame format=2 instr addr=319535ae
Stack from 30c1faec:
        00000000 00000020 00000000 00001000 00000000 01401000 30253928 300ffc00
        00a843ac 3026f640 00000000 00010000 0009e250 00d106c0 00011220 00000000
        00001000 301c6830 0009e32a 000000ff 00000009 3085c72c 00000000 00000000
        30c1fd14 00000000 00000020 00000000 30c1fd14 0009e26c 00000020 00000003
        00000000 0009dd8a 300b0b6c 30253928 00a843ac 00001000 00000000 00000000
        0000a008 3194e76a 30253928 00a843ac 00001000 00000000 00000000 00000002
Call Trace: [<00001000>] kernel_pg_dir+0x0/0x1000

    [...]

Code: 222e ff74 2a2e ff5c 2c2e ff60 4c45 1402 <2d40> ff64 2d41 ff68 2205 4c2e 1800 ff68 4c04 0800 2041 d1c0 2206 4c2e 1400 ff68

[Geert]

As diagnosed by Andreas, fs/btrfs/volumes.c:__btrfs_map_block()
calls

    do_div(stripe_nr, stripe_len);

with stripe_len u64, while do_div() assumes the divisor is a 32-bit number.

Due to the lack of truncation in the m68k-specific implementation of
do_div(), the division is performed using the upper 32-bit word of
stripe_len, which is zero.

This was introduced by commit 53b381b3abeb86f12787a6c40fee9b2f71edc23b
("Btrfs: RAID5 and RAID6"), which changed the divisor from
map->stripe_len (struct map_lookup.stripe_len is int) to a 64-bit temporary.

Reported-by: Thorsten Glaser <tg@debian.org>
Signed-off-by: Andreas Schwab <schwab@linux-m68k.org>
Tested-by: Thorsten Glaser <tg@debian.org>
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: stable@vger.kernel.org
11 years agom68k/atari: ARAnyM - Fix NatFeat module support
Geert Uytterhoeven [Thu, 25 Jul 2013 22:08:25 +0000 (00:08 +0200)]
m68k/atari: ARAnyM - Fix NatFeat module support

As pointed out by Andreas Schwab, pointers passed to ARAnyM NatFeat calls
should be physical addresses, not virtual addresses.

Fortunately on Atari, physical and virtual kernel addresses are the same,
as long as normal kernel memory is concerned, so this usually worked fine
without conversion.

But for modules, pointers to literal strings are located in vmalloc()ed
memory. Depending on the version of ARAnyM, this causes the nf_get_id()
call to just fail, or worse, crash ARAnyM itself with e.g.

    Gotcha! Illegal memory access. Atari PC = $968c

This is a big issue for distro kernels, who want to have all drivers as
loadable modules in an initrd.

Add a wrapper for nf_get_id() that copies the literal to the stack to
work around this issue.

Reported-by: Thorsten Glaser <tg@debian.org>
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: stable@vger.kernel.org
11 years agotun: compare with 0 instead of total_len
Weiping Pan [Tue, 13 Aug 2013 13:46:56 +0000 (21:46 +0800)]
tun: compare with 0 instead of total_len

Since we set "len = total_len" in the beginning of tun_get_user(),
so we should compare the new len with 0, instead of total_len,
or the if statement always returns false.

Signed-off-by: Weiping Pan <wpan@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agortnetlink: rtnl_bridge_getlink: Call nlmsg_find_attr() with ifinfomsg header
Asbjoern Sloth Toennesen [Mon, 12 Aug 2013 16:30:09 +0000 (16:30 +0000)]
rtnetlink: rtnl_bridge_getlink: Call nlmsg_find_attr() with ifinfomsg header

Fix the iproute2 command `bridge vlan show`, after switching from
rtgenmsg to ifinfomsg.

Let's start with a little history:

Feb 20:   Vlad Yasevich got his VLAN-aware bridge patchset included in
          the 3.9 merge window.
          In the kernel commit 6cbdceeb, he added attribute support to
          bridge GETLINK requests sent with rtgenmsg.

Mar 6th:  Vlad got this iproute2 reference implementation of the bridge
          vlan netlink interface accepted (iproute2 9eff0e5c)

Apr 25th: iproute2 switched from using rtgenmsg to ifinfomsg (63338dca)
          http://patchwork.ozlabs.org/patch/239602/
          http://marc.info/?t=136680900700007

Apr 28th: Linus released 3.9

Apr 30th: Stephen released iproute2 3.9.0

The `bridge vlan show` command haven't been working since the switch to
ifinfomsg, or in a released version of iproute2. Since the kernel side
only supports rtgenmsg, which iproute2 switched away from just prior to
the iproute2 3.9.0 release.

I haven't been able to find any documentation, about neither rtgenmsg
nor ifinfomsg, and in which situation to use which, but kernel commit
88c5b5ce seams to suggest that ifinfomsg should be used.

Fixing this in kernel will break compatibility, but I doubt that anybody
have been using it due to this bug in the user space reference
implementation, at least not without noticing this bug. That said the
functionality is still fully functional in 3.9, when reversing iproute2
commit 63338dca.

This could also be fixed in iproute2, but thats an ugly patch that would
reintroduce rtgenmsg in iproute2, and from searching in netdev it seams
like rtgenmsg usage is discouraged. I'm assuming that the only reason
that Vlad implemented the kernel side to use rtgenmsg, was because
iproute2 was using it at the time.

Signed-off-by: Asbjoern Sloth Toennesen <ast@fiberby.net>
Reviewed-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofs/proc/task_mmu.c: fix buffer overflow in add_page_map()
yonghua zheng [Tue, 13 Aug 2013 23:01:03 +0000 (16:01 -0700)]
fs/proc/task_mmu.c: fix buffer overflow in add_page_map()

Recently we met quite a lot of random kernel panic issues after enabling
CONFIG_PROC_PAGE_MONITOR.  After debuggind we found this has something
to do with following bug in pagemap:

In struct pagemapread:

  struct pagemapread {
      int pos, len;
      pagemap_entry_t *buffer;
      bool v2;
  };

pos is number of PM_ENTRY_BYTES in buffer, but len is the size of
buffer, it is a mistake to compare pos and len in add_page_map() for
checking buffer is full or not, and this can lead to buffer overflow and
random kernel panic issue.

Correct len to be total number of PM_ENTRY_BYTES in buffer.

[akpm@linux-foundation.org: document pagemapread.pos and .len units, fix PM_ENTRY_BYTES definition]
Signed-off-by: Yonghua Zheng <younghua.zheng@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoarch: *: Kconfig: add "kernel/Kconfig.freezer" to "arch/*/Kconfig"
Chen Gang [Tue, 13 Aug 2013 23:01:02 +0000 (16:01 -0700)]
arch: *: Kconfig: add "kernel/Kconfig.freezer" to "arch/*/Kconfig"

All architectures include "kernel/Kconfig.freezer" except three left, so
let them include it too, or 'allmodconfig' will report error.

The related errors: (with allmodconfig for openrisc):

    CC      kernel/cgroup_freezer.o
  kernel/cgroup_freezer.c: In function 'freezer_css_online':
  kernel/cgroup_freezer.c:133:15: error: 'system_freezing_cnt' undeclared (first use in this function)
  kernel/cgroup_freezer.c:133:15: note: each undeclared identifier is reported only once for each function it appears in
  kernel/cgroup_freezer.c: In function 'freezer_css_offline':
  kernel/cgroup_freezer.c:157:15: error: 'system_freezing_cnt' undeclared (first use in this function)
  kernel/cgroup_freezer.c: In function 'freezer_attach':
  kernel/cgroup_freezer.c:200:4: error: implicit declaration of function 'freeze_task'
  kernel/cgroup_freezer.c: In function 'freezer_apply_state':
  kernel/cgroup_freezer.c:371:16: error: 'system_freezing_cnt' undeclared (first use in this function)

Signed-off-by: Chen Gang <gang.chen@asianux.com>
Cc: Richard Kuo <rkuo@codeaurora.org>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Chen Liqin <liqin.chen@sunplusct.com>
Cc: Lennox Wu <lennox.wu@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoocfs2: fix null pointer dereference in ocfs2_dir_foreach_blk_id()
Jeff Liu [Tue, 13 Aug 2013 23:01:01 +0000 (16:01 -0700)]
ocfs2: fix null pointer dereference in ocfs2_dir_foreach_blk_id()

Fix a NULL pointer deference while removing an empty directory, which
was introduced by commit 3704412bdbf3 ("[readdir] convert ocfs2").

  BUG: unable to handle kernel NULL pointer dereference at (null)
  IP: [<(null)>]           (null)
  PGD 6da85067 PUD 6da89067 PMD 0
  Oops: 0010 [#1] SMP
  CPU: 0 PID: 6564 Comm: rmdir Tainted: G           O 3.11.0-rc1 #4
  RIP: 0010:[<0000000000000000>]  [<          (null)>]           (null)
  Call Trace:
    ocfs2_dir_foreach+0x49/0x50 [ocfs2]
    ocfs2_empty_dir+0x12c/0x3e0 [ocfs2]
    ocfs2_unlink+0x56e/0xc10 [ocfs2]
    vfs_rmdir+0xd5/0x140
    do_rmdir+0x1cb/0x1e0
    SyS_rmdir+0x16/0x20
    system_call_fastpath+0x16/0x1b
  Code:  Bad RIP value.
  RIP  [<          (null)>]           (null)
  RSP <ffff88006daddc10>
  CR2: 0000000000000000

[dan.carpenter@oracle.com: fix pointer math]
Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Reported-by: David Weber <wb@munzinger.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agox86 get_unmapped_area(): use proper mmap base for bottom-up direction
Radu Caragea [Tue, 13 Aug 2013 23:00:59 +0000 (16:00 -0700)]
x86 get_unmapped_area(): use proper mmap base for bottom-up direction

When the stack is set to unlimited, the bottomup direction is used for
mmap-ings but the mmap_base is not used and thus effectively renders
ASLR for mmapings along with PIE useless.

Cc: Michel Lespinasse <walken@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Adrian Sendroiu <molecula2788@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoocfs2: fix NULL pointer dereference in ocfs2_duplicate_clusters_by_page
Tiger Yang [Tue, 13 Aug 2013 23:00:58 +0000 (16:00 -0700)]
ocfs2: fix NULL pointer dereference in ocfs2_duplicate_clusters_by_page

Since ocfs2_cow_file_pos will invoke ocfs2_refcount_icow with a NULL as
the struct file pointer, it finally result in a null pointer dereference
in ocfs2_duplicate_clusters_by_page.

This patch replace file pointer with inode pointer in
cow_duplicate_clusters to fix this issue.

[jeff.liu@oracle.com: rebased patch against linux-next tree]
Signed-off-by: Tiger Yang <tiger.yang@oracle.com>
Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Acked-by: Tao Ma <tm@tao.ma>
Tested-by: David Weber <wb@munzinger.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoocfs2: Revert 40bd62e to avoid regression in extended allocation
Jie Liu [Tue, 13 Aug 2013 23:00:57 +0000 (16:00 -0700)]
ocfs2: Revert 40bd62e to avoid regression in extended allocation

Revert commit 40bd62eb7fb8 ("fs/ocfs2/journal.h: add bits_wanted while
calculating credits in ocfs2_calc_extend_credits").

Unfortunately this change broke fallocate even if there is insufficient
disk space for the preallocation, which is a serious problem.

  # df -h
  /dev/sda8        22G  1.2G   21G   6% /ocfs2
  # fallocate -o 0 -l 200M /ocfs2/testfile
  fallocate: /ocfs2/test: fallocate failed: No space left on device

and a kernel warning:

  CPU: 3 PID: 3656 Comm: fallocate Tainted: G        W  O 3.11.0-rc3 #2
  Call Trace:
    dump_stack+0x77/0x9e
    warn_slowpath_common+0xc4/0x110
    warn_slowpath_null+0x2a/0x40
    start_this_handle+0x6c/0x640 [jbd2]
    jbd2__journal_start+0x138/0x300 [jbd2]
    jbd2_journal_start+0x23/0x30 [jbd2]
    ocfs2_start_trans+0x166/0x300 [ocfs2]
    __ocfs2_extend_allocation+0x38f/0xdb0 [ocfs2]
    ocfs2_allocate_unwritten_extents+0x3c9/0x520
    __ocfs2_change_file_space+0x5e0/0xa60 [ocfs2]
    ocfs2_fallocate+0xb1/0xe0 [ocfs2]
    do_fallocate+0x1cb/0x220
    SyS_fallocate+0x6f/0xb0
    system_call_fastpath+0x16/0x1b
  JBD2: fallocate wants too many credits (51216 > 4381)

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Goldwyn Rodrigues <rgoldwyn@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agodrivers/rtc/rtc-stmp3xxx.c: provide timeout for potentially endless loop polling...
Lothar WaĂŸmann [Tue, 13 Aug 2013 23:00:56 +0000 (16:00 -0700)]
drivers/rtc/rtc-stmp3xxx.c: provide timeout for potentially endless loop polling a HW bit

It's always a bad idea to poll on HW bits without a timeout.

The i.MX28 RTC can be easily brought into a state in which the RTC is
not running (until after a power-on-reset) and thus the status bits
which are polled in the driver won't ever change.

This patch prevents the kernel from getting stuck in this case.

Signed-off-by: Lothar WaĂŸmann <LW@KARO-electronics.de>
Acked-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agohugetlb: fix lockdep splat caused by pmd sharing
Michal Hocko [Tue, 13 Aug 2013 23:00:55 +0000 (16:00 -0700)]
hugetlb: fix lockdep splat caused by pmd sharing

Dave has reported the following lockdep splat:

  =================================
  [ INFO: inconsistent lock state ]
  3.11.0-rc1+ #9 Not tainted
  ---------------------------------
  inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.
  kswapd0/49 [HC0[0]:SC0[0]:HE1:SE1] takes:
   (&mapping->i_mmap_mutex){+.+.?.}, at: [<c114971b>] page_referenced+0x87/0x5e3
  {RECLAIM_FS-ON-W} state was registered at:
     mark_held_locks+0x81/0xe7
     lockdep_trace_alloc+0x5e/0xbc
     __alloc_pages_nodemask+0x8b/0x9b6
     __get_free_pages+0x20/0x31
     get_zeroed_page+0x12/0x14
     __pmd_alloc+0x1c/0x6b
     huge_pmd_share+0x265/0x283
     huge_pte_alloc+0x5d/0x71
     hugetlb_fault+0x7c/0x64a
     handle_mm_fault+0x255/0x299
     __do_page_fault+0x142/0x55c
     do_page_fault+0xd/0x16
     error_code+0x6c/0x74
  irq event stamp: 3136917
  hardirqs last  enabled at (3136917):  _raw_spin_unlock_irq+0x27/0x50
  hardirqs last disabled at (3136916):  _raw_spin_lock_irq+0x15/0x78
  softirqs last  enabled at (3136180):  __do_softirq+0x137/0x30f
  softirqs last disabled at (3136175):  irq_exit+0xa8/0xaa
  other info that might help us debug this:
   Possible unsafe locking scenario:
         CPU0
         ----
    lock(&mapping->i_mmap_mutex);
    <Interrupt>
      lock(&mapping->i_mmap_mutex);

  *** DEADLOCK ***
  no locks held by kswapd0/49.

  stack backtrace:
  CPU: 1 PID: 49 Comm: kswapd0 Not tainted 3.11.0-rc1+ #9
  Hardware name: Dell Inc.                 Precision WorkStation 490    /0DT031, BIOS A08 04/25/2008
  Call Trace:
    dump_stack+0x4b/0x79
    print_usage_bug+0x1d9/0x1e3
    mark_lock+0x1e0/0x261
    __lock_acquire+0x623/0x17f2
    lock_acquire+0x7d/0x195
    mutex_lock_nested+0x6c/0x3a7
    page_referenced+0x87/0x5e3
    shrink_page_list+0x3d9/0x947
    shrink_inactive_list+0x155/0x4cb
    shrink_lruvec+0x300/0x5ce
    shrink_zone+0x53/0x14e
    kswapd+0x517/0xa75
    kthread+0xa8/0xaa
    ret_from_kernel_thread+0x1b/0x28

which is a false positive caused by hugetlb pmd sharing code which
allocates a new pmd from withing mapping->i_mmap_mutex.  If this
allocation causes reclaim then the lockdep detector complains that we
might self-deadlock.

This is not correct though, because hugetlb pages are not reclaimable so
their mapping will be never touched from the reclaim path.

The patch tells lockup detector that hugetlb i_mmap_mutex is special by
assigning it a separate lockdep class so it won't report possible
deadlocks on unrelated mappings.

[peterz@infradead.org: comment for annotation]
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Cc: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoaoe: adjust ref of head for compound page tails
Ed Cashin [Tue, 13 Aug 2013 23:00:53 +0000 (16:00 -0700)]
aoe: adjust ref of head for compound page tails

Fix a BUG which can trigger when direct-IO is used with AOE.

As discussed previously, the fact that some users of the block layer
provide bios that point to pages with a zero _count means that it is not
OK for the network layer to do a put_page on the skb frags during an
skb_linearize, so the aoe driver gets a reference to pages in bios and
puts the reference before ending the bio.  And because it cannot use
get_page on a page with a zero _count, it manipulates the value
directly.

It is not OK to increment the _count of a compound page tail, though,
since the VM layer will VM_BUG_ON a non-zero _count.  Block users that
do direct I/O can result in the aoe driver seeing compound page tails in
bios.  In that case, the same logic works as long as the head of the
compound page is used instead of the tails.  This patch handles compound
pages and does not BUG.

It relies on the block layer user leaving the relationship between the
page tail and its head alone for the duration between the submission of
the bio and its completion, whether successful or not.

Signed-off-by: Ed Cashin <ecashin@coraid.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomicroblaze: fix clone syscall
Michal Simek [Tue, 13 Aug 2013 23:00:53 +0000 (16:00 -0700)]
microblaze: fix clone syscall

Fix inadvertent breakage in the clone syscall ABI for Microblaze that
was introduced in commit f3268edbe6fe ("microblaze: switch to generic
fork/vfork/clone").

The Microblaze syscall ABI for clone takes the parent tid address in the
4th argument; the third argument slot is used for the stack size.  The
incorrectly-used CLONE_BACKWARDS type assigned parent tid to the 3rd
slot.

This commit restores the original ABI so that existing userspace libc
code will work correctly.

All kernel versions from v3.8-rc1 were affected.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: save soft-dirty bits on file pages
Cyrill Gorcunov [Tue, 13 Aug 2013 23:00:51 +0000 (16:00 -0700)]
mm: save soft-dirty bits on file pages

Andy reported that if file page get reclaimed we lose the soft-dirty bit
if it was there, so save _PAGE_BIT_SOFT_DIRTY bit when page address get
encoded into pte entry.  Thus when #pf happens on such non-present pte
we can restore it back.

Reported-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomm: save soft-dirty bits on swapped pages
Cyrill Gorcunov [Tue, 13 Aug 2013 23:00:49 +0000 (16:00 -0700)]
mm: save soft-dirty bits on swapped pages

Andy Lutomirski reported that if a page with _PAGE_SOFT_DIRTY bit set
get swapped out, the bit is getting lost and no longer available when
pte read back.

To resolve this we introduce _PTE_SWP_SOFT_DIRTY bit which is saved in
pte entry for the page being swapped out.  When such page is to be read
back from a swap cache we check for bit presence and if it's there we
clear it and restore the former _PAGE_SOFT_DIRTY bit back.

One of the problem was to find a place in pte entry where we can save
the _PTE_SWP_SOFT_DIRTY bit while page is in swap.  The _PAGE_PSE was
chosen for that, it doesn't intersect with swap entry format stored in
pte.

Reported-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agomemcg: don't initialize kmem-cache destroying work for root caches
Andrey Vagin [Tue, 13 Aug 2013 23:00:47 +0000 (16:00 -0700)]
memcg: don't initialize kmem-cache destroying work for root caches

struct memcg_cache_params has a union.  Different parts of this union
are used for root and non-root caches.  A part with destroying work is
used only for non-root caches.

I fixed the same problem in another place v3.9-rc1-16204-gf101a94, but
didn't notice this one.

This patch fixes the kernel panic:

[   46.848187] BUG: unable to handle kernel paging request at 000000fffffffeb8
[   46.849026] IP: [<ffffffff811a484c>] kmem_cache_destroy_memcg_children+0x6c/0xc0
[   46.849092] PGD 0
[   46.849092] Oops: 0000 [#1] SMP
...

Signed-off-by: Andrey Vagin <avagin@openvz.org>
Cc: Glauber Costa <glommer@openvz.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: <stable@vger.kernel.org> [3.9.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoethernet/arc/arc_emac - fix NAPI "work > weight" warning
Alexey Brodkin [Tue, 13 Aug 2013 13:04:36 +0000 (17:04 +0400)]
ethernet/arc/arc_emac - fix NAPI "work > weight" warning

Initially I improperly set a boundary for maximum number of input
packets to process on NAPI poll ("work") so it might be more than
expected amount ("weight").

This was really harmless but seeing WARN_ON_ONCE on every device boot is
not nice. So trivial fix ("<" instead of "<=") is here.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Mischa Jonker <mjonker@synopsys.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 13 Aug 2013 23:58:17 +0000 (16:58 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "Docbook fixes that make 99% of the diffstat, plus a oneliner fix"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched: Ensure update_cfs_shares() is called for parents of continuously-running tasks
  sched: Fix some kernel-doc warnings

11 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 13 Aug 2013 23:57:40 +0000 (16:57 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "Two small fixlets"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86: Add Haswell ULT model number used in Macbook Air and other systems
  perf/x86: Fix intel QPI uncore event definitions

11 years agoperf/arm: Fix armpmu_map_hw_event()
Stephen Boyd [Wed, 7 Aug 2013 23:18:08 +0000 (16:18 -0700)]
perf/arm: Fix armpmu_map_hw_event()

Fix constraint check in armpmu_map_hw_event().

Reported-and-tested-by: Vince Weaver <vincent.weaver@maine.edu>
Cc: <stable@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoip_tunnel: Do not use inner ip-header-id for tunnel ip-header-id.
Pravin B Shelar [Tue, 13 Aug 2013 08:41:06 +0000 (01:41 -0700)]
ip_tunnel: Do not use inner ip-header-id for tunnel ip-header-id.

Using inner-id for tunnel id is not safe in some rare cases.
E.g. packets coming from multiple sources entering same tunnel
can have same id. Therefore on tunnel packet receive we
could have packets from two different stream but with same
source and dst IP with same ip-id which could confuse ip packet
reassembly.

Following patch reverts optimization from commit
490ab08127 (IP_GRE: Fix IP-Identification.)

CC: Jarno Rajahalme <jrajahalme@nicira.com>
CC: Ansis Atteka <aatteka@nicira.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'bnx2x'
David S. Miller [Tue, 13 Aug 2013 23:04:38 +0000 (16:04 -0700)]
Merge branch 'bnx2x'

Dmitry Kravkov says:

====================
Please consider applying the series of bnx2x fixes to net:
* statistics may cause FW assert
* missing fairness configuration in DCB flow
* memory leak in sriov related part
* Illegal PTE access
* Pagefault crash in shutdown flow with cnic
v1->v2
* fixed sparse error pointed by Joe Perches
* added missing signed-off from Sergei Shtylyov
v2->v3
* added missing signed-off from Sergei Shtylyov
* fixed formatting from Sergei Shtylyov
v3->v4
* patch 1/6: fixed declaration order
* patch 2/6 replaced with: protect flows using set_bit constraints
v4->v5
* patch 2/6: replace proprietary locking with semaphore
* droped 1/6: since adds redundant code from Benjamin Poirier
The following patchset contains four netfilter fixes, they are:

* Fix possible invalid access and mangling of the TCPMSS option in
  xt_TCPMSS. This was spotted by Julian Anastasov.

* Fix possible off by one access and mangling of the TCP packet in
  xt_TCPOPTSTRIP, also spotted by Julian Anastasov.

* Fix possible information leak due to missing initialization of one
  padding field of several structures that are included in nfqueue and
  nflog netlink messages, from Dan Carpenter.

* Fix TCP window tracking with Fast Open, from Yuchung Cheng.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: prevent crash in shutdown flow with CNIC
Yuval Mintz [Mon, 12 Aug 2013 23:25:03 +0000 (02:25 +0300)]
bnx2x: prevent crash in shutdown flow with CNIC

There might be a crash as during shutdown flow CNIC might try
to access resources already freed by bnx2x.
Change bnx2x_close() into dev_close() in __bnx2x_remove (shutdown flow)
to guarantee CNIC is notified of the device's change of status.

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: fix PTE write access error
Barak Witkowsky [Mon, 12 Aug 2013 23:25:02 +0000 (02:25 +0300)]
bnx2x: fix PTE write access error

PTE write access error  might occur in MF_ALLOWED mode when IOMMU
is active. The patch adds rmmod HSI indicating to MFW to stop
running queries which might trigger this failure.

Signed-off-by: Barak Witkowsky <barak@broadcom.com>
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: fix memory leak in VF
Ariel Elior [Mon, 12 Aug 2013 23:25:01 +0000 (02:25 +0300)]
bnx2x: fix memory leak in VF

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: update fairness parameters following DCB negotiation
Dmitry Kravkov [Mon, 12 Aug 2013 23:25:00 +0000 (02:25 +0300)]
bnx2x: update fairness parameters following DCB negotiation

ETS can be enabled as a result of DCB negotiation, then
fairness must be recalculated after each negotiation.

Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobnx2x: protect different statistics flows
Dmitry Kravkov [Mon, 12 Aug 2013 23:24:59 +0000 (02:24 +0300)]
bnx2x: protect different statistics flows

Add locking to protect different statistics flows from
running simultaneously.
This in order to serialize statistics requests sent to FW,
otherwise two outstanding queries may cause FW assert.

Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoPM / QoS: Fix workqueue deadlock when using pm_qos_update_request_timeout()
Stephen Boyd [Tue, 13 Aug 2013 21:12:40 +0000 (14:12 -0700)]
PM / QoS: Fix workqueue deadlock when using pm_qos_update_request_timeout()

pm_qos_update_request_timeout() updates a qos and then schedules
a delayed work item to bring the qos back down to the default
after the timeout. When the work item runs, pm_qos_work_fn() will
call pm_qos_update_request() and deadlock because it tries to
cancel itself via cancel_delayed_work_sync(). Future callers of
that qos will also hang waiting to cancel the work that is
canceling itself. Let's extract the little bit of code that does
the real work of pm_qos_update_request() and call it from the
work function so that we don't deadlock.

Before ed1ac6e (PM: don't use [delayed_]work_pending()) this didn't
happen because the work function wouldn't try to cancel itself.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Reviewed-by: Tejun Heo <tj@kernel.org>
Cc: 3.9+ <stable@vger.kernel.org> # 3.9+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agomyri10ge: Update MAINTAINERS
Hyong-Youb Kim [Mon, 12 Aug 2013 09:20:02 +0000 (02:20 -0700)]
myri10ge: Update MAINTAINERS

Remove Andrew Gallatin, as he is no longer with Myricom. Add
Hyong-Youb Kim as the new maintainer. Update the website URL.

Signed-off-by: Hyong-Youb Kim <hykim@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoskge: dma_sync the whole receive buffer
stephen hemminger [Sat, 10 Aug 2013 22:02:07 +0000 (15:02 -0700)]
skge: dma_sync the whole receive buffer

The DMA sync should sync the whole receive buffer, not just
part of it. Fixes log messages dma_sync_check.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'security-fixes' into fixes
Russell King [Tue, 13 Aug 2013 19:23:28 +0000 (20:23 +0100)]
Merge branch 'security-fixes' into fixes

11 years agoARM: 7807/1: kexec: validate CPU hotplug support
Stephen Warren [Fri, 2 Aug 2013 19:52:49 +0000 (20:52 +0100)]
ARM: 7807/1: kexec: validate CPU hotplug support

Architectures should fully validate whether kexec is possible as part of
machine_kexec_prepare(), so that user-space's kexec_load() operation can
report any problems. Performing validation in machine_kexec() itself is
too late, since it is not allowed to return.

Prior to this patch, ARM's machine_kexec() was testing after-the-fact
whether machine_kexec_prepare() was able to disable all but one CPU.
Instead, modify machine_kexec_prepare() to validate all conditions
necessary for machine_kexec_prepare()'s to succeed. BUG if the validation
succeeded, yet disabling the CPUs didn't actually work.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoARM: 7812/1: rwlocks: retry trylock operation if strex fails on free lock
Will Deacon [Mon, 12 Aug 2013 17:04:05 +0000 (18:04 +0100)]
ARM: 7812/1: rwlocks: retry trylock operation if strex fails on free lock

Commit 15e7e5c1ebf5 ("ARM: 7749/1: spinlock: retry trylock operation if
strex fails on free lock") modifying our arch_spin_trylock to retry the
acquisition if the lock appeared uncontended, but the strex failed.

This patch does the same for rwlocks, which were missed by the original
patch.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoARM: 7811/1: locks: use early clobber in arch_spin_trylock
Will Deacon [Mon, 12 Aug 2013 17:03:26 +0000 (18:03 +0100)]
ARM: 7811/1: locks: use early clobber in arch_spin_trylock

The res variable is written before we've finished with the input
operands (namely the lock address), so ensure that we mark it as `early
clobber' to avoid unintended register sharing.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoARM: 7810/1: perf: Fix array out of bounds access in armpmu_map_hw_event()
Stephen Boyd [Thu, 8 Aug 2013 17:41:59 +0000 (18:41 +0100)]
ARM: 7810/1: perf: Fix array out of bounds access in armpmu_map_hw_event()

Vince Weaver reports an oops in the ARM perf event code while
running his perf_fuzzer tool on a pandaboard running v3.11-rc4.

Unable to handle kernel paging request at virtual address 73fd14cc
pgd = eca6c000
[73fd14cc] *pgd=00000000
Internal error: Oops: 5 [#1] SMP ARM
Modules linked in: snd_soc_omap_hdmi omapdss snd_soc_omap_abe_twl6040 snd_soc_twl6040 snd_soc_omap snd_soc_omap_hdmi_card snd_soc_omap_mcpdm snd_soc_omap_mcbsp snd_soc_core snd_compress regmap_spi snd_pcm snd_page_alloc snd_timer snd soundcore
CPU: 1 PID: 2790 Comm: perf_fuzzer Not tainted 3.11.0-rc4 #6
task: eddcab80 ti: ed892000 task.ti: ed892000
PC is at armpmu_map_event+0x20/0x88
LR is at armpmu_event_init+0x38/0x280
pc : [<c001c3e4>]    lr : [<c001c17c>]    psr: 60000013
sp : ed893e40  ip : ecececec  fp : edfaec00
r10: 00000000  r9 : 00000000  r8 : ed8c3ac0
r7 : ed8c3b5c  r6 : edfaec00  r5 : 00000000  r4 : 00000000
r3 : 000000ff  r2 : c0496144  r1 : c049611c  r0 : edfaec00
Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
Control: 10c5387d  Table: aca6c04a  DAC: 00000015
Process perf_fuzzer (pid: 2790, stack limit = 0xed892240)
Stack: (0xed893e40 to 0xed894000)
3e40: 00000800 c001c17c 00000002 c008a748 00000001 00000000 00000000 c00bf078
3e60: 00000000 edfaee50 00000000 00000000 00000000 edfaec00 ed8c3ac0 edfaec00
3e80: 00000000 c073ffac ed893f20 c00bf180 00000001 00000000 c00bf078 ed893f20
3ea0: 00000000 ed8c3ac0 00000000 00000000 00000000 c0cb0818 eddcab80 c00bf440
3ec0: ed893f20 00000000 eddcab80 eca76800 00000000 eca76800 00000000 00000000
3ee0: 00000000 ec984c80 eddcab80 c00bfe68 00000000 00000000 00000000 00000080
3f00: 00000000 ed892000 00000000 ed892030 00000004 ecc7e3c8 ecc7e3c8 00000000
3f20: 00000000 00000048 ecececec 00000000 00000000 00000000 00000000 00000000
3f40: 00000000 00000000 00297810 00000000 00000000 00000000 00000000 00000000
3f60: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
3f80: 00000002 00000002 000103a4 00000002 0000016c c00128e8 ed892000 00000000
3fa0: 00090998 c0012700 00000002 000103a4 00090ab8 00000000 00000000 0000000f
3fc0: 00000002 000103a4 00000002 0000016c 00090ab0 00090ab8 000107a0 00090998
3fe0: bed92be0 bed92bd0 0000b785 b6e8f6d0 40000010 00090ab8 00000000 00000000
[<c001c3e4>] (armpmu_map_event+0x20/0x88) from [<c001c17c>] (armpmu_event_init+0x38/0x280)
[<c001c17c>] (armpmu_event_init+0x38/0x280) from [<c00bf180>] (perf_init_event+0x108/0x180)
[<c00bf180>] (perf_init_event+0x108/0x180) from [<c00bf440>] (perf_event_alloc+0x248/0x40c)
[<c00bf440>] (perf_event_alloc+0x248/0x40c) from [<c00bfe68>] (SyS_perf_event_open+0x4f4/0x8fc)
[<c00bfe68>] (SyS_perf_event_open+0x4f4/0x8fc) from [<c0012700>] (ret_fast_syscall+0x0/0x48)
Code: 0a000005 e3540004 0a000016 e3540000 (0791010c)

This is because event->attr.config in armpmu_event_init()
contains a very large number copied directly from userspace and
is never checked against the size of the array indexed in
armpmu_map_hw_event(). Fix the problem by checking the value of
config before indexing the array and rejecting invalid config
values.

Reported-by: Vince Weaver <vincent.weaver@maine.edu>
Tested-by: Vince Weaver <vincent.weaver@maine.edu>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoARM: 7809/1: perf: fix event validation for software group leaders
Will Deacon [Wed, 7 Aug 2013 22:39:41 +0000 (23:39 +0100)]
ARM: 7809/1: perf: fix event validation for software group leaders

It is possible to construct an event group with a software event as a
group leader and then subsequently add a hardware event to the group.
This results in the event group being validated by adding all members
of the group to a fake PMU and attempting to allocate each event on
their respective PMU.

Unfortunately, for software events wthout a corresponding arm_pmu, this
results in a kernel crash attempting to dereference the ->get_event_idx
function pointer.

This patch fixes the problem by checking explicitly for software events
and ignoring those in event validation (since they can always be
scheduled). We will probably want to revisit this for 3.12, since the
validation checks don't appear to work correctly when dealing with
multiple hardware PMUs anyway.

Cc: <stable@vger.kernel.org>
Reported-by: Vince Weaver <vincent.weaver@maine.edu>
Tested-by: Vince Weaver <vincent.weaver@maine.edu>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
11 years agoclk: exynos4: Add CLK_GET_RATE_NOCACHE flag for the Exynos4x12 ISP clocks
Sylwester Nawrocki [Thu, 25 Jul 2013 21:07:05 +0000 (23:07 +0200)]
clk: exynos4: Add CLK_GET_RATE_NOCACHE flag for the Exynos4x12 ISP clocks

The ISP clock registers belong to the ISP power domain and may change
their values if this power domain is switched off/on. Add
CLK_GET_RATE_NOCACHE flags to ensure we do not rely on invalid cached
data when setting or getting frequency of those clocks.

Without this fix the FIMC-IS Cortex-A5 core and AXI bus clocks have
incorrect frequencies, which breaks the ISP operation and starting the
video pipeline fails with timeouts reported by the FIMC-IS firmware.

See related commit 722a860ecb29aa34ec6f7d7f32b949209e8 "[media]
exynos4-is: Fix FIMC-IS clocks initialization" for more details.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
11 years agoclk/zynq/clkc: Add CLK_SET_RATE_PARENT flag to ethernet muxes
Soren Brinkmann [Mon, 17 Jun 2013 22:47:40 +0000 (15:47 -0700)]
clk/zynq/clkc: Add CLK_SET_RATE_PARENT flag to ethernet muxes

Zynq's Ethernet clocks are created by the following hierarchy:
mux0 ---> div0 ---> div1 ---> mux1 ---> gate
Rate change requests on the gate have to propagate all the way up to
div0 to properly leverage all dividers. Mux1 was missing the
CLK_SET_RATE_PARENT flag, which is required to achieve this.

This does not fix a specific regression but the clock driver was merged
for 3.11-rc1, so best to fix the known bugs before the release.

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
[mturquette@linaro.org: added to changelog]

11 years agoclk/zynq/clkc: Add dedicated spinlock for the SWDT
Soren Brinkmann [Mon, 17 Jun 2013 22:03:46 +0000 (15:03 -0700)]
clk/zynq/clkc: Add dedicated spinlock for the SWDT

The clk_mux for the system watchdog timer reused the register lock
dedicated to the Ethernet module - for no apparent reason.
Add a lock dedicated to the SWDT's clock register to remove this
wrong dependency.

This does not fix a specific regression but the clock driver was merged
for 3.11-rc1, so best to fix the known bugs before the release.

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
[mturquette@linaro.org: added to changelog]

11 years agosched: fix the theoretical signal_wake_up() vs schedule() race
Oleg Nesterov [Mon, 12 Aug 2013 16:14:00 +0000 (18:14 +0200)]
sched: fix the theoretical signal_wake_up() vs schedule() race

This is only theoretical, but after try_to_wake_up(p) was changed
to check p->state under p->pi_lock the code like

__set_current_state(TASK_INTERRUPTIBLE);
schedule();

can miss a signal. This is the special case of wait-for-condition,
it relies on try_to_wake_up/schedule interaction and thus it does
not need mb() between __set_current_state() and if(signal_pending).

However, this __set_current_state() can move into the critical
section protected by rq->lock, now that try_to_wake_up() takes
another lock we need to ensure that it can't be reordered with
"if (signal_pending(current))" check inside that section.

The patch is actually one-liner, it simply adds smp_wmb() before
spin_lock_irq(rq->lock). This is what try_to_wake_up() already
does by the same reason.

We turn this wmb() into the new helper, smp_mb__before_spinlock(),
for better documentation and to allow the architectures to change
the default implementation.

While at it, kill smp_mb__after_lock(), it has no callers.

Perhaps we can also add smp_mb__before/after_spinunlock() for
prepare_to_wait().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agocpuset: fix the return value of cpuset_write_u64()
Li Zefan [Tue, 13 Aug 2013 02:05:59 +0000 (10:05 +0800)]
cpuset: fix the return value of cpuset_write_u64()

Writing to this file always returns -ENODEV:

  # echo 1 > cpuset.memory_pressure_enabled
  -bash: echo: write error: No such device

Signed-off-by: Li Zefan <lizefan@huawei.com>
Cc: <stable@vger.kernel.org> # 3.9+
Signed-off-by: Tejun Heo <tj@kernel.org>
11 years agogenetlink: fix family dump race
Johannes Berg [Tue, 13 Aug 2013 07:04:05 +0000 (09:04 +0200)]
genetlink: fix family dump race

When dumping generic netlink families, only the first dump call
is locked with genl_lock(), which protects the list of families,
and thus subsequent calls can access the data without locking,
racing against family addition/removal. This can cause a crash.
Fix it - the locking needs to be conditional because the first
time around it's already locked.

A similar bug was reported to me on an old kernel (3.4.47) but
the exact scenario that happened there is no longer possible,
on those kernels the first round wasn't locked either. Looking
at the current code I found the race described above, which had
also existed on the old kernel.

Cc: stable@vger.kernel.org
Reported-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: sctp: sctp_transport_destroy{, _rcu}: fix potential pointer corruption
Daniel Borkmann [Fri, 9 Aug 2013 14:25:21 +0000 (16:25 +0200)]
net: sctp: sctp_transport_destroy{, _rcu}: fix potential pointer corruption

Probably this one is quite unlikely to be triggered, but it's more safe
to do the call_rcu() at the end after we have dropped the reference on
the asoc and freed sctp packet chunks. The reason why is because in
sctp_transport_destroy_rcu() the transport is being kfree()'d, and if
we're unlucky enough we could run into corrupted pointers. Probably
that's more of theoretical nature, but it's safer to have this simple fix.

Introduced by commit 8c98653f ("sctp: sctp_close: fix release of bindings
for deferred call_rcu's"). I also did the 8c98653f regression test and
it's fine that way.

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>