alistair23-linux/kernel
Roman Gushchin 09dc4ab039 sched/fair: Fix tg_set_cfs_bandwidth() deadlock on rq->lock
tg_set_cfs_bandwidth() sets cfs_b->timer_active to 0 to
force the period timer restart. It's not safe, because
can lead to deadlock, described in commit 927b54fccb:
"__start_cfs_bandwidth calls hrtimer_cancel while holding rq->lock,
waiting for the hrtimer to finish. However, if sched_cfs_period_timer
runs for another loop iteration, the hrtimer can attempt to take
rq->lock, resulting in deadlock."

Three CPUs must be involved:

  CPU0               CPU1                         CPU2
  take rq->lock      period timer fired
  ...                take cfs_b lock
  ...                ...                          tg_set_cfs_bandwidth()
  throttle_cfs_rq()  release cfs_b lock           take cfs_b lock
  ...                distribute_cfs_runtime()     timer_active = 0
  take cfs_b->lock   wait for rq->lock            ...
  __start_cfs_bandwidth()
  {wait for timer callback
   break if timer_active == 1}

So, CPU0 and CPU1 are deadlocked.

Instead of resetting cfs_b->timer_active, tg_set_cfs_bandwidth can
wait for period timer callbacks (ignoring cfs_b->timer_active) and
restart the timer explicitly.

Signed-off-by: Roman Gushchin <klamm@yandex-team.ru>
Reviewed-by: Ben Segall <bsegall@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/87wqdi9g8e.wl\%klamm@yandex-team.ru
Cc: pjt@google.com
Cc: chris.j.arges@canonical.com
Cc: gregkh@linuxfoundation.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-06-05 11:51:34 +02:00
..
debug mm: per-thread vma caching 2014-04-07 16:35:53 -07:00
events Merge branch 'for-linus' of git://ftp.arm.linux.org.uk/~rmk/linux-arm 2014-04-05 13:20:43 -07:00
gcov
irq genirq: Export symbol no_action() 2014-03-22 11:33:09 +01:00
locking Merge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-04-16 16:35:18 -07:00
power kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
printk printk: fix one circular lockdep warning about console_lock 2014-04-03 16:21:08 -07:00
rcu Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-03-31 11:21:19 -07:00
sched sched/fair: Fix tg_set_cfs_bandwidth() deadlock on rq->lock 2014-06-05 11:51:34 +02:00
time tick-sched: Check tick_nohz_enabled in tick_nohz_switch_to_nohz() 2014-04-15 20:26:58 +02:00
trace This contains two fixes. 2014-04-18 10:16:43 -07:00
.gitignore
acct.c
async.c
audit.c Merge git://git.infradead.org/users/eparis/audit 2014-04-12 12:38:53 -07:00
audit.h audit: Use struct net not pid_t to remember the network namespce to reply in 2014-03-20 10:10:53 -04:00
audit_tree.c
audit_watch.c
auditfilter.c Merge git://git.infradead.org/users/eparis/audit 2014-04-12 12:38:53 -07:00
auditsc.c Merge git://git.infradead.org/users/eparis/audit 2014-04-12 12:38:53 -07:00
backtracetest.c
bounds.c
capability.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security 2014-04-03 09:26:18 -07:00
cgroup.c cgroup: newly created dirs and files should be owned by the creator 2014-04-07 16:44:47 -04:00
cgroup_freezer.c cgroup: drop const from @buffer of cftype->write_string() 2014-03-19 10:23:54 -04:00
compat.c Merge branch 'x86-x32-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-04-02 12:51:41 -07:00
configs.c
context_tracking.c
cpu.c sched: Fix hotplug vs. set_cpus_allowed_ptr() 2014-05-22 10:21:31 +02:00
cpu_pm.c
cpuset.c Merge branch 'akpm' (incoming from Andrew) 2014-04-03 16:22:16 -07:00
crash_dump.c
cred.c
delayacct.c
dma.c
elfcore.c
exec_domain.c
exit.c wait: WSTOPPED|WCONTINUED doesn't work if a zombie leader is traced by another process 2014-04-07 16:36:06 -07:00
extable.c
fork.c kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
freezer.c
futex.c futex: update documentation for ordering guarantees 2014-04-12 17:57:51 -07:00
futex_compat.c
groups.c kernel/groups.c: remove return value of set_groups 2014-04-03 16:21:05 -07:00
hrtimer.c timer: Remove code redundancy while calling get_nohz_timer_target() 2014-03-20 12:35:46 +01:00
hung_task.c kernel: audit/fix non-modular users of module_init in core code 2014-04-03 16:21:07 -07:00
irq_work.c perf/x86: Warn to early_printk() in case irq_work is too slow 2014-02-21 21:49:07 +01:00
itimer.c
jump_label.c
kallsyms.c kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
kcmp.c
Kconfig.freezer
Kconfig.hz
Kconfig.locks
Kconfig.preempt
kexec.c kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
kmod.c
kprobes.c
ksysfs.c kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
kthread.c kthread: ensure locality of task_struct allocations 2014-04-03 16:20:49 -07:00
latencytop.c
Makefile Merge branch 'x86-asmlinkage-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-03-31 14:13:25 -07:00
module-internal.h
module.c modules: use raw_cpu_write for initialization of per cpu refcount. 2014-04-07 16:36:14 -07:00
module_signing.c
notifier.c notifier: Substitute rcu_access_pointer() for rcu_dereference_raw() 2014-02-26 06:35:13 -08:00
nsproxy.c
padata.c
panic.c kernel/panic.c: display reason at end + pr_emerg 2014-04-07 16:36:08 -07:00
params.c
pid.c
pid_namespace.c pid_namespace: pidns_get() should check task_active_pid_ns() != NULL 2014-04-02 16:20:21 -07:00
posix-cpu-timers.c
posix-timers.c
profile.c CPU hotplug notifiers registration fixes for 3.15-rc1 2014-04-07 14:55:46 -07:00
ptrace.c kernel/compat: convert to COMPAT_SYSCALL_DEFINE 2014-03-06 15:35:10 +01:00
range.c
reboot.c
relay.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2014-04-12 14:49:50 -07:00
res_counter.c res_counter: remove interface for locked charging and uncharging 2014-04-07 16:35:54 -07:00
resource.c kernel/resource.c: make reallocate_resource() static 2014-04-03 16:21:07 -07:00
seccomp.c seccomp: fix memory leak on filter attach 2014-04-16 15:25:53 -04:00
signal.c kernel: use macros from compiler.h instead of __attribute__((...)) 2014-04-07 16:36:11 -07:00
smp.c smp: Rename __smp_call_function_single() to smp_call_function_single_async() 2014-02-24 14:47:15 -08:00
smpboot.c
smpboot.h
softirq.c softirq: Add linux/irq.h to make it compile again 2014-03-19 11:28:14 +01:00
stacktrace.c
stop_machine.c stop_machine: Fix^2 race between stop_two_cpus() and stop_cpus() 2014-03-11 11:33:47 +01:00
sys.c mm, thp: add VM_INIT_DEF_MASK and PRCTL_THP_DISABLE 2014-04-07 16:35:52 -07:00
sys_ni.c fs, kernel: permit disabling the uselib syscall 2014-04-03 16:21:05 -07:00
sysctl.c hung_task: check the value of "sysctl_hung_task_timeout_sec" 2014-04-07 16:36:07 -07:00
sysctl_binary.c
system_certificates.S
system_keyring.c
task_work.c
taskstats.c
test_kprobes.c
time.c
timeconst.bc
timer.c Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-04-01 11:00:07 -07:00
torture.c rcutorture: Gracefully handle NULL cleanup hooks 2014-02-23 09:04:39 -08:00
tracepoint.c This includes the final patch to clean up and fix the issue with the 2014-04-12 13:06:10 -07:00
tsacct.c
uid16.c
up.c smp: Rename __smp_call_function_single() to smp_call_function_single_async() 2014-02-24 14:47:15 -08:00
user-return-notifier.c
user.c kernel: audit/fix non-modular users of module_init in core code 2014-04-03 16:21:07 -07:00
user_namespace.c user namespace: fix incorrect memory barriers 2014-04-14 16:03:02 -07:00
utsname.c
utsname_sysctl.c
watchdog.c kernel/watchdog.c:touch_softlockup_watchdog(): use raw_cpu_write() 2014-04-18 16:40:08 -07:00
workqueue.c Merge branch 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-04-01 11:00:07 -07:00
workqueue_internal.h