Paul E. McKenney
83f5b01ffb
rcu: Fix long-grace-period race between forcing and initialization
...
Very long RCU read-side critical sections (50 milliseconds or
so) can cause a race between force_quiescent_state() and
rcu_start_gp() as follows on kernel builds with multi-level
rcu_node hierarchies:
1. CPU 0 calls force_quiescent_state(), sees that there is a
grace period in progress, and acquires ->fsqlock.
2. CPU 1 detects the end of the grace period, and so
cpu_quiet_msk_finish() sets rsp->completed to rsp->gpnum.
This operation is carried out under the root rnp->lock,
but CPU 0 has not yet acquired that lock. Note that
rsp->signaled is still RCU_SAVE_DYNTICK from the last
grace period.
3. CPU 1 calls rcu_start_gp(), but no one wants a new grace
period, so it drops the root rnp->lock and returns.
4. CPU 0 acquires the root rnp->lock and picks up rsp->completed
and rsp->signaled, then drops rnp->lock. It then enters the
RCU_SAVE_DYNTICK leg of the switch statement.
5. CPU 2 invokes call_rcu(), and now needs a new grace period.
It calls rcu_start_gp(), which acquires the root rnp->lock, sets
rsp->signaled to RCU_GP_INIT (too bad that CPU 0 is already in
the RCU_SAVE_DYNTICK leg of the switch statement!) and starts
initializing the rcu_node hierarchy. If there are multiple
levels to the hierarchy, it will drop the root rnp->lock and
initialize the lower levels of the hierarchy.
6. CPU 0 notes that rsp->completed has not changed, which permits
both CPU 2 and CPU 0 to try updating it concurrently. If CPU 0's
update prevails, later calls to force_quiescent_state() can
count old quiescent states against the new grace period, which
can in turn result in premature ending of grace periods.
Not good.
This patch adds an RCU_GP_IDLE state for rsp->signaled that is
set initially at boot time and any time a grace period ends.
This prevents CPU 0 from getting into the workings of
force_quiescent_state() in step 4. Additional locking and
checks prevent the concurrent update of rsp->signaled in step 6.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: laijs@cn.fujitsu.com
Cc: dipankar@in.ibm.com
Cc: mathieu.desnoyers@polymtl.ca
Cc: josh@joshtriplett.org
Cc: dvhltc@us.ibm.com
Cc: niv@us.ibm.com
Cc: peterz@infradead.org
Cc: rostedt@goodmis.org
Cc: Valdis.Kletnieks@vt.edu
Cc: dhowells@redhat.com
LKML-Reference: <1256742889199-git-send-email->
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-11-02 16:06:21 +01:00
..
gcov
microblaze: Enable GCOV_PROFILE_ALL
2009-09-21 14:29:21 +02:00
irq
Merge branch 'irq-threaded-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-09-11 13:21:31 -07:00
power
headers: utsname.h redux
2009-09-23 18:13:10 -07:00
time
NOHZ: update idle state also when NOHZ is inactive
2009-10-07 13:05:05 +02:00
trace
Merge branch 'tracing-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-10-08 12:06:09 -07:00
.gitignore
acct.c
bsdacct: switch credentials for writing to the accounting file
2009-08-24 11:33:40 +10:00
async.c
audit.c
Audit: send signal info if selinux is disabled
2009-09-24 03:50:26 -04:00
audit.h
audit_tree.c
audit_watch.c
Audit: reorganize struct audit_watch to save 8 bytes
2009-09-24 03:50:25 -04:00
auditfilter.c
auditsc.c
Audit: rearrange audit_context to save 16 bytes per struct
2009-09-24 03:50:26 -04:00
backtracetest.c
bounds.c
capability.c
cgroup.c
cgroup: catch bad css refcnt at css_put
2009-10-01 16:11:12 -07:00
cgroup_freezer.c
cgroups: let ss->can_attach and ss->attach do whole threadgroups at a time
2009-09-24 07:20:58 -07:00
compat.c
configs.c
cpu.c
Merge branch 'x86-pat-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-09-15 09:19:38 -07:00
cpuset.c
cgroups: let ss->can_attach and ss->attach do whole threadgroups at a time
2009-09-24 07:20:58 -07:00
cred-internals.h
cred.c
creds_are_invalid() needs to be exported for use by modules:
2009-09-23 11:02:26 -07:00
delayacct.c
headers: taskstats_kern.h trim
2009-09-18 09:48:52 -07:00
dma.c
exec_domain.c
exit.c
Merge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-10-08 12:16:35 -07:00
extable.c
fork.c
Merge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-10-08 12:16:35 -07:00
freezer.c
sched: fix nr_uninterruptible accounting of frozen tasks really
2009-07-18 14:19:53 +02:00
futex.c
futex: Fix spurious wakeup for requeue_pi really
2009-10-28 20:34:34 +01:00
futex_compat.c
futex: Fix compat_futex to be same as futex for REQUEUE_PI
2009-08-10 15:41:12 +02:00
groups.c
hrtimer.c
Merge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-10-05 12:04:16 -07:00
hung_task.c
sysctl: remove "struct file *" argument of ->proc_handler
2009-09-24 07:21:04 -07:00
itimer.c
itimers: Add tracepoints for itimer
2009-08-29 14:10:07 +02:00
kallsyms.c
kallsyms: use new arch_is_kernel_text()
2009-09-23 07:39:30 -07:00
Kconfig.freezer
Kconfig.hz
Kconfig.preempt
kexec.c
kexec: fix omitting offset in extended crashkernel syntax
2009-07-29 19:10:34 -07:00
kfifo.c
kfifo: Use "const" definitions
2009-09-19 13:13:17 -07:00
kgdb.c
kmod.c
Revert "kmod: fix race in usermodehelper code"
2009-09-23 18:12:10 -07:00
kprobes.c
const: constify remaining file_operations
2009-10-01 16:11:11 -07:00
ksysfs.c
kthread.c
sched: Keep kthreads at default priority
2009-09-09 17:30:06 +02:00
latencytop.c
lockdep.c
lockdep: Use cpu_clock() for lockstat
2009-10-09 15:56:44 +02:00
lockdep_internals.h
lockdep: BFS cleanup
2009-07-24 10:53:29 +02:00
lockdep_proc.c
seq_file: constify seq_operations
2009-09-23 07:39:29 -07:00
lockdep_states.h
Makefile
cgroups: move the cgroup debug subsys into cgroup.c to access internal state
2009-09-24 07:20:57 -07:00
module.c
module: fix up CONFIG_KALLSYMS=n build.
2009-10-01 16:11:11 -07:00
mutex-debug.c
mutex-debug.h
mutex.c
mutex.h
notifier.c
ns_cgroup.c
cgroups: let ss->can_attach and ss->attach do whole threadgroups at a time
2009-09-24 07:20:58 -07:00
nsproxy.c
panic.c
Merge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-10-08 12:16:35 -07:00
params.c
param: allow whitespace as kernel parameter separator
2009-09-25 00:32:58 +09:30
perf_event.c
perf_event: Provide vmalloc() based mmap() backing
2009-10-06 14:21:50 +02:00
pid.c
mm: also use alloc_large_system_hash() for the PID hash table
2009-09-22 07:17:38 -07:00
pid_namespace.c
pidns: deny CLONE_PARENT|CLONE_NEWPID combination
2009-09-24 07:21:04 -07:00
pm_qos_params.c
posix-cpu-timers.c
itimers: Add tracepoints for itimer
2009-08-29 14:10:07 +02:00
posix-timers.c
time: Introduce CLOCK_REALTIME_COARSE
2009-08-21 21:43:46 +02:00
printk.c
printk: add printk_delay to make messages readable for some scenarios
2009-09-23 07:39:28 -07:00
profile.c
kernel/profile.c: Switch /proc/irq/prof_cpu_mask to seq_file
2009-09-20 20:15:40 +02:00
ptrace.c
ptrace: __ptrace_detach: do __wake_up_parent() if we reap the tracee
2009-09-24 07:20:59 -07:00
rcupdate.c
rcu: Move rcu_barrier() to rcutree
2009-10-07 08:11:20 +02:00
rcutorture.c
rcu: Clean up code to address Ingo's checkpatch feedback
2009-09-23 19:46:30 +02:00
rcutree.c
rcu: Fix long-grace-period race between forcing and initialization
2009-11-02 16:06:21 +01:00
rcutree.h
rcu: Fix long-grace-period race between forcing and initialization
2009-11-02 16:06:21 +01:00
rcutree_plugin.h
rcu: Fix TREE_PREEMPT_RCU CPU_HOTPLUG bad-luck hang
2009-10-15 20:33:01 +02:00
rcutree_trace.c
rcu: Make hot-unplugged CPU relinquish its own RCU callbacks
2009-10-07 08:11:20 +02:00
relay.c
const: mark struct vm_struct_operations
2009-09-27 11:39:25 -07:00
res_counter.c
memcg: some modification to softlimit under hierarchical memory reclaim.
2009-10-01 16:11:13 -07:00
resource.c
walk system ram range
2009-09-23 07:39:41 -07:00
rtmutex-debug.c
rtmutex-debug.h
rtmutex-tester.c
rtmutex.c
rtmutex: Avoid deadlock in rt_mutex_start_proxy_lock()
2009-08-06 05:50:21 +02:00
rtmutex.h
rtmutex_common.h
rwsem.c
sched.c
sched: Set correct normal_prio and prio values in sched_fork()
2009-10-05 13:42:20 +02:00
sched_clock.c
sched_clock: Fix atomicity/continuity bug by using cmpxchg64()
2009-09-30 22:56:10 +02:00
sched_cpupri.c
sched: Add new prio to cpupri before removing old prio
2009-08-02 14:26:09 +02:00
sched_cpupri.h
sched_debug.c
sched: Add new wakeup preemption mode: WAKEUP_RUNNING
2009-09-17 10:17:25 +02:00
sched_fair.c
sysctl: remove "struct file *" argument of ->proc_handler
2009-09-24 07:21:04 -07:00
sched_features.h
sched: Add new wakeup preemption mode: WAKEUP_RUNNING
2009-09-17 10:17:25 +02:00
sched_idletask.c
sched: Simplify sys_sched_rr_get_interval() system call
2009-09-21 09:53:55 +02:00
sched_rt.c
sched: Simplify sys_sched_rr_get_interval() system call
2009-09-21 09:53:55 +02:00
sched_stats.h
seccomp.c
semaphore.c
signal.c
signals: inline __fatal_signal_pending
2009-09-24 07:21:01 -07:00
slow-work.c
sysctl: remove "struct file *" argument of ->proc_handler
2009-09-24 07:21:04 -07:00
smp.c
cpumask: remove arch_send_call_function_ipi
2009-09-24 09:34:47 +09:30
softirq.c
softirq: add BLOCK_IOPOLL to softirq_to_name
2009-09-17 15:53:44 -04:00
softlockup.c
sysctl: remove "struct file *" argument of ->proc_handler
2009-09-24 07:21:04 -07:00
spinlock.c
locking: Allow arch-inlined spinlocks
2009-08-31 18:08:50 +02:00
srcu.c
stacktrace.c
stop_machine.c
sys.c
Merge branch 'hwpoison' of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-mce-2.6
2009-09-24 07:53:22 -07:00
sys_ni.c
Merge branch 'master' of /home/davem/src/GIT/linux-2.6/
2009-09-24 15:13:11 -07:00
sysctl.c
Merge branch 'hwpoison' of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-mce-2.6
2009-09-24 07:53:22 -07:00
sysctl_check.c
taskstats.c
genetlink: make netns aware
2009-07-12 14:03:27 -07:00
test_kprobes.c
time.c
time: Prevent 32 bit overflow with set_normalized_timespec()
2009-09-15 10:17:30 +02:00
timeconst.pl
timer.c
Merge branch 'timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-09-23 09:46:15 -07:00
tracepoint.c
trivial: fix typo "to to" in multiple files
2009-09-21 15:14:55 +02:00
tsacct.c
uid16.c
headers: utsname.h redux
2009-09-23 18:13:10 -07:00
up.c
user.c
uids: Prevent tear down race
2009-11-02 16:02:39 +01:00
user_namespace.c
utsname.c
utsname_sysctl.c
sysctl: remove "struct file *" argument of ->proc_handler
2009-09-24 07:21:04 -07:00
wait.c
locking, sched: Give waitqueue spinlocks their own lockdep classes
2009-08-10 14:43:09 +02:00
workqueue.c
Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
2009-09-11 13:23:18 -07:00