linux/include/trace/events
Paul E. McKenney 21e52e1566 rcu: Make RCU_FAST_NO_HZ handle timer migration
The current RCU_FAST_NO_HZ assumes that timers do not migrate unless a
CPU goes offline, in which case it assumes that the CPU will have to come
out of dyntick-idle mode (cancelling the timer) in order to go offline.
This is important because when RCU_FAST_NO_HZ permits a CPU to enter
dyntick-idle mode despite having RCU callbacks pending, it posts a timer
on that CPU to force a wakeup on that CPU.  This wakeup ensures that the
CPU will eventually handle the end of the grace period, including invoking
its RCU callbacks.

However, Pascal Chapperon's test setup shows that the timer handler
rcu_idle_gp_timer_func() really does get invoked in some cases.  This is
problematic because this can cause the CPU that entered dyntick-idle
mode despite still having RCU callbacks pending to remain in
dyntick-idle mode indefinitely, which means that its RCU callbacks might
never be invoked.  This situation can result in grace-period delays or
even system hangs, which matches Pascal's observations of slow boot-up
and shutdown (https://lkml.org/lkml/2012/4/5/142).  See also the bugzilla:

	https://bugzilla.redhat.com/show_bug.cgi?id=806548

This commit therefore causes the "should never be invoked" timer handler
rcu_idle_gp_timer_func() to use smp_call_function_single() to wake up
the CPU for which the timer was intended, allowing that CPU to invoke
its RCU callbacks in a timely manner.

Reported-by: Pascal Chapperon <pascal.chapperon@wanadoo.fr>
Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2012-05-09 14:26:56 -07:00
..
9p.h net/9p: Convert net/9p protocol dumps to tracepoints 2011-10-24 11:13:12 -05:00
asoc.h ASoC: Add another DAPM stat for neighbour checks 2011-09-22 17:24:40 +01:00
block.h blktrace: add FLUSH/FUA support 2011-08-11 10:36:05 +02:00
btrfs.h btrfs: return void in functions without error conditions 2012-03-22 01:45:34 +01:00
compaction.h mm: compaction: add trace events for memory compaction activity 2011-01-13 17:32:33 -08:00
ext3.h ext3: Add fixed tracepoints 2011-06-25 17:29:51 +02:00
ext4.h ext4: do not reference pa_inode from group_pa 2011-12-18 15:49:54 -05:00
gfpflags.h mm: tracing: add missing GFP flags to tracing 2011-05-11 18:50:45 -07:00
gpio.h gpio: add trace events for setting direction and value 2011-05-20 00:40:19 -06:00
irq.h rcu: Use softirq to address performance regression 2011-06-14 15:25:39 -07:00
jbd2.h jbd2: issue cache flush after checkpointing even with internal journal 2012-03-13 22:22:54 -04:00
jbd.h jbd: Add fixed tracepoints 2011-06-25 17:29:51 +02:00
kmem.h mm-tracepoint: rename page-free events 2012-01-10 16:30:41 -08:00
kvm.h KVM: cleanup async_pf tracepoints 2011-01-12 11:28:57 +02:00
lock.h tracing: Factorize lock events in a lock class 2010-05-09 13:45:35 +02:00
mce.h tracing: Fix event alignment: mce:mce_record 2011-03-10 10:34:28 -05:00
module.h include: replace linux/module.h with "struct module" wherever possible 2011-10-31 19:32:32 -04:00
napi.h napi: Convert trace_napi_poll to TRACE_EVENT 2010-09-07 17:51:01 +02:00
net.h net: tracepoint of net_dev_xmit sees freed skb and causes panic 2011-06-02 14:06:31 -07:00
oom.h tracepoint: add tracepoints for debugging oom_score_adj 2012-01-10 16:30:44 -08:00
power.h Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-03-20 21:12:50 -07:00
printk.h printk/tracing: Add console output tracing 2012-02-13 13:46:05 -05:00
rcu.h rcu: Make RCU_FAST_NO_HZ handle timer migration 2012-05-09 14:26:56 -07:00
regmap.h The following text was taken from the original review request: 2012-03-24 10:41:37 -07:00
regulator.h regulator: Add basic trace facilities 2011-01-12 14:33:00 +00:00
rpm.h device.h: audit and cleanup users in main include dir 2012-03-16 10:38:24 -04:00
sched.h tracing, sched, vfs: Fix 'old_pid' usage in trace_sched_process_exec() 2012-03-31 11:53:22 +02:00
scsi.h [SCSI] Include protection operation in SCSI command trace 2011-03-14 18:36:02 -05:00
signal.h tracing: let trace_signal_generate() report more info, kill overflow_fail/lose_info 2012-01-13 18:48:50 +01:00
skb.h tracing: Fix event alignment: skb:kfree_skb 2011-03-10 10:34:31 -05:00
sock.h core: add tracepoints for queueing skb to rcvbuf 2011-06-21 16:06:10 -07:00
sunrpc.h SUNRPC: Adding status trace points 2012-02-06 10:37:53 -05:00
syscalls.h tracing: Allow raw syscall trace events for non privileged users 2010-11-18 14:37:43 +01:00
task.h tracepoint: add tracepoints for debugging oom_score_adj 2012-01-10 16:30:44 -08:00
timer.h tracing: Fix timer tracing 2010-08-19 13:00:41 +02:00
udp.h udp: add tracepoints for queueing skb to rcvbuf 2011-06-21 16:06:10 -07:00
vmscan.h vmscan/trace: Add 'file' info to trace_mm_vmscan_lru_isolate() 2012-01-12 20:13:08 -08:00
workqueue.h workqueue: add queue_work and activate_work trace points 2010-10-05 10:49:55 +02:00
writeback.h The following text was taken from the original review request: 2012-03-24 10:41:37 -07:00
xen.h xen/tracing: fix compile errors when tracing is disabled. 2011-07-25 15:51:02 -07:00