linux/tools/perf/util
Shaohua Li 09223371de rcu: Use softirq to address performance regression
Commit a26ac2455ffcf3(rcu: move TREE_RCU from softirq to kthread)
introduced performance regression. In an AIM7 test, this commit degraded
performance by about 40%.

The commit runs rcu callbacks in a kthread instead of softirq. We observed
high rate of context switch which is caused by this. Out test system has
64 CPUs and HZ is 1000, so we saw more than 64k context switch per second
which is caused by RCU's per-CPU kthread.  A trace showed that most of
the time the RCU per-CPU kthread doesn't actually handle any callbacks,
but instead just does a very small amount of work handling grace periods.
This means that RCU's per-CPU kthreads are making the scheduler do quite
a bit of work in order to allow a very small amount of RCU-related
processing to be done.

Alex Shi's analysis determined that this slowdown is due to lock
contention within the scheduler.  Unfortunately, as Peter Zijlstra points
out, the scheduler's real-time semantics require global action, which
means that this contention is inherent in real-time scheduling.  (Yes,
perhaps someone will come up with a workaround -- otherwise, -rt is not
going to do well on large SMP systems -- but this patch will work around
this issue in the meantime.  And "the meantime" might well be forever.)

This patch therefore re-introduces softirq processing to RCU, but only
for core RCU work.  RCU callbacks are still executed in kthread context,
so that only a small amount of RCU work runs in softirq context in the
common case.  This should minimize ksoftirqd execution, allowing us to
skip boosting of ksoftirqd for CONFIG_RCU_BOOST=y kernels.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Tested-by: "Alex,Shi" <alex.shi@intel.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
2011-06-14 15:25:39 -07:00
..
include perf tools: Fix build on older systems 2011-05-26 11:16:29 -03:00
scripting-engines perf session: Pass evsel in event_ops->sample() 2011-03-23 19:28:58 -03:00
ui perf: Only include annotate.h once in tools/perf/util/ui/browsers/annotate.c 2011-05-10 10:20:07 +02:00
abspath.c perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
alias.c
annotate.c perf symbols: Rename dso->origin to dso->symtab_type 2011-03-11 13:28:45 -03:00
annotate.h perf top: Live TUI Annotation 2011-02-22 12:02:07 -03:00
bitmap.c perf tools: Don't use code surrounded by __KERNEL__ 2010-05-02 12:00:44 -03:00
build-id.c perf session: Pass evsel in event_ops->sample() 2011-03-23 19:28:58 -03:00
build-id.h perf annotate: Use build-ids to find the right DSO 2010-05-20 12:15:33 -03:00
cache.h perf tools: Don't fallback to setup_pager unconditionally 2011-01-31 18:08:39 -02:00
callchain.c perf tools: Kill event_t typedef, use 'union perf_event' instead 2011-01-29 16:25:37 -02:00
callchain.h perf tools: Kill event_t typedef, use 'union perf_event' instead 2011-01-29 16:25:37 -02:00
cgroup.c perf: Fix a build error with some GCC versions 2011-04-08 17:40:21 +02:00
cgroup.h perf tool: Add cgroup support 2011-02-16 13:30:48 +01:00
color.c perf hist: Replace ->print() routines by ->snprintf() equivalents 2010-04-02 16:28:15 -03:00
color.h perf hist: Replace ->print() routines by ->snprintf() equivalents 2010-04-02 16:28:15 -03:00
config.c perf buildid: add perfconfig option to specify buildid cache dir 2010-06-05 09:34:04 -03:00
cpumap.c perf tools: Add missing cpu_map__delete() 2011-01-22 19:56:30 -02:00
cpumap.h perf tools: Add missing cpu_map__delete() 2011-01-22 19:56:30 -02:00
ctype.c
debug.c perf tools: Fixup exit path when not able to open events 2011-03-29 13:40:27 -03:00
debug.h perf tools: Fixup exit path when not able to open events 2011-03-29 13:40:27 -03:00
debugfs.c
debugfs.h
environment.c
event.c perf symbols: Handle /proc/sys/kernel/kptr_restrict 2011-05-26 11:15:25 -03:00
event.h perf tools: Pre-check sample size before parsing 2011-05-22 03:38:36 +02:00
evlist.c Merge branch 'perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing into perf/urgent 2011-05-22 10:10:01 +02:00
evlist.h Merge branch 'perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing into perf/urgent 2011-05-22 10:10:01 +02:00
evsel.c perf events: initialize fd array to -1 instead of 0 2011-05-27 16:02:12 -03:00
evsel.h perf evsel: Fix use of inherit 2011-04-15 12:52:28 -03:00
exec_cmd.c perf tools: Makefile: Remove various and sundry cruft 2011-02-18 07:43:06 -02:00
exec_cmd.h perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
generate-cmdlist.sh
header.c perf symbols: Handle /proc/sys/kernel/kptr_restrict 2011-05-26 11:15:25 -03:00
header.h perf tools: Move evlist sample helpers to evlist area 2011-05-22 03:12:29 +02:00
help.c perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
help.h
hist.c perf tools: Improve support for sessions with multiple events 2011-03-06 13:13:40 -03:00
hist.h perf session: Pass evsel in event_ops->sample() 2011-03-23 19:28:58 -03:00
hweight.c perf tools: Don't use code surrounded by __KERNEL__ 2010-05-02 12:00:44 -03:00
levenshtein.c
levenshtein.h
map.c perf tools: Fix 64 bit integer format strings 2011-01-22 23:41:57 -02:00
map.h perf probe: Add basic module support 2010-10-21 16:11:44 -02:00
pager.c
parse-events.c perf: Fix multi-event parsing bug 2011-05-17 20:45:36 +02:00
parse-events.h perf script: Add support for H/W and S/W events 2011-03-14 17:07:20 -03:00
parse-options.c perf options: Type check all the remaining OPT_ variants 2010-05-17 16:22:41 -03:00
parse-options.h perf options: add OPT_CALLBACK_DEFAULT_NOOPT 2010-12-06 15:33:29 -02:00
path.c perf tools: add test for strlcpy() 2010-08-21 11:22:47 -03:00
PERF-VERSION-GEN perf tools: Version incorrect with some versions of grep 2011-03-16 08:59:50 -03:00
probe-event.c perf probe: Fix multiple --vars options behavior 2011-04-05 15:36:04 -03:00
probe-event.h perf probe: Add filters support for available functions 2011-01-28 09:20:25 -02:00
probe-finder.c perf probe: Fix the missed parameter initialization 2011-05-10 17:06:23 +02:00
probe-finder.h perf probe: Add fastpath to do lookup by function name 2011-03-29 13:40:27 -03:00
pstack.c perf newt: Make <- zoom out filters 2010-05-14 20:05:21 -03:00
pstack.h perf ui: Move hists browser to util/ui/browsers/ 2010-08-10 16:11:08 -03:00
python.c Merge branch 'perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing into perf/urgent 2011-05-22 10:10:01 +02:00
quote.c perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
quote.h perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
run-command.c perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
run-command.h perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
session.c perf tools: Fix ommitted mmap data update on remap 2011-05-23 13:22:57 +02:00
session.h perf tools: Pre-check sample size before parsing 2011-05-22 03:38:36 +02:00
setup.py perf tools: Fix NO_NEWT=1 python build error 2011-03-29 16:46:57 -03:00
sigchain.c perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
sigchain.h perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
sort.c perf hist: Better displaying of unresolved DSOs and symbols 2010-12-06 15:12:34 -02:00
sort.h perf: Keep track of the max depth of a callchain 2010-08-22 20:43:17 +02:00
strbuf.c perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
strbuf.h perf tools: Remove some unused functions 2010-05-18 23:03:35 -03:00
strfilter.c perf: Fix missing strndup declaration 2011-03-04 01:17:18 +01:00
strfilter.h perf tools: Add strfilter for general purpose string filter 2011-01-28 09:19:38 -02:00
string.c Fix common misspellings 2011-03-31 11:26:23 -03:00
strlist.c
strlist.h
svghelper.c perf timechart: Fix black idle boxes in the title 2011-02-28 08:56:14 +01:00
svghelper.h
symbol.c perf symbols: Handle /proc/sys/kernel/kptr_restrict 2011-05-26 11:15:25 -03:00
symbol.h perf symbols: Handle /proc/sys/kernel/kptr_restrict 2011-05-26 11:15:25 -03:00
thread_map.c perf threads: Move thread_map to separate file 2011-01-24 10:59:00 -02:00
thread_map.h perf threads: Move thread_map to separate file 2011-01-24 10:59:00 -02:00
thread.c perf threads: Move thread_map to separate file 2011-01-24 10:59:00 -02:00
thread.h perf threads: Move thread_map to separate file 2011-01-24 10:59:00 -02:00
top.c perf top: Remove redundant syme->origin field 2011-03-11 13:28:45 -03:00
top.h perf symbol: Move sym_entry->skip to symbol->ignore 2011-03-11 13:36:01 -03:00
trace-event-info.c perf tools: Introduce event selectors 2011-01-03 16:39:04 -02:00
trace-event-parse.c rcu: Use softirq to address performance regression 2011-06-14 15:25:39 -07:00
trace-event-read.c perf: Use read() instead of lseek() in trace_event_read.c:skip() 2010-05-20 08:37:17 +02:00
trace-event-scripting.c perf session: Pass evsel in event_ops->sample() 2011-03-23 19:28:58 -03:00
trace-event.h perf session: Pass evsel in event_ops->sample() 2011-03-23 19:28:58 -03:00
types.h perf tools: Fix 64 bit integer format strings 2011-01-22 23:41:57 -02:00
usage.c
util.c perf util: Move do_read from session to util 2011-01-03 16:50:55 -02:00
util.h perf tools: Makefile: Remove platform-specific cruft 2011-02-18 07:42:07 -02:00
values.c perf tools: Fix 64 bit integer format strings 2011-01-22 23:41:57 -02:00
values.h
wrapper.c perf tools: remove xstrndup, xmalloc, xzalloc 2010-05-18 23:05:28 -03:00
xyarray.c perf tools: Introduce event selectors 2011-01-03 16:39:04 -02:00
xyarray.h perf tools: Introduce event selectors 2011-01-03 16:39:04 -02:00