For some events it's useful to weight sample with a hardware
provided number. This expresses how expensive the action the
sample represent was. This allows the profiler to scale
the samples to be more informative to the programmer.
There is already the period which is used similarly, but it
means something different, so I chose to not overload it.
Instead a new sample type for WEIGHT is added.
Can be used for multiple things. Initially it is used for TSX
abort costs and profiling by memory latencies (so to make
expensive load appear higher up in the histograms). The concept
is quite generic and can be extended to many other kinds of
events or architectures, as long as the hardware provides
suitable auxillary values. In principle it could be also used
for software tracepoints.
This adds the generic glue. A new optional sample format for a
64-bit weight value.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Cc: acme@redhat.com
Cc: jolsa@redhat.com
Cc: namhyung.kim@lge.com
Link: http://lkml.kernel.org/r/1359040242-8269-5-git-send-email-eranian@google.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This patch adds the --per-core option to perf stat.
This option is used to aggregate system-wide counts
on a per physical core basis. On processors with
hyperthreading, this means counts of all HT threads
running on a physical core are aggregated.
This mode is useful to find imblance between physical
cores running an uniform workload. Cores are identified
by socket: S0-C1, means physical core 1 on socket 0. Note
that cores are identified using their physical core id,
thus their numbering may not be continuous.
Per core aggregation can be combined with interval printing:
# perf stat -a --per-core -I 1000 -e cycles sleep 1000
# time core cpus counts events
1.000090030 S0-C0 1 4,765,747 cycles
1.000090030 S0-C1 1 5,580,647 cycles
1.000090030 S0-C2 1 221,181 cycles
1.000090030 S0-C3 1 266,092 cycles
Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1360846649-6411-4-git-send-email-eranian@google.com
[ committer note: Remove parts already applied on 86ee6e1 to keep bisectability ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Conflicts:
tools/Makefile
Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:
. Honor parallel jobs, fix from Borislav Petkov
. Introduce tools/lib/lk library, initially just removing duplication
among tools/perf and tools/vm. from Borislav Petkov
. Fix build on non-glibc systems due to libio.h absence, from Cody P Schafer.
. Remove some perf_session and tracing dead code, from David Ahern.
. Introduce perf stat --repeat forever, from Frederik Deweerdt.
. Add perf test entries for checking --cpu in record and stat, from Jiri Olsa.
. Add perf test entries for checking breakpoint overflow signal handler issues,
from Jiri Olsa.
. Add perf test entry for for checking number of EXIT events, from Namhyung Kim.
. Simplify some perf_evlist methods and to allow 'stat' to share code with
'record' and 'trace'.
. Remove dead code in related to libtraceevent integration, from Namhyung Kim.
. Event group view for 'annotate' in --stdio, --tui and --gtk, from Namhyung Kim.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
[ resolved the trivial merge conflict with upstream ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Merge in all pending fixes, before pulling the latest development
bits from Arnaldo - which will involve merge conflicts.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently kprobes check whether the copied instruction modifies
IF (interrupt flag) on each probe hit. This results not only in
introducing overhead but also involving
inat_get_opcode_attribute into the kprobes hot path, and it can
cause an infinite recursive call (and kernel panic in the end).
Actually, since the copied instruction itself can never be modified
on the buffer, it is needless to analyze the instruction on every
probe hit.
To fix this issue, we check it only once when registering probe
and store the result on ainsn->if_modifier.
Reported-by: Timo Juhani Lindfors <timo.lindfors@iki.fi>
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Acked-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: yrl.pp-manager.tt@hitachi.com
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20130314115242.19690.33573.stgit@mhiramat-M0-7522
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull perf/urgent fixes from Arnaldo Carvalho de Melo:
. perf probe: Fix segfault due to testing the wrong pointer for NULL,
from Ananth N Mavinakayanahalli.
. libtraceevent: Remove hard coded include to /usr/local/include in
Makefile, which causes cross builds to include host header files,
fix from Jack Mitchell.
. perf record: Use the right target interface for synthesizing
threads when --cpu/-C option is used, fix from Jiri Olsa.
. Check if -DFORTIFY_SOURCE=2 is allowed, as gcc 4.7.2 defines
it and then the build is broken when it is redefined in perf,
fix from Marcin Slusarz.
. Fix build with NO_NEWT=1, that can happen explicitely or when
the newt-devel package is not installed, from Michael Ellerman.
. perf/POWER7: Create a sysfs format entry for Power7 events, missing
patch from a patchseries already merged, from Sukadev Bhattiprolu.
. Fix LIBNUMA build with glibc 2.12 and older, from Vinson Lee.
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
perf_event_task_event() iterates pmu list and generate events
for each eligible pmu context. But if task_event has task_ctx
like in EXIT it'll generate events even though the pmu doesn't
have an eligible one. Fix it by moving the code to proper
places.
Before this patch:
$ perf record -n true
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.006 MB perf.data (~248 samples) ]
$ perf report -D | tail
Aggregated stats:
TOTAL events: 73
MMAP events: 67
COMM events: 2
EXIT events: 4
cycles stats:
TOTAL events: 73
MMAP events: 67
COMM events: 2
EXIT events: 4
After this patch:
$ perf report -D | tail
Aggregated stats:
TOTAL events: 70
MMAP events: 67
COMM events: 2
EXIT events: 1
cycles stats:
TOTAL events: 70
MMAP events: 67
COMM events: 2
EXIT events: 1
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1363332433-7637-1-git-send-email-namhyung@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When cpu/task clock events are initialized, their sampling
frequencies are converted to have a fixed value. However it
missed to update the hwc->last_period which was set to 1 for
initial sampling frequency calibration.
Because this hwc->last_period value is used as a period in
perf_swevent_ hrtime(), every recorded sample will have an
incorrected period of 1.
$ perf record -e task-clock noploop 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.158 MB perf.data (~6919 samples) ]
$ perf report -n --show-total-period --stdio
# Samples: 4K of event 'task-clock'
# Event count (approx.): 4000
#
# Overhead Samples Period Command Shared Object Symbol
# ........ ............ ............ ....... ............. ..................
#
99.95% 3998 3998 noploop noploop [.] main
0.03% 1 1 noploop libc-2.15.so [.] init_cacheinfo
0.03% 1 1 noploop ld-2.15.so [.] open_verify
Note that it doesn't affect the non-sampling event so that the
perf stat still gets correct value with or without this patch.
$ perf stat -e task-clock noploop 1
Performance counter stats for 'noploop 1':
1000.272525 task-clock # 1.000 CPUs utilized
1.000560605 seconds time elapsed
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1363574507-18808-1-git-send-email-namhyung@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Adding automated test for breakpoint event signal handler checking if
it's executed properly.
The test is related to the proper handling of the RF EFLAGS bit on
x86_64, but it's generic for all archs.
First we check the signal handler is properly called and that the
following debug exception return to user space wouldn't trigger
recursive breakpoint.
This is related to x86_64 RF EFLAGS bit being managed in a wrong way.
Second we check that we can set breakpoint in signal handler, which is
not possible on x86_64 if the signal handler is executed with RF EFLAG
set.
This test is inpired by overflow tests done by Vince Weaver.
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: http://lkml.kernel.org/r/1362940871-24486-6-git-send-email-jolsa@redhat.com
[ committer note: s/pr_err/pr_debug/g i.e. print just OK or FAILED in non verbose mode ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
