- 12 Jul, 2013 21 commits
-
-
David Ahern authored
No need to malloc the memory for it. Signed-off-by:
David Ahern <dsahern@gmail.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1372793245-4136-6-git-send-email-dsahern@gmail.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
David Ahern authored
Function should only be freeing the entries in the list in case of failure, as those were allocated there, not the list_head itself. Signed-off-by:
-
David Ahern authored
Fixes valgrind complaint: ==1870== Syscall param write(buf) points to uninitialised byte(s) ==1870== at 0x4E3F5B0: __write_nocancel (in /lib64/libpthread-2.14.90.so) ==1870== by 0x449D7C: perf_evlist__start_workload (evlist.c:846) ==1870== by 0x427BC1: cmd_record (builtin-record.c:561) ==1870== by 0x419D72: run_builtin (perf.c:319) ==1870== by 0x4195F2: main (perf.c:376) ==1870== Address 0x7feffcdd7 is on thread 1's stack Signed-off-by:
-
Runzhen Wang authored
Power7 supports over 530 different perf events but only a small subset of these can be specified by name, for the remaining events, we must specify them by their raw code: perf stat -e r2003c <application> This patch makes all the POWER7 events available in sysfs. So we can instead specify these as: perf stat -e 'cpu/PM_CMPLU_STALL_DFU/' <application> where PM_CMPLU_STALL_DFU is the r2003c in previous example. Before this patch is applied, the size of power7-pmu.o is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 3073 2720 0 5793 16a1 arch/powerpc/perf/power7-pmu.o and after the patch is applied, it is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 15950 31112 0 47062 b7d6 arch/powerpc/perf/power7-pmu.o For the run time overhead, I use two scripts, one is "event_name.sh", which contains 50 event names, it looks like: # ./perf record -e 'cpu/PM_CMPLU_STALL_DFU/' -e ..... /bin/sleep 1 the other one is named "event_code.sh" which use corresponding events raw code instead of events names, it looks like: # ./perf record -e r2003c -e ...... /bin/sleep 1 below is the result. Using events name: [root@localhost perf]# time ./event_name.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ] real 0m1.192s user 0m0.028s sys 0m0.106s Using events raw code: [root@localhost perf]# time ./event_code.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.003 MB perf.data (~112 samples) ] real 0m1.198s user 0m0.028s sys 0m0.105s Signed-off-by:Runzhen Wang <runzhen@linux.vnet.ibm.com> Acked-by:
Michael Ellerman <michael@ellerman.id.au> Cc: icycoder@gmail.com Cc: linuxppc-dev@lists.ozlabs.org Cc: Michael Ellerman <michael@ellerman.id.au> Cc: Paul Mackerras <paulus@samba.org> Cc: Runzhen Wang <runzhew@clemson.edu> Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/1372407297-6996-3-git-send-email-runzhen@linux.vnet.ibm.comSigned-off-by:
-
Jiri Olsa authored
There's no point of having out_delete label with perf_session__delete call within __cmd_report function, because it's called at the end of the cmd_report function. The speed up due to commenting out the perf_session__delete at the end does not seem relevant anymore. Measured speedup for ~1GB data file with 222466 FORKS events is around 0.5%. $ perf report -i perf.data.delete -P perf_session__delete -s parent + 99.51% [other] + 0.49% perf_session__delete Signed-off-by:
Jiri Olsa <jolsa@redhat.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/1372161253-22081-6-git-send-email-jolsa@redhat.comSigned-off-by:
-
Jiri Olsa authored
I found the parent symbol column data interesting even if there's another sorting enabled. Switching it on. Previous behaviour: $ perf report -i perf.data.delete -p perf_session__delete -x + 3.60% perf perf [.] __rb_change_child + 1.89% perf perf [.] rb_erase + 1.89% perf perf [.] rb_erase + 1.83% perf perf [.] free@plt Current behaviour: $ perf report -i perf.data.delete -p perf_session__delete -x + 3.60% perf perf [.] __rb_change_child perf_session__delete + 1.89% perf perf [.] rb_erase perf_session__delete_dead_threads + 1.89% perf perf [.] rb_erase perf_session__delete_threads + 1.83% perf perf [.] free@plt perf_session__delete Signed-off-by:
-
Namhyung Kim authored
Most tracepoint events already have their system and event name in ->name field so that searching whole event tracing directory for each evsel to match given id is suboptimal. Factor out this routine into tracepoint_name_to_path(). In case of en invalid name, it'll try to find path using id again. Signed-off-by:
Namhyung Kim <namhyung@kernel.org> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Stephane Eranian <eranian@google.com> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/1372230862-15861-3-git-send-email-namhyung@kernel.orgSigned-off-by:
-
Namhyung Kim authored
Since they're generic helpers move them to util.c so that they can be used by others. Signed-off-by:
-
Jiri Olsa authored
Removing callchain_cursor_reset call as it is called in subsequent machine__resolve_callchain_sample function. Signed-off-by:
-
David Ahern authored
Signed-off-by:
-
Robert Richter authored
Header files of libtraceevent or no longer local headers. Thus, use default path notation for them. Also removing extra traceevent include path and instead handle this similar to liblk. Signed-off-by:
Robert Richter <robert.richter@linaro.org> Signed-off-by:
Robert Richter <rric@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Robert Richter <rric@kernel.org> Link: http://lkml.kernel.org/r/1370964558-8599-1-git-send-email-rric@kernel.orgSigned-off-by:
-
Jiri Olsa authored
Making TEST_ASSERT_VAL global as it's used in multiple objects. Signed-off-by:
-
Jiri Olsa authored
Omitting end of the function check failure for test 1, since there's no way to get exact symbol end via kallsyms. Leaving the debug message. Signed-off-by:
-
Jiri Olsa authored
Removing 'cwd' from perf_session struct as it's no longer needed. Signed-off-by:
-
Sukadev Bhattiprolu authored
perf: Add objdump option to 'perf top' Like with 'perf annotate' add the --objdump option to perf top so users can specify an alternate path to the /usr/bin/objdump binary. Reported-by:
David A. Gilbert <DavidAGilbert@uk.ibm.com> Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: DavidAGilbert@uk.ibm.com Cc: Jiri Olsa <jolsa@redhat.com> Cc: Maynard Johnson <mpjohn@us.ibm.com> Link: http://lkml.kernel.org/r/20130515055651.GA9985@us.ibm.comSigned-off-by:
-
Sukadev Bhattiprolu authored
The CPU map is in an "empty" (or not-applicable) state when monitoring specific threads. cpu_map__all() returns true if the CPU map is in this empty state (i.e for the 'empty_cpu_map' or if we created the map via cpu_map__dummy_new(). The name, cpu_map__all(), is misleading, because even when monitoring all CPUs, (eg: perf record -a), cpu_map__all() returns false. Rename cpu_map__all() to cpu_map__empty(). Signed-off-by:
-
Jiri Olsa authored
It gives the following benefits: - only one function pointer is passed along the way - the 'match' function is called within output function and could be inlined by the compiler Suggested-by:Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by:
Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1373388991-9711-1-git-send-email-jolsa@redhat.comSigned-off-by:
Ingo Molnar <mingo@kernel.org>
-
Peter Zijlstra authored
Jiri managed to trigger this warning: [] ====================================================== [] [ INFO: possible circular locking dependency detected ] [] 3.10.0+ #228 Tainted: G W [] ------------------------------------------------------- [] p/6613 is trying to acquire lock: [] (rcu_node_0){..-...}, at: [<ffffffff810ca797>] rcu_read_unlock_special+0xa7/0x250 [] [] but task is already holding lock: [] (&ctx->lock){-.-...}, at: [<ffffffff810f2879>] perf_lock_task_context+0xd9/0x2c0 [] [] which lock already depends on the new lock. [] [] the existing dependency chain (in reverse order) is: [] [] -> #4 (&ctx->lock){-.-...}: [] -> #3 (&rq->lock){-.-.-.}: [] -> #2 (&p->pi_lock){-.-.-.}: [] -> #1 (&rnp->nocb_gp_wq[1]){......}: [] -> #0 (rcu_node_0){..-...}: Paul was quick to explain that due to preemptible RCU we cannot call rcu_read_unlock() while holding scheduler (or nested) locks when part of the read side critical section was preemptible. Therefore solve it by making the entire RCU read side non-preemptible. Also pull out the retry from under the non-preempt to play nice with RT. Reported-by:Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: <stable@kernel.org> Signed-off-by:
-
Jiri Olsa authored
The '!ctx->is_active' check has a valid scenario, so there's no need for the warning. The reason is that there's a time window between the 'ctx->is_active' check in the perf_event_enable() function and the __perf_event_enable() function having: - IRQs on - ctx->lock unlocked where the task could be killed and 'ctx' deactivated by perf_event_exit_task(), ending up with the warning below. So remove the WARN_ON_ONCE() check and add comments to explain it all. This addresses the following warning reported by Vince Weaver: [ 324.983534] ------------[ cut here ]------------ [ 324.984420] WARNING: at kernel/events/core.c:1953 __perf_event_enable+0x187/0x190() [ 324.984420] Modules linked in: [ 324.984420] CPU: 19 PID: 2715 Comm: nmi_bug_snb Not tainted 3.10.0+ #246 [ 324.984420] Hardware name: Supermicro X8DTN/X8DTN, BIOS 4.6.3 01/08/2010 [ 324.984420] 0000000000000009 ffff88043fce3ec8 ffffffff8160ea0b ffff88043fce3f00 [ 324.984420] ffffffff81080ff0 ffff8802314fdc00 ffff880231a8f800 ffff88043fcf7860 [ 324.984420] 0000000000000286 ffff880231a8f800 ffff88043fce3f10 ffffffff8108103a [ 324.984420] Call Trace: [ 324.984420] <IRQ> [<ffffffff8160ea0b>] dump_stack+0x19/0x1b [ 324.984420] [<ffffffff81080ff0>] warn_slowpath_common+0x70/0xa0 [ 324.984420] [<ffffffff8108103a>] warn_slowpath_null+0x1a/0x20 [ 324.984420] [<ffffffff81134437>] __perf_event_enable+0x187/0x190 [ 324.984420] [<ffffffff81130030>] remote_function+0x40/0x50 [ 324.984420] [<ffffffff810e51de>] generic_smp_call_function_single_interrupt+0xbe/0x130 [ 324.984420] [<ffffffff81066a47>] smp_call_function_single_interrupt+0x27/0x40 [ 324.984420] [<ffffffff8161fd2f>] call_function_single_interrupt+0x6f/0x80 [ 324.984420] <EOI> [<ffffffff816161a1>] ? _raw_spin_unlock_irqrestore+0x41/0x70 [ 324.984420] [<ffffffff8113799d>] perf_event_exit_task+0x14d/0x210 [ 324.984420] [<ffffffff810acd04>] ? switch_task_namespaces+0x24/0x60 [ 324.984420] [<ffffffff81086946>] do_exit+0x2b6/0xa40 [ 324.984420] [<ffffffff8161615c>] ? _raw_spin_unlock_irq+0x2c/0x30 [ 324.984420] [<ffffffff81087279>] do_group_exit+0x49/0xc0 [ 324.984420] [<ffffffff81096854>] get_signal_to_deliver+0x254/0x620 [ 324.984420] [<ffffffff81043057>] do_signal+0x57/0x5a0 [ 324.984420] [<ffffffff8161a164>] ? __do_page_fault+0x2a4/0x4e0 [ 324.984420] [<ffffffff8161665c>] ? retint_restore_args+0xe/0xe [ 324.984420] [<ffffffff816166cd>] ? retint_signal+0x11/0x84 [ 324.984420] [<ffffffff81043605>] do_notify_resume+0x65/0x80 [ 324.984420] [<ffffffff81616702>] retint_signal+0x46/0x84 [ 324.984420] ---[ end trace 442ec2f04db3771a ]--- Reported-by:
Vince Weaver <vincent.weaver@maine.edu> Signed-off-by:
-
Jiri Olsa authored
Currently when the child context for inherited events is created, it's based on the pmu object of the first event of the parent context. This is wrong for the following scenario: - HW context having HW and SW event - HW event got removed (closed) - SW event stays in HW context as the only event and its pmu is used to clone the child context The issue starts when the cpu context object is touched based on the pmu context object (__get_cpu_context). In this case the HW context will work with SW cpu context ending up with following WARN below. Fixing this by using parent context pmu object to clone from child context. Addresses the following warning reported by Vince Weaver: [ 2716.472065] ------------[ cut here ]------------ [ 2716.476035] WARNING: at kernel/events/core.c:2122 task_ctx_sched_out+0x3c/0x) [ 2716.476035] Modules linked in: nfsd auth_rpcgss oid_registry nfs_acl nfs locn [ 2716.476035] CPU: 0 PID: 3164 Comm: perf_fuzzer Not tainted 3.10.0-rc4 #2 [ 2716.476035] Hardware name: AOpen DE7000/nMCP7ALPx-DE R1.06 Oct.19.2012, BI2 [ 2716.476035] 0000000000000000 ffffffff8102e215 0000000000000000 ffff88011fc18 [ 2716.476035] ffff8801175557f0 0000000000000000 ffff880119fda88c ffffffff810ad [ 2716.476035] ffff880119fda880 ffffffff810af02a 0000000000000009 ffff880117550 [ 2716.476035] Call Trace: [ 2716.476035] [<ffffffff8102e215>] ? warn_slowpath_common+0x5b/0x70 [ 2716.476035] [<ffffffff810ab2bd>] ? task_ctx_sched_out+0x3c/0x5f [ 2716.476035] [<ffffffff810af02a>] ? perf_event_exit_task+0xbf/0x194 [ 2716.476035] [<ffffffff81032a37>] ? do_exit+0x3e7/0x90c [ 2716.476035] [<ffffffff810cd5ab>] ? __do_fault+0x359/0x394 [ 2716.476035] [<ffffffff81032fe6>] ? do_group_exit+0x66/0x98 [ 2716.476035] [<ffffffff8103dbcd>] ? get_signal_to_deliver+0x479/0x4ad [ 2716.476035] [<ffffffff810ac05c>] ? __perf_event_task_sched_out+0x230/0x2d1 [ 2716.476035] [<ffffffff8100205d>] ? do_signal+0x3c/0x432 [ 2716.476035] [<ffffffff810abbf9>] ? ctx_sched_in+0x43/0x141 [ 2716.476035] [<ffffffff810ac2ca>] ? perf_event_context_sched_in+0x7a/0x90 [ 2716.476035] [<ffffffff810ac311>] ? __perf_event_task_sched_in+0x31/0x118 [ 2716.476035] [<ffffffff81050dd9>] ? mmdrop+0xd/0x1c [ 2716.476035] [<ffffffff81051a39>] ? finish_task_switch+0x7d/0xa6 [ 2716.476035] [<ffffffff81002473>] ? do_notify_resume+0x20/0x5d [ 2716.476035] [<ffffffff813654f5>] ? retint_signal+0x3d/0x78 [ 2716.476035] ---[ end trace 827178d8a5966c3d ]--- Reported-by: -
Ingo Molnar authored
Merge tag 'perf-urgent-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent Pull perf/urgent fixes from Arnaldo Carvalho de Melo: * Fix some freeing bugs on the parsing error paths, from Adrian Hunter. * Update symbol_conf.nr_events when processing attribute events, fix from Adrian Hunter. * Fix missing increment in sample parsing when PERF_SAMPLE_STACK_USER is present, from Adrian Hunt. * Fix count parameter to read call in event_format__new, from David Ahern. * Remove -A/--append option, not working for a long time, from Jiri Olsa. * Remove -f/--force option, was a no-op for quite some time, from Jiri Olsa. * Fix -x/--exclude-other option for report command, from Jiri Olsa. * Cross build fixes, at least one for Android, from Joonsoo Kim. * Fix memory allocation fail check in mem{set,cpy} 'perf bench' workloads, from Kirill A. Shutemov. * Revert regression in configuration of Python support, from Michael Witten. * Fix -ldw/-lelf link test when static linking, from Mike Frysinger. * Fix issues with multiple children processing in perf_evlist__start_workload(), from Namhyung Kim. * Fix broken include in Context.xs ('perf script'), from Ramkumar Ramachandra. * Fixes for build problems, from Robert Richter. * Fix a typo of a Power7 event name, from Runzhen Wang. * Avoid sending SIGTERM to random processes in 'perf stat', fix from Stephane Eranian. * Fix per-socket output bug for uncore events in 'perf stat', from Stephane Eranian. * Fix vdso list searching, from Waiman Long. Signed-off-by:
-
- 10 Jul, 2013 2 commits
-
-
Ramkumar Ramachandra authored
765532c8 (perf script: Finish the rename from trace to script, 2010-12-23) made a mistake during find-and-replace replacing "../../../util/trace-event.h" with "../../../util/script-event.h", a non-existent file. Fix this include. Signed-off-by:
Ramkumar Ramachandra <artagnon@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: http://lkml.kernel.org/r/1373364033-7918-3-git-send-email-artagnon@gmail.comSigned-off-by:
-
Mike Frysinger authored
Since libelf sometimes uses libpthread, we have to list that after -lelf when someone tries to build statically. Else things go boom: Makefile:479: *** No libelf.h/libelf found, please install \ libelf-dev/elfutils-libelf-devel. Stop. Similarly, the -ldw test fails as it often uses -lz: Makefile:462: No libdw.h found or old libdw.h found or elfutils is older \ than 0.138, disables dwarf support. Please install new elfutils-devel/libdw-dev And if we add debugging to try-cc, we see: + echo '#include <dwarf.h> int main(void) { Dwarf *dbg = dwarf_begin(0, DWARF_C_READ); return (long)dbg; }' + i686-pc-linux-gnu-gcc -x c - -O2 -pipe -march=atom -mtune=atom -mfpmath=sse -g \ -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE \ -ldw -lelf -static -lpthread -lrt -lelf -lm -o .24368 /usr/lib/libdw.a(dwarf_begin_elf.o):function check_section.isra.1: error: undefined reference to 'inflateInit_' /usr/lib/libdw.a(dwarf_begin_elf.o):function check_section.isra.1: error: undefined reference to 'inflate' /usr/lib/libdw.a(dwarf_begin_elf.o):function check_section.isra.1: error: undefined reference to 'inflateReset' /usr/lib/libdw.a(dwarf_begin_elf.o):function check_section.isra.1: error: undefined reference to 'inflateEnd' + echo '#include <libelf.h> int main(void) { Elf *elf = elf_begin(0, ELF_C_READ, 0); return (long)elf; }' + i686-pc-linux-gnu-gcc -x c - -O2 -pipe -march=atom -mtune=atom -mfpmath=sse -g \ -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE \ -static -lpthread -lrt -lelf -lm -o .19216 /usr/lib/libelf.a(elf_begin.o):function file_read_elf: error: undefined reference to 'pthread_rwlock_init' /usr/lib/libelf.a(elf_begin.o):function __libelf_read_mmaped_file: error: undefined reference to 'pthread_rwlock_init' /usr/lib/libelf.a(elf_begin.o):function __libelf_read_mmaped_file: error: undefined reference to 'pthread_rwlock_init' /usr/lib/libelf.a(elf_begin.o):function read_file: error: undefined reference to 'pthread_rwlock_init' /usr/lib/libelf.a(elf_begin.o):function lock_dup_elf.8072: error: undefined reference to 'pthread_rwlock_unlock' /usr/lib/libelf.a(elf_begin.o):function lock_dup_elf.8072: error: undefined reference to 'pthread_rwlock_wrlock' /usr/lib/libelf.a(elf_begin.o):function elf_begin: error: undefined reference to 'pthread_rwlock_rdlock' /usr/lib/libelf.a(elf_begin.o):function elf_begin: error: undefined reference to 'pthread_rwlock_unlock' Signed-off-by:Mike Frysinger <vapier@gentoo.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1368073064-18276-1-git-send-email-vapier@gentoo.orgSigned-off-by:
-
- 09 Jul, 2013 1 commit
-
-
Michael Witten authored
Among other things, the following: commit 31160d7f Date: Tue Jan 8 16:22:36 2013 -0500 perf tools: Fix GNU make v3.80 compatibility issue attempts to aid the user by tapping into an existing error message, as described in the commit message: ... Also fix an issue where _get_attempt was called with only one argument. This prevented the error message from printing the name of the variable that can be used to fix the problem. or more precisely: -$(if $($(1)),$(call _ge_attempt,$($(1)),$(1)),$(call _ge_attempt,$(2))) +$(if $($(1)),$(call _ge_attempt,$($(1)),$(1)),$(call _ge_attempt,$(2),$(1))) However, The "missing" argument was in fact missing on purpose; it's absence is a signal that the error message should be skipped, because the failure would be due to the default value, not any user-supplied value. This can be seen in how `_ge_attempt' uses `gea_err' (in the config/utilities.mak file): _ge_attempt = $(if $(get-executable),$(get-executable),$(_gea_warn)$(call _gea_err,$(2))) _gea_warn = $(warning The path '$(1)' is not executable.) _gea_err = $(if $(1),$(error Please set '$(1)' appropriately)) That is, because the argument is no longer missing, the value `$(1)' (associated with `_gea_err') always evaluates to true, thus always triggering the error condition that is meant to be reserved for only the case when a user explicitly supplies an invalid value. Concretely, the result is a regression in the Makefile's configuration of python support; rather than gracefully disable support when the relevant executables cannot be found according to default values, the build process halts in error as though the user explicitly supplied the values. This new commit simply reverts the offending one-line change. Reported-by:
Pekka Enberg <penberg@kernel.org> Link: http://lkml.kernel.org/r/CAOJsxLHv17Ys3M7P5q25imkUxQW6LE_vABxh1N3Tt7Mv6Ho4iw@mail.gmail.comSigned-off-by:
Michael Witten <mfwitten@gmail.com>
-
- 08 Jul, 2013 16 commits
-
-
Robert Richter authored
The tag of the perf version is wrongly determined, always the latest tag is taken regardless of the HEAD commit: $ perf --version perf version 3.9.rc8.gd7f5d3 $ git describe d7f5d3 v3.9-rc7-154-gd7f5d33 $ head -n 4 Makefile VERSION = 3 PATCHLEVEL = 9 SUBLEVEL = 0 EXTRAVERSION = -rc7 In other cases no tag might be found. This patch fixes this. This new implementation handles also the case if there are no tags at all found in the git repo but there is a commit id. Signed-off-by:
Robert Richter <robert.richter@calxeda.com> Link: http://lkml.kernel.org/r/1368006214-12912-1-git-send-email-rric@kernel.orgSigned-off-by:
-
Stephane Eranian authored
This patch fixes a problem reported by Andi Kleen on perf stat when measuring uncore events: # perf stat --per-socket -e uncore_pcu/event=0x0/ -I1000 -a sleep 2 It would not report counts for the second socket. That was due to a cpu mapping bug in print_aggr(). This patch also fixes the socket numbering bug for <not counted> events. Reported-by:
Andi Kleen <ak@linux.intel.com> Signed-off-by:
Stephane Eranian <eranian@google.com> Tested-by:
-
Waiman Long authored
When "perf record" was used on a large machine with a lot of CPUs, the perf post-processing time (the time after the workload was done until the perf command itself exited) could take a lot of minutes and even hours depending on how large the resulting perf.data file was. While running AIM7 1500-user high_systime workload on a 80-core x86-64 system with a 3.9 kernel (with only the -s -a options used), the workload itself took about 2 minutes to run and the perf.data file had a size of 1108.746 MB. However, the post-processing step took more than 10 minutes. With a gprof-profiled perf binary, the time spent by perf was as follows: % cumulative self self total time seconds seconds calls s/call s/call name 96.90 822.10 822.10 192156 0.00 0.00 dsos__find 0.81 828.96 6.86 172089958 0.00 0.00 rb_next 0.41 832.44 3.48 48539289 0.00 0.00 rb_erase So 97% (822 seconds) of the time was spent in a single dsos_find() function. After analyzing the call-graph data below: ----------------------------------------------- 0.00 822.12 192156/192156 map__new [6] [7] 96.9 0.00 822.12 192156 vdso__dso_findnew [7] 822.10 0.00 192156/192156 dsos__find [8] 0.01 0.00 192156/192156 dsos__add [62] 0.01 0.00 192156/192366 dso__new [61] 0.00 0.00 1/45282525 memdup [31] 0.00 0.00 192156/192230 dso__set_long_name [91] ----------------------------------------------- 822.10 0.00 192156/192156 vdso__dso_findnew [7] [8] 96.9 822.10 0.00 192156 dsos__find [8] ----------------------------------------------- It was found that the vdso__dso_findnew() function failed to locate VDSO__MAP_NAME ("[vdso]") in the dso list and have to insert a new entry at the end for 192156 times. This problem is due to the fact that there are 2 types of name in the dso entry - short name and long name. The initial dso__new() adds "[vdso]" to both the short and long names. After that, vdso__dso_findnew() modifies the long name to something like /tmp/perf-vdso.so-NoXkDj. The dsos__find() function only compares the long name. As a result, the same vdso entry is duplicated many time in the dso list. This bug increases memory consumption as well as slows the symbol processing time to a crawl. To resolve this problem, the dsos__find() function interface was modified to enable searching either the long name or the short name. The vdso__dso_findnew() will now search only the short name while the other call sites search for the long name as before. With this change, the cpu time of perf was reduced from 848.38s to 15.77s and dsos__find() only accounted for 0.06% of the total time. 0.06 15.73 0.01 192151 0.00 0.00 dsos__find Signed-off-by:Waiman Long <Waiman.Long@hp.com> Acked-by:
-
Adrian Hunter authored
The final sample format bit used to be PERF_SAMPLE_STACK_USER which neglected to do a final increment of the array pointer. The result is that the following parsing might start at the wrong place. Signed-off-by:
Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/r/1372944040-32690-16-git-send-email-adrian.hunter@intel.comSigned-off-by:
-
Adrian Hunter authored
Signed-off-by:
-
Adrian Hunter authored
On the error path, newly allocated 'term' must be freed. Signed-off-by:
-
Adrian Hunter authored
On the error path, 'data.terms' may not have been initialised. Signed-off-by:
-
David Ahern authored
per realloc above the length of the buffer is alloc_size, not BUFSIZ. Adjust length per size as done for buf start. Addresses some valgrind complaints: ==1870== Syscall param read(buf) points to unaddressable byte(s) ==1870== at 0x4E3F610: __read_nocancel (in /lib64/libpthread-2.14.90.so) ==1870== by 0x44AEE1: event_format__new (unistd.h:45) ==1870== by 0x44B025: perf_evsel__newtp (evsel.c:158) ==1870== by 0x451919: add_tracepoint_event (parse-events.c:395) ==1870== by 0x479815: parse_events_parse (parse-events.y:292) ==1870== by 0x45463A: parse_events_option (parse-events.c:861) ==1870== by 0x44FEE4: get_value (parse-options.c:113) ==1870== by 0x450767: parse_options_step (parse-options.c:192) ==1870== by 0x450C40: parse_options (parse-options.c:422) ==1870== by 0x42735F: cmd_record (builtin-record.c:918) ==1870== by 0x419D72: run_builtin (perf.c:319) ==1870== by 0x4195F2: main (perf.c:376) ==1870== Address 0xcffebf0 is 0 bytes after a block of size 8,192 alloc'd ==1870== at 0x4C2A62F: malloc (vg_replace_malloc.c:270) ==1870== by 0x4C2A7A3: realloc (vg_replace_malloc.c:662) ==1870== by 0x44AF07: event_format__new (evsel.c:121) ==1870== by 0x44B025: perf_evsel__newtp (evsel.c:158) ==1870== by 0x451919: add_tracepoint_event (parse-events.c:395) ==1870== by 0x479815: parse_events_parse (parse-events.y:292) ==1870== by 0x45463A: parse_events_option (parse-events.c:861) ==1870== by 0x44FEE4: get_value (parse-options.c:113) ==1870== by 0x450767: parse_options_step (parse-options.c:192) ==1870== by 0x450C40: parse_options (parse-options.c:422) ==1870== by 0x42735F: cmd_record (builtin-record.c:918) ==1870== by 0x419D72: run_builtin (perf.c:319) Signed-off-by:
-
Runzhen Wang authored
In the Power7 PMU guide: https://www.power.org/documentation/commonly-used-metrics-for-performance-analysis/ PM_BRU_MPRED is referred to as PM_BR_MPRED. It fixed the typo by changing the name of the event in kernel and documentation accordingly. This patch changes the ABI, there are some reasons I think it's ok: - It is relatively new interface, specific to the Power7 platform. - No tools that we know of actually use this interface at this point (none are listed near the interface). - Users of this interface (eg oprofile users migrating to perf) would be more used to the "PM_BR_MPRED" rather than "PM_BRU_MPRED". - These are in the ABI/testing at this point rather than ABI/stable, so hoping we have some wiggle room. Signed-off-by:
-
Jiri Olsa authored
Currently we have symbol_conf.exclude_other being set as true every time so the -x/--exclude-other has nothing to do. Also we have no way to see the data with symbol_conf.exclude_other being false which is useful sometimes. Fixing it by making symbol_conf.exclude_other false by default. 1) Example without -x option: $ perf report -i perf.data.delete -p perf_session__delete -s parent + 99.91% [other] + 0.08% perf_session__delete + 0.00% perf_session__delete_dead_threads + 0.00% perf_session__delete_threads 2) Example with -x option: $ ./perf report -i perf.data.delete -p perf_session__delete -s parent -x + 96.22% perf_session__delete + 1.89% perf_session__delete_dead_threads + 1.89% perf_session__delete_threads In Example 1) we get the sorted out data together with the rest "[other]". This could help us estimate how much time we spent in the sorted data. In Example 2) the total is just the sorted data. Signed-off-by:
-
Namhyung Kim authored
When perf tries to start a workload, it relies on a pipe which the workload was blocked for reading. After closing the pipe on the parent, the workload (child) can start the actual work via exec(). However, if another process was forked after creating a workload, this mechanism cannot work since the other process (child) also inherits the pipe, so that closing the pipe in parent cannot unblock the workload. Fix it by using explicit write call can then closing it. For similar reason, the pipe fd on parent should be marked as CLOEXEC so that it can be closed after another child exec'ed. Signed-off-by:
-
Jiri Olsa authored
It no longer have any affect on the processing and is marked as obsolete anyway. Signed-off-by:
-
Jiri Olsa authored
It's no longer working and needed. Quite straightforward discussion/vote was in here: http://marc.info/?t=137028288300004&r=1&w=2Signed-off-by:
-
Stephane Eranian authored
This patch fixes a problem with perf stat whereby on termination it may send a SIGTERM signal to random processes on systems with high PID recycling. I got some actual bug reports on this. There is race between the SIGCHLD and sig_atexit() handlers. This patch addresses this problem by clearing child_pid in the SIGCHLD handler. Signed-off-by:
-
Joonsoo Kim authored
Currently, lib lk doesn't use CROSS_COMPILE environment variable, so cross build always fails. This is a quick fix for this problem. Signed-off-by:
Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1371603750-15053-2-git-send-email-iamjoonsoo.kim@lge.comSigned-off-by:
-
Joonsoo Kim authored
Building perf for android fails because it can't find the definition of struct winsize. This definition is in termios.h, so I add this header to util.h to solve the problem. It is missed by commit '2c803e52' which moves get_term_dimensions() from builtin-top.c to util.c, but missed to move termios.h header. Signed-off-by:
-
