You can't pass demangled name into "perf probe", because of special chars:
./perf probe -f -x /tmp/a.out 'foo(int)'
Semantic error :There is non-digit char in line number.
And you can't even pass without demangling (because it search symbol in
DSO with demangle=true):
./perf probe -f -x /tmp/a.out _Z3fooi
no symbols found in /tmp/a.out, maybe install a debug package?
However:
nm /tmp/a.out | grep foo
000000000040056d T _Z3fooi
After this patch, using the next command:
./perf probe -f --no-demangle -x /tmp/a.out _Z3fooi
probe will be successfully added.
Signed-off-by: Azat Khuzhin <a3at.mail@gmail.com>
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/1382947464-31266-1-git-send-email-a3at.mail@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When introducing the PERF_RECORD_MMAP2 in:
5c5e854bc7 perf tools: Add attr->mmap2 support
A check for the number of entries parsed by sscanf was introduced that
assumed all of the 8 fields needed to be correctly parsed so that
particular /proc/pid/maps line would be considered synthesizable.
That broke anon records synthesizing, as it doesn't have the 'execname'
field.
Fix it by keeping the sscanf return check, changing it to not require
that the 'execname' variable be parsed, so that the preexisting logic
can kick in and set it to '//anon'.
This should get things like JIT profiling working again.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Cc: Bill Gray <bgray@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Joe Mario <jmario@redhat.com>
Cc: Richard Fowles <rfowles@redhat.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/n/tip-bo4akalno7579shpz29u867j@git.kernel.org
[ commit log message is mine, dzickus reported the problem with a patch ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Set feature-libunwind-debug-frame. We don't want it in
CORE_FEATURE_TESTS because it's not the generic case, but we
need to set it in the !feature-libunwind case.
Also, because x86 distributions typically don't have
dwarf_find_debug_frame() unwinding method:
test-libunwind-debug-frame.c:(.text+0x31): undefined reference to `_Ux86_64_dwarf_find_debug_frame'
Restrict this new API to ARM for the time being.
With this patch test-all.c works again, so repeat perf builds
are fast again:
comet:~/tip> perf stat --null --repeat 5 make -C tools/perf/
[...]
0,452899660 seconds time elapsed ( +- 0,11% )
While with before it was:
comet:~/tip> perf stat --null --repeat 5 make -C tools/perf/
[...]
1,674001829 seconds time elapsed ( +- 0,16% )
[ Includes fix to config/feature-checks/Makefile from Will Deacon. ]
Tested-by: Will Deacon <will.deacon@arm.com>
Tested-by: Jean Pihet <jean.pihet@linaro.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/n/tip-scsoctqzmou3rpkixCHezy9e@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
'feature_timerfd' is checked all the time and calculated explicitly,
in a serial fashion. Add it to CORE_FEATURE_TESTS which causes it to
be built in parallel, using the newfangled parallel build autodetection
code.
This shaves 137 msecs off the perf build time on my system, which
speeds up the common case cached build by 43%:
Before:
comet:~/tip> perf stat --null --repeat 5 make -C tools/perf/
[...]
0,453771441 seconds time elapsed ( +- 0,09% )
After:
comet:~/tip> perf stat --null --repeat 5 make -C tools/perf/
[...]
0,316290185 seconds time elapsed ( +- 0,24% )
Cc: David Ahern <dsahern@gmail.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/n/tip-bb92CmexihopoSyqnkqepvsy@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull ARM updates from Russell King:
"Included in this series are:
1. BE8 (modern big endian) changes for ARM from Ben Dooks
2. big.Little support from Nicolas Pitre and Dave Martin
3. support for LPAE systems with all system memory above 4GB
4. Perf updates from Will Deacon
5. Additional prefetching and other performance improvements from Will.
6. Neon-optimised AES implementation fro Ard.
7. A number of smaller fixes scattered around the place.
There is a rather horrid merge conflict in tools/perf - I was never
notified of the conflict because it originally occurred between Will's
tree and other stuff. Consequently I have a resolution which Will
forwarded me, which I'll forward on immediately after sending this
mail.
The other notable thing is I'm expecting some build breakage in the
crypto stuff on ARM only with Ard's AES patches. These were merged
into a stable git branch which others had already pulled, so there's
little I can do about this. The problem is caused because these
patches have a dependency on some code in the crypto git tree - I
tried requesting a branch I can pull to resolve these, and all I got
each time from the crypto people was "we'll revert our patches then"
which would only make things worse since I still don't have the
dependent patches. I've no idea what's going on there or how to
resolve that, and since I can't split these patches from the rest of
this pull request, I'm rather stuck with pushing this as-is or
reverting Ard's patches.
Since it should "come out in the wash" I've left them in - the only
build problems they seem to cause at the moment are with randconfigs,
and since it's a new feature anyway. However, if by -rc1 the
dependencies aren't in, I think it'd be best to revert Ard's patches"
I resolved the perf conflict roughly as per the patch sent by Russell,
but there may be some differences. Any errors are likely mine. Let's
see how the crypto issues work out..
* 'for-linus' of git://git.linaro.org/people/rmk/linux-arm: (110 commits)
ARM: 7868/1: arm/arm64: remove atomic_clear_mask() in "include/asm/atomic.h"
ARM: 7867/1: include: asm: use 'int' instead of 'unsigned long' for 'oldval' in atomic_cmpxchg().
ARM: 7866/1: include: asm: use 'long long' instead of 'u64' within atomic.h
ARM: 7871/1: amba: Extend number of IRQS
ARM: 7887/1: Don't smp_cross_call() on UP devices in arch_irq_work_raise()
ARM: 7872/1: Support arch_irq_work_raise() via self IPIs
ARM: 7880/1: Clear the IT state independent of the Thumb-2 mode
ARM: 7878/1: nommu: Implement dummy early_paging_init()
ARM: 7876/1: clear Thumb-2 IT state on exception handling
ARM: 7874/2: bL_switcher: Remove cpu_hotplug_driver_{lock,unlock}()
ARM: footbridge: fix build warnings for netwinder
ARM: 7873/1: vfp: clear vfp_current_hw_state for dying cpu
ARM: fix misplaced arch_virt_to_idmap()
ARM: 7848/1: mcpm: Implement cpu_kill() to synchronise on powerdown
ARM: 7847/1: mcpm: Factor out logical-to-physical CPU translation
ARM: 7869/1: remove unused XSCALE_PMU Kconfig param
ARM: 7864/1: Handle 64-bit memory in case of 32-bit phys_addr_t
ARM: 7863/1: Let arm_add_memory() always use 64-bit arguments
ARM: 7862/1: pcpu: replace __get_cpu_var_uses
ARM: 7861/1: cacheflush: consolidate single-CPU ARMv7 cache disabling code
...
Currently perf requires the -m / --mmap_pages option to be a power of 2.
To be more user friendly perf should automatically round this up to the
next power of 2.
Currently:
$ perf record -m 3 -a -- sleep 1
--mmap_pages/-m value must be a power of two.sleep: Terminated
With patch:
$ perf record -m 3 -a -- sleep 1
rounding mmap pages size to 16384 (4 pages)
...
v2: Add bytes units to rounding message per Ingo's request. Other
suggestions (e.g., prefixing INFO) should be addressed by wrapping
pr_info to catch all instances.
Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1384267617-3446-3-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We were using it at 10 kHz, which doesn't work in machines where somehow
the max freq was auto reduced by the kernel:
[root@ssdandy ~]# perf test 19
19: Test software clock events have valid period values : FAILED!
[root@ssdandy ~]# perf test -v 19
19: Test software clock events have valid period values :
--- start ---
Couldn't open evlist: Invalid argument
---- end ----
Test software clock events have valid period values: FAILED!
[root@ssdandy ~]#
[root@ssdandy ~]# cat /proc/sys/kernel/perf_event_max_sample_rate
7000
Reducing it to 500 Hz should be good enough for this test and also
shouldn't affect what it is testing.
But warn the user if it fails, informing the knob and the freq tried.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-548rhj1uo6xbwnxa95kw3hqe@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
We were not checking if we successfully opened the counters, i.e. if
sys_perf_event_open worked, when it doesn't in this test, we were
continuing anyway and then segfaulting when trying to access the file
descriptor array, that at that point had been freed in perf_evlist__open
error path:
[root@ssdandy ~]# perf test -v 19
19: Test software clock events have valid period values :
--- start ---
Segmentation fault (core dumped)
[root@ssdandy ~]#
Do the check and bail out instead.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6qy8ljkn0e9hm7bh7keo5z68@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
If given sort keys are all elided there'll be no output except for the
overhead column - actually the TUI shows a noisy output. In this case
it'd be better to show up the sort keys rather than elide.
Before:
$ perf report -s comm -c perf
(...)
# Overhead
# ........
#
100.00%
After:
$ perf report -s comm -c perf
(...)
# Overhead Command
# ........ .......
#
100.00% perf
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1383900822-14609-1-git-send-email-namhyung@kernel.org
[ Us curly braces around multi-line statements, as requested by Ingo Molnar ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Most uses of the evsel constructor are followed by a call to
perf_evlist__add with an idex of evlist->nr_entries, so make rename
the current constructor to perf_evsel__new_idx and remove the need
for passing the constructor for the common case.
We still need the new_idx variant because the way groups are handled,
with evsel->nr_members holding the number of entries in an evlist,
partitioning the evlist into sublists inside a single linked list.
This asks for a clarifying refactoring, but for now simplify the non
parser cases, so that tool writers don't have to bother with evsel idx
setting.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-zy9tskx6jqm2rmw7468zze2a@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Each call to tui_progress__update() would forcibly refresh the entire
screen. This is somewhat inefficient and causes noticable flickering
during the startup of perf-report, especially on large/slow terminals.
It looks like the force-refresh in tui_progress__update() serves no
purpose other than to clear the screen so that the progress bar of a
previous operation does not subsume that of a subsequent operation. But
we can do just that in a much more efficient manner by clearing only the
region that a previous progress bar may have occupied before repainting
the new progress bar. Then the force-refresh could be removed with no
change in visuals.
This patch disables the slow force-refresh in tui_progress__update() and
instead calls SLsmg_fill_region() on the entire area that the progress
bar may occupy before repainting it. This change makes the startup of
perf-report much faster and appear much "smoother".
It turns out that this was a big bottleneck in the startup speed of
perf-report -- with this patch, perf-report starts up ~2x faster (1.1s
vs 0.55s) on my machines. (These numbers were measured by running "time
perf report" on an 8MB perf.data and pressing 'q' immediately.)
Signed-off-by: Patrick Palka <patrick@parcs.ath.cx>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
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/1382747149-9716-1-git-send-email-patrick@parcs.ath.cx
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Instead do the lookups just when creating the tracepoints, initially for
the most common, raw_syscalls:sys_{enter,exit}.
It works by having evsel->priv have a per tracepoint structure with
entries for the fields, for direct access, with the offset and a
function to get the value from the sample, doing the swap if needed.
Using a simple workload that does M millions write syscalls, we go from:
# perf stat -i -e cycles /tmp/oldperf trace ./sc_hello 100 > /dev/null
Performance counter stats for '/tmp/oldperf trace ./sc_hello 100':
8,366,771,459 cycles
2.668025928 seconds time elapsed
# perf stat -i -e cycles perf trace ./sc_hello 100 > /dev/null
Performance counter stats for 'perf trace ./sc_hello 100':
8,345,187,650 cycles
2.631748425 seconds time elapsed
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-eyfhvoo510a5i10b27dnvm88@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When building perf out of tree:
$ make perf-tar-src-pkg
$ tar -xf perf-<ver>.tar -C /tmp
$ cd /tmp/perf<ver>
$ make -C tools/perf
you get this warning message:
make[1]: *** No rule to make target `kernelversion'. Stop.
Fix it by saving the perf version in the tar file and using that for the
out of tree builds.
v2: removed short form request and fixed up version string from usual output.
Signed-off-by: David Ahern <dsahern@gmail.com>
Suggested-by: Ingo Molnar <mingo@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1383753335-25782-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Adding the check for maximum allowed frequency rate defined in following
file:
/proc/sys/kernel/perf_event_max_sample_rate
When we cross the maximum value we fail and display detailed error
message with advise.
$ perf record -F 3000 ls
Maximum frequency rate (2000) reached.
Please use -F freq option with lower value or consider
tweaking /proc/sys/kernel/perf_event_max_sample_rate.
In case user does not specify the frequency and the default value cross
the maximum, we display warning and set the frequency value to the
current maximum.
$ perf record ls
Lowering default frequency rate to 2000.
Please consider tweaking /proc/sys/kernel/perf_event_max_sample_rate.
Same messages are used for 'perf top'.
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Adrian Hunter <adrian.hunter@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>
Link: http://lkml.kernel.org/r/1383660887-1734-4-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
