Bug 104971 – Use 'perf' to profile on linux instead of google-pprof by default

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Eric Seidel (no email)]

Use 'perf' to profile on linux instead of google-pprof by default

[Eric Seidel (no email)]

Created attachment 179364 [details]
Patch

[Dirk Pranke]

Comment on attachment 179364 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=179364&action=review

> Tools/Scripts/webkitpy/common/system/executive.py:313
> +        while seconds_left > 0:

this could be just 'while seconds_left > 0 and self.check_running_pid(pid):'

> Tools/Scripts/webkitpy/common/system/profiler.py:104
> +        self._target_pid = None

Given that you're passing the pids to the executive, this isn't a target_pid, this is either a host_pid or just a pid (I'd use just a pid)

[Eric Seidel (no email)]

Comment on attachment 179364 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=179364&action=review

>> Tools/Scripts/webkitpy/common/system/executive.py:313
>> +        while seconds_left > 0:
> 
> this could be just 'while seconds_left > 0 and self.check_running_pid(pid):'

Yes, sorry, this was more complicated originally.  Will fix.

>> Tools/Scripts/webkitpy/common/system/profiler.py:104
>> +        self._target_pid = None
> 
> Given that you're passing the pids to the executive, this isn't a target_pid, this is either a host_pid or just a pid (I'd use just a pid)

Oh, I was trying to differentiate it from the perf pid.  I'm open to other names.  I wasn't trying to make the host/target distinction, but I can see how that's confusing here.

Suggestions?

[Eric Seidel (no email)]

updated.  Works:

%run-perf-tests --chromium --profile                                                                   [/src/WebKit]
Running 113 tests
Running Animation/balls.html (1 of 113)
WARNING: Kernel address maps (/proc/{kallsyms,modules}) are restricted,
check /proc/sys/kernel/kptr_restrict.

Samples in kernel functions may not be resolved if a suitable vmlinux
file is not found in the buildid cache or in the vmlinux path.

Samples in kernel modules won't be resolved at all.

If some relocation was applied (e.g. kexec) symbols may be misresolved
even with a suitable vmlinux or kallsyms file.

Finished: 3.059183 s

[ perf record: Woken up 3 times to write data ]
[ perf record: Captured and wrote 0.661 MB /src/WebKit/Source/WebKit/chromium/webkit/Release/layout-test-results/test-41.data (~28895 samples) ]
perf report -g none -i /src/WebKit/Source/WebKit/chromium/webkit/Release/layout-test-results/test-41.data
     5.04%  DumpRenderTree  perf-6532.map            [.] 0x29bc5a44b1e7  
     2.19%  DumpRenderTree  DumpRenderTree           [.] v8::internal::FastDtoa(double, v8::internal::FastDtoaMode, int, v8::internal::Vector<char>, int*, int*)
     2.18%  DumpRenderTree  DumpRenderTree           [.] WebCore::Length::initFromLength(WebCore::Length const&)
     1.95%  DumpRenderTree  libfreetype.so.6.8.0     [.] 0x248db         
     1.64%  DumpRenderTree  DumpRenderTree           [.] tc_free
     1.64%  DumpRenderTree  DumpRenderTree           [.] WebCore::RenderStyle::diff(WebCore::RenderStyle const*, unsigned int&) const
     1.47%  DumpRenderTree  [kernel.kallsyms]        [k] 0xffffffff8103b51a
     1.46%  DumpRenderTree  DumpRenderTree           [.] tc_malloc
     1.28%  DumpRenderTree  DumpRenderTree           [.] v8::internal::JSObject::SetPropertyWithInterceptor(v8::internal::String*, v8::internal::Object*, PropertyAttributes, v8::internal::StrictModeFlag)
     1.26%  DumpRenderTree  DumpRenderTree           [.] WebCore::RenderStyle::create()

To view the full profile, run:
perf report -i /src/WebKit/Source/WebKit/chromium/webkit/Release/layout-test-results/test-41.data

[Eric Seidel (no email)]

Created attachment 179399 [details]
Patch

[WebKit Review Bot]

Comment on attachment 179399 [details]
Patch

Clearing flags on attachment: 179399

Committed r137706: <http://trac.webkit.org/changeset/137706>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Eric Seidel (no email)]

This caused a test-webkitpy failure, but I fixed in http://trac.webkit.org/changeset/137711

[Eric Seidel (no email)]

*** Bug 104609 has been marked as a duplicate of this bug. ***