Commit Graph

25 Commits

Author SHA1 Message Date
mikhail.naganov@gmail.com
5ebc1a68ef Fix tests after r3811.
TBR=sgjesse@chromium.org

Review URL: http://codereview.chromium.org/574027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3813 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-02-08 08:55:01 +00:00
mikhail.naganov@gmail.com
f332ab9615 Profiler tick processor: exploit d8's readline to avoid reading
entire file prior to parsing it. This allows processing of huge
log files (over 200 MB.)

Review URL: http://codereview.chromium.org/574015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3811 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-02-08 08:15:18 +00:00
mikhail.naganov@gmail.com
e51a05f338 Fix issue 597: builtins and stubs are missing in profiler log when using snapshots.
After this fix, profiles of non-snapshotted VMs are now equivalent to
profiles of snapshotted VMs (having that --log-snapshot-positions is used,
and mksnapshot's log is given to the tick processor script.)

BUG=597

Review URL: http://codereview.chromium.org/574005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3802 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-02-04 21:34:03 +00:00
mikhail.naganov@gmail.com
999e3fca90 Fix issue 553: function frame is skipped in profile when compare stub is called.
The problem appeared due to a fact that stubs doesn't create a stack
frame, reusing the stack frame of the caller function. When building
stack traces, the current function is retrieved from PC, and its
callees are retrieved by traversing the stack backwards. Thus, for
stubs, the stub itself was discovered via PC, and then stub's caller's
caller was retrieved from stack.

To fix this problem, a pointer to JSFunction object is now captured
from the topmost stack frame, and is saved into stack trace log
record. Then a simple heuristics is applied whether a referred
function should be added to decoded stack, or not, to avoid reporting
the same function twice (from PC and from the pointer.)

BUG=553
TEST=added to mjsunit/tools/tickprocessor

Review URL: http://codereview.chromium.org/546089

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3673 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-01-21 16:42:41 +00:00
mikhail.naganov@gmail.com
37d3972429 Fix issue 571: display descriptive names for code objects from snapshot.
As this is only needed for internal profiling (not for DevTools),
the following approach had been chosen:

 - during snapshot creation, positions of serialized objects inside
   a snapshot are logged;

 - then during V8 initialization, positions of deserealized objects
   are logged;

 - those positions are used for retrieving code objects names from
   snapshot creation log, which needs to be supplied to tick processor
   script.

Positions logging is controlled with the new flag: --log_snapshot_positions.
This flag is turned off by default, and this adds no startup penalty.

To plug this fix to Golem, the following actions are needed:

 - logs created using 'mksnapshot' need to be stored along with VM images;

 - tick processor script needs to be run with '--snapshot-log=...' cmdline
   argument.

BUG=571

Review URL: http://codereview.chromium.org/551062

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-01-18 16:04:25 +00:00
mikhail.naganov@gmail.com
efb8cee324 Add initial semi-working producers profile.
Turned on with '--log-producers' flag, also needs '--noinline-new' (this is temporarily), '--log-code', '--log-gc'. Not all allocations are traced (I'm investigating.)

Stacks are stored using weak handles. Thus, when an object is collected, its allocation stack is deleted.

Review URL: http://codereview.chromium.org/267077

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3069 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-10-15 07:50:23 +00:00
mikhail.naganov@gmail.com
c8d358aa91 Fix issue 434: make OS::LogSharedLibraryAddresses work on x64 Mac.
Also, enable tick processor to process 64-bit profiler logs on Mac.

BUG=http://code.google.com/p/v8/issues/detail?id=434

Review URL: http://codereview.chromium.org/180063

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2810 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-09-02 08:18:27 +00:00
mikhail.naganov@gmail.com
d87f83c5f8 Fix issue 427: JS tick processor now works out-of-the-box for Chromium on Windows.
MSVS names '.map' file using only module's name, so both 'a.exe' and 'a.dll' will have 'a.map' file. To distinguish an originating module, we're now checking for image base which is always 00400000 for .exe files, and not 00400000 for .dlls.

Verified that windows-tick-processor can now process logs from Chromium using .map file generated for 'chrome.dll', an that it still works for V8's 'shell.exe'.

Review URL: http://codereview.chromium.org/172044

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2699 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-08-17 13:45:03 +00:00
mikhail.naganov@gmail.com
ee340a52ff X64: Fixes to enable C/C++ functions processing in profiler.
- rewrote Linux version of LogSharedLibraryAddresses to work correctly with 64-bit libs;
 - fixed address length restriction in JS tickprofiler script.

Review URL: http://codereview.chromium.org/160273


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2563 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-07-28 15:37:05 +00:00
mikhail.naganov@gmail.com
e0b829b1a5 TickProcessor: more accurate mapping of statically compiled code on Linux.
'nm' is now called with an option to report function code sizes. Static code entries are restricted to the sizes reported, and the remaining unnamed code is attributed to a library as a whole. This makes reports more accurate, as some functions are tiny, but has chunks of unnamed code behind them.

This change doesn't affect reporting on Windows, as in .map files function code sizes aren't specified.

Review URL: http://codereview.chromium.org/149513


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2455 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-07-14 09:47:44 +00:00
mikhail.naganov@gmail.com
73710e057a Implement shared libraries logging on Mac OS X, added required support in Tick Processor.
Review URL: http://codereview.chromium.org/155437

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2452 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-07-14 05:01:06 +00:00
mikhail.naganov@gmail.com
32b0e47fa5 Add automatic tests for Tick Processor, take two.
Now tests can be run from any directory. Location of test data is now determined using test file location provided by 'testcfg.py' script.

Tested under Linux, Mac, and Windows.

Review URL: http://codereview.chromium.org/155161


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2407 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-07-09 06:39:38 +00:00
kasperl@chromium.org
0684291c89 Revert r2372 to get the tree green again.
TBR=mikhail.naganov@gmail.com
Review URL: http://codereview.chromium.org/155137

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2377 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-07-07 13:06:40 +00:00
mikhail.naganov@gmail.com
0f7b263bdd Add automatic tests for Tick Processor.
Added tests for cmdline args parsing, symbols processing, and the whole process.

Tick Processor code was refactored to make it testable.

Review URL: http://codereview.chromium.org/149195


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2373 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-07-07 12:11:12 +00:00
mikhail.naganov@gmail.com
1e18e55c15 Two requested changes to tick processor.
1. If D8_PATH isn't specified, first try to locate 'd8' shell in path,
   and if not found, fallback to the one in tools_path/..
2. Add '--nm=<nm_exec>' parameter to specify 'nm' executable to use.

Review URL: http://codereview.chromium.org/132021


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2225 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-20 12:54:02 +00:00
mikhail.naganov@gmail.com
6b03961d4c Fix 'nm' results parsing in tickprocessor.
Without an explicit check if a function belongs to shared library address space, "finishing" a library symbols processing with 'addPrevEntry(libEnd);' can cause emission of code entries which cover almost the entire address space, shadowing other code.

Review URL: http://codereview.chromium.org/131033


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2221 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-19 07:08:31 +00:00
mikhail.naganov@gmail.com
a1a962f65e Implemented processing of compressed log files.
- fixed address delta calculation;
- code creations are also compressed to be in sync with other events;
- factored out a base class from TickProcessor to reuse code in DevTools profiler.

Review URL: http://codereview.chromium.org/125256


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2216 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-18 07:59:13 +00:00
mikhail.naganov@gmail.com
4ca9836bc7 Include weak symbols from 'nm' report. This fixes "C++" part of profiling results.
Review URL: http://codereview.chromium.org/125243


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2199 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-17 08:27:19 +00:00
mikhail.naganov@gmail.com
2cc6fef9d5 Fix the problem with JS entries in tickprocessor.
The main problem was due to the following: after Erik had fixed the logger to report library addresses, tickprocessor started to add to the code map entries that covered almost entire memory. This happened because tickprocessor contains a heuristic to bias addresses of functions from dynamic libraries:

    if (funcInfo.start < libStart && funcInfo.start < libEnd - libStart) {
      funcInfo.start += libStart;
    }

And, as tickprocessor tried to process all symbols from the library, including data entries, which can be outside reported library addresses range, the second condition failed, and funcInfo.start remained unbiased.

Review URL: http://codereview.chromium.org/125192


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2194 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-16 18:57:26 +00:00
erik.corry@gmail.com
92bbf75971 Don't panic if tickprocessor can't find a shared library.
Don't swallow exceptions so we can't see where they are really
thrown.
Review URL: http://codereview.chromium.org/126200

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2192 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-16 13:44:48 +00:00
mikhail.naganov@gmail.com
d7cccf6b8d Add log compression ability.
This is a trivial per-row compression:
- short aliases are introduced for events and code creation tags;
- in tick events, offsets are used instead of absolute addresses;
- removed 'code-allocation' event, as it seems not used.

The first two options are depend on the new flag: 'compress-log', which is off by default.

On benchmarks run w/o snapshot, this gives 45% log size reduction.

Review URL: http://codereview.chromium.org/119304


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2122 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-08 13:39:48 +00:00
mikhail.naganov@gmail.com
703fcecaaf Make tickprocessor's ProfileView extensible and move out DevTools-only stuff.
Review URL: http://codereview.chromium.org/119076


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2092 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-06-03 11:21:23 +00:00
mikhail.naganov@gmail.com
7d260e5f8c Don't keep data about JS code that is never executed.
This reduces memory usage of tickprocessor. Thanks to William Hesse for pointing out this issue.

Also speed up static symbols loading.

Review URL: http://codereview.chromium.org/113101

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1902 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-05-08 11:27:02 +00:00
mikhail.naganov@gmail.com
178a656429 Enhancing profiling data processing code with functionality needed for the Dev Tools Profiler.
Details:

 - added properties / functions in view objects needed for WebKit's ProfileView;

 - added ability to count profiles for specific functions.

The tickprocessor functionality does not affected.

Review URL: http://codereview.chromium.org/99181

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1823 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-04-30 08:10:27 +00:00
mikhail.naganov@gmail.com
aa2c33126c TickProcessor script reimplemented in JavaScript.
This is an effort to reuse profiler data processing code both in
TickProcessor and Dev Tools Profiler. The old Python implementation
will be removed.

The new TickProcessor works almost identical to the previous one.
However, it has some differences:

1. Not very useful "Call profile" section is replaced with a new
   WebKit-like "Bottom up (heavy) profile" which shows the most
   expensive functions together with their callers. I used it
   personally in order to find and remove bottlenecks in the
   tickprocessor script itself, and found it quite helpful.

2. Code entries with duplicate names (they occur for RegExes, stubs
   and sometimes for anonymous Function objects) are now distinguished
   by adding an occurence number inside curly brackets.

3. (Address -> code entry) mapping is more precise in boundary cases.

4. Windows version no more requires specifying .map file location.

5. Works faster.

Review URL: http://codereview.chromium.org/99054

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1802 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2009-04-27 13:50:42 +00:00