Review URL: http://codereview.chromium.org/125245

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2201 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

certain size.
Review URL: http://codereview.chromium.org/126265

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2200 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/125243


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2199 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

out for now.
Review URL: http://codereview.chromium.org/125241

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2197 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

mode and on actual ARM hardware.

TBR=erik.corry@gmail.com
Review URL: http://codereview.chromium.org/125240

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2196 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126262

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2195 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

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

handling addresses.
Review URL: http://codereview.chromium.org/125187

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2193 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

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

- Added fast case to String.prototype.split (Mads's idea).
- Made minor other optimizations in String.prototype.split.

Review URL: http://codereview.chromium.org/126201

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2191 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

two symbols, they're only equal if the objects are identical.
Review URL: http://codereview.chromium.org/125184

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2188 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

(Android does this).  Fix logging for executable mappings that have no
file associated.  Be more consistent with use of uintptr_t.
Review URL: http://codereview.chromium.org/125183

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2187 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

To profile running the JavaScript file test.js using the V8 release mode shell (assuming it is build passing prof=oprofile to the SCons build). The following commands can be used:

$ tools/oprofile/start
$ tools/oprofile/run test.js
$ tools/oprofile/report | less
$ tools/oprofile/annotate | less
$ tools/oprofile/shutdown

Here is a summary of the commands.

For all the commands taking an executable the executable is expected to be a binary using V8. If no executable is specified the release mode V8 shell is assumed.

By default the --session-dir=/tmp/oprofv8 is passed to all oprofile commands. This walue can be changed by setting environment variable OPROFILE_SESSION_DIR.

When using the defaulf executable (V8 shell in release mode) it is assumed to be located in ../.. relative from the oprofile utility scripts. This default location can be overridden using the V8_SHELL_DIR environment variable.

start
-----
Start the oprofiling daemon.

run [executable] [parameters]
-----------------------------
Profile a V8 executable. Running this will reset oprofile samples, run the command and do an oprofile dump to flush samples and write ELF binaries for the generated code. The parameters are passed to the executable together with the --oprofile option.

report [executable] [parameters]
--------------------------------
Print the report for a profile run. The parameters are passed to opreport. E.g report --callgraph.

annotate [executable] [parameters]
----------------------------------
Print annotated assembly for a profile run. The parameters are passed to opannotate. E.g annotate -threshold 1.

reset
-----
Reset oprofile samples.

dump
----
Flush oprofile samples and write ELF binaries for the generated code.

shutdown
--------
Shutdown oprofile daemon.

Added a warning which is printed if option --oprofile is passed to a V8 which has not been compiled with oprofile support.
Review URL: http://codereview.chromium.org/125181

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2186 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Don't infer name for a function if a result of its call is assigned to a variable / property. E.g., in this case:

  a = function() { ... } ();

the function must remain anonymous because 'a' doesn't receive a function reference, but instead a result of its call.

BUG=http://code.google.com/p/v8/issues/detail?id=380
TEST=cctest/test-func-name-inference/Issue380

Review URL: http://codereview.chromium.org/126195


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2185 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2184 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126193

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2183 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126192

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2182 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/125177

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2180 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

when the left operand lets us shortcut the computation.
Review URL: http://codereview.chromium.org/125176

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2179 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126191

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2178 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/125131

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2177 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126188

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2176 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126189

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2175 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Only send the inscructions part of a code object to oprofile when reporting dynamically generated code. Before the code object header was also reported to oprofile as code which caused strange disassembly output when using opannotate.
Review URL: http://codereview.chromium.org/125126

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2174 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=kmillikin@chromium.org
Review URL: http://codereview.chromium.org/125129

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2173 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

It seems that when calling a method that has two overloaded versions like this:
  f(char* format, ...)
  f(char* format, va_list args)
with a second pointer argument: f("format", pointer), the second version is picked up.

I've found a description of a similar issue here: http://bugs.gentoo.org/63112

So, to resolve this ambiguity, I've named such LogMessageBuilder's Append functions differently.

Review URL: http://codereview.chromium.org/125125


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2172 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This fixes assertion failures in debug mode tests.
Review URL: http://codereview.chromium.org/126128

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2171 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

as deep on newer VMs in debug mode.
Review URL: http://codereview.chromium.org/126125

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2170 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Because of varying floating-point precision, the slow case is hard to
test with explicit values.  Instead, we check that sine and cosine do
not return the same value (the regression was that the slow case of
cosine accidentally did sine instead of cosine).

Review URL: http://codereview.chromium.org/126123

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2169 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

under Mac OS.
Review URL: http://codereview.chromium.org/125123

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2168 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/125122

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2167 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/125121

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2166 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126121

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2165 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126116

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2164 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

avoiding excessive ToNumber calls and by dealing with NaNs
in BIT_AND and SAR.
Review URL: http://codereview.chromium.org/125118

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2163 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Code addresses are now written as an offset from the previous address for ticks, code move and delete events. Employed backreference and RLE compression for code move and delete events. This gives additional 30% log size reduction for benchmarks run w/o snapshot.

Overall compression results (compared with the revision of V8 having no compression):
 - V8: 70% size reduction for benchmarks run w/o snapshot (for reference, gzip gives 87%)
 - Chromium: 65% size reduction for public html version of benchmarks (v4) (for reference, gzip gives 90%)

The one obvious opportunity for improving compression results in Chromium is to compress URLs of scripts.

Review URL: http://codereview.chromium.org/125114


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2162 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126115

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2161 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/126114

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2160 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Marsaglia's multiply-with-carry instead of mixing the
bits obtained from calling the system random() twice.

This seems to be a bit faster and gives a better 
distribution than the system random() in particular on 
Windows. 
Review URL: http://codereview.chromium.org/126113

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2159 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

expanded the macro assembler owns the buffer and needs to dispose it
in the destructor.
Review URL: http://codereview.chromium.org/126111

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2158 ce2b1a6d-e550-0410-aec6-3dcde31c8c00