diff options
| author | evan@chromium.org <evan@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2011-05-18 20:37:47 +0000 |
|---|---|---|
| committer | evan@chromium.org <evan@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2011-05-18 20:37:47 +0000 |
| commit | 50b6c01fb0ed1e95f6e0922d6c44ea01f482ae17 (patch) | |
| tree | 2fb5686a3b54ce86f9a671b605904afe45faa1ea /base/sys_info_linux.cc | |
| parent | c78f18881208f776cd4f065356924e9036b57b07 (diff) | |
| download | chromium_src-50b6c01fb0ed1e95f6e0922d6c44ea01f482ae17.zip chromium_src-50b6c01fb0ed1e95f6e0922d6c44ea01f482ae17.tar.gz chromium_src-50b6c01fb0ed1e95f6e0922d6c44ea01f482ae17.tar.bz2 | |
memory: don't reserve trace events unless we're using tracing
Even when TraceEvent is disabled, we still poke at it unconditionally
via the tracing macros. The ctor would reserve 1024 TraceEvents,
allocating ~100kb, and TraceEvents have a ctor that touch the
newly-reserved memory, which means it's not just virtual address
space, we actually were wasting 100kb per process.
Fix it by only reserving space for the TraceEvents once tracing
is enabled. (It would be nice if we could avoid paying for any
of this stuff in the common case of not using tracing, but this
fix is simple.)
Review URL: http://codereview.chromium.org/7037034
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@85810 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'base/sys_info_linux.cc')
0 files changed, 0 insertions, 0 deletions