Factor out (and optimize) MessageInTransit.
This is mainly so that MessageInTransit has room for a header (for use in other
code), but while at it I optimized it a bit. The 10- and 100-byte
SystemPerftest.MessagePipe_WriteAndRead performance tests go from ~2.3 million
iterations/second to ~2.6; in perf, |free()| is no longer the top hotspot in
SystemPerftest.MessagePipe_WriteAndRead (all sizes) -- instead, |memcpy()| is,
as you'd hope.
R=darin
Review URL: https://codereview.chromium.org/25443011
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@226839 0039d316-1c4b-4281-b951-d872f2087c98
diff --git a/mojo/mojo.gyp b/mojo/mojo.gyp
index 567452f..d8f3121 100644
--- a/mojo/mojo.gyp
+++ b/mojo/mojo.gyp
@@ -78,6 +78,8 @@
'system/limits.h',
'system/memory.cc',
'system/memory.h',
+ 'system/message_in_transit.cc',
+ 'system/message_in_transit.h',
'system/message_pipe.cc',
'system/message_pipe.h',
'system/message_pipe_dispatcher.cc',
