summaryrefslogtreecommitdiffstats
path: root/chrome/app/chrome_dll_main.cc
diff options
context:
space:
mode:
authoragl@chromium.org <agl@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>2009-08-24 22:10:43 +0000
committeragl@chromium.org <agl@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>2009-08-24 22:10:43 +0000
commit5101c0073bf723b6e6bc81273264eaf7ac124937 (patch)
treea8ee1593c3f419acab1e04507db25cbffc29a1f3 /chrome/app/chrome_dll_main.cc
parentd23a94d2a6c2717afeeb8b3fe1769f2c66a94797 (diff)
downloadchromium_src-5101c0073bf723b6e6bc81273264eaf7ac124937.zip
chromium_src-5101c0073bf723b6e6bc81273264eaf7ac124937.tar.gz
chromium_src-5101c0073bf723b6e6bc81273264eaf7ac124937.tar.bz2
Linux: exit render processes with the correct code.
When using a zygote, we don't set the |rv| variable in ChromeMain. Thus, renderers exit with a code of 255. There appears to be a harmless race where, when closing a render process, sometimes we'll see the close event from the IPC channel before we believe that the process has exited. In that case, the 255 exit code makes us believe that the renderer crashed. This is probably the cause of the huge renderer crash numbers in UMA. BUG=20149 http://codereview.chromium.org/173302 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@24157 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'chrome/app/chrome_dll_main.cc')
-rw-r--r--chrome/app/chrome_dll_main.cc4
1 files changed, 3 insertions, 1 deletions
diff --git a/chrome/app/chrome_dll_main.cc b/chrome/app/chrome_dll_main.cc
index 8bacd59..792d0ad 100644
--- a/chrome/app/chrome_dll_main.cc
+++ b/chrome/app/chrome_dll_main.cc
@@ -545,7 +545,9 @@ int ChromeMain(int argc, const char** argv) {
*CommandLine::ForCurrentProcess();
MainFunctionParams main_params(parsed_command_line, sandbox_wrapper,
&autorelease_pool);
- RendererMain(main_params);
+ rv = RendererMain(main_params);
+ } else {
+ rv = 0;
}
#else
NOTIMPLEMENTED();