From 229db93bc9232e2d5ea26f3b026c337e8ed00298 Mon Sep 17 00:00:00 2001 From: "ananta@chromium.org" Date: Tue, 15 Sep 2009 01:00:57 +0000 Subject: Disabling the AsyncIO.IPCChannelClose UMA macro due to a potential race condition causing it to be registered from multiple threads. Logged a bug http://code.google.com/p/chromium/issues/detail?id=21827 against this and assigned it to rvargas. Workaround for http://b/issue?id=2111432 for now. Bug=2111432 Review URL: http://codereview.chromium.org/204014 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@26189 0039d316-1c4b-4281-b951-d872f2087c98 --- ipc/ipc_channel_win.cc | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/ipc/ipc_channel_win.cc b/ipc/ipc_channel_win.cc index b4fc2d0..8aba74f 100644 --- a/ipc/ipc_channel_win.cc +++ b/ipc/ipc_channel_win.cc @@ -71,7 +71,13 @@ void Channel::ChannelImpl::Close() { } if (waited) { // We want to see if we block the message loop for too long. - UMA_HISTOGRAM_TIMES("AsyncIO.IPCChannelClose", base::Time::Now() - start); + // There is a potential race condition here as this function can run from + // multiple threads, which causes a DCHECK to fire at times from the + // StatisticsRecorder object indicating that the histogram is being + // registered multiple times. + // Commenting out this UMA call for now. + // Bug http://code.google.com/p/chromium/issues/detail?id=21827 + // UMA_HISTOGRAM_TIMES("AsyncIO.IPCChannelClose", base::Time::Now() - start); } while (!output_queue_.empty()) { -- cgit v1.1