IPC: Restrict 'toplevel' to truly toplevel events.

IPC: Restrict 'toplevel' to truly toplevel events.

These traces shouldn't ever be top level:
- "SyncChannel::SendWithTimeout"
- "ChannelProxy::Context::OnDispatchMessage"

One is sending an IPC (contained within RunTask), and one is
recieving an IPC that was forwarded (also contained within
RunTask).

This brings back the 'ipc' label for all ipc related
events, and just adds 'toplevel' to the one IPC trace that
is actually toplevel, "ChannelReader::DispatchInputData".

BUG=None.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=267923

[epennerAtGoogle, 2014-04-14 20:23:55]

Ptal.

[jam, 2014-04-14 23:00:51]

rubberstamp lgtm

[epennerAtGoogle, 2014-04-15 02:20:24]

On 2014/04/14 23:00:51, jam wrote:
> rubberstamp lgtm

jam@, The IPC review is actually the canonical review in this case. Does the
description sound correct regarding "SyncChannel::SendWithTimeout" and
"ChannelProxy::Context::OnDispatchMessage"? It certainly seems to be the case
that those will always be called within another piece of code (eg RunTask), and
that's what I've seen in all my traces, but it would be good to confirm with IPC
reviewer.

[jam, 2014-04-15 19:16:51]

On 2014/04/15 02:20:24, epennerAtGoogle wrote:
> On 2014/04/14 23:00:51, jam wrote:
> > rubberstamp lgtm
> 
> jam@, The IPC review is actually the canonical review in this case. Does the
> description sound correct regarding "SyncChannel::SendWithTimeout" and
> "ChannelProxy::Context::OnDispatchMessage"? It certainly seems to be the case
> that those will always be called within another piece of code (eg RunTask),
and
> that's what I've seen in all my traces, but it would be good to confirm with
IPC
> reviewer.

I'm not really familiar with tracing, so I don't know how you classify these
calls. That's why I deferred to others.

[epennerAtGoogle, 2014-04-15 22:05:05]

On 2014/04/15 19:16:51, jam wrote:
> On 2014/04/15 02:20:24, epennerAtGoogle wrote:
> > On 2014/04/14 23:00:51, jam wrote:
> > > rubberstamp lgtm
> > 
> > jam@, The IPC review is actually the canonical review in this case. Does the
> > description sound correct regarding "SyncChannel::SendWithTimeout" and
> > "ChannelProxy::Context::OnDispatchMessage"? It certainly seems to be the
case
> > that those will always be called within another piece of code (eg RunTask),
> and
> > that's what I've seen in all my traces, but it would be good to confirm with
> IPC
> > reviewer.
> 
> I'm not really familiar with tracing, so I don't know how you classify these
> calls. That's why I deferred to others.

Okay no worries. In this case the traces depend a bit on the semantics of the
underlying code. However I'm quite confident of this being correct in this case
(only the ChannelReader::DispatchInputData can actually create work on a thread
outside of the message loop).

[epenner, 2014-04-29 01:35:28]

The CQ bit was checked by epenner@chromium.org

[commit-bot: I haz the power, 2014-04-29 01:36:29]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/epenner@chromium.org/237113006/1

[commit-bot: I haz the power, 2014-04-29 01:36:35]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-04-29 01:36:35]

Failed to apply patch for ipc/ipc_sync_channel.cc:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file ipc/ipc_sync_channel.cc
  Hunk #1 FAILED at 442.
  1 out of 1 hunk FAILED -- saving rejects to file ipc/ipc_sync_channel.cc.rej

Patch:       ipc/ipc_sync_channel.cc
Index: ipc/ipc_sync_channel.cc
diff --git a/ipc/ipc_sync_channel.cc b/ipc/ipc_sync_channel.cc
index
4b706e25363132e9c3dbce92ba976a874cb44d8a..588b66eff806de3583be6f1acdeaf7d4fe14ad26
100644
--- a/ipc/ipc_sync_channel.cc
+++ b/ipc/ipc_sync_channel.cc
@@ -442,10 +442,10 @@ bool SyncChannel::SendWithTimeout(Message* message, int
timeout_ms) {
   Logging* logger = Logging::GetInstance();
   std::string name;
   logger->GetMessageText(message->type(), &name, message, NULL);
-  TRACE_EVENT1("toplevel", "SyncChannel::SendWithTimeout",
+  TRACE_EVENT1("ipc", "SyncChannel::SendWithTimeout",
                "name", name);
 #else
-  TRACE_EVENT2("toplevel", "SyncChannel::SendWithTimeout",
+  TRACE_EVENT2("ipc", "SyncChannel::SendWithTimeout",
                "class", IPC_MESSAGE_ID_CLASS(message->type()),
                "line", IPC_MESSAGE_ID_LINE(message->type()));
 #endif

[epenner, 2014-05-02 20:00:45]

The CQ bit was checked by epenner@chromium.org

[commit-bot: I haz the power, 2014-05-02 20:01:25]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/epenner@chromium.org/237113006/20001

[commit-bot: I haz the power, 2014-05-02 21:29:25]

Change committed as 267923