GPU: Add trace for real SwapBuffers.

GPU: Add trace for real SwapBuffers.

There is too many abstractions in our surface classes to
determine 'real' SwapBuffers calls in traces. This adds
traces in all the relevant places where we do 'real' swaps.

BUG=334063
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=245107

[epenner, 2014-01-14 04:33:00]

Ptal. I think I got all of them. Let me know if there is any others you know of.

[rjkroege, 2014-01-14 15:24:06]

On 2014/01/14 04:33:00, epenner wrote:
> Ptal. I think I got all of them. Let me know if there is any others you know
of.

apologies for drive-by: is it worthwhile in the interests of completeness to
mark swaps of hardware buffers underlying software rendering? If you think so
but don't feel like doing it, feel free to assign the bug to me.

[Ken Russell (switch to Gerrit), 2014-01-14 20:01:16]

lgtm

https://codereview.chromium.org/137823002/diff/1/ui/gl/gl_surface_egl.cc
File ui/gl/gl_surface_egl.cc (right):

https://codereview.chromium.org/137823002/diff/1/ui/gl/gl_surface_egl.cc#newc...
ui/gl/gl_surface_egl.cc:398: TRACE_EVENT2("gpu", "SwapBuffers",
You might want to name these events more precisely, like "gpu",
"NativeViewGLSurfaceEGL::SwapBuffers". Otherwise it might be difficult to track
down who generates them in the future. I see that some trace events have a
src_file and src_line but don't know what causes that. If that's generated for
these then never mind.

[epenner, 2014-01-14 21:38:20]

I've also removed the "Onscreen" trace that was always confusing (since several
non-real swaps can still be considered "Onscreen").

https://codereview.chromium.org/137823002/diff/1/ui/gl/gl_surface_egl.cc
File ui/gl/gl_surface_egl.cc (right):

https://codereview.chromium.org/137823002/diff/1/ui/gl/gl_surface_egl.cc#newc...
ui/gl/gl_surface_egl.cc:398: TRACE_EVENT2("gpu", "SwapBuffers",
On 2014/01/14 20:01:16, Ken Russell wrote:
> You might want to name these events more precisely, like "gpu",
> "NativeViewGLSurfaceEGL::SwapBuffers". Otherwise it might be difficult to
track
> down who generates them in the future. I see that some trace events have a
> src_file and src_line but don't know what causes that. If that's generated for
> these then never mind.

Good point. Done.

[epennerAtGoogle, 2014-01-14 21:40:35]

Landing. rjkroege, it does sound good but I'm just not familiar enough with
software rendering to know where to put those traces. If it's simple you can
disable cq and just let me know where to put them. Otherwise it can be another
patch.

[no sievers, 2014-01-14 22:22:18]

lgtm. Not sure it needs the 'Real' prefixes but if it makes telemetry/parsing
easier, I don't care :)

https://codereview.chromium.org/137823002/diff/120001/ui/gl/gl_surface_egl.cc
File ui/gl/gl_surface_egl.cc (right):

https://codereview.chromium.org/137823002/diff/120001/ui/gl/gl_surface_egl.cc...
ui/gl/gl_surface_egl.cc:15: #include "base/debug/trace_event.h"
nit: order

[epennerAtGoogle, 2014-01-14 22:40:53]

Yeah I snuck in the real to make the parsing easier since we plan to parse it.
That's not really a good solution but we'll need to discuss naming constants
somewhere if parsing becomes more prevalent.

https://codereview.chromium.org/137823002/diff/120001/ui/gl/gl_surface_egl.cc
File ui/gl/gl_surface_egl.cc (right):

https://codereview.chromium.org/137823002/diff/120001/ui/gl/gl_surface_egl.cc...
ui/gl/gl_surface_egl.cc:15: #include "base/debug/trace_event.h"
On 2014/01/14 22:22:18, sievers wrote:
> nit: order

Oy. I figured presubmit was referring to the osmesa. Looks like I was also
wrong!

[commit-bot: I haz the power, 2014-01-14 23:45:30]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/epenner@chromium.org/137823002/120001

[commit-bot: I haz the power, 2014-01-14 23:45:37]

Failed to trigger a try job on win_rel
HTTP Error 400: Bad Request

[commit-bot: I haz the power, 2014-01-15 04:35:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/epenner@chromium.org/137823002/130007

[commit-bot: I haz the power, 2014-01-16 04:23:26]

Change committed as 245107