Make the number of raster threads appear in chrome://gpu.

Make the number of raster threads appear in chrome://gpu.

This add methods to compositor_util.cc to compute the number of
raster threads to be used in the renderer process. Then this number
is passed to the renderer process explicitly instead of just forwarding
a command line flag blindly.

If the renderer will use more than one thread, chrome://gpu will report
that "Multiple Raster Threads" is enabled, otherwise it shows disabled.
If the --num-raster-threads command line argument is used to force more
than one thread, then it will show Force enabled.

There is no change in behaviour with this patch, it still uses one thread
unless forced otherwise on the command line.

BUG=237669
Committed: https://crrev.com/8a91b03e231894791a81d45cc6b050890e4f4e9d
Cr-Commit-Position: refs/heads/master@{#292726}

[danakj, 2014-08-29 18:46:43]

danakj@chromium.org changed reviewers:
+ piman@chromium.org

[danakj, 2014-08-29 19:47:21]

danakj@chromium.org changed reviewers:
+ jamesr@chromium.org, kbr@chromium.org

[danakj, 2014-08-29 19:47:22]

+kbr for content/browser stuff
+jamesr for content/renderer stuff

[jamesr, 2014-08-29 19:50:39]

lgtm

https://codereview.chromium.org/519923002/diff/1/content/renderer/render_thre...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/519923002/diff/1/content/renderer/render_thre...
content/renderer/render_thread_impl.cc:558: int num_raster_threads;
initialize this var, please. if the StringToInt call fails and we aren't in a
debug build this will call SetNumRasterThreads() with whatever's sitting on the
stack here which'll probably lead to a difficult to understand crash later

https://codereview.chromium.org/519923002/diff/1/content/renderer/render_thre...
content/renderer/render_thread_impl.cc:562: DCHECK(ok) << string_value;
if you give this var a better name the error message will be slightly easier to
parse when this trips. DCHECK(num_raster_threads_parsed) or something?

[danakj, 2014-08-29 19:53:31]

https://codereview.chromium.org/519923002/diff/1/content/renderer/render_thre...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/519923002/diff/1/content/renderer/render_thre...
content/renderer/render_thread_impl.cc:558: int num_raster_threads;
On 2014/08/29 19:50:39, jamesr wrote:
> initialize this var, please. if the StringToInt call fails and we aren't in a
> debug build this will call SetNumRasterThreads() with whatever's sitting on
the
> stack here which'll probably lead to a difficult to understand crash later

Done.

https://codereview.chromium.org/519923002/diff/1/content/renderer/render_thre...
content/renderer/render_thread_impl.cc:562: DCHECK(ok) << string_value;
On 2014/08/29 19:50:38, jamesr wrote:
> if you give this var a better name the error message will be slightly easier
to
> parse when this trips. DCHECK(num_raster_threads_parsed) or something?

Done.

[Ken Russell (switch to Gerrit), 2014-08-29 22:41:06]

LGTM if this has been tested. One question.

https://codereview.chromium.org/519923002/diff/20001/content/browser/renderer...
File content/browser/renderer_host/render_process_host_impl.cc (left):

https://codereview.chromium.org/519923002/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_process_host_impl.cc:1182:
switches::kNumRasterThreads,
Why is this being removed from this list? Isn't it needed here to get this
argument propagated down to the renderer?

[danakj, 2014-08-29 22:42:20]

https://codereview.chromium.org/519923002/diff/20001/content/browser/renderer...
File content/browser/renderer_host/render_process_host_impl.cc (left):

https://codereview.chromium.org/519923002/diff/20001/content/browser/renderer...
content/browser/renderer_host/render_process_host_impl.cc:1182:
switches::kNumRasterThreads,
On 2014/08/29 22:41:05, Ken Russell wrote:
> Why is this being removed from this list? Isn't it needed here to get this
> argument propagated down to the renderer?

because the other code adds it to the renderer command line explicitly, we don't
just forward it along blindly. this list is flags to just copy from the browser
command line.

[danakj, 2014-08-29 22:42:57]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-08-29 22:43:43]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/danakj@chromium.org/519923002/20001

[commit-bot: I haz the power, 2014-08-30 00:05:31]

Committed patchset #2 (id:20001) as fd8046161872adb6bcbabeb582fbafd989b1123d

[commit-bot: I haz the power, 2014-09-10 03:12:06]

Patchset 2 (id:??) landed as
https://crrev.com/8a91b03e231894791a81d45cc6b050890e4f4e9d
Cr-Commit-Position: refs/heads/master@{#292726}