Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(371)

Issue 22320019: cc: Start commit early and block on tree activation (Closed)

Created:
7 years, 4 months ago by brianderson
Modified:
7 years, 3 months ago
Reviewers:
CC:
chromium-reviews, cc-bugs_chromium.org, nduca, Sami, jdduke (slow), enne (OOO), danakj, boliu, joth, piman
Base URL:
http://git.chromium.org/chromium/src.git@usedl26
Visibility:
Public.

Description

cc: Start commit early and block on tree activation For discussion not review yet. This patch allows us to record the next commit while we are rasterizing the previous commit. BUG=none yet

Patch Set 1 #

Total comments: 2
Unified diffs Side-by-side diffs Delta from patch set Stats (+51 lines, -11 lines) Patch
M cc/scheduler/scheduler_state_machine.h View 2 chunks +3 lines, -1 line 0 comments Download
M cc/scheduler/scheduler_state_machine.cc View 9 chunks +24 lines, -9 lines 0 comments Download
M cc/trees/layer_tree_settings.cc View 1 chunk +1 line, -1 line 0 comments Download
M cc/trees/thread_proxy.h View 1 chunk +3 lines, -0 lines 0 comments Download
M cc/trees/thread_proxy.cc View 2 chunks +20 lines, -0 lines 2 comments Download

Messages

Total messages: 3 (0 generated)
brianderson
This sometimes starts commits way too early, blocking the main thread longer than necessary, but ...
7 years, 4 months ago (2013-08-10 02:19:15 UTC) #1
Sami
I like the idea for increased parallelism. I wonder if we should make it tweakable ...
7 years, 4 months ago (2013-08-13 15:56:02 UTC) #2
brianderson
7 years, 3 months ago (2013-09-07 01:18:54 UTC) #3
Replacing this patch with: https://codereview.chromium.org/23907006.

https://codereview.chromium.org/22320019/diff/1/cc/trees/thread_proxy.cc
File cc/trees/thread_proxy.cc (right):

https://codereview.chromium.org/22320019/diff/1/cc/trees/thread_proxy.cc#newc...
cc/trees/thread_proxy.cc:890: // We defer forced commits while there is a
pending tree here.
On 2013/08/13 15:56:02, Sami wrote:
> Could this be done in the state machine instead? Seems like it has all the
> necessary bits of information already.

Hmm, I'll give that a try. I did it this way because I knew for sure that we
haven't affected state on the impl-thread at this point yet. Doing it in the
scheduler would allow us to defer the commit even later into the process.

Powered by Google App Engine
This is Rietveld 408576698