Make SandboxIPCProcess a thread.

Make SandboxIPCProcess a thread.

This is a reland of https://chromiumcodereview.appspot.com/255693002/

Race conditions in Skia have been fixed.

BUG=322185
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=272493

[Jorge Lucangeli Obes, 2014-05-22 20:29:46]

Hi guys,

This is a reland of the original SandboxIPCProcess to thread patch. The race
conditions in Skia have been fixed.

This patch applies on top of https://chromiumcodereview.appspot.com/291173009/

Antoine for content/ OWNERS,
Lei for chrome/browser/memory_details.cc OWNERS.

[Lei Zhang, 2014-05-22 20:33:17]

chrome/ is exactly the same as last time, so lgtm

[Jorge Lucangeli Obes, 2014-05-22 20:35:24]

On 2014/05/22 20:33:17, Lei Zhang wrote:
> chrome/ is exactly the same as last time, so lgtm

And I promise I will run the TSan bots this time =D

[mdempsky, 2014-05-22 20:41:57]

https://chromiumcodereview.appspot.com/286903021/diff/1/content/browser/rende...
File content/browser/renderer_host/sandbox_ipc_linux.cc (right):

https://chromiumcodereview.appspot.com/286903021/diff/1/content/browser/rende...
content/browser/renderer_host/sandbox_ipc_linux.cc:614:
SandboxIPCHandler::~SandboxIPCHandler() {
I think we're leaking lifeline_fd_ and browser_socket_ here.  It probably
doesn't matter much since the browser is going to be exiting soon anyway, but I
think it's worth fixing.

Even better would be to use base::ScopedFD's to store the FDs in
SandboxIPCHandler and RenderSandboxHostLinux, but I won't begrudge you if you
want to just add close() calls here and switch to ScopedFD in a separate CL.

[piman, 2014-05-22 20:56:51]

LGTM % mdempsky's comment

[Jorge Lucangeli Obes, 2014-05-22 22:09:56]

Thanks for the comments! PTAL.

https://chromiumcodereview.appspot.com/286903021/diff/1/content/browser/rende...
File content/browser/renderer_host/sandbox_ipc_linux.cc (right):

https://chromiumcodereview.appspot.com/286903021/diff/1/content/browser/rende...
content/browser/renderer_host/sandbox_ipc_linux.cc:614:
SandboxIPCHandler::~SandboxIPCHandler() {
On 2014/05/22 20:41:57, mdempsky wrote:
> I think we're leaking lifeline_fd_ and browser_socket_ here.  It probably
> doesn't matter much since the browser is going to be exiting soon anyway, but
I
> think it's worth fixing.
> 
> Even better would be to use base::ScopedFD's to store the FDs in
> SandboxIPCHandler and RenderSandboxHostLinux, but I won't begrudge you if you
> want to just add close() calls here and switch to ScopedFD in a separate CL.

Added the close() calls, will follow-up with ScopedFDs.

[mdempsky, 2014-05-22 22:29:47]

lgtm

[Jorge Lucangeli Obes, 2014-05-23 04:13:35]

The CQ bit was checked by jorgelo@chromium.org

[commit-bot: I haz the power, 2014-05-23 04:15:10]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/jorgelo@chromium.org/286903021/20001

[commit-bot: I haz the power, 2014-05-23 07:49:10]

FYI, CQ is re-trying this CL (attempt #1).
Please consider checking whether the failures are real,
and report flakes to chrome-troopers@google.com.
The failing builders are:
  mac_chromium_rel on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/mac_chromium_rel/buil...)

[commit-bot: I haz the power, 2014-05-23 11:31:25]

Change committed as 272493