Create crash dump if a bad IPC message is received.

Create crash dump if a bad IPC message is received.

BUG=473370
Committed: https://crrev.com/af32d0b1819312fb2b0c1276ebe9e008069cf82e
Cr-Commit-Position: refs/heads/master@{#324003}

[Will Harris, 2015-04-04 21:02:18]

wfh@chromium.org changed reviewers:
+ jam@chromium.org

[Will Harris, 2015-04-04 21:15:48]

PTAL.

RESULT_CODE_KILLED_BAD_MESSAGE is the fourth largest cause of renderer crashes
(behind hangs/crashes/debugbreak) but we don't get any crash reports for these.

Now we've branched I'd like to get some data on what these are.

[jam, 2015-04-06 16:00:05]

https://codereview.chromium.org/1058423002/diff/20001/content/browser/browser...
File content/browser/browser_child_process_host_impl.cc (right):

https://codereview.chromium.org/1058423002/diff/20001/content/browser/browser...
content/browser/browser_child_process_host_impl.cc:259: // If content is hosted
within a process that has set a dump without crashing
I don't understand what this comment means? i.e. this code always runs in the
browser process, so why does the comment make it seem like what will happen next
depends on the process?

[Will Harris, 2015-04-06 16:34:21]

On 2015/04/06 16:00:05, jam wrote:
>
https://codereview.chromium.org/1058423002/diff/20001/content/browser/browser...
> File content/browser/browser_child_process_host_impl.cc (right):
> 
>
https://codereview.chromium.org/1058423002/diff/20001/content/browser/browser...
> content/browser/browser_child_process_host_impl.cc:259: // If content is
hosted
> within a process that has set a dump without crashing
> I don't understand what this comment means? i.e. this code always runs in the
> browser process, so why does the comment make it seem like what will happen
next
> depends on the process?

DumpWithoutCrashing doesn't do anything unless the content embedder has
previously called SetDumpWithoutCrashingFunction

e.g. Chrome does this here:

https://code.google.com/p/chromium/codesearch#chromium/src/chrome/app/chrome_...

maybe I should use 'embedded' rather than 'hosted'.

[jam, 2015-04-06 21:26:06]

lgtm

https://codereview.chromium.org/1058423002/diff/20001/content/browser/browser...
File content/browser/browser_child_process_host_impl.cc (right):

https://codereview.chromium.org/1058423002/diff/20001/content/browser/browser...
content/browser/browser_child_process_host_impl.cc:259: // If content is hosted
within a process that has set a dump without crashing
On 2015/04/06 16:00:04, jam wrote:
> I don't understand what this comment means? i.e. this code always runs in the
> browser process, so why does the comment make it seem like what will happen
next
> depends on the process?

yeah, that would be clearer. or also take out the first sentence entirely since
it's redundant (we don't want to have to retype that each time code in content
or base calls DumpWithoutCrashing)

[Will Harris, 2015-04-07 01:08:50]

The CQ bit was checked by wfh@chromium.org

[Will Harris, 2015-04-07 01:08:51]

The patchset sent to the CQ was uploaded after l-g-t-m from jam@chromium.org
Link to the patchset: https://codereview.chromium.org/1058423002/#ps40001
(title: "change comment")

[commit-bot: I haz the power, 2015-04-07 01:08:56]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1058423002/40001

[commit-bot: I haz the power, 2015-04-07 02:21:48]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-04-07 02:23:29]

Patchset 3 (id:??) landed as
https://crrev.com/af32d0b1819312fb2b0c1276ebe9e008069cf82e
Cr-Commit-Position: refs/heads/master@{#324003}