Handle systemic child process launch failures/crashes more gracefully |
|||
Issue descriptionIf child processes cannot be started or are terminating very quickly after launch we should try and handle this more gracefully with one or more of the following: 1. Stopping or throttling the sending of crash reports if the crash is identical. 2. Stopping the (attempted) spawning of child processes. 3. A notification to the user on the sad tab page that something is wrong, rather than it just being black/grey and empty, or a normal sad tab. Link to a help center article or something. 4. Terminating Chrome entirely?
,
Sep 12 2016
This was raised in the context of EMET but the cause of the child process startup failure (or immediate crash) would not necessarily be EMET in the future, so any solution we come up with would have to be flexible enough to cope with another root cause e.g. a real bug in Chrome or another third party software interaction.
,
Sep 13 2016
,
Sep 22 2016
See http://crbug.com/648736 for Seb's investigation for how the fork loop manifests.
,
Aug 23
|
|||
►
Sign in to add a comment |
|||
Comment 1 by yyushkina@chromium.org
, Sep 12 2016