1265036 - Change Buffer::try_realloc() crash to an OOM abort

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Andrew McCreight [:mccr8] (PTO until 12-2)]

I see 8 of these crashes in the content process in the last week. Bill, is this something that we should be concerned about? Maybe it is a dupe of one of the other regressions from your patch.

Comment 1

[Bill McCloskey [inactive unless it's an emergency] (:billm)]

It looks like an OOM. We're asking for more memory and not getting it. Maybe I should use NS_ABORT_OOM here?

Comment 2

[Andrew McCreight [:mccr8] (PTO until 12-2)]

I'd just switch from realloc to moz_xrealloc, which I think is infallible, so it will annotate everything correctly. That's what Kan-ru used in Pickle::Resize().

Comment 3

[Andrew McCreight [:mccr8] (PTO until 12-2)]

Oh, right, but you want to survive a failure sometimes. Yeah, NS_ABORT_OOM would be good.

Comment 4

[Andrew McCreight [:mccr8] (PTO until 12-2)]

Attachment: Use NS_ABORT_OOM() if try_realloc() fails.

Comment 5

[Andrew McCreight [:mccr8] (PTO until 12-2)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/1a309a682f19

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/1a309a682f19