1270147 - [e10s] remote nsIOService::SpeculativeConnect to the parent

Status: RESOLVED FIXED

(Core :: Networking: HTTP, defect, P3)

Description

[Julian Hector [:tedd] [:jhector]]

After a discussion with :mcmanus regarding Bug 1259508, it seems that this should be remoted anyways, regardless of the sandboxing issue. 

See Bug 1259508 Comment 16, for further information.

Comment 1

[Julian Hector [:tedd] [:jhector]]

:bagder, I was told that you are probably the right person to work on this. I don't know how much work is required and how busy you are.

This is currently one of the last bugs blocking us from enabling seccomp on nightly. If you are too busy to work on this, I would write a workaround patch for Bug 1259508 which can then be backed out again once this bug is resolved.

Comment 2

[Daniel Stenberg [:bagder]]

I'm moving this up to the top of my list. Just these national holidays (Sweden) in the way...

Comment 3

[Julian Hector [:tedd] [:jhector]]

Great thanks Daniel.

Comment 4

[Daniel Stenberg [:bagder]]

Attachment: 0001-Bug-1270147-remote-the-SpeculativeConnect-call-r-mcm.patch

Were you thinking about something as simple as this?

I'm not sure exactly how I should proceed to test this specifically. Any suggestions? (I'll send it off on a try-run of course.)

Comment 5

[Patrick McManus [:mcmanus]]

to test this I think its sufficient to put print/log in the new code as well as the parent execution of the remoted call.. and then try and trigger it based on the stack in https://bug1259508.bmoattachments.org/attachment.cgi?id=8735522

Comment 6

[Daniel Stenberg [:bagder]]

This is the try-run: https://treeherder.mozilla.org/#/jobs?repo=try&revision=8b1b092b5cf9

Comment 7

[Daniel Stenberg [:bagder]]

Wow, it wasn't that easy to actually trigger that code path but I've now tested it manually and the above try-run seems to agree. Good to merge methinks.

Comment 8

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/aa2785a532d7

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/aa2785a532d7