Closed Bug 1497270 Opened 2 years ago Closed 1 year ago

[socket-proc] HttpTransactionParent must have its own suspend counter and queue messages

Categories

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

enhancement

Tracking

()

RESOLVED FIXED
Tracking Status
firefox64 --- affected

People

(Reporter: mayhemer, Assigned: dragana)

References

(Blocks 1 open bug)

Details

(Whiteboard: [necko-triaged])

Attachments

(1 file, 1 obsolete file)

Call to HttpTransactionParent::Suspend/Resume must have its own suspend counter.  Any notifications that are received by the parent trans while suspend counter is >0 must be queued.  It may happen that RecvSuspend is received too late on the child and e.g. OnData/OnStop has already been sent - there can be a (valid) race.
Working on top of bug 1485355.
Assignee: nobody → honzab.moz
Status: NEW → ASSIGNED
No, rather on top of bug 1496464 to save Kershaw some merging and for basing on more complete code.
Depends on: 1496464
No longer depends on: 1485355
Attached patch v1 (obsolete) — Splinter Review
Backing up work that's been done.
Whiteboard: [necko-triaged]
fixing for bonus points
Summary: [scoket-proc] HttpTransactionParent must have its own suspend counter and queue messages → [socket-proc] HttpTransactionParent must have its own suspend counter and queue messages
Priority: P1 → P3
Assignee: honzab.moz → nobody
Status: ASSIGNED → NEW
Assignee: nobody → dd.mozilla
Status: NEW → ASSIGNED

There are some r+ patches which didn't land and no activity in this bug for 2 weeks.
:dragana, could you have a look please?

Flags: needinfo?(dd.mozilla)

This is a separate project that will not land on m.-c.. I was waiting for rebasing before landing...

Flags: needinfo?(dd.mozilla)
Attachment #9015317 - Attachment is obsolete: true
Status: ASSIGNED → RESOLVED
Closed: 1 year ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.