Closed Bug 1399541 Opened 7 years ago Closed 7 years ago

Add assertion that we rcwn is enabled when mRaceCacheWithNetwork == true

Categories

(Core :: Networking, enhancement, P1)

enhancement

Tracking

()

RESOLVED WONTFIX

People

(Reporter: valentin, Assigned: valentin)

References

(Blocks 1 open bug)

Details

(Whiteboard: [necko-active])

Attachments

(2 files)

(In reply to Michal Novotny (:michal) from bug 1396307 comment #16)
> So add MOZ_RELEASE_ASSERT(sRCWNEnabled) where we set mRaceCacheWithNetwork =
> true. Then we would know for sure there is some problem.
MozReview-Commit-ID: 8UDaIklOOmw
Attachment #8907684 - Flags: review?(michal.novotny)
Attachment #8907684 - Flags: review?(michal.novotny) → review+
Pushed by ryanvm@gmail.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/4198b9d273b9
Add assertion that we rcwn is enabled when mRaceCacheWithNetwork is true. r=michal
Keywords: checkin-needed
https://hg.mozilla.org/mozilla-central/rev/4198b9d273b9
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla57
Depends on: 1399899
Backed out for causing bug 1399899.
https://hg.mozilla.org/mozilla-central/rev/943c2c2a1daf5cd0d3287c1974eb510615d5ed5b
Status: RESOLVED → REOPENED
Flags: needinfo?(valentin.gosu)
Resolution: FIXED → ---
Target Milestone: mozilla57 → ---
Attached file Ubuntu Opt Log
Bughunter found

Assertion failure: sRCWNEnabled (Racing should be enabled), at /builds/worker/workspace/build/src/netwerk/protocol/http/nsHttpChannel.cpp:9395

and

==5294==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc 0x7fc7df9217c6 bp 0x7ffdb2962a50 sp 0x7ffdb2962a30 T0)
==5294==The signal is caused by a WRITE memory access.
==5294==Hint: address points to the zero page.
    #0 0x7fc7df9217c5 in mozilla::net::nsHttpChannel::TriggerNetwork() /builds/worker/workspace/build/src/netwerk/protocol/http/nsHttpChannel.cpp:9395:9

in an *opt* asan build on https://www.gamasutra.com/view/news/128836/InDepth_Static_Code_Analysis.php on Ubuntu 16.04 but not Fedora 26. I don't know how an Assertion failure fired in an opt build...
(In reply to Bob Clary [:bc:] from comment #7)
> Created attachment 8908303 [details]
> Ubuntu Opt Log
> 
> Bughunter found
> 
> Assertion failure: sRCWNEnabled (Racing should be enabled), at
> /builds/worker/workspace/build/src/netwerk/protocol/http/nsHttpChannel.cpp:
> 9395
> 
> and
> 
> ==5294==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000 (pc
> 0x7fc7df9217c6 bp 0x7ffdb2962a50 sp 0x7ffdb2962a30 T0)
> ==5294==The signal is caused by a WRITE memory access.
> ==5294==Hint: address points to the zero page.
>     #0 0x7fc7df9217c5 in mozilla::net::nsHttpChannel::TriggerNetwork()
> /builds/worker/workspace/build/src/netwerk/protocol/http/nsHttpChannel.cpp:
> 9395:9
> 
> in an *opt* asan build on
> https://www.gamasutra.com/view/news/128836/InDepth_Static_Code_Analysis.php
> on Ubuntu 16.04 but not Fedora 26. I don't know how an Assertion failure
> fired in an opt build...

This was a RELEASE_ASSERT - it would have crashed in any build
Flags: needinfo?(valentin.gosu)
So, this bug has served its purpose, and confirmed that we sometimes set the flag incorrectly.
A fix has landed in bug 1396307. Marking as WONTFIX since the assert would be useless now.
Status: REOPENED → RESOLVED
Closed: 7 years ago7 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.