crash when using reverse-cont in Firefox replay session
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 10.7k
- Forks
- 662
- Avg merge
- 2d 3h
- Merged PRs (30d)
- 2
Description
Debugging Firefox, I can repeatedly get the following assert/crash:
[FATAL /home/mfroman/rr/rr/src/Task.cc:2021:write_bytes_helper() errno: EIO](task 19627 %28rec:19195%29 at time 820829)
-> Assertion `nwritten == buf_size' failed to hold. Should have written 614400 bytes to 0x7f30c1e68000, but only wrote -1
Linux 4.2.0-36-generic #42~14.04.1-Ubuntu SMP Fri May 13 17:27:22 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
I'm on the lastest git commit for rr (3710623f9fbec963a8f8096092de1240f82a0091).
Debugging on mozilla-central changeset: d98f20c25fee
./mach run --debugger=rr https://web.ciscospark.com
I start a call on spark (currently debugging a video freeze), and then quit after I see the freeze.
rr replay
b rtcp_utility.cc:1516
b rtcp_receiver.cc:1107
c
c
reverse-cont
boom!
Let me know if you need additional info, or if I can clarify anything.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The failure is reported at rr/src/Task.cc:2021 in write_bytes_helper(); start there and reproduce with the supplied Firefox/rr recording. Then inspect the reverse-cont sequence and the breakpoints in rtcp_utility.cc:1516 and rtcp_receiver.cc:1107. Done means the reported replay no longer asserts; the issue provides no named regression test.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp, linux
- Domain
- operating-systems, tooling
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100