-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
prog: panic: no result
during serialization
#4715
Comments
panic: no result
panic: no result
during serialization
Hi Christoph, How frequently/soon does it crash? I don't remember seeing anything similar on syzbot nor locally. |
We do this trick to dump more info when decoding programs: But it won't work for encoding, since if encoding panics we cannot encode to dump it :) Maybe if you flip this var to true, it may give some useful info earlier:
With this flag program manipulation will do more checks every now and then. |
Been running with that suggested change for a while now and I'm no longer seeing the crash reported above. What I AM seeing instead is a lot of this:
I'll keep this running over night to see if that is really the difference in behavior or just another artifact of the debug variable being set. |
What's in the crash log files for these? They should contain unmasked values for these NUM/ADDR + maybe some additional info (stack, programs). |
It's really interesting that we don't see this happen on syzbot and when we run syzkaller locally. If it's easily reproducible on your side, I wonder what we are doing differently. I think it would be great to find some minimized setup (Linux kernel revision + config, syzkaller revision + config, host and target arches) that would let us reproduce the problem too. |
It also may be some bad programs in the corpus. |
It's a separate problem from the one that @cpaasch described and I remember seeing it too when I ran syz-manger with prog validation enabled. We use negative vma addresses for our test program to emulate protected memory areas around the data buffer, and the validator is not particularly happy about this. |
Sorry. Should I move this to a separate issue or do we have a fix? |
Yes, let's create a separate issue for that. To not let it stop debugging of the original problem, does it work if you remove the following line and recompile syzkaller? syzkaller/sys/targets/common.go Line 43 in ddfc15a
|
Moving my posts to #4750 |
Strange this wasn't caught in tests. Tests should enable debug=true. Would be good to check and fix as well. |
I think I got an instrumented backtrace:
|
Then again, 6 minutes later I get the backtrace from the top of this issue, though I've been running with the |
No concrete expectation, it was just a simple way to try to get something more useful. I've tried to run prog package test for tens of CPU hours in a loop, but no success so far. |
No progress with |
I've run The problem is fixed in #4802. |
Forgot to get back to this, but just wanted to thank you as my syzkaller instances were running fine again last week! |
For quite a while now my syz-manager instances stopped running for long. Very soon they end up with an error and terminate (see below).
Let me know what kind of information I can provide to debug this. (go version: go1.22.2)
The text was updated successfully, but these errors were encountered: