Skip to content
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

Ocassionally line number in stacktrace is 0 #1173

Closed
alex opened this issue Feb 17, 2018 · 7 comments
Closed

Ocassionally line number in stacktrace is 0 #1173

alex opened this issue Feb 17, 2018 · 7 comments

Comments

@alex
Copy link
Contributor

alex commented Feb 17, 2018

I'm not sure where this bug exists, it might be a bug in ASAN or the symbolizer, filing here as a first step :-)

If you look at https://oss-fuzz.com/v2/testcase-detail/5638304557957120?noredirect=1 at frame #5 you can see it points to line 0. I'm not sure what the cause might be, none of the potential locations are in macros or anything else that I might expect to cause problems.

@kcc
Copy link
Contributor

kcc commented Feb 17, 2018

These are usually bugs in the compiler-generated dwarf or (less frequently) in the symbolizer.
Unless this is something very frequent and very annoying I'd just ignore it.
Bugs like this are periodically fixed, and getting the reproducer from imagemagick will require disproportionately large work.

@alex
Copy link
Contributor Author

alex commented Feb 17, 2018

It's definitely not ruining my life or anything :-) Sounds like the best thing to do is just hope LLVM fixes whatever the bug is.

@kcc
Copy link
Contributor

kcc commented Feb 17, 2018

Yep. As usual, if you have a smaller repro the bug is more likely to be fixed.

@kcc kcc closed this as completed Feb 17, 2018
@alex
Copy link
Contributor Author

alex commented Feb 17, 2018

Indeed.

@alex
Copy link
Contributor Author

alex commented Feb 17, 2018

Is there a good way to attempt to test whether an attempted repro is effected without actually triggering an ASAN failure?

@kcc
Copy link
Contributor

kcc commented Feb 17, 2018

just use addr2line and llbm-symbolizer on the binary with the affected PC.
If addr2line also shows bogus value, it's likely a bug in emitted dwarf.
If llvm-symbolizer shows the bogus value and addr2line shows a good one, most likely the LLVM dwarf reader is wrong.
(You can also dump dwarf with elfdump and some such, but I never remember the incantations)

@alex
Copy link
Contributor Author

alex commented Feb 17, 2018 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants