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
Stackwalk issue on arm64 #303
Comments
It should be fully functional on master, and that code hasn't changed other than your proccontrol changes in months (at least in any arm-specific way). Can you get a mutator-side stack trace for this crash? |
@wrwilliams how do I do that? |
Run the mutator under gdb and get the backtrace from the crash? It shouldn't be different from any other crashing program. |
I've reproduced what looks like this issue locally. Third party stack walking looks okay from here; it's only first party that's broken. I would guess that we're missing -fno-omit-frame-pointer or that the gcc flag for forcing a frame pointer has changed; the actual logic here should be stable. I'm checking the build system now but I'd suggest disabling cotire for stackwalker and testing again. The code as written should work if that flag is functional. ETA: the frame pointer flag is correct, per a clean/make VERBOSE=1. @ssunny7 do you mind taking a look at this one? Reproducible with:
Not sure why r29/r30 are not pointing to where the caller frame is saved, but they evidently aren't... |
Should be fixed by #307. I have no clue how this ever worked except by accident. |
This is observed on Mustang.
|
@mxz297 Does this match a plausible code path for the libdwarf double-frees you've seen? |
This does not look like crashing during opening a file with SymtabAPI, which is what I encountered. But I agree that running valgrind can help further. |
@wrwilliams @mxz297 please have a look at https://gist.github.com/ikitayama/fba4d7f0ff1f913b1a2143b97a5b62f6 |
Running it with DYNISINT_DEBUG_TRANSLATE=1 |
@wrwilliams @mxz297 does the below look reasonable?
|
Even PR #308 applied, running just ./test_basic -self ends up wth this message:
|
The uses of |
PR #309 fixes this issue. |
@wrwilliams I'm wondering whether the StackwalkerAPI is fully functional on someone's branch other than master. As of this writing I get:
The text was updated successfully, but these errors were encountered: