-
Notifications
You must be signed in to change notification settings - Fork 55
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
Potential Unsound: 1 out-of-bound read and 5 unaligned memory access. #52
Comments
Thank you! My results so far:
Based on reading the error messages and based on my previous discussion with @RalfJung, I believe these to be instances of the Miri issue mentioned in the error messages. Due to predating Also, as I understand it, migrating to Does this analysis look correct to you? |
@hsivonen what you say makes sense for unaligned accesses, but should not lead to out-of-bounds errors. I should really prioritize fixing that alignment problem... |
Hi @hsivonen. Thank you for the analysis on alignment. I agree with you that these are safe behaviors given the nature of the library. With regard to out-of-bound, I should correct myself that what MIRI reports is not a read per-se, but rather creating a pointer that is outside the bounds of allocated memory. If you run the
Is the Since For your information, my toolchain versions are:
and the tarball imports Finally, thank you for the help @RalfJung. Thanks, |
Since this issue was reported, Miri has been changed to properly support hand-rolled alignment math, at the expense of having to run alignment-related tests a dozen times or so to be sure that they do not pass by chance. |
@hsivonen I can reproduce an OOB error, though it seems to be slightly different. With latest Miri, when I run
|
@RalfJung @YoshikiTakashima Thank you. The OOB case indeed materialized an OOB pointer though didn't dereference it. Fixed in #53 . I'm leaving this issue open to remind me to check the SIMD code for the same pattern. |
For clarity, closing this issue and spinning revision of the SIMD code into a distinct issue. |
Hello.
I'm Yoshiki, a PhD student at CMU.
We are testing a tool to automatically generate test cases from API data and existing tests.
A few of our generated test cases were reported as "unsound" by Miri, mostly due to unaligned or out-of-bound memory. I've attached a Tarball that contains the test cases that induce this behavior.
Please note that, because the framework leverages existing tests as templates, some of the test cases overlap with existing test cases for the library. In particular,
also shows up in the manually written test cases.
In case this is intended behavior, or you would prefer if I focused on other parts of the code, please let me know.
Thanks.
~Yoshiki
The text was updated successfully, but these errors were encountered: