[strace/strace] sysvipc handling is time64-incompatible (#116)

[Rich Felker]

Now that fix is in progress on the kernel side, what should the action on the strace side be? I think the choices are roughly:

1. Default to using kernel uapi headers with a test for whether they're broken (only needed on x32 and sparc32), falling back to libc ones if so. Users of a libc with time64 will have to ensure they have fixed uapi headers.

2. Default to using libc headers unless arch is 32-bit and time_t is 64-bit. In that case, fallback to kernel uapi headers.

I like approach 1 better because it's semantically correct, and would also work if libc were just using different structures to interface with applications independent of anything to do with time_t. The definitions strace wants here really are the ones defining the ABI between user and kernel space, not between application and libc.


-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/strace/strace/issues/116#issuecomment-574386421
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20200114/dd7283eb/attachment.html>