AFAICT, there are no tests for objdump's -R option. It was added in https://reviews.llvm.org/D47493 (also clearly missing some tests).
Certainly missing a test.
LGTM as long as they are not used, but I'd like to check with @sunfish to be sure as well.
Thanks. The runtime check is now done earlier.
- If we can fold it to strdup, we should (strndup does more things than strdup).
- Annotation mechanism. (Works for strdup well).
p.s: strdup and strndup are part of C 20 :)
Test added. Thanks for the suggestion!
Nevermind that long report, it actually turned out to be a bug in NetBSD kernel that resulted in the thread not being resumed.
When we optimise for size, and need to emit runtime checks to disambiguate memory, we should nicely bail, don't vectorise, and not run in an assert like we currently do.
Looks like CUDA test-suite is triggering the assertion added by this patch:
I am taking a look.