[fedora-arm] quick update on clang/llvm

Jon Masters jcm at redhat.com
Mon May 7 06:44:09 UTC 2012


Hey,

So I can't make my TrimSlice stall in the testsuite under mock. It fails
the testsuite, but that's expected (and caught in the SPEC file, which
ignores the overall failure on ARM as it does on ppc64/s390x). What it
does not do is go out to lunch. Now, I ran it a couple of times and it
didn't fall over. It might be my board, it might be the F17 host, it
might be the recent kernel :) Lots of possibilities. Still, perhaps
someone can try a mock build on a F17 host+upstream kernel and let me
know if they get any stalls.

What I can do is get the root filesystem dsd sent to stall the exact
test he was experiencing, on my same system. So hmm...makes me think
some weird code generation - I'll check the binaries. I can at least use
that filesystem to figure out what's happening. I didn't get chance over
the weekend to throw a lot more time at it (you know, it being a weekend
and all) but I did spend some time reading up on LLVM and Clang design,
setup some stuff, and I'm in a good place to work this a bit this week.
If someone else has more time, please let me know.

Jon.


More information about the arm mailing list