My previous commits were reverted in D28683, this patch reapplies them along with the fix from D28678 which has already been accepted. The only original addition that this patch introduces is a check in setARMSubArch to ensure that the CPU_arch attribute exists before trying to read it, as it would seem to be causing the sanitizer failure.
Details
Details
Diff Detail
Diff Detail
- Repository
- rL LLVM
Event Timeline
Comment Actions
Hi Sam,
Now that we have a bit more time, would you mind running it through an LLD check?
--renato
Comment Actions
Hi Renato,
Yes, I've added LLD to my local build so I know that passes now. But I'm not sure about the issue raised by the sanitizer buildbot, I've fixed the obvious bug on the line reported by the sanitizer and I hope that is all that needs doing.
Cheers,
sam