I've now tracked this down to a problem with LEApcrel rematerialization where the rematerialized LEApcrel can address a different literal pool to the original. I've raised https://bugs.llvm.org/show_bug.cgi?id=33074
This is actually a bug that already existed before my patches, but because my patches made LEApcrel be rematerialized in more situations they made it more likely to trigger the bug. I'll continue looking into this to see if I can figure out how to fix it. John > -----Original Message----- > From: Renato Golin [mailto:renato.go...@linaro.org] > Sent: 16 May 2017 19:13 > To: John Brawn > Cc: James Molloy; Diana Picus; LLVM Commits; Clang Commits; nd > Subject: Re: Potential self-hosting failure > > On 16 May 2017 at 18:26, John Brawn <john.br...@arm.com> wrote: > > I've managed to reproduce this, but no luck so far in figuring out > > what exactly is going wrong. I'll continue looking into it tomorrow. > > Thanks John, > > I have reverted it for now on r303193, to get the bots green. > > cheers, > --renato _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
