2014-06-18 15:11 GMT+02:00 Dominique Dhumieres <domi...@lps.ens.fr>: > >> The following change in predicates.md seems to be a bit premature. >> There is still the point about Darwin's PIC issue for unspec-gotpcrel. > > The change is indeed incompatible with the patch in pr61387 comment 9. > And without it the failures are back!-( > > Kai, what is wrong with Iain's patch in comment 4? > > TIA > > Dominique
Well, I don't see there patch, just a comment about a possible code-change. I think it is getting fixed at wrong place. Reverting fnaddr back to its original seems to be wrong solution here. It would be better to do something in the lines 'if (!flag_pic || targetm.binds_local_p (function) || TARGET_PECOFF)' instead. Why shall we make it here an unspec, and later on revert that change? If Darwin isn't able to handle such an UNSPEC_GOTPCREL as address for a sibcall pattern, we should avoid it in general, and not just papering over at place of mi__tunk. Kai
