I agree that it's pretty edge-case-y :-) The other approach is at http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20140714/226276.html ; it's not much longer and keeps us more compatible with gcc, so I suppose it's slightly nicer.
On Thu, Jul 17, 2014 at 12:41 PM, Eric Christopher <[email protected]> wrote: > Yeah, I think either/or works. There's really no reason why they're > not there right now, but if we want to keep maximum compatibility then > we can do it Reid's way for now. > > -eric > > On Thu, Jul 17, 2014 at 12:31 PM, Nico Weber <[email protected]> wrote: > > Some more background: This list is used to check the clobber list, and > since > > ms inline asm doesn't have explicit clobber lists, > > AsmParser::parseMSInlineAsm() adds all touched registers to the clobber > > list. > > > > This change would also permit the segment registers in the clobber lists > for > > gcc registers. A different fix would be to change parseMSInlineAsm() to > not > > add segment registers as clobbers. rnk tells me he likes that approach > > better, so I'll do that instead. > > > > > > On Thu, Jul 17, 2014 at 12:06 PM, Eric Christopher <[email protected]> > > wrote: > >> > >> Seems reasonable. > >> > >> -eric > >> > >> On Thu, Jul 17, 2014 at 11:55 AM, Nico Weber <[email protected]> > wrote: > >> > Hi, > >> > > >> > the attached patch fixes PR20343 by adding the segment registers to > the > >> > x86 > >> > register list in Targets.cpp. > >> > > >> > Nico > >> > > >> > _______________________________________________ > >> > cfe-commits mailing list > >> > [email protected] > >> > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits > >> > > > > > >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
