On Wed, Nov 13, 2013 at 12:33 PM, Eric Botcazou <ebotca...@adacore.com> wrote: >> Changing the coalesce cost to MUST_COALESCE_COST for >> PARM_DECLs and RESULT_DECLs should add the required >> checking (we'll ICE if we cannot honor that coalescing request). >> Note that the code also has a !DECL_IGNORED_P check, so >> eventually the expansion code change should restrict itself to >> that as well. >> >> But the patch is ok, you can followup with the coalesce cost change >> and the !DECL_IGNORED_P checking (if you are not confident enough it >> works as-is). > > Even if I were, I strongly believe in consistency and the changes you sketched > here are the ones I should have come up with, had I known my way around the > coalescing machinery... So, thanks again for the tip, patch attached, tested > with no regressions on x86_64-suse-linux.
Ok. Thanks, Richard. > > 2013-11-13 Eric Botcazou <ebotca...@adacore.com> > > * cfgexpand.c (expand_used_vars): Allocate space for partitions based > on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info > or if optimization is enabled. > * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is > disabled, > require that all the names based on a PARM_DECL or a RESULT_DECL that > isn't ignored for debug info be coalesced. > > > -- > Eric Botcazou
