Hi! On Thu, Jul 17, 2025 at 11:16:07AM +0530, Kishan Parmar wrote: > I would like to backport patch r15-2928-gbf891fcabca7a5 which is > available from GCC-15. > > In general, if -mcpu=power9, float128 hardware is available, but in the > case the user explicitly does -mno-float128-hardware or runs on a machine > that doesn't enable float128 by default (i.e. big endian Linux and AIX), > you don't want these insns to be generated.
Terminology: double-double *is* a valid float128 type. What we care about is if the hardware IEEE QP stuff is available, here. > Applyig r15-2928-gbf891fcabca7a5 causes 3 hunks and below is the rejected > diff, > Because three patterns modified by the patch, "isinf<mode>2", > "*xststdc_<mode>" > and xststdc_<mode> are not present in the GCC 14 source code. As these > patterns > were introduced to the master branch after the GCC 14 release. So, how do you resolve it in this backport, simply leaving out those hunks? If so: okay for 14. You don't want to backport further, right? More stuff becomes missing :-) Thanks! Segher
