https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82855
--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Author: jakub Date: Tue Nov 7 20:48:35 2017 New Revision: 254509 URL: https://gcc.gnu.org/viewcvs?rev=254509&root=gcc&view=rev Log: PR target/82855 * config/i386/i386.c (ix86_swap_binary_operands_p): Treat RTX_COMM_COMPARE as commutative as well. (ix86_binary_operator_ok): Formatting fix. * config/i386/sse.md (*mul<mode>3<mask_name><round_name>, *<code><mode>3<mask_name><round_saeonly_name>, *<code><mode>3<mask_name>, *<code>tf3, *mul<mode>3<mask_name>, *<s>mul<mode>3_highpart<mask_name>, *vec_widen_umult_even_v16si<mask_name>, *vec_widen_umult_even_v8si<mask_name>, *vec_widen_umult_even_v4si<mask_name>, *vec_widen_smult_even_v16si<mask_name>, *vec_widen_smult_even_v8si<mask_name>, *sse4_1_mulv2siv2di3<mask_name>, *avx2_pmaddwd, *sse2_pmaddwd, *<sse4_1_avx2>_mul<mode>3<mask_name>, *avx2_<code><mode>3, *avx512f_<code><mode>3<mask_name>, *sse4_1_<code><mode>3<mask_name>, *<code>v8hi3, *sse4_1_<code><mode>3<mask_name>, *<code>v16qi3, *avx2_eq<mode>3, <avx512>_eq<mode>3<mask_scalar_merge_name>_1, *sse4_1_eqv2di3, *sse2_eq<mode>3, <mask_codefor><code><mode>3<mask_name>, *<code><mode>3, *<sse2_avx2>_uavg<mode>3<mask_name>, *<ssse3_avx2>_pmulhrsw<mode>3<mask_name>, *ssse3_pmulhrswv4hi3): Use !(MEM_P (operands[1]) && MEM_P (operands[2])) condition instead of ix86_binary_operator_ok. Formatting fixes. (*<plusminus_insn><mode>3<mask_name><round_name>, *<plusminus_insn><mode>3, *<plusminus_insn><mode>3_m): Formatting fixes. Modified: trunk/gcc/ChangeLog trunk/gcc/config/i386/i386.c trunk/gcc/config/i386/sse.md
