[aarch64] PR107920 - Fix incorrect handling of virtual operands in svld1rq_impl::fold

2022-12-01 Thread Prathamesh Kulkarni via Gcc-patches
Hi, The following test: #include "arm_sve.h" svint8_t test_s8(int8_t *x) { return svld1rq_s8 (svptrue_b8 (), &x[0]); } ICE's with -march=armv8.2-a+sve -O1 -fno-tree-ccp -fno-tree-forwprop: during GIMPLE pass: fre pr107920.c: In function ‘test_s8’: pr107920.c:7:1: internal compiler error: in ex

Re: [aarch64] PR107920 - Fix incorrect handling of virtual operands in svld1rq_impl::fold

2022-12-05 Thread Richard Sandiford via Gcc-patches
Prathamesh Kulkarni writes: > Hi, > The following test: > > #include "arm_sve.h" > > svint8_t > test_s8(int8_t *x) > { > return svld1rq_s8 (svptrue_b8 (), &x[0]); > } > > ICE's with -march=armv8.2-a+sve -O1 -fno-tree-ccp -fno-tree-forwprop: > during GIMPLE pass: fre > pr107920.c: In function ‘te

Re: [aarch64] PR107920 - Fix incorrect handling of virtual operands in svld1rq_impl::fold

2022-12-05 Thread Prathamesh Kulkarni via Gcc-patches
On Tue, 6 Dec 2022 at 00:08, Richard Sandiford wrote: > > Prathamesh Kulkarni writes: > > Hi, > > The following test: > > > > #include "arm_sve.h" > > > > svint8_t > > test_s8(int8_t *x) > > { > > return svld1rq_s8 (svptrue_b8 (), &x[0]); > > } > > > > ICE's with -march=armv8.2-a+sve -O1 -fno-t

Re: [aarch64] PR107920 - Fix incorrect handling of virtual operands in svld1rq_impl::fold

2022-12-05 Thread Richard Sandiford via Gcc-patches
Prathamesh Kulkarni writes: > On Tue, 6 Dec 2022 at 00:08, Richard Sandiford > wrote: >> >> Prathamesh Kulkarni writes: >> > Hi, >> > The following test: >> > >> > #include "arm_sve.h" >> > >> > svint8_t >> > test_s8(int8_t *x) >> > { >> > return svld1rq_s8 (svptrue_b8 (), &x[0]); >> > } >> >