https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97085
Bug ID: 97085
Summary: [11 Regression] aarch64, SVE: ICE in
gimple_expand_vec_cond_expr since r11-2610-ga1ee6d507b
Product: gcc
Version: 11.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: acoplan at gcc dot gnu.org
Target Milestone: ---
For the following testcase:
int a, b, c, d;
short e, g;
unsigned short f;
void h() {
for (; d; d++) {
g = d;
e = b == 0 ? 1 : a % b;
c ^= (f = e) > (g == 5);
}
}
AArch64 GCC ICEs at -O3 with -march=armv8.2-a+sve since
r11-2610-ga1ee6d507b0c26466be519d177f5a08b22f63647:
Author: Marc Glisse <marc.gli...@inria.fr>
Date: Fri Aug 7 17:49:04 2020
Disable some VEC_COND_EXPR transformations after vector lowering
To reproduce:
$ aarch64-none-elf-gcc -c test.c -O3 -march=armv8.2-a+sve
during GIMPLE pass: isel
test.c: In function 'h':
test.c:4:6: internal compiler error: in gimple_expand_vec_cond_expr, at
gimple-isel.cc:135
4 | void h() {
| ^
0x102c7bd gimple_expand_vec_cond_expr
/home/alecop01/toolchain/src/gcc/gcc/gimple-isel.cc:133
0x102cbe1 gimple_expand_vec_cond_exprs
/home/alecop01/toolchain/src/gcc/gcc/gimple-isel.cc:191
0x102cbe1 execute
/home/alecop01/toolchain/src/gcc/gcc/gimple-isel.cc:240
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
