diff options
| author | Robin Dapp <rdapp@oss.qualcomm.com> | 2026-02-19 18:01:28 +0100 |
|---|---|---|
| committer | Robin Dapp <rdapp@oss.qualcomm.com> | 2026-02-24 21:48:48 +0100 |
| commit | ad5905160560d1abecfb90e10bcafbf19361649a (patch) | |
| tree | 86fe126ec2474748537396bfe763808831cae87d /libcpp | |
| parent | ce6d82d19d0a746f63166fa2eccb2091144e5dc8 (diff) | |
| download | gcc-master.zip gcc-master.tar.gz gcc-master.tar.bz2 | |
This fixes a latent bug on rv32 after the lra patch for PR123381.
We have those _lra insn_and_splits
(define_insn_and_split "*mov<V_FRACT:mode><P:mode>_lra"
[(set (match_operand:V_FRACT 0 "reg_or_mem_operand" "=vr, m,vr")
(match_operand:V_FRACT 1 "reg_or_mem_operand" " m,vr,vr"))
(clobber (match_scratch:P 2 "=&r,&r,X"))]
where the clobber constraint for the third alternative is 'X'.
This is supposed to express that we accept anything because we don't
need the scratch for that alternative anyway.
However, in mask_gather_load_run-11.c we create a mem
(insn 302 228 218 (parallel [
(set (reg:RVVMF8BI 124 v28 [orig:214 mask__62.15 ] [214])
(reg:RVVMF8BI 120 v24 [orig:354 MEM <vector(16) long long int> [(int64_t *)cond_32(D) + 768B] ] [354]))
(clobber (mem/c:SI (plus:SI (reg/f:SI 2 sp)
(const_int 12 [0xc])) [3 %sfp+-324 S4 A32]))
that match_scratch does not handle, so we never split the insn.
This patch changes the constraint to 'r', this should work as we don't
make use of it anyway.
gcc/ChangeLog:
* config/riscv/vector.md: Change 'X' to 'r' in _lra insns.
Diffstat (limited to 'libcpp')
0 files changed, 0 insertions, 0 deletions
