https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91753
Tamar Christina <tnfchris at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |tnfchris at gcc dot gnu.org Resolution|--- |FIXED Target Milestone|--- |12.0 --- Comment #9 from Tamar Christina <tnfchris at gcc dot gnu.org> --- Fixed in GCC 12 where we generate at -O1/2/3 neon_transform_nada(uint8x16x4_t, unsigned char*, int): ldp q7, q6, [x0] ldp q5, q4, [x0, 32] cmp w1, 0 ble .L2 mov w2, 0 .L3: add w2, w2, 1 tbx v7.16b, {v0.16b - v3.16b}, v7.16b tbx v6.16b, {v0.16b - v3.16b}, v6.16b tbx v5.16b, {v0.16b - v3.16b}, v5.16b tbx v4.16b, {v0.16b - v3.16b}, v4.16b cmp w1, w2 bne .L3 .L2: stp q7, q6, [x0] stp q5, q4, [x0, 32] ret