ir3: Fix shared IMAD24 lowering.
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run

Caught by
dEQP-VK.spirv_assembly.instruction.compute.opsdotkhr.all_us_v2i16_out32
and friends in !41178

Fixes: b4874aa5cf ("ir3: Use scalar ALU instructions when possible")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/41278>
This commit is contained in:
Emma Anholt 2026-04-29 16:58:23 -07:00 committed by Marge Bot
parent 608bc0e593
commit d8b525466c

View file

@ -870,7 +870,7 @@ emit_alu(struct ir3_context *ctx, nir_alu_instr *alu)
case nir_op_imad24_ir3:
if (use_shared) {
dst = ir3_ADD_U_rpt(b, dst_sz,
ir3_MUL_U24_rpt(b, dst_sz, src[0], 0, src[1], 0),
ir3_MUL_S24_rpt(b, dst_sz, src[0], 0, src[1], 0),
0, src[2], 0);
} else {
dst = ir3_MAD_S24_rpt(b, dst_sz, src[0], 0, src[1], 0, src[2], 0);