aco: fix carry-out size for wave32 v_add_co_u32_e64

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Fixes: e0bcefc3a0a ('aco/wave32: Use lane mask regclass for exec/vcc.')
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3902>

diff --git a/src/amd/compiler/aco_builder_h.py b/src/amd/compiler/aco_builder_h.py
index 97c03ac8adf0740f3edf49a029dd9a356d15c4a5..65eee5ebd1d29e70040eff13f95d454281180f41 100644 (file)
--- a/src/amd/compiler/aco_builder_h.py
+++ b/src/amd/compiler/aco_builder_h.py
@@ -392,7 +392,7 @@ public:
       if (!carry_in.op.isUndefined())
          return vop2(aco_opcode::v_addc_co_u32, Definition(dst), hint_vcc(def(lm)), a, b, carry_in);
       else if (program->chip_class >= GFX10 && carry_out)
-         return vop3(aco_opcode::v_add_co_u32_e64, Definition(dst), def(s2), a, b);
+         return vop3(aco_opcode::v_add_co_u32_e64, Definition(dst), def(lm), a, b);
       else if (program->chip_class < GFX9 || carry_out)
          return vop2(aco_opcode::v_add_co_u32, Definition(dst), hint_vcc(def(lm)), a, b);
       else