aco: fix partial copies on GFX6/7

While we don't allow partial subdword copies,
we still need to be able to split 64bit registers

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5663>

diff --git a/src/amd/compiler/aco_lower_to_hw_instr.cpp b/src/amd/compiler/aco_lower_to_hw_instr.cpp
index 1973cca027efe9b96b74fbbfa77a9570521545d4..f052f7ed79224f3a2a1a5ad379a4813bb025a044 100644 (file)
--- a/src/amd/compiler/aco_lower_to_hw_instr.cpp
+++ b/src/amd/compiler/aco_lower_to_hw_instr.cpp
@@ -1323,7 +1323,7 @@ void handle_operands(std::map<PhysReg, copy_operation>& copy_map, lower_context*
          return;
       }
       if (it == copy_map.end()) {
-         if (!skip_partial_copies || ctx->program->chip_class < GFX8)
+         if (!skip_partial_copies)
             break;
          skip_partial_copies = false;
          it = copy_map.begin();
@@ -1423,7 +1423,8 @@ void handle_operands(std::map<PhysReg, copy_operation>& copy_map, lower_context*
              * a partial copy allows further copies, it should be done instead. */
             bool partial_copy = (has_zero_use_bytes == 0xf) || (has_zero_use_bytes == 0xf0);
             for (std::pair<const PhysReg, copy_operation>& copy : copy_map) {
-               if (partial_copy)
+               /* on GFX6/7, we can only do copies with full registers */
+               if (partial_copy || ctx->program->chip_class <= GFX7)
                   break;
                for (uint16_t i = 0; i < copy.second.bytes; i++) {
                   /* distance might underflow */