intel/fs/gen12: Fixup/simplify SWSB annotations of SIMD32 scratch writes.

Found by inspection.  Existing code was trying to avoid assuming that
an SBID had been assigned to the virtual instruction, but
synchronizing the header setup with respect to the previous SIMD16
SEND by using SYNC.ALLRD doesn't really seem possible unless the SEND
instruction had been assigned an SBID.  Assert-fail instead if no SBID
has been allocated.

Fixes: 15e3a0d9d264becc "intel/eu/gen12: Set SWSB annotations in hand-crafted assembly."
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Cc: 20.0 <mesa-stable@lists.freedesktop.org>

diff --git a/src/intel/compiler/brw_fs_generator.cpp b/src/intel/compiler/brw_fs_generator.cpp
index 00b0f29ef7b9193162be6a4250ed9e66ca50a7c5..163ddbc45d9dad7be0fa0df77d1e9e8fcd760eb1 100644 (file)
--- a/src/intel/compiler/brw_fs_generator.cpp
+++ b/src/intel/compiler/brw_fs_generator.cpp
@@ -1378,8 +1378,8 @@ fs_generator::generate_scratch_write(fs_inst *inst, struct brw_reg src)
       brw_set_default_group(p, inst->group + lower_size * i);
 
       if (i > 0) {
-         brw_set_default_swsb(p, tgl_swsb_null());
-         brw_SYNC(p, TGL_SYNC_ALLRD);
+         assert(swsb.mode & TGL_SBID_SET);
+         brw_set_default_swsb(p, tgl_swsb_sbid(TGL_SBID_SRC, swsb.sbid));
       } else {
          brw_set_default_swsb(p, tgl_swsb_src_dep(swsb));
       }
@@ -1387,11 +1387,7 @@ fs_generator::generate_scratch_write(fs_inst *inst, struct brw_reg src)
       brw_MOV(p, brw_uvec_mrf(lower_size, inst->base_mrf + 1, 0),
               retype(offset(src, block_size * i), BRW_REGISTER_TYPE_UD));
 
-      if (i + 1 < inst->exec_size / lower_size)
-         brw_set_default_swsb(p, tgl_swsb_regdist(1));
-      else
-         brw_set_default_swsb(p, tgl_swsb_dst_dep(swsb, 1));
-
+      brw_set_default_swsb(p, tgl_swsb_dst_dep(swsb, 1));
       brw_oword_block_write_scratch(p, brw_message_reg(inst->base_mrf),
                                     block_size,
                                     inst->offset + block_size * REG_SIZE * i);