From: Alyssa Rosenzweig Date: Wed, 13 Nov 2019 14:00:37 +0000 (-0500) Subject: pan/midgard: Prioritize texture registers X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=095654e3c23b255e6b29c63a8e5519539c43cffb;p=mesa.git pan/midgard: Prioritize texture registers On newer GPUs, this is a no-op. On older GPUs, this prevents needless spilling since texture registers are shared with a subset of work registers. Signed-off-by: Alyssa Rosenzweig Reviewed-by: Tomeu Vizoso Tested-by: Andre Heider --- diff --git a/src/panfrost/midgard/midgard_schedule.c b/src/panfrost/midgard/midgard_schedule.c index 588f7c6302e..2e794687e66 100644 --- a/src/panfrost/midgard/midgard_schedule.c +++ b/src/panfrost/midgard/midgard_schedule.c @@ -1118,7 +1118,8 @@ find_or_allocate_temp(compiler_context *ctx, unsigned hash) return temp; } -/* Reassigns numbering to get rid of gaps in the indices */ +/* Reassigns numbering to get rid of gaps in the indices and to prioritize + * smaller register classes */ static void mir_squeeze_index(compiler_context *ctx) @@ -1128,8 +1129,18 @@ mir_squeeze_index(compiler_context *ctx) /* TODO don't leak old hash_to_temp */ ctx->hash_to_temp = _mesa_hash_table_u64_create(NULL); + /* We need to prioritize texture registers on older GPUs so we don't + * fail RA trying to assign to work registers r0/r1 when a work + * register is already there */ + + mir_foreach_instr_global(ctx, ins) { + if (ins->type == TAG_TEXTURE_4) + ins->dest = find_or_allocate_temp(ctx, ins->dest); + } + mir_foreach_instr_global(ctx, ins) { - ins->dest = find_or_allocate_temp(ctx, ins->dest); + if (ins->type != TAG_TEXTURE_4) + ins->dest = find_or_allocate_temp(ctx, ins->dest); for (unsigned i = 0; i < ARRAY_SIZE(ins->src); ++i) ins->src[i] = find_or_allocate_temp(ctx, ins->src[i]);