[Mesa-dev] [PATCH 1/2] radeonsi: add a debug flag that disables optimized shader variants

[Nicolai Hähnle]

Thanks for this. The series is

Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>

On 21.11.2016 22:41, Marek Olšák wrote:
> From: Marek Olšák <marek.olsak at amd.com>
>
> ---
>  src/gallium/drivers/radeon/r600_pipe_common.c   | 1 +
>  src/gallium/drivers/radeon/r600_pipe_common.h   | 1 +
>  src/gallium/drivers/radeonsi/si_state_shaders.c | 5 +++++
>  3 files changed, 7 insertions(+)
>
> diff --git a/src/gallium/drivers/radeon/r600_pipe_common.c b/src/gallium/drivers/radeon/r600_pipe_common.c
> index 1a7a712..15e8a12 100644
> --- a/src/gallium/drivers/radeon/r600_pipe_common.c
> +++ b/src/gallium/drivers/radeon/r600_pipe_common.c
> @@ -664,20 +664,21 @@ static const struct debug_named_value common_debug_options[] = {
>  	{ "gs", DBG_GS, "Print geometry shaders" },
>  	{ "ps", DBG_PS, "Print pixel shaders" },
>  	{ "cs", DBG_CS, "Print compute shaders" },
>  	{ "tcs", DBG_TCS, "Print tessellation control shaders" },
>  	{ "tes", DBG_TES, "Print tessellation evaluation shaders" },
>  	{ "noir", DBG_NO_IR, "Don't print the LLVM IR"},
>  	{ "notgsi", DBG_NO_TGSI, "Don't print the TGSI"},
>  	{ "noasm", DBG_NO_ASM, "Don't print disassembled shaders"},
>  	{ "preoptir", DBG_PREOPT_IR, "Print the LLVM IR before initial optimizations" },
>  	{ "checkir", DBG_CHECK_IR, "Enable additional sanity checks on shader IR" },
> +	{ "nooptvariant", DBG_NO_OPT_VARIANT, "Disable compiling optimized shader variants." },
>
>  	{ "testdma", DBG_TEST_DMA, "Invoke SDMA tests and exit." },
>
>  	/* features */
>  	{ "nodma", DBG_NO_ASYNC_DMA, "Disable asynchronous DMA" },
>  	{ "nohyperz", DBG_NO_HYPERZ, "Disable Hyper-Z" },
>  	/* GL uses the word INVALIDATE, gallium uses the word DISCARD */
>  	{ "noinvalrange", DBG_NO_DISCARD_RANGE, "Disable handling of INVALIDATE_RANGE map flags" },
>  	{ "no2d", DBG_NO_2D_TILING, "Disable 2D tiling" },
>  	{ "notiling", DBG_NO_TILING, "Disable tiling" },
> diff --git a/src/gallium/drivers/radeon/r600_pipe_common.h b/src/gallium/drivers/radeon/r600_pipe_common.h
> index 941fc37..25d40da 100644
> --- a/src/gallium/drivers/radeon/r600_pipe_common.h
> +++ b/src/gallium/drivers/radeon/r600_pipe_common.h
> @@ -72,20 +72,21 @@
>  #define DBG_GS			(1 << 7)
>  #define DBG_PS			(1 << 8)
>  #define DBG_CS			(1 << 9)
>  #define DBG_TCS			(1 << 10)
>  #define DBG_TES			(1 << 11)
>  #define DBG_NO_IR		(1 << 12)
>  #define DBG_NO_TGSI		(1 << 13)
>  #define DBG_NO_ASM		(1 << 14)
>  #define DBG_PREOPT_IR		(1 << 15)
>  #define DBG_CHECK_IR		(1 << 16)
> +#define DBG_NO_OPT_VARIANT	(1 << 17)
>  /* gaps */
>  #define DBG_TEST_DMA		(1 << 20)
>  /* Bits 21-31 are reserved for the r600g driver. */
>  /* features */
>  #define DBG_NO_ASYNC_DMA	(1llu << 32)
>  #define DBG_NO_HYPERZ		(1llu << 33)
>  #define DBG_NO_DISCARD_RANGE	(1llu << 34)
>  #define DBG_NO_2D_TILING	(1llu << 35)
>  #define DBG_NO_TILING		(1llu << 36)
>  #define DBG_SWITCH_ON_EOP	(1llu << 37)
> diff --git a/src/gallium/drivers/radeonsi/si_state_shaders.c b/src/gallium/drivers/radeonsi/si_state_shaders.c
> index cd4b339..0f5a920 100644
> --- a/src/gallium/drivers/radeonsi/si_state_shaders.c
> +++ b/src/gallium/drivers/radeonsi/si_state_shaders.c
> @@ -1114,20 +1114,25 @@ static void si_build_shader_variant(void *job, int thread_index)
>  /* Select the hw shader variant depending on the current state. */
>  static int si_shader_select_with_key(struct si_screen *sscreen,
>  				     struct si_shader_ctx_state *state,
>  				     struct si_shader_key *key,
>  				     int thread_index)
>  {
>  	static const struct si_shader_key zeroed;
>  	struct si_shader_selector *sel = state->cso;
>  	struct si_shader *current = state->current;
>  	struct si_shader *iter, *shader = NULL;
> +
> +	if (unlikely(sscreen->b.chip_class & DBG_NO_OPT_VARIANT)) {
> +		memset(&key->opt, 0, sizeof(key->opt));
> +	}
> +
>  again:
>  	/* Check if we don't need to change anything.
>  	 * This path is also used for most shaders that don't need multiple
>  	 * variants, it will cost just a computation of the key and this
>  	 * test. */
>  	if (likely(current &&
>  		   memcmp(&current->key, key, sizeof(*key)) == 0 &&
>  		   (!current->is_optimized ||
>  		    util_queue_fence_is_signalled(&current->optimized_ready))))
>  		return 0;
>