[Mesa-dev] [PATCH] i965/fs: Add an allow_spilling flag to brw_compile_fs
Eduardo Lima Mitev
elima at igalia.com
Tue May 17 09:40:44 UTC 2016
On 05/16/2016 11:33 PM, Jason Ekstrand wrote:
> This allows us to disable spilling for blorp shaders since blorp state
> setup doesn't handle spilling. Without this, blorp faisl hard if you run
>
s/faisl/fails
> with INTEL_DEBUG=spill.
>
> Cc: Curro <currojerez at riseup.net>
> ---
> src/intel/vulkan/anv_pipeline.c | 3 ++-
> src/mesa/drivers/dri/i965/brw_blorp.c | 2 +-
> src/mesa/drivers/dri/i965/brw_compiler.h | 1 +
> src/mesa/drivers/dri/i965/brw_fs.cpp | 29 +++++++++++++----------
> src/mesa/drivers/dri/i965/brw_fs.h | 6 ++---
> src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp | 4 ++--
> src/mesa/drivers/dri/i965/brw_wm.c | 3 ++-
> 7 files changed, 28 insertions(+), 20 deletions(-)
>
> diff --git a/src/intel/vulkan/anv_pipeline.c b/src/intel/vulkan/anv_pipeline.c
> index a8e31b1..7d265d8 100644
> --- a/src/intel/vulkan/anv_pipeline.c
> +++ b/src/intel/vulkan/anv_pipeline.c
> @@ -675,7 +675,8 @@ anv_pipeline_compile_fs(struct anv_pipeline *pipeline,
> unsigned code_size;
> const unsigned *shader_code =
> brw_compile_fs(compiler, NULL, mem_ctx, &key, &prog_data, nir,
> - NULL, -1, -1, pipeline->use_repclear, &code_size, NULL);
> + NULL, -1, -1, true, pipeline->use_repclear,
> + &code_size, NULL);
> if (shader_code == NULL) {
> ralloc_free(mem_ctx);
> return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> diff --git a/src/mesa/drivers/dri/i965/brw_blorp.c b/src/mesa/drivers/dri/i965/brw_blorp.c
> index 09a0fd1..9590968 100644
> --- a/src/mesa/drivers/dri/i965/brw_blorp.c
> +++ b/src/mesa/drivers/dri/i965/brw_blorp.c
> @@ -223,7 +223,7 @@ brw_blorp_compile_nir_shader(struct brw_context *brw, struct nir_shader *nir,
>
> const unsigned *program =
> brw_compile_fs(compiler, brw, mem_ctx, wm_key, &wm_prog_data, nir,
> - NULL, -1, -1, use_repclear, program_size, NULL);
> + NULL, -1, -1, false, use_repclear, program_size, NULL);
>
> /* Copy the relavent bits of wm_prog_data over into the blorp prog data */
> prog_data->dispatch_8 = wm_prog_data.dispatch_8;
> diff --git a/src/mesa/drivers/dri/i965/brw_compiler.h b/src/mesa/drivers/dri/i965/brw_compiler.h
> index a2148ae..15f32be 100644
> --- a/src/mesa/drivers/dri/i965/brw_compiler.h
> +++ b/src/mesa/drivers/dri/i965/brw_compiler.h
> @@ -789,6 +789,7 @@ brw_compile_fs(const struct brw_compiler *compiler, void *log_data,
> struct gl_program *prog,
> int shader_time_index8,
> int shader_time_index16,
> + bool allow_spilling,
> bool use_rep_send,
> unsigned *final_assembly_size,
> char **error_str);
> diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
> index 2de5533..55d10fc 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
> @@ -5462,7 +5462,7 @@ fs_visitor::fixup_3src_null_dest()
> }
>
> void
> -fs_visitor::allocate_registers()
> +fs_visitor::allocate_registers(bool allow_spilling)
> {
> bool allocated_without_spills;
>
> @@ -5472,6 +5472,8 @@ fs_visitor::allocate_registers()
> SCHEDULE_PRE_LIFO,
> };
>
> + bool spill_all = allow_spilling && (INTEL_DEBUG & DEBUG_SPILL_FS);
> +
> /* Try each scheduling heuristic to see if it can successfully register
> * allocate without spilling. They should be ordered by decreasing
> * performance but increasing likelihood of allocating.
> @@ -5483,7 +5485,7 @@ fs_visitor::allocate_registers()
> assign_regs_trivial();
> allocated_without_spills = true;
> } else {
> - allocated_without_spills = assign_regs(false);
> + allocated_without_spills = assign_regs(false, spill_all);
> }
> if (allocated_without_spills)
> break;
> @@ -5508,12 +5510,14 @@ fs_visitor::allocate_registers()
> /* Since we're out of heuristics, just go spill registers until we
> * get an allocation.
> */
> - while (!assign_regs(true)) {
> + while (!assign_regs(true, spill_all)) {
> if (failed)
> break;
> }
> }
>
> + assert(last_scratch == 0 || allow_spilling);
> +
> /* This must come after all optimization and register allocation, since
> * it inserts dead code that happens to have side effects, and it does
> * so based on the actual physical registers in use.
> @@ -5559,7 +5563,7 @@ fs_visitor::run_vs(gl_clip_plane *clip_planes)
> assign_vs_urb_setup();
>
> fixup_3src_null_dest();
> - allocate_registers();
> + allocate_registers(true);
>
> return !failed;
> }
> @@ -5641,7 +5645,7 @@ fs_visitor::run_tcs_single_patch()
> assign_tcs_single_patch_urb_setup();
>
> fixup_3src_null_dest();
> - allocate_registers();
> + allocate_registers(true);
>
> return !failed;
> }
> @@ -5675,7 +5679,7 @@ fs_visitor::run_tes()
> assign_tes_urb_setup();
>
> fixup_3src_null_dest();
> - allocate_registers();
> + allocate_registers(true);
>
> return !failed;
> }
> @@ -5724,13 +5728,13 @@ fs_visitor::run_gs()
> assign_gs_urb_setup();
>
> fixup_3src_null_dest();
> - allocate_registers();
> + allocate_registers(true);
>
> return !failed;
> }
>
> bool
> -fs_visitor::run_fs(bool do_rep_send)
> +fs_visitor::run_fs(bool allow_spilling, bool do_rep_send)
> {
> brw_wm_prog_data *wm_prog_data = (brw_wm_prog_data *) this->prog_data;
> brw_wm_prog_key *wm_key = (brw_wm_prog_key *) this->key;
> @@ -5794,7 +5798,7 @@ fs_visitor::run_fs(bool do_rep_send)
> assign_urb_setup();
>
> fixup_3src_null_dest();
> - allocate_registers();
> + allocate_registers(allow_spilling);
>
> if (failed)
> return false;
> @@ -5837,7 +5841,7 @@ fs_visitor::run_cs()
> assign_curb_setup();
>
> fixup_3src_null_dest();
> - allocate_registers();
> + allocate_registers(true);
>
> if (failed)
> return false;
> @@ -5962,6 +5966,7 @@ brw_compile_fs(const struct brw_compiler *compiler, void *log_data,
> const nir_shader *src_shader,
> struct gl_program *prog,
> int shader_time_index8, int shader_time_index16,
> + bool allow_spilling,
> bool use_rep_send,
> unsigned *final_assembly_size,
> char **error_str)
> @@ -6005,7 +6010,7 @@ brw_compile_fs(const struct brw_compiler *compiler, void *log_data,
> fs_visitor v8(compiler, log_data, mem_ctx, key,
> &prog_data->base, prog, shader, 8,
> shader_time_index8);
> - if (!v8.run_fs(false /* do_rep_send */)) {
> + if (!v8.run_fs(allow_spilling, false /* do_rep_send */)) {
> if (error_str)
> *error_str = ralloc_strdup(mem_ctx, v8.fail_msg);
>
> @@ -6023,7 +6028,7 @@ brw_compile_fs(const struct brw_compiler *compiler, void *log_data,
> &prog_data->base, prog, shader, 16,
> shader_time_index16);
> v16.import_uniforms(&v8);
> - if (!v16.run_fs(use_rep_send)) {
> + if (!v16.run_fs(allow_spilling, use_rep_send)) {
> compiler->shader_perf_log(log_data,
> "SIMD16 shader failed to compile: %s",
> v16.fail_msg);
> diff --git a/src/mesa/drivers/dri/i965/brw_fs.h b/src/mesa/drivers/dri/i965/brw_fs.h
> index d4eb8fb..b6fe564 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs.h
> +++ b/src/mesa/drivers/dri/i965/brw_fs.h
> @@ -105,14 +105,14 @@ public:
> uint32_t const_offset);
> void DEP_RESOLVE_MOV(const brw::fs_builder &bld, int grf);
>
> - bool run_fs(bool do_rep_send);
> + bool run_fs(bool allow_spilling, bool do_rep_send);
> bool run_vs(gl_clip_plane *clip_planes);
> bool run_tcs_single_patch();
> bool run_tes();
> bool run_gs();
> bool run_cs();
> void optimize();
> - void allocate_registers();
> + void allocate_registers(bool allow_spilling);
> void setup_fs_payload_gen4();
> void setup_fs_payload_gen6();
> void setup_vs_payload();
> @@ -127,7 +127,7 @@ public:
> void assign_tcs_single_patch_urb_setup();
> void assign_tes_urb_setup();
> void assign_gs_urb_setup();
> - bool assign_regs(bool allow_spilling);
> + bool assign_regs(bool allow_spilling, bool spill_all);
> void assign_regs_trivial();
> void calculate_payload_ranges(int payload_node_count,
> int *payload_last_use_ip);
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp b/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp
> index 2347cd5..e65f73b 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_reg_allocate.cpp
> @@ -542,7 +542,7 @@ setup_mrf_hack_interference(fs_visitor *v, struct ra_graph *g,
> }
>
> bool
> -fs_visitor::assign_regs(bool allow_spilling)
> +fs_visitor::assign_regs(bool allow_spilling, bool spill_all)
> {
> /* Most of this allocation was written for a reg_width of 1
> * (dispatch_width == 8). In extending to SIMD16, the code was
> @@ -668,7 +668,7 @@ fs_visitor::assign_regs(bool allow_spilling)
> }
>
> /* Debug of register spilling: Go spill everything. */
> - if (unlikely(INTEL_DEBUG & DEBUG_SPILL_FS)) {
> + if (unlikely(spill_all)) {
> int reg = choose_spill_reg(g);
>
> if (reg != -1) {
> diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
> index 395b0b8..8af1ff7 100644
> --- a/src/mesa/drivers/dri/i965/brw_wm.c
> +++ b/src/mesa/drivers/dri/i965/brw_wm.c
> @@ -137,7 +137,8 @@ brw_codegen_wm_prog(struct brw_context *brw,
> program = brw_compile_fs(brw->intelScreen->compiler, brw, mem_ctx,
> key, &prog_data, fp->program.Base.nir,
> &fp->program.Base, st_index8, st_index16,
> - brw->use_rep_send, &program_size, &error_str);
> + true, brw->use_rep_send,
> + &program_size, &error_str);
> if (program == NULL) {
> if (prog) {
> prog->LinkStatus = false;
>
More information about the mesa-dev
mailing list