1 /*
2 * Copyright © 2020 Intel Corporation
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
13 * Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
21 * IN THE SOFTWARE.
22 */
23
24 #include "nir.h"
25 #include "nir_builder.h"
26 #include "nir_phi_builder.h"
27 #include "util/u_math.h"
28
29 static bool
move_system_values_to_top(nir_shader * shader)30 move_system_values_to_top(nir_shader *shader)
31 {
32 nir_function_impl *impl = nir_shader_get_entrypoint(shader);
33
34 bool progress = false;
35 nir_foreach_block(block, impl) {
36 nir_foreach_instr_safe(instr, block) {
37 if (instr->type != nir_instr_type_intrinsic)
38 continue;
39
40 /* These intrinsics not only can't be re-materialized but aren't
41 * preserved when moving to the continuation shader. We have to move
42 * them to the top to ensure they get spilled as needed.
43 */
44 nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(instr);
45 switch (intrin->intrinsic) {
46 case nir_intrinsic_load_shader_record_ptr:
47 case nir_intrinsic_load_btd_local_arg_addr_intel:
48 nir_instr_remove(instr);
49 nir_instr_insert(nir_before_cf_list(&impl->body), instr);
50 progress = true;
51 break;
52
53 default:
54 break;
55 }
56 }
57 }
58
59 if (progress) {
60 nir_metadata_preserve(impl, nir_metadata_block_index |
61 nir_metadata_dominance);
62 } else {
63 nir_metadata_preserve(impl, nir_metadata_all);
64 }
65
66 return progress;
67 }
68
69 static bool
instr_is_shader_call(nir_instr * instr)70 instr_is_shader_call(nir_instr *instr)
71 {
72 if (instr->type != nir_instr_type_intrinsic)
73 return false;
74
75 nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(instr);
76 return intrin->intrinsic == nir_intrinsic_trace_ray ||
77 intrin->intrinsic == nir_intrinsic_report_ray_intersection ||
78 intrin->intrinsic == nir_intrinsic_execute_callable;
79 }
80
81 /* Previously named bitset, it had to be renamed as FreeBSD defines a struct
82 * named bitset in sys/_bitset.h required by pthread_np.h which is included
83 * from src/util/u_thread.h that is indirectly included by this file.
84 */
85 struct brw_bitset {
86 BITSET_WORD *set;
87 unsigned size;
88 };
89
90 static struct brw_bitset
bitset_create(void * mem_ctx,unsigned size)91 bitset_create(void *mem_ctx, unsigned size)
92 {
93 return (struct brw_bitset) {
94 .set = rzalloc_array(mem_ctx, BITSET_WORD, BITSET_WORDS(size)),
95 .size = size,
96 };
97 }
98
99 static bool
src_is_in_bitset(nir_src * src,void * _set)100 src_is_in_bitset(nir_src *src, void *_set)
101 {
102 struct brw_bitset *set = _set;
103 assert(src->is_ssa);
104
105 /* Any SSA values which were added after we generated liveness information
106 * are things generated by this pass and, while most of it is arithmetic
107 * which we could re-materialize, we don't need to because it's only used
108 * for a single load/store and so shouldn't cross any shader calls.
109 */
110 if (src->ssa->index >= set->size)
111 return false;
112
113 return BITSET_TEST(set->set, src->ssa->index);
114 }
115
116 static void
add_ssa_def_to_bitset(nir_ssa_def * def,struct brw_bitset * set)117 add_ssa_def_to_bitset(nir_ssa_def *def, struct brw_bitset *set)
118 {
119 if (def->index >= set->size)
120 return;
121
122 BITSET_SET(set->set, def->index);
123 }
124
125 static bool
can_remat_instr(nir_instr * instr,struct brw_bitset * remat)126 can_remat_instr(nir_instr *instr, struct brw_bitset *remat)
127 {
128 /* Set of all values which are trivially re-materializable and we shouldn't
129 * ever spill them. This includes:
130 *
131 * - Undef values
132 * - Constants
133 * - Uniforms (UBO or push constant)
134 * - ALU combinations of any of the above
135 * - Derefs which are either complete or casts of any of the above
136 *
137 * Because this pass rewrites things in-order and phis are always turned
138 * into register writes, We can use "is it SSA?" to answer the question
139 * "can my source be re-materialized?".
140 */
141 switch (instr->type) {
142 case nir_instr_type_alu:
143 if (!nir_instr_as_alu(instr)->dest.dest.is_ssa)
144 return false;
145
146 return nir_foreach_src(instr, src_is_in_bitset, remat);
147
148 case nir_instr_type_deref:
149 return nir_foreach_src(instr, src_is_in_bitset, remat);
150
151 case nir_instr_type_intrinsic: {
152 nir_intrinsic_instr *intrin = nir_instr_as_intrinsic(instr);
153 switch (intrin->intrinsic) {
154 case nir_intrinsic_load_ubo:
155 case nir_intrinsic_vulkan_resource_index:
156 case nir_intrinsic_vulkan_resource_reindex:
157 case nir_intrinsic_load_vulkan_descriptor:
158 case nir_intrinsic_load_push_constant:
159 /* These intrinsics don't need to be spilled as long as they don't
160 * depend on any spilled values.
161 */
162 return nir_foreach_src(instr, src_is_in_bitset, remat);
163
164 case nir_intrinsic_load_scratch_base_ptr:
165 case nir_intrinsic_load_ray_launch_id:
166 case nir_intrinsic_load_btd_dss_id_intel:
167 case nir_intrinsic_load_btd_global_arg_addr_intel:
168 case nir_intrinsic_load_btd_resume_sbt_addr_intel:
169 case nir_intrinsic_load_ray_base_mem_addr_intel:
170 case nir_intrinsic_load_ray_hw_stack_size_intel:
171 case nir_intrinsic_load_ray_sw_stack_size_intel:
172 case nir_intrinsic_load_ray_num_dss_rt_stacks_intel:
173 case nir_intrinsic_load_ray_hit_sbt_addr_intel:
174 case nir_intrinsic_load_ray_hit_sbt_stride_intel:
175 case nir_intrinsic_load_ray_miss_sbt_addr_intel:
176 case nir_intrinsic_load_ray_miss_sbt_stride_intel:
177 case nir_intrinsic_load_callable_sbt_addr_intel:
178 case nir_intrinsic_load_callable_sbt_stride_intel:
179 /* Notably missing from the above list is btd_local_arg_addr_intel.
180 * This is because the resume shader will have a different local
181 * argument pointer because it has a different BSR. Any access of
182 * the original shader's local arguments needs to be preserved so
183 * that pointer has to be saved on the stack.
184 *
185 * TODO: There may be some system values we want to avoid
186 * re-materializing as well but we have to be very careful
187 * to ensure that it's a system value which cannot change
188 * across a shader call.
189 */
190 return true;
191
192 default:
193 return false;
194 }
195 }
196
197 case nir_instr_type_ssa_undef:
198 case nir_instr_type_load_const:
199 return true;
200
201 default:
202 return false;
203 }
204 }
205
206 static bool
can_remat_ssa_def(nir_ssa_def * def,struct brw_bitset * remat)207 can_remat_ssa_def(nir_ssa_def *def, struct brw_bitset *remat)
208 {
209 return can_remat_instr(def->parent_instr, remat);
210 }
211
212 static nir_ssa_def *
remat_ssa_def(nir_builder * b,nir_ssa_def * def)213 remat_ssa_def(nir_builder *b, nir_ssa_def *def)
214 {
215 nir_instr *clone = nir_instr_clone(b->shader, def->parent_instr);
216 nir_builder_instr_insert(b, clone);
217 return nir_instr_ssa_def(clone);
218 }
219
220 struct pbv_array {
221 struct nir_phi_builder_value **arr;
222 unsigned len;
223 };
224
225 static struct nir_phi_builder_value *
get_phi_builder_value_for_def(nir_ssa_def * def,struct pbv_array * pbv_arr)226 get_phi_builder_value_for_def(nir_ssa_def *def,
227 struct pbv_array *pbv_arr)
228 {
229 if (def->index >= pbv_arr->len)
230 return NULL;
231
232 return pbv_arr->arr[def->index];
233 }
234
235 static nir_ssa_def *
get_phi_builder_def_for_src(nir_src * src,struct pbv_array * pbv_arr,nir_block * block)236 get_phi_builder_def_for_src(nir_src *src, struct pbv_array *pbv_arr,
237 nir_block *block)
238 {
239 assert(src->is_ssa);
240
241 struct nir_phi_builder_value *pbv =
242 get_phi_builder_value_for_def(src->ssa, pbv_arr);
243 if (pbv == NULL)
244 return NULL;
245
246 return nir_phi_builder_value_get_block_def(pbv, block);
247 }
248
249 static bool
rewrite_instr_src_from_phi_builder(nir_src * src,void * _pbv_arr)250 rewrite_instr_src_from_phi_builder(nir_src *src, void *_pbv_arr)
251 {
252 nir_block *block;
253 if (src->parent_instr->type == nir_instr_type_phi) {
254 nir_phi_src *phi_src = exec_node_data(nir_phi_src, src, src);
255 block = phi_src->pred;
256 } else {
257 block = src->parent_instr->block;
258 }
259
260 nir_ssa_def *new_def = get_phi_builder_def_for_src(src, _pbv_arr, block);
261 if (new_def != NULL)
262 nir_instr_rewrite_src(src->parent_instr, src, nir_src_for_ssa(new_def));
263 return true;
264 }
265
266 static nir_ssa_def *
spill_fill(nir_builder * before,nir_builder * after,nir_ssa_def * def,unsigned offset,nir_address_format address_format,unsigned stack_alignment)267 spill_fill(nir_builder *before, nir_builder *after, nir_ssa_def *def, unsigned offset,
268 nir_address_format address_format, unsigned stack_alignment)
269 {
270 const unsigned comp_size = def->bit_size / 8;
271
272 switch(address_format) {
273 case nir_address_format_32bit_offset:
274 nir_store_scratch(before, def, nir_imm_int(before, offset),
275 .align_mul = MIN2(comp_size, stack_alignment), .write_mask = ~0);
276 def = nir_load_scratch(after, def->num_components, def->bit_size,
277 nir_imm_int(after, offset), .align_mul = MIN2(comp_size, stack_alignment));
278 break;
279 case nir_address_format_64bit_global: {
280 nir_ssa_def *addr = nir_iadd_imm(before, nir_load_scratch_base_ptr(before, 1, 64, 1), offset);
281 nir_store_global(before, addr, MIN2(comp_size, stack_alignment), def, ~0);
282 addr = nir_iadd_imm(after, nir_load_scratch_base_ptr(after, 1, 64, 1), offset);
283 def = nir_load_global(after, addr, MIN2(comp_size, stack_alignment),
284 def->num_components, def->bit_size);
285 break;
286 }
287 default:
288 unreachable("Unimplemented address format");
289 }
290 return def;
291 }
292
293 static void
spill_ssa_defs_and_lower_shader_calls(nir_shader * shader,uint32_t num_calls,nir_address_format address_format,unsigned stack_alignment)294 spill_ssa_defs_and_lower_shader_calls(nir_shader *shader, uint32_t num_calls,
295 nir_address_format address_format,
296 unsigned stack_alignment)
297 {
298 /* TODO: If a SSA def is filled more than once, we probably want to just
299 * spill it at the LCM of the fill sites so we avoid unnecessary
300 * extra spills
301 *
302 * TODO: If a SSA def is defined outside a loop but live through some call
303 * inside the loop, we probably want to spill outside the loop. We
304 * may also want to fill outside the loop if it's not used in the
305 * loop.
306 *
307 * TODO: Right now, we only re-materialize things if their immediate
308 * sources are things which we filled. We probably want to expand
309 * that to re-materialize things whose sources are things we can
310 * re-materialize from things we filled. We may want some DAG depth
311 * heuristic on this.
312 */
313
314 /* This happens per-shader rather than per-impl because we mess with
315 * nir_shader::scratch_size.
316 */
317 nir_function_impl *impl = nir_shader_get_entrypoint(shader);
318
319 nir_metadata_require(impl, nir_metadata_live_ssa_defs |
320 nir_metadata_dominance |
321 nir_metadata_block_index);
322
323 void *mem_ctx = ralloc_context(shader);
324
325 const unsigned num_ssa_defs = impl->ssa_alloc;
326 const unsigned live_words = BITSET_WORDS(num_ssa_defs);
327 struct brw_bitset trivial_remat = bitset_create(mem_ctx, num_ssa_defs);
328
329 /* Array of all live SSA defs which are spill candidates */
330 nir_ssa_def **spill_defs =
331 rzalloc_array(mem_ctx, nir_ssa_def *, num_ssa_defs);
332
333 /* For each spill candidate, an array of every time it's defined by a fill,
334 * indexed by call instruction index.
335 */
336 nir_ssa_def ***fill_defs =
337 rzalloc_array(mem_ctx, nir_ssa_def **, num_ssa_defs);
338
339 /* For each call instruction, the liveness set at the call */
340 const BITSET_WORD **call_live =
341 rzalloc_array(mem_ctx, const BITSET_WORD *, num_calls);
342
343 /* For each call instruction, the block index of the block it lives in */
344 uint32_t *call_block_indices = rzalloc_array(mem_ctx, uint32_t, num_calls);
345
346 /* Walk the call instructions and fetch the liveness set and block index
347 * for each one. We need to do this before we start modifying the shader
348 * so that liveness doesn't complain that it's been invalidated. Don't
349 * worry, we'll be very careful with our live sets. :-)
350 */
351 unsigned call_idx = 0;
352 nir_foreach_block(block, impl) {
353 nir_foreach_instr(instr, block) {
354 if (!instr_is_shader_call(instr))
355 continue;
356
357 call_block_indices[call_idx] = block->index;
358
359 /* The objective here is to preserve values around shader call
360 * instructions. Therefore, we use the live set after the
361 * instruction as the set of things we want to preserve. Because
362 * none of our shader call intrinsics return anything, we don't have
363 * to worry about spilling over a return value.
364 *
365 * TODO: This isn't quite true for report_intersection.
366 */
367 call_live[call_idx] =
368 nir_get_live_ssa_defs(nir_after_instr(instr), mem_ctx);
369
370 call_idx++;
371 }
372 }
373
374 nir_builder before, after;
375 nir_builder_init(&before, impl);
376 nir_builder_init(&after, impl);
377
378 call_idx = 0;
379 unsigned max_scratch_size = shader->scratch_size;
380 nir_foreach_block(block, impl) {
381 nir_foreach_instr_safe(instr, block) {
382 nir_ssa_def *def = nir_instr_ssa_def(instr);
383 if (def != NULL) {
384 if (can_remat_ssa_def(def, &trivial_remat)) {
385 add_ssa_def_to_bitset(def, &trivial_remat);
386 } else {
387 spill_defs[def->index] = def;
388 }
389 }
390
391 if (!instr_is_shader_call(instr))
392 continue;
393
394 const BITSET_WORD *live = call_live[call_idx];
395
396 /* Make a copy of trivial_remat that we'll update as we crawl through
397 * the live SSA defs and unspill them.
398 */
399 struct brw_bitset remat = bitset_create(mem_ctx, num_ssa_defs);
400 memcpy(remat.set, trivial_remat.set, live_words * sizeof(BITSET_WORD));
401
402 /* Before the two builders are always separated by the call
403 * instruction, it won't break anything to have two of them.
404 */
405 before.cursor = nir_before_instr(instr);
406 after.cursor = nir_after_instr(instr);
407
408 unsigned offset = shader->scratch_size;
409 for (unsigned w = 0; w < live_words; w++) {
410 BITSET_WORD spill_mask = live[w] & ~trivial_remat.set[w];
411 while (spill_mask) {
412 int i = u_bit_scan(&spill_mask);
413 assert(i >= 0);
414 unsigned index = w * BITSET_WORDBITS + i;
415 assert(index < num_ssa_defs);
416
417 nir_ssa_def *def = spill_defs[index];
418 if (can_remat_ssa_def(def, &remat)) {
419 /* If this SSA def is re-materializable or based on other
420 * things we've already spilled, re-materialize it rather
421 * than spilling and filling. Anything which is trivially
422 * re-materializable won't even get here because we take
423 * those into account in spill_mask above.
424 */
425 def = remat_ssa_def(&after, def);
426 } else {
427 bool is_bool = def->bit_size == 1;
428 if (is_bool)
429 def = nir_b2b32(&before, def);
430
431 const unsigned comp_size = def->bit_size / 8;
432 offset = ALIGN(offset, comp_size);
433
434 def = spill_fill(&before, &after, def, offset,
435 address_format,stack_alignment);
436
437 if (is_bool)
438 def = nir_b2b1(&after, def);
439
440 offset += def->num_components * comp_size;
441 }
442
443 /* Mark this SSA def as available in the remat set so that, if
444 * some other SSA def we need is computed based on it, we can
445 * just re-compute instead of fetching from memory.
446 */
447 BITSET_SET(remat.set, index);
448
449 /* For now, we just make a note of this new SSA def. We'll
450 * fix things up with the phi builder as a second pass.
451 */
452 if (fill_defs[index] == NULL) {
453 fill_defs[index] =
454 rzalloc_array(mem_ctx, nir_ssa_def *, num_calls);
455 }
456 fill_defs[index][call_idx] = def;
457 }
458 }
459
460 nir_builder *b = &before;
461
462 offset = ALIGN(offset, stack_alignment);
463 max_scratch_size = MAX2(max_scratch_size, offset);
464
465 /* First thing on the called shader's stack is the resume address
466 * followed by a pointer to the payload.
467 */
468 nir_intrinsic_instr *call = nir_instr_as_intrinsic(instr);
469
470 /* Lower to generic intrinsics with information about the stack & resume shader. */
471 switch (call->intrinsic) {
472 case nir_intrinsic_trace_ray: {
473 nir_rt_trace_ray(b, call->src[0].ssa, call->src[1].ssa,
474 call->src[2].ssa, call->src[3].ssa,
475 call->src[4].ssa, call->src[5].ssa,
476 call->src[6].ssa, call->src[7].ssa,
477 call->src[8].ssa, call->src[9].ssa,
478 call->src[10].ssa,
479 .call_idx = call_idx, .stack_size = offset);
480 break;
481 }
482
483 case nir_intrinsic_report_ray_intersection:
484 unreachable("Any-hit shaders must be inlined");
485
486 case nir_intrinsic_execute_callable: {
487 nir_rt_execute_callable(b, call->src[0].ssa, call->src[1].ssa, .call_idx = call_idx, .stack_size = offset);
488 break;
489 }
490
491 default:
492 unreachable("Invalid shader call instruction");
493 }
494
495 nir_rt_resume(b, .call_idx = call_idx, .stack_size = offset);
496
497 nir_instr_remove(&call->instr);
498
499 call_idx++;
500 }
501 }
502 assert(call_idx == num_calls);
503 shader->scratch_size = max_scratch_size;
504
505 struct nir_phi_builder *pb = nir_phi_builder_create(impl);
506 struct pbv_array pbv_arr = {
507 .arr = rzalloc_array(mem_ctx, struct nir_phi_builder_value *,
508 num_ssa_defs),
509 .len = num_ssa_defs,
510 };
511
512 const unsigned block_words = BITSET_WORDS(impl->num_blocks);
513 BITSET_WORD *def_blocks = ralloc_array(mem_ctx, BITSET_WORD, block_words);
514
515 /* Go through and set up phi builder values for each spillable value which
516 * we ever needed to spill at any point.
517 */
518 for (unsigned index = 0; index < num_ssa_defs; index++) {
519 if (fill_defs[index] == NULL)
520 continue;
521
522 nir_ssa_def *def = spill_defs[index];
523
524 memset(def_blocks, 0, block_words * sizeof(BITSET_WORD));
525 BITSET_SET(def_blocks, def->parent_instr->block->index);
526 for (unsigned call_idx = 0; call_idx < num_calls; call_idx++) {
527 if (fill_defs[index][call_idx] != NULL)
528 BITSET_SET(def_blocks, call_block_indices[call_idx]);
529 }
530
531 pbv_arr.arr[index] = nir_phi_builder_add_value(pb, def->num_components,
532 def->bit_size, def_blocks);
533 }
534
535 /* Walk the shader one more time and rewrite SSA defs as needed using the
536 * phi builder.
537 */
538 nir_foreach_block(block, impl) {
539 nir_foreach_instr_safe(instr, block) {
540 nir_ssa_def *def = nir_instr_ssa_def(instr);
541 if (def != NULL) {
542 struct nir_phi_builder_value *pbv =
543 get_phi_builder_value_for_def(def, &pbv_arr);
544 if (pbv != NULL)
545 nir_phi_builder_value_set_block_def(pbv, block, def);
546 }
547
548 if (instr->type == nir_instr_type_phi)
549 continue;
550
551 nir_foreach_src(instr, rewrite_instr_src_from_phi_builder, &pbv_arr);
552
553 if (instr->type != nir_instr_type_intrinsic)
554 continue;
555
556 nir_intrinsic_instr *resume = nir_instr_as_intrinsic(instr);
557 if (resume->intrinsic != nir_intrinsic_rt_resume)
558 continue;
559
560 call_idx = nir_intrinsic_call_idx(resume);
561
562 /* Technically, this is the wrong place to add the fill defs to the
563 * phi builder values because we haven't seen any of the load_scratch
564 * instructions for this call yet. However, we know based on how we
565 * emitted them that no value ever gets used until after the load
566 * instruction has been emitted so this should be safe. If we ever
567 * fail validation due this it likely means a bug in our spilling
568 * code and not the phi re-construction code here.
569 */
570 for (unsigned index = 0; index < num_ssa_defs; index++) {
571 if (fill_defs[index] && fill_defs[index][call_idx]) {
572 nir_phi_builder_value_set_block_def(pbv_arr.arr[index], block,
573 fill_defs[index][call_idx]);
574 }
575 }
576 }
577
578 nir_if *following_if = nir_block_get_following_if(block);
579 if (following_if) {
580 nir_ssa_def *new_def =
581 get_phi_builder_def_for_src(&following_if->condition,
582 &pbv_arr, block);
583 if (new_def != NULL)
584 nir_if_rewrite_condition(following_if, nir_src_for_ssa(new_def));
585 }
586
587 /* Handle phi sources that source from this block. We have to do this
588 * as a separate pass because the phi builder assumes that uses and
589 * defs are processed in an order that respects dominance. When we have
590 * loops, a phi source may be a back-edge so we have to handle it as if
591 * it were one of the last instructions in the predecessor block.
592 */
593 nir_foreach_phi_src_leaving_block(block,
594 rewrite_instr_src_from_phi_builder,
595 &pbv_arr);
596 }
597
598 nir_phi_builder_finish(pb);
599
600 ralloc_free(mem_ctx);
601
602 nir_metadata_preserve(impl, nir_metadata_block_index |
603 nir_metadata_dominance);
604 }
605
606 static nir_instr *
find_resume_instr(nir_function_impl * impl,unsigned call_idx)607 find_resume_instr(nir_function_impl *impl, unsigned call_idx)
608 {
609 nir_foreach_block(block, impl) {
610 nir_foreach_instr(instr, block) {
611 if (instr->type != nir_instr_type_intrinsic)
612 continue;
613
614 nir_intrinsic_instr *resume = nir_instr_as_intrinsic(instr);
615 if (resume->intrinsic != nir_intrinsic_rt_resume)
616 continue;
617
618 if (nir_intrinsic_call_idx(resume) == call_idx)
619 return &resume->instr;
620 }
621 }
622 unreachable("Couldn't find resume instruction");
623 }
624
625 /* Walk the CF tree and duplicate the contents of every loop, one half runs on
626 * resume and the other half is for any post-resume loop iterations. We are
627 * careful in our duplication to ensure that resume_instr is in the resume
628 * half of the loop though a copy of resume_instr will remain in the other
629 * half as well in case the same shader call happens twice.
630 */
631 static bool
duplicate_loop_bodies(nir_function_impl * impl,nir_instr * resume_instr)632 duplicate_loop_bodies(nir_function_impl *impl, nir_instr *resume_instr)
633 {
634 nir_register *resume_reg = NULL;
635 for (nir_cf_node *node = resume_instr->block->cf_node.parent;
636 node->type != nir_cf_node_function; node = node->parent) {
637 if (node->type != nir_cf_node_loop)
638 continue;
639
640 nir_loop *loop = nir_cf_node_as_loop(node);
641
642 if (resume_reg == NULL) {
643 /* We only create resume_reg if we encounter a loop. This way we can
644 * avoid re-validating the shader and calling ssa_to_regs in the case
645 * where it's just if-ladders.
646 */
647 resume_reg = nir_local_reg_create(impl);
648 resume_reg->num_components = 1;
649 resume_reg->bit_size = 1;
650
651 nir_builder b;
652 nir_builder_init(&b, impl);
653
654 /* Initialize resume to true */
655 b.cursor = nir_before_cf_list(&impl->body);
656 nir_store_reg(&b, resume_reg, nir_imm_true(&b), 1);
657
658 /* Set resume to false right after the resume instruction */
659 b.cursor = nir_after_instr(resume_instr);
660 nir_store_reg(&b, resume_reg, nir_imm_false(&b), 1);
661 }
662
663 /* Before we go any further, make sure that everything which exits the
664 * loop or continues around to the top of the loop does so through
665 * registers. We're about to duplicate the loop body and we'll have
666 * serious trouble if we don't do this.
667 */
668 nir_convert_loop_to_lcssa(loop);
669 nir_lower_phis_to_regs_block(nir_loop_first_block(loop));
670 nir_lower_phis_to_regs_block(
671 nir_cf_node_as_block(nir_cf_node_next(&loop->cf_node)));
672
673 nir_cf_list cf_list;
674 nir_cf_list_extract(&cf_list, &loop->body);
675
676 nir_if *_if = nir_if_create(impl->function->shader);
677 _if->condition = nir_src_for_reg(resume_reg);
678 nir_cf_node_insert(nir_after_cf_list(&loop->body), &_if->cf_node);
679
680 nir_cf_list clone;
681 nir_cf_list_clone(&clone, &cf_list, &loop->cf_node, NULL);
682
683 /* Insert the clone in the else and the original in the then so that
684 * the resume_instr remains valid even after the duplication.
685 */
686 nir_cf_reinsert(&cf_list, nir_before_cf_list(&_if->then_list));
687 nir_cf_reinsert(&clone, nir_before_cf_list(&_if->else_list));
688 }
689
690 if (resume_reg != NULL)
691 nir_metadata_preserve(impl, nir_metadata_none);
692
693 return resume_reg != NULL;
694 }
695
696 static bool
cf_node_contains_instr(nir_cf_node * node,nir_instr * instr)697 cf_node_contains_instr(nir_cf_node *node, nir_instr *instr)
698 {
699 for (nir_cf_node *n = &instr->block->cf_node; n != NULL; n = n->parent) {
700 if (n == node)
701 return true;
702 }
703
704 return false;
705 }
706
707 static void
rewrite_phis_to_pred(nir_block * block,nir_block * pred)708 rewrite_phis_to_pred(nir_block *block, nir_block *pred)
709 {
710 nir_foreach_instr(instr, block) {
711 if (instr->type != nir_instr_type_phi)
712 break;
713
714 nir_phi_instr *phi = nir_instr_as_phi(instr);
715
716 ASSERTED bool found = false;
717 nir_foreach_phi_src(phi_src, phi) {
718 if (phi_src->pred == pred) {
719 found = true;
720 assert(phi_src->src.is_ssa);
721 nir_ssa_def_rewrite_uses(&phi->dest.ssa, phi_src->src.ssa);
722 break;
723 }
724 }
725 assert(found);
726 }
727 }
728
729 /** Flattens if ladders leading up to a resume
730 *
731 * Given a resume_instr, this function flattens any if ladders leading to the
732 * resume instruction and deletes any code that cannot be encountered on a
733 * direct path to the resume instruction. This way we get, for the most part,
734 * straight-line control-flow up to the resume instruction.
735 *
736 * While we do this flattening, we also move any code which is in the remat
737 * set up to the top of the function or to the top of the resume portion of
738 * the current loop. We don't worry about control-flow as we do this because
739 * phis will never be in the remat set (see can_remat_instr) and so nothing
740 * control-dependent will ever need to be re-materialized. It is possible
741 * that this algorithm will preserve too many instructions by moving them to
742 * the top but we leave that for DCE to clean up. Any code not in the remat
743 * set is deleted because it's either unused in the continuation or else
744 * unspilled from a previous continuation and the unspill code is after the
745 * resume instruction.
746 *
747 * If, for instance, we have something like this:
748 *
749 * // block 0
750 * if (cond1) {
751 * // block 1
752 * } else {
753 * // block 2
754 * if (cond2) {
755 * // block 3
756 * resume;
757 * if (cond3) {
758 * // block 4
759 * }
760 * } else {
761 * // block 5
762 * }
763 * }
764 *
765 * then we know, because we know the resume instruction had to be encoutered,
766 * that cond1 = false and cond2 = true and we lower as follows:
767 *
768 * // block 0
769 * // block 2
770 * // block 3
771 * resume;
772 * if (cond3) {
773 * // block 4
774 * }
775 *
776 * As you can see, the code in blocks 1 and 5 was removed because there is no
777 * path from the start of the shader to the resume instruction which execute
778 * blocks 1 or 5. Any remat code from blocks 0, 2, and 3 is preserved and
779 * moved to the top. If the resume instruction is inside a loop then we know
780 * a priori that it is of the form
781 *
782 * loop {
783 * if (resume) {
784 * // Contents containing resume_instr
785 * } else {
786 * // Second copy of contents
787 * }
788 * }
789 *
790 * In this case, we only descend into the first half of the loop. The second
791 * half is left alone as that portion is only ever executed after the resume
792 * instruction.
793 */
794 static bool
flatten_resume_if_ladder(nir_function_impl * impl,nir_instr * cursor,struct exec_list * child_list,bool child_list_contains_cursor,nir_instr * resume_instr,struct brw_bitset * remat)795 flatten_resume_if_ladder(nir_function_impl *impl,
796 nir_instr *cursor,
797 struct exec_list *child_list,
798 bool child_list_contains_cursor,
799 nir_instr *resume_instr,
800 struct brw_bitset *remat)
801 {
802 nir_shader *shader = impl->function->shader;
803 nir_cf_list cf_list;
804
805 /* If our child list contains the cursor instruction then we start out
806 * before the cursor instruction. We need to know this so that we can skip
807 * moving instructions which are already before the cursor.
808 */
809 bool before_cursor = child_list_contains_cursor;
810
811 nir_cf_node *resume_node = NULL;
812 foreach_list_typed_safe(nir_cf_node, child, node, child_list) {
813 switch (child->type) {
814 case nir_cf_node_block: {
815 nir_block *block = nir_cf_node_as_block(child);
816 nir_foreach_instr_safe(instr, block) {
817 if (instr == cursor) {
818 assert(nir_cf_node_is_first(&block->cf_node));
819 assert(before_cursor);
820 before_cursor = false;
821 continue;
822 }
823
824 if (instr == resume_instr)
825 goto found_resume;
826
827 if (!before_cursor && can_remat_instr(instr, remat)) {
828 nir_instr_remove(instr);
829 nir_instr_insert(nir_before_instr(cursor), instr);
830
831 nir_ssa_def *def = nir_instr_ssa_def(instr);
832 BITSET_SET(remat->set, def->index);
833 }
834 }
835 break;
836 }
837
838 case nir_cf_node_if: {
839 assert(!before_cursor);
840 nir_if *_if = nir_cf_node_as_if(child);
841 if (flatten_resume_if_ladder(impl, cursor, &_if->then_list,
842 false, resume_instr, remat)) {
843 resume_node = child;
844 rewrite_phis_to_pred(nir_cf_node_as_block(nir_cf_node_next(child)),
845 nir_if_last_then_block(_if));
846 goto found_resume;
847 }
848
849 if (flatten_resume_if_ladder(impl, cursor, &_if->else_list,
850 false, resume_instr, remat)) {
851 resume_node = child;
852 rewrite_phis_to_pred(nir_cf_node_as_block(nir_cf_node_next(child)),
853 nir_if_last_else_block(_if));
854 goto found_resume;
855 }
856 break;
857 }
858
859 case nir_cf_node_loop: {
860 assert(!before_cursor);
861 nir_loop *loop = nir_cf_node_as_loop(child);
862
863 if (cf_node_contains_instr(&loop->cf_node, resume_instr)) {
864 /* Thanks to our loop body duplication pass, every level of loop
865 * containing the resume instruction contains exactly three nodes:
866 * two blocks and an if. We don't want to lower away this if
867 * because it's the resume selection if. The resume half is
868 * always the then_list so that's what we want to flatten.
869 */
870 nir_block *header = nir_loop_first_block(loop);
871 nir_if *_if = nir_cf_node_as_if(nir_cf_node_next(&header->cf_node));
872
873 /* We want to place anything re-materialized from inside the loop
874 * at the top of the resume half of the loop.
875 */
876 nir_instr *loop_cursor =
877 &nir_intrinsic_instr_create(shader, nir_intrinsic_nop)->instr;
878 nir_instr_insert(nir_before_cf_list(&_if->then_list), loop_cursor);
879
880 ASSERTED bool found =
881 flatten_resume_if_ladder(impl, loop_cursor, &_if->then_list,
882 true, resume_instr, remat);
883 assert(found);
884 resume_node = child;
885 goto found_resume;
886 } else {
887 ASSERTED bool found =
888 flatten_resume_if_ladder(impl, cursor, &loop->body,
889 false, resume_instr, remat);
890 assert(!found);
891 }
892 break;
893 }
894
895 case nir_cf_node_function:
896 unreachable("Unsupported CF node type");
897 }
898 }
899 assert(!before_cursor);
900
901 /* If we got here, we didn't find the resume node or instruction. */
902 return false;
903
904 found_resume:
905 /* If we got here then we found either the resume node or the resume
906 * instruction in this CF list.
907 */
908 if (resume_node) {
909 /* If the resume instruction is buried in side one of our children CF
910 * nodes, resume_node now points to that child.
911 */
912 if (resume_node->type == nir_cf_node_if) {
913 /* Thanks to the recursive call, all of the interesting contents of
914 * resume_node have been copied before the cursor. We just need to
915 * copy the stuff after resume_node.
916 */
917 nir_cf_extract(&cf_list, nir_after_cf_node(resume_node),
918 nir_after_cf_list(child_list));
919 } else {
920 /* The loop contains its own cursor and still has useful stuff in it.
921 * We want to move everything after and including the loop to before
922 * the cursor.
923 */
924 assert(resume_node->type == nir_cf_node_loop);
925 nir_cf_extract(&cf_list, nir_before_cf_node(resume_node),
926 nir_after_cf_list(child_list));
927 }
928 } else {
929 /* If we found the resume instruction in one of our blocks, grab
930 * everything after it in the entire list (not just the one block), and
931 * place it before the cursor instr.
932 */
933 nir_cf_extract(&cf_list, nir_after_instr(resume_instr),
934 nir_after_cf_list(child_list));
935 }
936 nir_cf_reinsert(&cf_list, nir_before_instr(cursor));
937
938 if (!resume_node) {
939 /* We want the resume to be the first "interesting" instruction */
940 nir_instr_remove(resume_instr);
941 nir_instr_insert(nir_before_cf_list(&impl->body), resume_instr);
942 }
943
944 /* We've copied everything interesting out of this CF list to before the
945 * cursor. Delete everything else.
946 */
947 if (child_list_contains_cursor) {
948 nir_cf_extract(&cf_list, nir_after_instr(cursor),
949 nir_after_cf_list(child_list));
950 } else {
951 nir_cf_list_extract(&cf_list, child_list);
952 }
953 nir_cf_delete(&cf_list);
954
955 return true;
956 }
957
958 static nir_instr *
lower_resume(nir_shader * shader,int call_idx)959 lower_resume(nir_shader *shader, int call_idx)
960 {
961 nir_function_impl *impl = nir_shader_get_entrypoint(shader);
962
963 nir_instr *resume_instr = find_resume_instr(impl, call_idx);
964
965 if (duplicate_loop_bodies(impl, resume_instr)) {
966 nir_validate_shader(shader, "after duplicate_loop_bodies in "
967 "brw_nir_lower_shader_calls");
968 /* If we duplicated the bodies of any loops, run regs_to_ssa to get rid
969 * of all those pesky registers we just added.
970 */
971 NIR_PASS_V(shader, nir_lower_regs_to_ssa);
972 }
973
974 /* Re-index nir_ssa_def::index. We don't care about actual liveness in
975 * this pass but, so we can use the same helpers as the spilling pass, we
976 * need to make sure that live_index is something sane. It's used
977 * constantly for determining if an SSA value has been added since the
978 * start of the pass.
979 */
980 nir_index_ssa_defs(impl);
981
982 void *mem_ctx = ralloc_context(shader);
983
984 /* Used to track which things may have been assumed to be re-materialized
985 * by the spilling pass and which we shouldn't delete.
986 */
987 struct brw_bitset remat = bitset_create(mem_ctx, impl->ssa_alloc);
988
989 /* Create a nop instruction to use as a cursor as we extract and re-insert
990 * stuff into the CFG.
991 */
992 nir_instr *cursor =
993 &nir_intrinsic_instr_create(shader, nir_intrinsic_nop)->instr;
994 nir_instr_insert(nir_before_cf_list(&impl->body), cursor);
995
996 ASSERTED bool found =
997 flatten_resume_if_ladder(impl, cursor, &impl->body,
998 true, resume_instr, &remat);
999 assert(found);
1000
1001 ralloc_free(mem_ctx);
1002
1003 nir_validate_shader(shader, "after flatten_resume_if_ladder in "
1004 "brw_nir_lower_shader_calls");
1005
1006 nir_metadata_preserve(impl, nir_metadata_none);
1007
1008 return resume_instr;
1009 }
1010
1011 static void
replace_resume_with_halt(nir_shader * shader,nir_instr * keep)1012 replace_resume_with_halt(nir_shader *shader, nir_instr *keep)
1013 {
1014 nir_function_impl *impl = nir_shader_get_entrypoint(shader);
1015
1016 nir_builder b;
1017 nir_builder_init(&b, impl);
1018
1019 nir_foreach_block_safe(block, impl) {
1020 nir_foreach_instr_safe(instr, block) {
1021 if (instr == keep)
1022 continue;
1023
1024 if (instr->type != nir_instr_type_intrinsic)
1025 continue;
1026
1027 nir_intrinsic_instr *resume = nir_instr_as_intrinsic(instr);
1028 if (resume->intrinsic != nir_intrinsic_rt_resume)
1029 continue;
1030
1031 /* If this is some other resume, then we've kicked off a ray or
1032 * bindless thread and we don't want to go any further in this
1033 * shader. Insert a halt so that NIR will delete any instructions
1034 * dominated by this call instruction including the scratch_load
1035 * instructions we inserted.
1036 */
1037 nir_cf_list cf_list;
1038 nir_cf_extract(&cf_list, nir_after_instr(&resume->instr),
1039 nir_after_block(block));
1040 nir_cf_delete(&cf_list);
1041 b.cursor = nir_instr_remove(&resume->instr);
1042 nir_jump(&b, nir_jump_halt);
1043 break;
1044 }
1045 }
1046 }
1047
1048 /** Lower shader call instructions to split shaders.
1049 *
1050 * Shader calls can be split into an initial shader and a series of "resume"
1051 * shaders. When the shader is first invoked, it is the initial shader which
1052 * is executed. At any point in the initial shader or any one of the resume
1053 * shaders, a shader call operation may be performed. The possible shader call
1054 * operations are:
1055 *
1056 * - trace_ray
1057 * - report_ray_intersection
1058 * - execute_callable
1059 *
1060 * When a shader call operation is performed, we push all live values to the
1061 * stack,call rt_trace_ray/rt_execute_callable and then kill the shader. Once
1062 * the operation we invoked is complete, a callee shader will return execution
1063 * to the respective resume shader. The resume shader pops the contents off
1064 * the stack and picks up where the calling shader left off.
1065 *
1066 * Stack management is assumed to be done after this pass. Call
1067 * instructions and their resumes get annotated with stack information that
1068 * should be enough for the backend to implement proper stack management.
1069 */
1070 bool
nir_lower_shader_calls(nir_shader * shader,nir_address_format address_format,unsigned stack_alignment,nir_shader *** resume_shaders_out,uint32_t * num_resume_shaders_out,void * mem_ctx)1071 nir_lower_shader_calls(nir_shader *shader,
1072 nir_address_format address_format,
1073 unsigned stack_alignment,
1074 nir_shader ***resume_shaders_out,
1075 uint32_t *num_resume_shaders_out,
1076 void *mem_ctx)
1077 {
1078 nir_function_impl *impl = nir_shader_get_entrypoint(shader);
1079
1080 nir_builder b;
1081 nir_builder_init(&b, impl);
1082
1083 int num_calls = 0;
1084 nir_foreach_block(block, impl) {
1085 nir_foreach_instr_safe(instr, block) {
1086 if (instr_is_shader_call(instr))
1087 num_calls++;
1088 }
1089 }
1090
1091 if (num_calls == 0) {
1092 nir_shader_preserve_all_metadata(shader);
1093 *num_resume_shaders_out = 0;
1094 return false;
1095 }
1096
1097 /* Some intrinsics not only can't be re-materialized but aren't preserved
1098 * when moving to the continuation shader. We have to move them to the top
1099 * to ensure they get spilled as needed.
1100 */
1101 {
1102 bool progress = false;
1103 NIR_PASS(progress, shader, move_system_values_to_top);
1104 if (progress)
1105 NIR_PASS(progress, shader, nir_opt_cse);
1106 }
1107
1108 NIR_PASS_V(shader, spill_ssa_defs_and_lower_shader_calls,
1109 num_calls, address_format, stack_alignment);
1110
1111 nir_opt_remove_phis(shader);
1112
1113 /* Make N copies of our shader */
1114 nir_shader **resume_shaders = ralloc_array(mem_ctx, nir_shader *, num_calls);
1115 for (unsigned i = 0; i < num_calls; i++)
1116 resume_shaders[i] = nir_shader_clone(mem_ctx, shader);
1117
1118 replace_resume_with_halt(shader, NULL);
1119 for (unsigned i = 0; i < num_calls; i++) {
1120 nir_instr *resume_instr = lower_resume(resume_shaders[i], i);
1121 replace_resume_with_halt(resume_shaders[i], resume_instr);
1122 nir_opt_remove_phis(resume_shaders[i]);
1123 }
1124
1125 *resume_shaders_out = resume_shaders;
1126 *num_resume_shaders_out = num_calls;
1127
1128 return true;
1129 }
1130