[Mesa-dev] [PATCH 8/8] nvc0: Add shader disk caching

[Mark Menzynski]

Adds shader disk caching for nvc0 to reduce the need to every time compile
shaders. Shaders are saved into disk_shader_cache from nvc0_screen structure.

It serializes the input nv50_ir_prog_info to compute the hash key and
also to do a byte compare between the original nv50_ir_prog_info and the one
saved in the cache. If keys match and also the byte compare returns they
are equal, shaders are same, and the compiled nv50_ir_prog_info_out from the
cache can be used instead of compiling input info.

Seems to be significantly improving loading times. Piglit tests seem
to be OK.

Signed-off-by: Mark Menzynski <mmenzyns at redhat.com>
---
 .../drivers/nouveau/nvc0/nvc0_context.h       |  1 +
 .../drivers/nouveau/nvc0/nvc0_program.c       | 49 ++++++++++++++++---
 .../drivers/nouveau/nvc0/nvc0_shader_state.c  |  3 +-
 src/gallium/drivers/nouveau/nvc0/nvc0_state.c |  2 +
 4 files changed, 46 insertions(+), 9 deletions(-)

diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_context.h b/src/gallium/drivers/nouveau/nvc0/nvc0_context.h
index 8a2a8f2797e..4b83d1afeb4 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_context.h
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_context.h
@@ -321,6 +321,7 @@ extern struct draw_stage *nvc0_draw_render_stage(struct nvc0_context *);
 
 /* nvc0_program.c */
 bool nvc0_program_translate(struct nvc0_program *, uint16_t chipset,
+                            struct disk_cache *,
                             struct pipe_debug_callback *);
 bool nvc0_program_upload(struct nvc0_context *, struct nvc0_program *);
 void nvc0_program_destroy(struct nvc0_context *, struct nvc0_program *);
diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
index 1a5073292e8..06b6f7b4db5 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
@@ -24,6 +24,7 @@
 
 #include "compiler/nir/nir.h"
 #include "tgsi/tgsi_ureg.h"
+#include "util/blob.h"
 
 #include "nvc0/nvc0_context.h"
 
@@ -568,11 +569,19 @@ nvc0_program_dump(struct nvc0_program *prog)
 
 bool
 nvc0_program_translate(struct nvc0_program *prog, uint16_t chipset,
+                       struct disk_cache *disk_shader_cache,
                        struct pipe_debug_callback *debug)
 {
+   struct blob blob;
    struct nv50_ir_prog_info *info;
    struct nv50_ir_prog_info_out info_out = {};
-   int ret;
+
+   void *cached_data = NULL;
+   size_t cached_size;
+   bool shader_found = false;
+
+   int ret = 0;
+   cache_key key;
 
    info = CALLOC_STRUCT(nv50_ir_prog_info);
    if (!info)
@@ -631,14 +640,38 @@ nvc0_program_translate(struct nvc0_program *prog, uint16_t chipset,
    info->assignSlots = nvc0_program_assign_varying_slots;
 
    /* these fields might be overwritten by the compiler */
-   info_out.bin.smemSize = prog->cp.smem_size;
-   info_out.io.genUserClip = prog->vp.num_ucps;
-
-   ret = nv50_ir_generate_code(info, &info_out);
-   if (ret) {
-      NOUVEAU_ERR("shader translation failed: %i\n", ret);
-      goto out;
+   info->bin.smemSize = prog->cp.smem_size;
+   info->io.genUserClip = prog->vp.num_ucps;
+
+   blob_init(&blob);
+   nv50_ir_prog_info_serialize(&blob, info);
+
+   if (disk_shader_cache) {
+      disk_cache_compute_key(disk_shader_cache, blob.data, blob.size, key);
+      cached_data = disk_cache_get(disk_shader_cache, key, &cached_size);
+
+      if (cached_data && cached_size >= blob.size) { // blob.size is the size of serialized "info"
+         if (memcmp(cached_data, blob.data, blob.size) == 0) {
+            shader_found = true;
+            /* Blob contains only "info". In disk cache, "info_out" comes right after it */
+            size_t offset = blob.size;
+            nv50_ir_prog_info_out_deserialize(cached_data, cached_size, offset, &info_out);
+         }
+      }
+      free(cached_data);
+   }
+   if (!shader_found) {
+      ret = nv50_ir_generate_code(info, &info_out);
+      if (ret) {
+         NOUVEAU_ERR("shader translation failed: %i\n", ret);
+         goto out;
+      }
+      if (disk_shader_cache) {
+         nv50_ir_prog_info_out_serialize(&blob, &info_out);
+         disk_cache_put(disk_shader_cache, key, blob.data, blob.size, NULL);
+      }
    }
+   blob_finish(&blob);
 
    prog->code = info_out.bin.code;
    prog->code_size = info_out.bin.codeSize;
diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_shader_state.c b/src/gallium/drivers/nouveau/nvc0/nvc0_shader_state.c
index 774c5648113..4327a89454b 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_shader_state.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_shader_state.c
@@ -54,7 +54,8 @@ nvc0_program_validate(struct nvc0_context *nvc0, struct nvc0_program *prog)
 
    if (!prog->translated) {
       prog->translated = nvc0_program_translate(
-         prog, nvc0->screen->base.device->chipset, &nvc0->base.debug);
+         prog, nvc0->screen->base.device->chipset,
+         nvc0->screen->base.disk_shader_cache, &nvc0->base.debug);
       if (!prog->translated)
          return false;
    }
diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_state.c b/src/gallium/drivers/nouveau/nvc0/nvc0_state.c
index 49546e5de68..e57d1ed39e9 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_state.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_state.c
@@ -619,6 +619,7 @@ nvc0_sp_state_create(struct pipe_context *pipe,
 
    prog->translated = nvc0_program_translate(
       prog, nvc0_context(pipe)->screen->base.device->chipset,
+      nvc0_context(pipe)->screen->base.disk_shader_cache,
       &nouveau_context(pipe)->debug);
 
    return (void *)prog;
@@ -758,6 +759,7 @@ nvc0_cp_state_create(struct pipe_context *pipe,
 
    prog->translated = nvc0_program_translate(
       prog, nvc0_context(pipe)->screen->base.device->chipset,
+      nvc0_context(pipe)->screen->base.disk_shader_cache,
       &nouveau_context(pipe)->debug);
 
    return (void *)prog;
-- 
2.21.1