[mesa.git] / src / compiler / glsl / shader_cache.cpp

/*
 * Copyright © 2014 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
 * DEALINGS IN THE SOFTWARE.
 */

/**
 * \file shader_cache.cpp
 *
 * GLSL shader cache implementation
 *
 * This uses disk_cache.c to write out a serialization of various
 * state that's required in order to successfully load and use a
 * binary written out by a drivers backend, this state is referred to as
 * "metadata" throughout the implementation.
 *
 * The hash key for glsl metadata is a hash of the hashes of each GLSL
 * source string as well as some API settings that change the final program
 * such as SSO, attribute bindings, frag data bindings, etc.
 *
 * In order to avoid caching any actual IR we use the put_key/get_key support
 * in the disk_cache to put the SHA-1 hash for each successfully compiled
 * shader into the cache, and optimisticly return early from glCompileShader
 * (if the identical shader had been successfully compiled in the past),
 * in the hope that the final linked shader will be found in the cache.
 * If anything goes wrong (shader variant not found, backend cache item is
 * corrupt, etc) we will use a fallback path to compile and link the IR.
 */

#include "blob.h"
#include "compiler/shader_info.h"
#include "glsl_symbol_table.h"
#include "glsl_parser_extras.h"
#include "ir.h"
#include "ir_optimization.h"
#include "ir_rvalue_visitor.h"
#include "ir_uniform.h"
#include "linker.h"
#include "link_varyings.h"
#include "main/core.h"
#include "nir.h"
#include "program.h"
#include "shader_cache.h"
#include "util/mesa-sha1.h"
#include "util/string_to_uint_map.h"

extern "C" {
#include "main/enums.h"
#include "main/shaderobj.h"
#include "program/program.h"
}

static void
compile_shaders(struct gl_context *ctx, struct gl_shader_program *prog) {
   for (unsigned i = 0; i < prog->NumShaders; i++) {
      _mesa_glsl_compile_shader(ctx, prog->Shaders[i], false, false, true);
   }
}

static void
get_struct_type_field_and_pointer_sizes(size_t *s_field_size,
                                        size_t *s_field_ptrs,
                                        unsigned num_fields)
{
   *s_field_size = sizeof(glsl_struct_field) * num_fields;
   *s_field_ptrs =
     sizeof(((glsl_struct_field *)0)->type) +
     sizeof(((glsl_struct_field *)0)->name);
}

static void
encode_type_to_blob(struct blob *blob, const glsl_type *type)
{
   uint32_t encoding;

   if (!type) {
      blob_write_uint32(blob, 0);
      return;
   }

   switch (type->base_type) {
   case GLSL_TYPE_UINT:
   case GLSL_TYPE_INT:
   case GLSL_TYPE_FLOAT:
   case GLSL_TYPE_BOOL:
   case GLSL_TYPE_DOUBLE:
   case GLSL_TYPE_UINT64:
   case GLSL_TYPE_INT64:
      encoding = (type->base_type << 24) |
         (type->vector_elements << 4) |
         (type->matrix_columns);
      break;
   case GLSL_TYPE_SAMPLER:
      encoding = (type->base_type) << 24 |
         (type->sampler_dimensionality << 4) |
         (type->sampler_shadow << 3) |
         (type->sampler_array << 2) |
         (type->sampled_type);
      break;
   case GLSL_TYPE_SUBROUTINE:
      encoding = type->base_type << 24;
      blob_write_uint32(blob, encoding);
      blob_write_string(blob, type->name);
      return;
   case GLSL_TYPE_IMAGE:
      encoding = (type->base_type) << 24 |
         (type->sampler_dimensionality << 3) |
         (type->sampler_array << 2) |
         (type->sampled_type);
      break;
   case GLSL_TYPE_ATOMIC_UINT:
      encoding = (type->base_type << 24);
      break;
   case GLSL_TYPE_ARRAY:
      blob_write_uint32(blob, (type->base_type) << 24);
      blob_write_uint32(blob, type->length);
      encode_type_to_blob(blob, type->fields.array);
      return;
   case GLSL_TYPE_STRUCT:
   case GLSL_TYPE_INTERFACE:
      blob_write_uint32(blob, (type->base_type) << 24);
      blob_write_string(blob, type->name);
      blob_write_uint32(blob, type->length);

      size_t s_field_size, s_field_ptrs;
      get_struct_type_field_and_pointer_sizes(&s_field_size, &s_field_ptrs,
                                              type->length);

      for (unsigned i = 0; i < type->length; i++) {
         encode_type_to_blob(blob, type->fields.structure[i].type);
         blob_write_string(blob, type->fields.structure[i].name);

         /* Write the struct field skipping the pointers */
         blob_write_bytes(blob,
                          ((char *)&type->fields.structure[i]) + s_field_ptrs,
                          s_field_size - s_field_ptrs);
      }

      if (type->is_interface()) {
         blob_write_uint32(blob, type->interface_packing);
         blob_write_uint32(blob, type->interface_row_major);
      }
      return;
   case GLSL_TYPE_VOID:
   case GLSL_TYPE_ERROR:
   default:
      assert(!"Cannot encode type!");
      encoding = 0;
      break;
   }

   blob_write_uint32(blob, encoding);
}

static const glsl_type *
decode_type_from_blob(struct blob_reader *blob)
{
   uint32_t u = blob_read_uint32(blob);

   if (u == 0) {
      return NULL;
   }

   glsl_base_type base_type = (glsl_base_type) (u >> 24);

   switch (base_type) {
   case GLSL_TYPE_UINT:
   case GLSL_TYPE_INT:
   case GLSL_TYPE_FLOAT:
   case GLSL_TYPE_BOOL:
   case GLSL_TYPE_DOUBLE:
   case GLSL_TYPE_UINT64:
   case GLSL_TYPE_INT64:
      return glsl_type::get_instance(base_type, (u >> 4) & 0x0f, u & 0x0f);
   case GLSL_TYPE_SAMPLER:
      return glsl_type::get_sampler_instance((enum glsl_sampler_dim) ((u >> 4) & 0x07),
                                             (u >> 3) & 0x01,
                                             (u >> 2) & 0x01,
                                             (glsl_base_type) ((u >> 0) & 0x03));
   case GLSL_TYPE_SUBROUTINE:
      return glsl_type::get_subroutine_instance(blob_read_string(blob));
   case GLSL_TYPE_IMAGE:
      return glsl_type::get_image_instance((enum glsl_sampler_dim) ((u >> 3) & 0x07),
                                             (u >> 2) & 0x01,
                                             (glsl_base_type) ((u >> 0) & 0x03));
   case GLSL_TYPE_ATOMIC_UINT:
      return glsl_type::atomic_uint_type;
   case GLSL_TYPE_ARRAY: {
      unsigned length = blob_read_uint32(blob);
      return glsl_type::get_array_instance(decode_type_from_blob(blob),
                                           length);
   }
   case GLSL_TYPE_STRUCT:
   case GLSL_TYPE_INTERFACE: {
      char *name = blob_read_string(blob);
      unsigned num_fields = blob_read_uint32(blob);

      size_t s_field_size, s_field_ptrs;
      get_struct_type_field_and_pointer_sizes(&s_field_size, &s_field_ptrs,
                                              num_fields);

      glsl_struct_field *fields =
         (glsl_struct_field *) malloc(s_field_size * num_fields);
      for (unsigned i = 0; i < num_fields; i++) {
         fields[i].type = decode_type_from_blob(blob);
         fields[i].name = blob_read_string(blob);

         blob_copy_bytes(blob, ((uint8_t *) &fields[i]) + s_field_ptrs,
                         s_field_size - s_field_ptrs);
      }

      const glsl_type *t;
      if (base_type == GLSL_TYPE_INTERFACE) {
         enum glsl_interface_packing packing =
            (glsl_interface_packing) blob_read_uint32(blob);
         bool row_major = blob_read_uint32(blob);
         t = glsl_type::get_interface_instance(fields, num_fields, packing,
                                               row_major, name);
      } else {
         t = glsl_type::get_record_instance(fields, num_fields, name);
      }

      free(fields);
      return t;
   }
   case GLSL_TYPE_VOID:
   case GLSL_TYPE_ERROR:
   default:
      assert(!"Cannot decode type!");
      return NULL;
   }
}

static void
write_subroutines(struct blob *metadata, struct gl_shader_program *prog)
{
   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (!sh)
         continue;

      struct gl_program *glprog = sh->Program;

      blob_write_uint32(metadata, glprog->sh.NumSubroutineUniforms);
      blob_write_uint32(metadata, glprog->sh.MaxSubroutineFunctionIndex);
      blob_write_uint32(metadata, glprog->sh.NumSubroutineFunctions);
      for (unsigned j = 0; j < glprog->sh.NumSubroutineFunctions; j++) {
         int num_types = glprog->sh.SubroutineFunctions[j].num_compat_types;

         blob_write_string(metadata, glprog->sh.SubroutineFunctions[j].name);
         blob_write_uint32(metadata, glprog->sh.SubroutineFunctions[j].index);
         blob_write_uint32(metadata, num_types);

         for (int k = 0; k < num_types; k++) {
            encode_type_to_blob(metadata,
                                glprog->sh.SubroutineFunctions[j].types[k]);
         }
      }
   }
}

static void
read_subroutines(struct blob_reader *metadata, struct gl_shader_program *prog)
{
   struct gl_subroutine_function *subs;

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (!sh)
         continue;

      struct gl_program *glprog = sh->Program;

      glprog->sh.NumSubroutineUniforms = blob_read_uint32(metadata);
      glprog->sh.MaxSubroutineFunctionIndex = blob_read_uint32(metadata);
      glprog->sh.NumSubroutineFunctions = blob_read_uint32(metadata);

      subs = rzalloc_array(prog, struct gl_subroutine_function,
                           glprog->sh.NumSubroutineFunctions);
      glprog->sh.SubroutineFunctions = subs;

      for (unsigned j = 0; j < glprog->sh.NumSubroutineFunctions; j++) {
         subs[j].name = ralloc_strdup(prog, blob_read_string (metadata));
         subs[j].index = (int) blob_read_uint32(metadata);
         subs[j].num_compat_types = (int) blob_read_uint32(metadata);

         subs[j].types = rzalloc_array(prog, const struct glsl_type *,
                                       subs[j].num_compat_types);
         for (int k = 0; k < subs[j].num_compat_types; k++) {
            subs[j].types[k] = decode_type_from_blob(metadata);
         }
      }
   }
}

static void
write_buffer_block(struct blob *metadata, struct gl_uniform_block *b)
{
   blob_write_string(metadata, b->Name);
   blob_write_uint32(metadata, b->NumUniforms);
   blob_write_uint32(metadata, b->Binding);
   blob_write_uint32(metadata, b->UniformBufferSize);
   blob_write_uint32(metadata, b->stageref);

   for (unsigned j = 0; j < b->NumUniforms; j++) {
      blob_write_string(metadata, b->Uniforms[j].Name);
      blob_write_string(metadata, b->Uniforms[j].IndexName);
      encode_type_to_blob(metadata, b->Uniforms[j].Type);
      blob_write_uint32(metadata, b->Uniforms[j].Offset);
   }
}

static void
write_buffer_blocks(struct blob *metadata, struct gl_shader_program *prog)
{
   blob_write_uint32(metadata, prog->data->NumUniformBlocks);
   blob_write_uint32(metadata, prog->data->NumShaderStorageBlocks);

   for (unsigned i = 0; i < prog->data->NumUniformBlocks; i++) {
      write_buffer_block(metadata, &prog->data->UniformBlocks[i]);
   }

   for (unsigned i = 0; i < prog->data->NumShaderStorageBlocks; i++) {
      write_buffer_block(metadata, &prog->data->ShaderStorageBlocks[i]);
   }

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (!sh)
         continue;

      struct gl_program *glprog = sh->Program;

      blob_write_uint32(metadata, glprog->info.num_ubos);
      blob_write_uint32(metadata, glprog->info.num_ssbos);

      for (unsigned j = 0; j < glprog->info.num_ubos; j++) {
         uint32_t offset =
            glprog->sh.UniformBlocks[j] - prog->data->UniformBlocks;
         blob_write_uint32(metadata, offset);
      }

      for (unsigned j = 0; j < glprog->info.num_ssbos; j++) {
         uint32_t offset = glprog->sh.ShaderStorageBlocks[j] -
            prog->data->ShaderStorageBlocks;
         blob_write_uint32(metadata, offset);
      }
   }
}

static void
read_buffer_block(struct blob_reader *metadata, struct gl_uniform_block *b,
                  struct gl_shader_program *prog)
{
      b->Name = ralloc_strdup(prog->data, blob_read_string (metadata));
      b->NumUniforms = blob_read_uint32(metadata);
      b->Binding = blob_read_uint32(metadata);
      b->UniformBufferSize = blob_read_uint32(metadata);
      b->stageref = blob_read_uint32(metadata);

      b->Uniforms =
         rzalloc_array(prog->data, struct gl_uniform_buffer_variable,
                       b->NumUniforms);
      for (unsigned j = 0; j < b->NumUniforms; j++) {
         b->Uniforms[j].Name = ralloc_strdup(prog->data,
                                             blob_read_string (metadata));

         char *index_name = blob_read_string(metadata);
         if (strcmp(b->Uniforms[j].Name, index_name) == 0) {
            b->Uniforms[j].IndexName = b->Uniforms[j].Name;
         } else {
            b->Uniforms[j].IndexName = ralloc_strdup(prog->data, index_name);
         }

         b->Uniforms[j].Type = decode_type_from_blob(metadata);
         b->Uniforms[j].Offset = blob_read_uint32(metadata);
      }
}

static void
read_buffer_blocks(struct blob_reader *metadata,
                   struct gl_shader_program *prog)
{
   prog->data->NumUniformBlocks = blob_read_uint32(metadata);
   prog->data->NumShaderStorageBlocks = blob_read_uint32(metadata);

   prog->data->UniformBlocks =
      rzalloc_array(prog->data, struct gl_uniform_block,
                    prog->data->NumUniformBlocks);

   prog->data->ShaderStorageBlocks =
      rzalloc_array(prog->data, struct gl_uniform_block,
                    prog->data->NumShaderStorageBlocks);

   for (unsigned i = 0; i < prog->data->NumUniformBlocks; i++) {
      read_buffer_block(metadata, &prog->data->UniformBlocks[i], prog);
   }

   for (unsigned i = 0; i < prog->data->NumShaderStorageBlocks; i++) {
      read_buffer_block(metadata, &prog->data->ShaderStorageBlocks[i], prog);
   }

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (!sh)
         continue;

      struct gl_program *glprog = sh->Program;

      glprog->info.num_ubos = blob_read_uint32(metadata);
      glprog->info.num_ssbos = blob_read_uint32(metadata);

      glprog->sh.UniformBlocks =
         rzalloc_array(glprog, gl_uniform_block *, glprog->info.num_ubos);
      glprog->sh.ShaderStorageBlocks =
         rzalloc_array(glprog, gl_uniform_block *, glprog->info.num_ssbos);

      for (unsigned j = 0; j < glprog->info.num_ubos; j++) {
         uint32_t offset = blob_read_uint32(metadata);
         glprog->sh.UniformBlocks[j] = prog->data->UniformBlocks + offset;
      }

      for (unsigned j = 0; j < glprog->info.num_ssbos; j++) {
         uint32_t offset = blob_read_uint32(metadata);
         glprog->sh.ShaderStorageBlocks[j] =
            prog->data->ShaderStorageBlocks + offset;
      }
   }
}

static void
write_atomic_buffers(struct blob *metadata, struct gl_shader_program *prog)
{
   blob_write_uint32(metadata, prog->data->NumAtomicBuffers);

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      if (prog->_LinkedShaders[i]) {
         struct gl_program *glprog = prog->_LinkedShaders[i]->Program;
         blob_write_uint32(metadata, glprog->info.num_abos);
      }
   }

   for (unsigned i = 0; i < prog->data->NumAtomicBuffers; i++) {
      blob_write_uint32(metadata, prog->data->AtomicBuffers[i].Binding);
      blob_write_uint32(metadata, prog->data->AtomicBuffers[i].MinimumSize);
      blob_write_uint32(metadata, prog->data->AtomicBuffers[i].NumUniforms);

      blob_write_bytes(metadata, prog->data->AtomicBuffers[i].StageReferences,
                       sizeof(prog->data->AtomicBuffers[i].StageReferences));

      for (unsigned j = 0; j < prog->data->AtomicBuffers[i].NumUniforms; j++) {
         blob_write_uint32(metadata, prog->data->AtomicBuffers[i].Uniforms[j]);
      }
   }
}

static void
read_atomic_buffers(struct blob_reader *metadata,
                     struct gl_shader_program *prog)
{
   prog->data->NumAtomicBuffers = blob_read_uint32(metadata);
   prog->data->AtomicBuffers =
      rzalloc_array(prog, gl_active_atomic_buffer,
                    prog->data->NumAtomicBuffers);

   struct gl_active_atomic_buffer **stage_buff_list[MESA_SHADER_STAGES];
   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      if (prog->_LinkedShaders[i]) {
         struct gl_program *glprog = prog->_LinkedShaders[i]->Program;

         glprog->info.num_abos = blob_read_uint32(metadata);
         glprog->sh.AtomicBuffers =
            rzalloc_array(glprog, gl_active_atomic_buffer *,
                          glprog->info.num_abos);
         stage_buff_list[i] = glprog->sh.AtomicBuffers;
      }
   }

   for (unsigned i = 0; i < prog->data->NumAtomicBuffers; i++) {
      prog->data->AtomicBuffers[i].Binding = blob_read_uint32(metadata);
      prog->data->AtomicBuffers[i].MinimumSize = blob_read_uint32(metadata);
      prog->data->AtomicBuffers[i].NumUniforms = blob_read_uint32(metadata);

      blob_copy_bytes(metadata,
                      (uint8_t *) &prog->data->AtomicBuffers[i].StageReferences,
                      sizeof(prog->data->AtomicBuffers[i].StageReferences));

      prog->data->AtomicBuffers[i].Uniforms = rzalloc_array(prog, unsigned,
         prog->data->AtomicBuffers[i].NumUniforms);

      for (unsigned j = 0; j < prog->data->AtomicBuffers[i].NumUniforms; j++) {
         prog->data->AtomicBuffers[i].Uniforms[j] = blob_read_uint32(metadata);
      }

      for (unsigned j = 0; j < MESA_SHADER_STAGES; j++) {
         if (prog->data->AtomicBuffers[i].StageReferences[j]) {
            *stage_buff_list[j] = &prog->data->AtomicBuffers[i];
            stage_buff_list[j]++;
         }
      }
   }
}

static void
write_xfb(struct blob *metadata, struct gl_shader_program *shProg)
{
   struct gl_program *prog = shProg->last_vert_prog;

   if (!prog) {
      blob_write_uint32(metadata, ~0u);
      return;
   }

   struct gl_transform_feedback_info *ltf = prog->sh.LinkedTransformFeedback;

   blob_write_uint32(metadata, prog->info.stage);

   blob_write_uint32(metadata, ltf->NumOutputs);
   blob_write_uint32(metadata, ltf->ActiveBuffers);
   blob_write_uint32(metadata, ltf->NumVarying);

   blob_write_bytes(metadata, ltf->Outputs,
                    sizeof(struct gl_transform_feedback_output) *
                       ltf->NumOutputs);

   for (int i = 0; i < ltf->NumVarying; i++) {
      blob_write_string(metadata, ltf->Varyings[i].Name);
      blob_write_uint32(metadata, ltf->Varyings[i].Type);
      blob_write_uint32(metadata, ltf->Varyings[i].BufferIndex);
      blob_write_uint32(metadata, ltf->Varyings[i].Size);
      blob_write_uint32(metadata, ltf->Varyings[i].Offset);
   }

   blob_write_bytes(metadata, ltf->Buffers,
                    sizeof(struct gl_transform_feedback_buffer) *
                       MAX_FEEDBACK_BUFFERS);
}

static void
read_xfb(struct blob_reader *metadata, struct gl_shader_program *shProg)
{
   unsigned xfb_stage = blob_read_uint32(metadata);

   if (xfb_stage == ~0u)
      return;

   struct gl_program *prog = shProg->_LinkedShaders[xfb_stage]->Program;
   struct gl_transform_feedback_info *ltf =
      rzalloc(prog, struct gl_transform_feedback_info);

   prog->sh.LinkedTransformFeedback = ltf;
   shProg->last_vert_prog = prog;

   ltf->NumOutputs = blob_read_uint32(metadata);
   ltf->ActiveBuffers = blob_read_uint32(metadata);
   ltf->NumVarying = blob_read_uint32(metadata);

   ltf->Outputs = rzalloc_array(prog, struct gl_transform_feedback_output,
                                ltf->NumOutputs);

   blob_copy_bytes(metadata, (uint8_t *) ltf->Outputs,
                   sizeof(struct gl_transform_feedback_output) *
                      ltf->NumOutputs);

   ltf->Varyings = rzalloc_array(prog,
                                 struct gl_transform_feedback_varying_info,
                                 ltf->NumVarying);

   for (int i = 0; i < ltf->NumVarying; i++) {
      ltf->Varyings[i].Name = ralloc_strdup(prog, blob_read_string(metadata));
      ltf->Varyings[i].Type = blob_read_uint32(metadata);
      ltf->Varyings[i].BufferIndex = blob_read_uint32(metadata);
      ltf->Varyings[i].Size = blob_read_uint32(metadata);
      ltf->Varyings[i].Offset = blob_read_uint32(metadata);
   }

   blob_copy_bytes(metadata, (uint8_t *) ltf->Buffers,
                   sizeof(struct gl_transform_feedback_buffer) *
                      MAX_FEEDBACK_BUFFERS);
}

static void
write_uniforms(struct blob *metadata, struct gl_shader_program *prog)
{
   blob_write_uint32(metadata, prog->SamplersValidated);
   blob_write_uint32(metadata, prog->data->NumUniformStorage);
   blob_write_uint32(metadata, prog->data->NumUniformDataSlots);

   for (unsigned i = 0; i < prog->data->NumUniformStorage; i++) {
      encode_type_to_blob(metadata, prog->data->UniformStorage[i].type);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].array_elements);
      blob_write_string(metadata, prog->data->UniformStorage[i].name);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].storage -
                                  prog->data->UniformDataSlots);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].builtin);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].remap_location);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].block_index);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].atomic_buffer_index);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].offset);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].array_stride);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].hidden);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].is_shader_storage);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].active_shader_mask);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].matrix_stride);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].row_major);
      blob_write_uint32(metadata, prog->data->UniformStorage[i].is_bindless);
      blob_write_uint32(metadata,
                        prog->data->UniformStorage[i].num_compatible_subroutines);
      blob_write_uint32(metadata,
                        prog->data->UniformStorage[i].top_level_array_size);
      blob_write_uint32(metadata,
                        prog->data->UniformStorage[i].top_level_array_stride);
      blob_write_bytes(metadata, prog->data->UniformStorage[i].opaque,
                       sizeof(prog->data->UniformStorage[i].opaque));
   }

   /* Here we cache all uniform values. We do this to retain values for
    * uniforms with initialisers and also hidden uniforms that may be lowered
    * constant arrays. We could possibly just store the values we need but for
    * now we just store everything.
    */
   blob_write_uint32(metadata, prog->data->NumHiddenUniforms);
   for (unsigned i = 0; i < prog->data->NumUniformStorage; i++) {
      if (!prog->data->UniformStorage[i].builtin &&
          !prog->data->UniformStorage[i].is_shader_storage &&
          prog->data->UniformStorage[i].block_index == -1) {
         unsigned vec_size =
            prog->data->UniformStorage[i].type->component_slots() *
            MAX2(prog->data->UniformStorage[i].array_elements, 1);
         blob_write_bytes(metadata, prog->data->UniformStorage[i].storage,
                          sizeof(union gl_constant_value) * vec_size);
      }
   }
}

static void
read_uniforms(struct blob_reader *metadata, struct gl_shader_program *prog)
{
   struct gl_uniform_storage *uniforms;
   union gl_constant_value *data;

   prog->SamplersValidated = blob_read_uint32(metadata);
   prog->data->NumUniformStorage = blob_read_uint32(metadata);
   prog->data->NumUniformDataSlots = blob_read_uint32(metadata);

   uniforms = rzalloc_array(prog, struct gl_uniform_storage,
                            prog->data->NumUniformStorage);
   prog->data->UniformStorage = uniforms;

   data = rzalloc_array(uniforms, union gl_constant_value,
                        prog->data->NumUniformDataSlots);
   prog->data->UniformDataSlots = data;

   prog->UniformHash = new string_to_uint_map;

   for (unsigned i = 0; i < prog->data->NumUniformStorage; i++) {
      uniforms[i].type = decode_type_from_blob(metadata);
      uniforms[i].array_elements = blob_read_uint32(metadata);
      uniforms[i].name = ralloc_strdup(prog, blob_read_string (metadata));
      uniforms[i].storage = data + blob_read_uint32(metadata);
      uniforms[i].builtin = blob_read_uint32(metadata);
      uniforms[i].remap_location = blob_read_uint32(metadata);
      uniforms[i].block_index = blob_read_uint32(metadata);
      uniforms[i].atomic_buffer_index = blob_read_uint32(metadata);
      uniforms[i].offset = blob_read_uint32(metadata);
      uniforms[i].array_stride = blob_read_uint32(metadata);
      uniforms[i].hidden = blob_read_uint32(metadata);
      uniforms[i].is_shader_storage = blob_read_uint32(metadata);
      uniforms[i].active_shader_mask = blob_read_uint32(metadata);
      uniforms[i].matrix_stride = blob_read_uint32(metadata);
      uniforms[i].row_major = blob_read_uint32(metadata);
      uniforms[i].is_bindless = blob_read_uint32(metadata);
      uniforms[i].num_compatible_subroutines = blob_read_uint32(metadata);
      uniforms[i].top_level_array_size = blob_read_uint32(metadata);
      uniforms[i].top_level_array_stride = blob_read_uint32(metadata);
      prog->UniformHash->put(i, uniforms[i].name);

      memcpy(uniforms[i].opaque,
             blob_read_bytes(metadata, sizeof(uniforms[i].opaque)),
             sizeof(uniforms[i].opaque));
   }

   /* Restore uniform values. */
   prog->data->NumHiddenUniforms = blob_read_uint32(metadata);
   for (unsigned i = 0; i < prog->data->NumUniformStorage; i++) {
      if (!prog->data->UniformStorage[i].builtin &&
          !prog->data->UniformStorage[i].is_shader_storage &&
          prog->data->UniformStorage[i].block_index == -1) {
         unsigned vec_size =
            prog->data->UniformStorage[i].type->component_slots() *
            MAX2(prog->data->UniformStorage[i].array_elements, 1);
         blob_copy_bytes(metadata,
                         (uint8_t *) prog->data->UniformStorage[i].storage,
                         sizeof(union gl_constant_value) * vec_size);

        assert(vec_size + prog->data->UniformStorage[i].storage <=
               data +  prog->data->NumUniformDataSlots);
      }
   }
}

enum uniform_remap_type
{
   remap_type_inactive_explicit_location,
   remap_type_null_ptr,
   remap_type_uniform_offset
};

static void
write_uniform_remap_table_entry(struct blob *metadata,
                                gl_uniform_storage *uniform_storage,
                                gl_uniform_storage *entry)
{
   if (entry == INACTIVE_UNIFORM_EXPLICIT_LOCATION) {
      blob_write_uint32(metadata, remap_type_inactive_explicit_location);
   } else if (entry == NULL) {
      blob_write_uint32(metadata, remap_type_null_ptr);
   } else {
      blob_write_uint32(metadata, remap_type_uniform_offset);

      uint32_t offset = entry - uniform_storage;
      blob_write_uint32(metadata, offset);
   }
}

static void
write_uniform_remap_tables(struct blob *metadata,
                           struct gl_shader_program *prog)
{
   blob_write_uint32(metadata, prog->NumUniformRemapTable);

   for (unsigned i = 0; i < prog->NumUniformRemapTable; i++) {
      write_uniform_remap_table_entry(metadata, prog->data->UniformStorage,
                                      prog->UniformRemapTable[i]);
   }

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (sh) {
         struct gl_program *glprog = sh->Program;
         blob_write_uint32(metadata, glprog->sh.NumSubroutineUniformRemapTable);

         for (unsigned j = 0; j < glprog->sh.NumSubroutineUniformRemapTable; j++) {
            write_uniform_remap_table_entry(metadata,
                                            prog->data->UniformStorage,
                                            glprog->sh.SubroutineUniformRemapTable[j]);
         }
      }
   }
}

static void
read_uniform_remap_table_entry(struct blob_reader *metadata,
                               gl_uniform_storage *uniform_storage,
                               gl_uniform_storage **entry,
                               enum uniform_remap_type type)
{
   if (type == remap_type_inactive_explicit_location) {
      *entry = INACTIVE_UNIFORM_EXPLICIT_LOCATION;
   } else if (type == remap_type_null_ptr) {
      *entry = NULL;
   } else {
      uint32_t uni_offset = blob_read_uint32(metadata);
      *entry = uniform_storage + uni_offset;
   }
}

static void
read_uniform_remap_tables(struct blob_reader *metadata,
                          struct gl_shader_program *prog)
{
   prog->NumUniformRemapTable = blob_read_uint32(metadata);

   prog->UniformRemapTable = rzalloc_array(prog, struct gl_uniform_storage *,
                                           prog->NumUniformRemapTable);

   for (unsigned i = 0; i < prog->NumUniformRemapTable; i++) {
      enum uniform_remap_type type =
         (enum uniform_remap_type) blob_read_uint32(metadata);

      read_uniform_remap_table_entry(metadata, prog->data->UniformStorage,
                                     &prog->UniformRemapTable[i], type);
   }

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (sh) {
         struct gl_program *glprog = sh->Program;
         glprog->sh.NumSubroutineUniformRemapTable = blob_read_uint32(metadata);

         glprog->sh.SubroutineUniformRemapTable =
            rzalloc_array(glprog, struct gl_uniform_storage *,
                          glprog->sh.NumSubroutineUniformRemapTable);

         for (unsigned j = 0; j < glprog->sh.NumSubroutineUniformRemapTable; j++) {
            enum uniform_remap_type type =
               (enum uniform_remap_type) blob_read_uint32(metadata);

            read_uniform_remap_table_entry(metadata,
                                           prog->data->UniformStorage,
                                           &glprog->sh.SubroutineUniformRemapTable[j],
                                           type);
         }
      }
   }
}

struct whte_closure
{
   struct blob *blob;
   size_t num_entries;
};

static void
write_hash_table_entry(const char *key, unsigned value, void *closure)
{
   struct whte_closure *whte = (struct whte_closure *) closure;

   blob_write_string(whte->blob, key);
   blob_write_uint32(whte->blob, value);

   whte->num_entries++;
}

static void
write_hash_table(struct blob *metadata, struct string_to_uint_map *hash)
{
   size_t offset;
   struct whte_closure whte;

   whte.blob = metadata;
   whte.num_entries = 0;

   offset = metadata->size;

   /* Write a placeholder for the hashtable size. */
   blob_write_uint32 (metadata, 0);

   hash->iterate(write_hash_table_entry, &whte);

   /* Overwrite with the computed number of entries written. */
   blob_overwrite_uint32 (metadata, offset, whte.num_entries);
}

static void
read_hash_table(struct blob_reader *metadata, struct string_to_uint_map *hash)
{
   size_t i, num_entries;
   const char *key;
   uint32_t value;

   num_entries = blob_read_uint32 (metadata);

   for (i = 0; i < num_entries; i++) {
      key = blob_read_string(metadata);
      value = blob_read_uint32(metadata);

      hash->put(value, key);
   }
}

static void
write_hash_tables(struct blob *metadata, struct gl_shader_program *prog)
{
   write_hash_table(metadata, prog->AttributeBindings);
   write_hash_table(metadata, prog->FragDataBindings);
   write_hash_table(metadata, prog->FragDataIndexBindings);
}

static void
read_hash_tables(struct blob_reader *metadata, struct gl_shader_program *prog)
{
   read_hash_table(metadata, prog->AttributeBindings);
   read_hash_table(metadata, prog->FragDataBindings);
   read_hash_table(metadata, prog->FragDataIndexBindings);
}

static void
write_shader_subroutine_index(struct blob *metadata,
                              struct gl_linked_shader *sh,
                              struct gl_program_resource *res)
{
   assert(sh);

   for (unsigned j = 0; j < sh->Program->sh.NumSubroutineFunctions; j++) {
      if (strcmp(((gl_subroutine_function *)res->Data)->name,
                 sh->Program->sh.SubroutineFunctions[j].name) == 0) {
         blob_write_uint32(metadata, j);
         break;
      }
   }
}

static void
get_shader_var_and_pointer_sizes(size_t *s_var_size, size_t *s_var_ptrs,
                                 const gl_shader_variable *var)
{
   *s_var_size = sizeof(gl_shader_variable);
   *s_var_ptrs =
      sizeof(var->type) +
      sizeof(var->interface_type) +
      sizeof(var->outermost_struct_type) +
      sizeof(var->name);
}

static void
write_program_resource_data(struct blob *metadata,
                            struct gl_shader_program *prog,
                            struct gl_program_resource *res)
{
   struct gl_linked_shader *sh;

   switch(res->Type) {
   case GL_PROGRAM_INPUT:
   case GL_PROGRAM_OUTPUT: {
      const gl_shader_variable *var = (gl_shader_variable *)res->Data;

      encode_type_to_blob(metadata, var->type);
      encode_type_to_blob(metadata, var->interface_type);
      encode_type_to_blob(metadata, var->outermost_struct_type);

      blob_write_string(metadata, var->name);

      size_t s_var_size, s_var_ptrs;
      get_shader_var_and_pointer_sizes(&s_var_size, &s_var_ptrs, var);

      /* Write gl_shader_variable skipping over the pointers */
      blob_write_bytes(metadata, ((char *)var) + s_var_ptrs,
                       s_var_size - s_var_ptrs);
      break;
   }
   case GL_UNIFORM_BLOCK:
      for (unsigned i = 0; i < prog->data->NumUniformBlocks; i++) {
         if (strcmp(((gl_uniform_block *)res->Data)->Name,
                    prog->data->UniformBlocks[i].Name) == 0) {
            blob_write_uint32(metadata, i);
            break;
         }
      }
      break;
   case GL_SHADER_STORAGE_BLOCK:
      for (unsigned i = 0; i < prog->data->NumShaderStorageBlocks; i++) {
         if (strcmp(((gl_uniform_block *)res->Data)->Name,
                    prog->data->ShaderStorageBlocks[i].Name) == 0) {
            blob_write_uint32(metadata, i);
            break;
         }
      }
      break;
   case GL_BUFFER_VARIABLE:
   case GL_VERTEX_SUBROUTINE_UNIFORM:
   case GL_GEOMETRY_SUBROUTINE_UNIFORM:
   case GL_FRAGMENT_SUBROUTINE_UNIFORM:
   case GL_COMPUTE_SUBROUTINE_UNIFORM:
   case GL_TESS_CONTROL_SUBROUTINE_UNIFORM:
   case GL_TESS_EVALUATION_SUBROUTINE_UNIFORM:
   case GL_UNIFORM:
      for (unsigned i = 0; i < prog->data->NumUniformStorage; i++) {
         if (strcmp(((gl_uniform_storage *)res->Data)->name,
                    prog->data->UniformStorage[i].name) == 0) {
            blob_write_uint32(metadata, i);
            break;
         }
      }
      break;
   case GL_ATOMIC_COUNTER_BUFFER:
      for (unsigned i = 0; i < prog->data->NumAtomicBuffers; i++) {
         if (((gl_active_atomic_buffer *)res->Data)->Binding ==
             prog->data->AtomicBuffers[i].Binding) {
            blob_write_uint32(metadata, i);
            break;
         }
      }
      break;
   case GL_TRANSFORM_FEEDBACK_BUFFER:
      for (unsigned i = 0; i < MAX_FEEDBACK_BUFFERS; i++) {
         if (((gl_transform_feedback_buffer *)res->Data)->Binding ==
             prog->last_vert_prog->sh.LinkedTransformFeedback->Buffers[i].Binding) {
            blob_write_uint32(metadata, i);
            break;
         }
      }
      break;
   case GL_TRANSFORM_FEEDBACK_VARYING:
      for (int i = 0; i < prog->last_vert_prog->sh.LinkedTransformFeedback->NumVarying; i++) {
         if (strcmp(((gl_transform_feedback_varying_info *)res->Data)->Name,
                    prog->last_vert_prog->sh.LinkedTransformFeedback->Varyings[i].Name) == 0) {
            blob_write_uint32(metadata, i);
            break;
         }
      }
      break;
   case GL_VERTEX_SUBROUTINE:
   case GL_TESS_CONTROL_SUBROUTINE:
   case GL_TESS_EVALUATION_SUBROUTINE:
   case GL_GEOMETRY_SUBROUTINE:
   case GL_FRAGMENT_SUBROUTINE:
   case GL_COMPUTE_SUBROUTINE:
      sh =
         prog->_LinkedShaders[_mesa_shader_stage_from_subroutine(res->Type)];
      write_shader_subroutine_index(metadata, sh, res);
      break;
   default:
      assert(!"Support for writing resource not yet implemented.");
   }
}

static void
read_program_resource_data(struct blob_reader *metadata,
                           struct gl_shader_program *prog,
                           struct gl_program_resource *res)
{
   struct gl_linked_shader *sh;

   switch(res->Type) {
   case GL_PROGRAM_INPUT:
   case GL_PROGRAM_OUTPUT: {
      gl_shader_variable *var = ralloc(prog, struct gl_shader_variable);

      var->type = decode_type_from_blob(metadata);
      var->interface_type = decode_type_from_blob(metadata);
      var->outermost_struct_type = decode_type_from_blob(metadata);

      var->name = ralloc_strdup(prog, blob_read_string(metadata));

      size_t s_var_size, s_var_ptrs;
      get_shader_var_and_pointer_sizes(&s_var_size, &s_var_ptrs, var);

      blob_copy_bytes(metadata, ((uint8_t *) var) + s_var_ptrs,
                      s_var_size - s_var_ptrs);

      res->Data = var;
      break;
   }
   case GL_UNIFORM_BLOCK:
      res->Data = &prog->data->UniformBlocks[blob_read_uint32(metadata)];
      break;
   case GL_SHADER_STORAGE_BLOCK:
      res->Data = &prog->data->ShaderStorageBlocks[blob_read_uint32(metadata)];
      break;
   case GL_BUFFER_VARIABLE:
   case GL_VERTEX_SUBROUTINE_UNIFORM:
   case GL_GEOMETRY_SUBROUTINE_UNIFORM:
   case GL_FRAGMENT_SUBROUTINE_UNIFORM:
   case GL_COMPUTE_SUBROUTINE_UNIFORM:
   case GL_TESS_CONTROL_SUBROUTINE_UNIFORM:
   case GL_TESS_EVALUATION_SUBROUTINE_UNIFORM:
   case GL_UNIFORM:
      res->Data = &prog->data->UniformStorage[blob_read_uint32(metadata)];
      break;
   case GL_ATOMIC_COUNTER_BUFFER:
      res->Data = &prog->data->AtomicBuffers[blob_read_uint32(metadata)];
      break;
   case GL_TRANSFORM_FEEDBACK_BUFFER:
      res->Data = &prog->last_vert_prog->
         sh.LinkedTransformFeedback->Buffers[blob_read_uint32(metadata)];
      break;
   case GL_TRANSFORM_FEEDBACK_VARYING:
      res->Data = &prog->last_vert_prog->
         sh.LinkedTransformFeedback->Varyings[blob_read_uint32(metadata)];
      break;
   case GL_VERTEX_SUBROUTINE:
   case GL_TESS_CONTROL_SUBROUTINE:
   case GL_TESS_EVALUATION_SUBROUTINE:
   case GL_GEOMETRY_SUBROUTINE:
   case GL_FRAGMENT_SUBROUTINE:
   case GL_COMPUTE_SUBROUTINE:
      sh =
         prog->_LinkedShaders[_mesa_shader_stage_from_subroutine(res->Type)];
      res->Data =
         &sh->Program->sh.SubroutineFunctions[blob_read_uint32(metadata)];
      break;
   default:
      assert(!"Support for reading resource not yet implemented.");
   }
}

static void
write_program_resource_list(struct blob *metadata,
                            struct gl_shader_program *prog)
{
   blob_write_uint32(metadata, prog->data->NumProgramResourceList);

   for (unsigned i = 0; i < prog->data->NumProgramResourceList; i++) {
      blob_write_uint32(metadata, prog->data->ProgramResourceList[i].Type);
      write_program_resource_data(metadata, prog,
                                  &prog->data->ProgramResourceList[i]);
      blob_write_bytes(metadata,
                       &prog->data->ProgramResourceList[i].StageReferences,
                       sizeof(prog->data->ProgramResourceList[i].StageReferences));
   }
}

static void
read_program_resource_list(struct blob_reader *metadata,
                           struct gl_shader_program *prog)
{
   prog->data->NumProgramResourceList = blob_read_uint32(metadata);

   prog->data->ProgramResourceList =
      ralloc_array(prog, gl_program_resource,
                   prog->data->NumProgramResourceList);

   for (unsigned i = 0; i < prog->data->NumProgramResourceList; i++) {
      prog->data->ProgramResourceList[i].Type = blob_read_uint32(metadata);
      read_program_resource_data(metadata, prog,
                                 &prog->data->ProgramResourceList[i]);
      blob_copy_bytes(metadata,
                      (uint8_t *) &prog->data->ProgramResourceList[i].StageReferences,
                      sizeof(prog->data->ProgramResourceList[i].StageReferences));
   }
}

static void
write_shader_parameters(struct blob *metadata,
                        struct gl_program_parameter_list *params)
{
   blob_write_uint32(metadata, params->NumParameters);
   uint32_t i = 0;

   while (i < params->NumParameters) {
      struct gl_program_parameter *param = &params->Parameters[i];

      blob_write_uint32(metadata, param->Type);
      blob_write_string(metadata, param->Name);
      blob_write_uint32(metadata, param->Size);
      blob_write_uint32(metadata, param->DataType);
      blob_write_bytes(metadata, param->StateIndexes,
                       sizeof(param->StateIndexes));

      i += (param->Size + 3) / 4;
   }

   blob_write_bytes(metadata, params->ParameterValues,
                    sizeof(gl_constant_value) * 4 * params->NumParameters);

   blob_write_uint32(metadata, params->StateFlags);
}

static void
read_shader_parameters(struct blob_reader *metadata,
                       struct gl_program_parameter_list *params)
{
   gl_state_index state_indexes[STATE_LENGTH];
   uint32_t i = 0;
   uint32_t num_parameters = blob_read_uint32(metadata);

   _mesa_reserve_parameter_storage(params, num_parameters);
   while (i < num_parameters) {
      gl_register_file type = (gl_register_file) blob_read_uint32(metadata);
      const char *name = blob_read_string(metadata);
      unsigned size = blob_read_uint32(metadata);
      unsigned data_type = blob_read_uint32(metadata);
      blob_copy_bytes(metadata, (uint8_t *) state_indexes,
                      sizeof(state_indexes));

      _mesa_add_parameter(params, type, name, size, data_type,
                          NULL, state_indexes);

      i += (size + 3) / 4;
   }

   blob_copy_bytes(metadata, (uint8_t *) params->ParameterValues,
                    sizeof(gl_constant_value) * 4 * params->NumParameters);

   params->StateFlags = blob_read_uint32(metadata);
}

static void
write_shader_metadata(struct blob *metadata, gl_linked_shader *shader)
{
   assert(shader->Program);
   struct gl_program *glprog = shader->Program;
   unsigned i;

   blob_write_bytes(metadata, glprog->TexturesUsed,
                    sizeof(glprog->TexturesUsed));
   blob_write_uint64(metadata, glprog->SamplersUsed);

   blob_write_bytes(metadata, glprog->SamplerUnits,
                    sizeof(glprog->SamplerUnits));
   blob_write_bytes(metadata, glprog->sh.SamplerTargets,
                    sizeof(glprog->sh.SamplerTargets));
   blob_write_uint32(metadata, glprog->ShadowSamplers);

   blob_write_bytes(metadata, glprog->sh.ImageAccess,
                    sizeof(glprog->sh.ImageAccess));
   blob_write_bytes(metadata, glprog->sh.ImageUnits,
                    sizeof(glprog->sh.ImageUnits));

   blob_write_uint32(metadata, glprog->sh.NumBindlessSamplers);
   blob_write_uint32(metadata, glprog->sh.HasBoundBindlessSampler);
   for (i = 0; i < glprog->sh.NumBindlessSamplers; i++) {
      blob_write_bytes(metadata, &glprog->sh.BindlessSamplers[i],
                       sizeof(struct gl_bindless_sampler));
   }

   blob_write_uint32(metadata, glprog->sh.NumBindlessImages);
   blob_write_uint32(metadata, glprog->sh.HasBoundBindlessImage);
   for (i = 0; i < glprog->sh.NumBindlessImages; i++) {
      blob_write_bytes(metadata, &glprog->sh.BindlessImages[i],
                       sizeof(struct gl_bindless_image));
   }

   write_shader_parameters(metadata, glprog->Parameters);
}

static void
read_shader_metadata(struct blob_reader *metadata,
                     struct gl_program *glprog,
                     gl_linked_shader *linked)
{
   unsigned i;

   blob_copy_bytes(metadata, (uint8_t *) glprog->TexturesUsed,
                   sizeof(glprog->TexturesUsed));
   glprog->SamplersUsed = blob_read_uint64(metadata);

   blob_copy_bytes(metadata, (uint8_t *) glprog->SamplerUnits,
                   sizeof(glprog->SamplerUnits));
   blob_copy_bytes(metadata, (uint8_t *) glprog->sh.SamplerTargets,
                   sizeof(glprog->sh.SamplerTargets));
   glprog->ShadowSamplers = blob_read_uint32(metadata);

   blob_copy_bytes(metadata, (uint8_t *) glprog->sh.ImageAccess,
                   sizeof(glprog->sh.ImageAccess));
   blob_copy_bytes(metadata, (uint8_t *) glprog->sh.ImageUnits,
                   sizeof(glprog->sh.ImageUnits));

   glprog->sh.NumBindlessSamplers = blob_read_uint32(metadata);
   glprog->sh.HasBoundBindlessSampler = blob_read_uint32(metadata);
   if (glprog->sh.NumBindlessSamplers > 0) {
      glprog->sh.BindlessSamplers =
         rzalloc_array(glprog, gl_bindless_sampler,
                       glprog->sh.NumBindlessSamplers);

      for (i = 0; i < glprog->sh.NumBindlessSamplers; i++) {
         blob_copy_bytes(metadata, (uint8_t *) &glprog->sh.BindlessSamplers[i],
                         sizeof(struct gl_bindless_sampler));
      }
   }

   glprog->sh.NumBindlessImages = blob_read_uint32(metadata);
   glprog->sh.HasBoundBindlessImage = blob_read_uint32(metadata);
   if (glprog->sh.NumBindlessImages > 0) {
      glprog->sh.BindlessImages =
         rzalloc_array(glprog, gl_bindless_image,
                       glprog->sh.NumBindlessImages);

      for (i = 0; i < glprog->sh.NumBindlessImages; i++) {
         blob_copy_bytes(metadata, (uint8_t *) &glprog->sh.BindlessImages[i],
                        sizeof(struct gl_bindless_image));
      }
   }

   glprog->Parameters = _mesa_new_parameter_list();
   read_shader_parameters(metadata, glprog->Parameters);
}

static void
create_binding_str(const char *key, unsigned value, void *closure)
{
   char **bindings_str = (char **) closure;
   ralloc_asprintf_append(bindings_str, "%s:%u,", key, value);
}

static void
create_linked_shader_and_program(struct gl_context *ctx,
                                 gl_shader_stage stage,
                                 struct gl_shader_program *prog,
                                 struct blob_reader *metadata)
{
   struct gl_program *glprog;

   struct gl_linked_shader *linked = rzalloc(NULL, struct gl_linked_shader);
   linked->Stage = stage;

   glprog = ctx->Driver.NewProgram(ctx, _mesa_shader_stage_to_program(stage),
                                   prog->Name, false);
   glprog->info.stage = stage;
   linked->Program = glprog;

   read_shader_metadata(metadata, glprog, linked);

   /* Restore shader info */
   blob_copy_bytes(metadata, (uint8_t *) &glprog->info, sizeof(shader_info));
   if (glprog->info.name)
      glprog->info.name = ralloc_strdup(glprog, blob_read_string(metadata));
   if (glprog->info.label)
      glprog->info.label = ralloc_strdup(glprog, blob_read_string(metadata));

   _mesa_reference_shader_program_data(ctx, &glprog->sh.data, prog->data);
   _mesa_reference_program(ctx, &linked->Program, glprog);
   prog->_LinkedShaders[stage] = linked;
}

void
shader_cache_write_program_metadata(struct gl_context *ctx,
                                    struct gl_shader_program *prog)
{
   struct disk_cache *cache = ctx->Cache;
   if (!cache)
      return;

   /* Exit early when we are dealing with a ff shader with no source file to
    * generate a source from.
    *
    * TODO: In future we should use another method to generate a key for ff
    * programs.
    */
   static const char zero[sizeof(prog->data->sha1)] = {0};
   if (memcmp(prog->data->sha1, zero, sizeof(prog->data->sha1)) == 0)
      return;

   struct blob *metadata = blob_create();

   write_uniforms(metadata, prog);

   write_hash_tables(metadata, prog);

   blob_write_uint32(metadata, prog->data->Version);
   blob_write_uint32(metadata, prog->data->linked_stages);

   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
      struct gl_linked_shader *sh = prog->_LinkedShaders[i];
      if (sh) {
         write_shader_metadata(metadata, sh);

         /* Store nir shader info */
         blob_write_bytes(metadata, &sh->Program->info, sizeof(shader_info));

         if (sh->Program->info.name)
            blob_write_string(metadata, sh->Program->info.name);

         if (sh->Program->info.label)
            blob_write_string(metadata, sh->Program->info.label);
      }
   }

   write_xfb(metadata, prog);

   write_uniform_remap_tables(metadata, prog);

   write_atomic_buffers(metadata, prog);

   write_buffer_blocks(metadata, prog);

   write_subroutines(metadata, prog);

   write_program_resource_list(metadata, prog);

   char sha1_buf[41];
   for (unsigned i = 0; i < prog->NumShaders; i++) {
      disk_cache_put_key(cache, prog->Shaders[i]->sha1);
      if (ctx->_Shader->Flags & GLSL_CACHE_INFO) {
         _mesa_sha1_format(sha1_buf, prog->Shaders[i]->sha1);
         fprintf(stderr, "marking shader: %s\n", sha1_buf);
      }
   }

   disk_cache_put(cache, prog->data->sha1, metadata->data, metadata->size);

   blob_destroy(metadata);

   if (ctx->_Shader->Flags & GLSL_CACHE_INFO) {
      _mesa_sha1_format(sha1_buf, prog->data->sha1);
      fprintf(stderr, "putting program metadata in cache: %s\n", sha1_buf);
   }
}

bool
shader_cache_read_program_metadata(struct gl_context *ctx,
                                   struct gl_shader_program *prog)
{
   /* Fixed function programs generated by Mesa are not cached. So don't
    * try to read metadata for them from the cache.
    */
   if (prog->Name == 0)
      return false;

   struct disk_cache *cache = ctx->Cache;
   if (!cache || prog->data->cache_fallback || prog->data->skip_cache)
      return false;

   /* Include bindings when creating sha1. These bindings change the resulting
    * binary so they are just as important as the shader source.
    */
   char *buf = ralloc_strdup(NULL, "vb: ");
   prog->AttributeBindings->iterate(create_binding_str, &buf);
   ralloc_strcat(&buf, "fb: ");
   prog->FragDataBindings->iterate(create_binding_str, &buf);
   ralloc_strcat(&buf, "fbi: ");
   prog->FragDataIndexBindings->iterate(create_binding_str, &buf);

   /* SSO has an effect on the linked program so include this when generating
    * the sha also.
    */
   ralloc_asprintf_append(&buf, "sso: %s\n",
                          prog->SeparateShader ? "T" : "F");

   /* A shader might end up producing different output depending on the glsl
    * version supported by the compiler. For example a different path might be
    * taken by the preprocessor, so add the version to the hash input.
    */
   ralloc_asprintf_append(&buf, "api: %d glsl: %d fglsl: %d\n",
                          ctx->API, ctx->Const.GLSLVersion,
                          ctx->Const.ForceGLSLVersion);

   /* We run the preprocessor on shaders after hashing them, so we need to
    * add any extension override vars to the hash. If we don't do this the
    * preprocessor could result in different output and we could load the
    * wrong shader.
    */
   char *ext_override = getenv("MESA_EXTENSION_OVERRIDE");
   if (ext_override) {
      ralloc_asprintf_append(&buf, "ext:%s", ext_override);
   }

   /* DRI config options may also change the output from the compiler so
    * include them as an input to sha1 creation.
    */
   char sha1buf[41];
   _mesa_sha1_format(sha1buf, ctx->Const.dri_config_options_sha1);
   ralloc_strcat(&buf, sha1buf);

   for (unsigned i = 0; i < prog->NumShaders; i++) {
      struct gl_shader *sh = prog->Shaders[i];
      _mesa_sha1_format(sha1buf, sh->sha1);
      ralloc_asprintf_append(&buf, "%s: %s\n",
                             _mesa_shader_stage_to_abbrev(sh->Stage), sha1buf);
   }
   disk_cache_compute_key(cache, buf, strlen(buf), prog->data->sha1);
   ralloc_free(buf);

   size_t size;
   uint8_t *buffer = (uint8_t *) disk_cache_get(cache, prog->data->sha1,
                                                &size);
   if (buffer == NULL) {
      /* Cached program not found. We may have seen the individual shaders
       * before and skipped compiling but they may not have been used together
       * in this combination before. Fall back to linking shaders but first
       * re-compile the shaders.
       *
       * We could probably only compile the shaders which were skipped here
       * but we need to be careful because the source may also have been
       * changed since the last compile so for now we just recompile
       * everything.
       */
      compile_shaders(ctx, prog);
      return false;
   }

   if (ctx->_Shader->Flags & GLSL_CACHE_INFO) {
      _mesa_sha1_format(sha1buf, prog->data->sha1);
      fprintf(stderr, "loading shader program meta data from cache: %s\n",
              sha1buf);
   }

   struct blob_reader metadata;
   blob_reader_init(&metadata, buffer, size);

   assert(prog->data->UniformStorage == NULL);

   read_uniforms(&metadata, prog);

   read_hash_tables(&metadata, prog);

   prog->data->Version = blob_read_uint32(&metadata);
   prog->data->linked_stages = blob_read_uint32(&metadata);

   unsigned mask = prog->data->linked_stages;
   while (mask) {
      const int j = u_bit_scan(&mask);
      create_linked_shader_and_program(ctx, (gl_shader_stage) j, prog,
                                       &metadata);
   }

   read_xfb(&metadata, prog);

   read_uniform_remap_tables(&metadata, prog);

   read_atomic_buffers(&metadata, prog);

   read_buffer_blocks(&metadata, prog);

   read_subroutines(&metadata, prog);

   read_program_resource_list(&metadata, prog);

   if (metadata.current != metadata.end || metadata.overrun) {
      /* Something has gone wrong discard the item from the cache and rebuild
       * from source.
       */
      assert(!"Invalid GLSL shader disk cache item!");

      if (ctx->_Shader->Flags & GLSL_CACHE_INFO) {
         fprintf(stderr, "Error reading program from cache (invalid GLSL "
                 "cache item)\n");
      }

      disk_cache_remove(cache, prog->data->sha1);
      compile_shaders(ctx, prog);
      free(buffer);
      return false;
   }

   /* This is used to flag a shader retrieved from cache */
   prog->data->LinkStatus = linking_skipped;

   /* Since the program load was successful, CompileStatus of all shaders at
    * this point should normally be compile_skipped. However because of how
    * the eviction works, it may happen that some of the individual shader keys
    * have been evicted, resulting in unnecessary recompiles on this load, so
    * mark them again to skip such recompiles next time.
    */
   char sha1_buf[41];
   for (unsigned i = 0; i < prog->NumShaders; i++) {
      if (prog->Shaders[i]->CompileStatus == compiled_no_opts) {
         disk_cache_put_key(cache, prog->Shaders[i]->sha1);
         if (ctx->_Shader->Flags & GLSL_CACHE_INFO) {
            _mesa_sha1_format(sha1_buf, prog->Shaders[i]->sha1);
            fprintf(stderr, "re-marking shader: %s\n", sha1_buf);
         }
      }
   }

   free (buffer);

   return true;
}