freedreno/a6xx/fd6_texture.c

/*
 * Copyright (C) 2016 Rob Clark <robclark@freedesktop.org>
 * Copyright © 2018 Google, Inc.
 *
 * 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.
 *
 * Authors:
 *    Rob Clark <robclark@freedesktop.org>
 */

#include "pipe/p_state.h"
#include "util/format/u_format.h"
#include "util/hash_table.h"
#include "util/u_inlines.h"
#include "util/u_memory.h"
#include "util/u_string.h"

#include "fd6_emit.h"
#include "fd6_format.h"
#include "fd6_resource.h"
#include "fd6_texture.h"

static void
remove_tex_entry(struct fd6_context *fd6_ctx, struct hash_entry *entry)
{
   struct fd6_texture_state *tex = entry->data;
   _mesa_hash_table_remove(fd6_ctx->tex_cache, entry);
   fd6_texture_state_reference(&tex, NULL);
}

static enum a6xx_tex_clamp
tex_clamp(unsigned wrap, bool *needs_border)
{
   switch (wrap) {
   case PIPE_TEX_WRAP_REPEAT:
      return A6XX_TEX_REPEAT;
   case PIPE_TEX_WRAP_CLAMP_TO_EDGE:
      return A6XX_TEX_CLAMP_TO_EDGE;
   case PIPE_TEX_WRAP_CLAMP_TO_BORDER:
      *needs_border = true;
      return A6XX_TEX_CLAMP_TO_BORDER;
   case PIPE_TEX_WRAP_MIRROR_CLAMP_TO_EDGE:
      /* only works for PoT.. need to emulate otherwise! */
      return A6XX_TEX_MIRROR_CLAMP;
   case PIPE_TEX_WRAP_MIRROR_REPEAT:
      return A6XX_TEX_MIRROR_REPEAT;
   case PIPE_TEX_WRAP_MIRROR_CLAMP:
   case PIPE_TEX_WRAP_MIRROR_CLAMP_TO_BORDER:
      /* these two we could perhaps emulate, but we currently
       * just don't advertise PIPE_CAP_TEXTURE_MIRROR_CLAMP
       */
   default:
      DBG("invalid wrap: %u", wrap);
      return 0;
   }
}

static enum a6xx_tex_filter
tex_filter(unsigned filter, bool aniso)
{
   switch (filter) {
   case PIPE_TEX_FILTER_NEAREST:
      return A6XX_TEX_NEAREST;
   case PIPE_TEX_FILTER_LINEAR:
      return aniso ? A6XX_TEX_ANISO : A6XX_TEX_LINEAR;
   default:
      DBG("invalid filter: %u", filter);
      return 0;
   }
}

static void *
fd6_sampler_state_create(struct pipe_context *pctx,
                         const struct pipe_sampler_state *cso)
{
   struct fd6_sampler_stateobj *so = CALLOC_STRUCT(fd6_sampler_stateobj);
   unsigned aniso = util_last_bit(MIN2(cso->max_anisotropy >> 1, 8));
   bool miplinear = false;

   if (!so)
      return NULL;

   so->base = *cso;
   so->seqno = ++fd6_context(fd_context(pctx))->tex_seqno;

   if (cso->min_mip_filter == PIPE_TEX_MIPFILTER_LINEAR)
      miplinear = true;

   so->needs_border = false;
   so->texsamp0 =
      COND(miplinear, A6XX_TEX_SAMP_0_MIPFILTER_LINEAR_NEAR) |
      A6XX_TEX_SAMP_0_XY_MAG(tex_filter(cso->mag_img_filter, aniso)) |
      A6XX_TEX_SAMP_0_XY_MIN(tex_filter(cso->min_img_filter, aniso)) |
      A6XX_TEX_SAMP_0_ANISO(aniso) |
      A6XX_TEX_SAMP_0_WRAP_S(tex_clamp(cso->wrap_s, &so->needs_border)) |
      A6XX_TEX_SAMP_0_WRAP_T(tex_clamp(cso->wrap_t, &so->needs_border)) |
      A6XX_TEX_SAMP_0_WRAP_R(tex_clamp(cso->wrap_r, &so->needs_border));

   so->texsamp1 =
      COND(cso->min_mip_filter == PIPE_TEX_MIPFILTER_NONE,
           A6XX_TEX_SAMP_1_MIPFILTER_LINEAR_FAR) |
      COND(!cso->seamless_cube_map, A6XX_TEX_SAMP_1_CUBEMAPSEAMLESSFILTOFF) |
      COND(!cso->normalized_coords, A6XX_TEX_SAMP_1_UNNORM_COORDS);

   so->texsamp0 |= A6XX_TEX_SAMP_0_LOD_BIAS(cso->lod_bias);
   so->texsamp1 |= A6XX_TEX_SAMP_1_MIN_LOD(cso->min_lod) |
                   A6XX_TEX_SAMP_1_MAX_LOD(cso->max_lod);

   if (cso->compare_mode)
      so->texsamp1 |=
         A6XX_TEX_SAMP_1_COMPARE_FUNC(cso->compare_func); /* maps 1:1 */

   return so;
}

static void
fd6_sampler_state_delete(struct pipe_context *pctx, void *hwcso)
{
   struct fd_context *ctx = fd_context(pctx);
   struct fd6_context *fd6_ctx = fd6_context(ctx);
   struct fd6_sampler_stateobj *samp = hwcso;

   fd_screen_lock(ctx->screen);

   hash_table_foreach (fd6_ctx->tex_cache, entry) {
      struct fd6_texture_state *state = entry->data;

      for (unsigned i = 0; i < ARRAY_SIZE(state->key.samp); i++) {
         if (samp->seqno == state->key.samp[i].seqno) {
            remove_tex_entry(fd6_ctx, entry);
            break;
         }
      }
   }

   fd_screen_unlock(ctx->screen);

   free(hwcso);
}

static struct pipe_sampler_view *
fd6_sampler_view_create(struct pipe_context *pctx, struct pipe_resource *prsc,
                        const struct pipe_sampler_view *cso)
{
   struct fd6_pipe_sampler_view *so = CALLOC_STRUCT(fd6_pipe_sampler_view);

   if (!so)
      return NULL;

   so->base = *cso;
   pipe_reference(NULL, &prsc->reference);
   so->base.texture = prsc;
   so->base.reference.count = 1;
   so->base.context = pctx;
   so->needs_validate = true;

   return &so->base;
}

static void
fd6_set_sampler_views(struct pipe_context *pctx, enum pipe_shader_type shader,
                      unsigned start, unsigned nr,
                      unsigned unbind_num_trailing_slots,
                      bool take_ownership,
                      struct pipe_sampler_view **views) in_dt
{
   struct fd_context *ctx = fd_context(pctx);

   fd_set_sampler_views(pctx, shader, start, nr, unbind_num_trailing_slots,
                        take_ownership, views);

   if (!views)
      return;

   for (unsigned i = 0; i < nr; i++) {
      struct fd6_pipe_sampler_view *so = fd6_pipe_sampler_view(views[i]);

      if (!(so && so->needs_validate))
         continue;

      struct fd_resource *rsc = fd_resource(so->base.texture);

      fd6_validate_format(ctx, rsc, so->base.format);
      fd6_sampler_view_update(ctx, so);

      so->needs_validate = false;
   }
}

void
fd6_sampler_view_update(struct fd_context *ctx,
                        struct fd6_pipe_sampler_view *so)
{
   const struct pipe_sampler_view *cso = &so->base;
   struct pipe_resource *prsc = cso->texture;
   struct fd_resource *rsc = fd_resource(prsc);
   enum pipe_format format = cso->format;
   bool ubwc_enabled = false;
   unsigned lvl, layers = 0;

   fd6_validate_format(ctx, rsc, cso->format);

   if (format == PIPE_FORMAT_X32_S8X24_UINT) {
      rsc = rsc->stencil;
      format = rsc->b.b.format;
   }

   so->seqno = ++fd6_context(ctx)->tex_seqno;
   so->ptr1 = rsc;
   so->rsc_seqno = rsc->seqno;

   if (cso->target == PIPE_BUFFER) {
      unsigned elements = cso->u.buf.size / util_format_get_blocksize(format);

      lvl = 0;
      so->texconst1 = A6XX_TEX_CONST_1_WIDTH(elements & MASK(15)) |
                      A6XX_TEX_CONST_1_HEIGHT(elements >> 15);
      so->texconst2 = A6XX_TEX_CONST_2_UNK4 | A6XX_TEX_CONST_2_UNK31;
      so->offset1 = cso->u.buf.offset;
   } else {
      unsigned miplevels;

      lvl = fd_sampler_first_level(cso);
      miplevels = fd_sampler_last_level(cso) - lvl;
      layers = cso->u.tex.last_layer - cso->u.tex.first_layer + 1;

      so->texconst0 |= A6XX_TEX_CONST_0_MIPLVLS(miplevels);
      so->texconst1 = A6XX_TEX_CONST_1_WIDTH(u_minify(prsc->width0, lvl)) |
                      A6XX_TEX_CONST_1_HEIGHT(u_minify(prsc->height0, lvl));
      so->texconst2 = A6XX_TEX_CONST_2_PITCHALIGN(rsc->layout.pitchalign - 6) |
                      A6XX_TEX_CONST_2_PITCH(fd_resource_pitch(rsc, lvl));

      ubwc_enabled = fd_resource_ubwc_enabled(rsc, lvl);

      if (rsc->b.b.format == PIPE_FORMAT_R8_G8B8_420_UNORM) {
         struct fd_resource *next = fd_resource(rsc->b.b.next);

         /* In case of biplanar R8_G8B8, the UBWC metadata address in
          * dwords 7 and 8, is instead the pointer to the second plane.
          */
         so->ptr2 = next;
         so->texconst6 =
            A6XX_TEX_CONST_6_PLANE_PITCH(fd_resource_pitch(next, lvl));

         if (ubwc_enabled) {
            /* Further, if using UBWC with R8_G8B8, we only point to the
             * UBWC header and the color data is expected to follow immediately.
             */
            so->offset1 =
               fd_resource_ubwc_offset(rsc, lvl, cso->u.tex.first_layer);
            so->offset2 =
               fd_resource_ubwc_offset(next, lvl, cso->u.tex.first_layer);
         } else {
            so->offset1 = fd_resource_offset(rsc, lvl, cso->u.tex.first_layer);
            so->offset2 = fd_resource_offset(next, lvl, cso->u.tex.first_layer);
         }
      } else {
         so->offset1 = fd_resource_offset(rsc, lvl, cso->u.tex.first_layer);
         if (ubwc_enabled) {
            so->ptr2 = rsc;
            so->offset2 =
               fd_resource_ubwc_offset(rsc, lvl, cso->u.tex.first_layer);
         }
      }
   }

   so->texconst0 |=
      fd6_tex_const_0(prsc, lvl, cso->format, cso->swizzle_r, cso->swizzle_g,
                      cso->swizzle_b, cso->swizzle_a);

   so->texconst2 |= A6XX_TEX_CONST_2_TYPE(fd6_tex_type(cso->target));

   switch (cso->target) {
   case PIPE_TEXTURE_RECT:
   case PIPE_TEXTURE_1D:
   case PIPE_TEXTURE_2D:
      so->texconst3 = A6XX_TEX_CONST_3_ARRAY_PITCH(rsc->layout.layer_size);
      so->texconst5 = A6XX_TEX_CONST_5_DEPTH(1);
      break;
   case PIPE_TEXTURE_1D_ARRAY:
   case PIPE_TEXTURE_2D_ARRAY:
      so->texconst3 = A6XX_TEX_CONST_3_ARRAY_PITCH(rsc->layout.layer_size);
      so->texconst5 = A6XX_TEX_CONST_5_DEPTH(layers);
      break;
   case PIPE_TEXTURE_CUBE:
   case PIPE_TEXTURE_CUBE_ARRAY:
      so->texconst3 = A6XX_TEX_CONST_3_ARRAY_PITCH(rsc->layout.layer_size);
      so->texconst5 = A6XX_TEX_CONST_5_DEPTH(layers / 6);
      break;
   case PIPE_TEXTURE_3D:
      so->texconst3 =
         A6XX_TEX_CONST_3_MIN_LAYERSZ(
            fd_resource_slice(rsc, prsc->last_level)->size0) |
         A6XX_TEX_CONST_3_ARRAY_PITCH(fd_resource_slice(rsc, lvl)->size0);
      so->texconst5 = A6XX_TEX_CONST_5_DEPTH(u_minify(prsc->depth0, lvl));
      break;
   default:
      break;
   }

   if (rsc->layout.tile_all)
      so->texconst3 |= A6XX_TEX_CONST_3_TILE_ALL;

   if (ubwc_enabled) {
      uint32_t block_width, block_height;
      fdl6_get_ubwc_blockwidth(&rsc->layout, &block_width, &block_height);

      so->texconst3 |= A6XX_TEX_CONST_3_FLAG;
      so->texconst9 |= A6XX_TEX_CONST_9_FLAG_BUFFER_ARRAY_PITCH(
         rsc->layout.ubwc_layer_size >> 2);
      so->texconst10 |=
         A6XX_TEX_CONST_10_FLAG_BUFFER_PITCH(
            fdl_ubwc_pitch(&rsc->layout, lvl)) |
         A6XX_TEX_CONST_10_FLAG_BUFFER_LOGW(util_logbase2_ceil(
            DIV_ROUND_UP(u_minify(prsc->width0, lvl), block_width))) |
         A6XX_TEX_CONST_10_FLAG_BUFFER_LOGH(util_logbase2_ceil(
            DIV_ROUND_UP(u_minify(prsc->height0, lvl), block_height)));
   }
}

/* NOTE this can be called in either driver thread or frontend thread
 * depending on where the last unref comes from
 */
static void
fd6_sampler_view_destroy(struct pipe_context *pctx,
                         struct pipe_sampler_view *_view)
{
   struct fd_context *ctx = fd_context(pctx);
   struct fd6_context *fd6_ctx = fd6_context(ctx);
   struct fd6_pipe_sampler_view *view = fd6_pipe_sampler_view(_view);

   fd_screen_lock(ctx->screen);

   hash_table_foreach (fd6_ctx->tex_cache, entry) {
      struct fd6_texture_state *state = entry->data;

      for (unsigned i = 0; i < ARRAY_SIZE(state->key.view); i++) {
         if (view->seqno == state->key.view[i].seqno) {
            remove_tex_entry(fd6_ctx, entry);
            break;
         }
      }
   }

   fd_screen_unlock(ctx->screen);

   pipe_resource_reference(&view->base.texture, NULL);

   free(view);
}

static uint32_t
key_hash(const void *_key)
{
   const struct fd6_texture_key *key = _key;
   return XXH32(key, sizeof(*key), 0);
}

static bool
key_equals(const void *_a, const void *_b)
{
   const struct fd6_texture_key *a = _a;
   const struct fd6_texture_key *b = _b;
   return memcmp(a, b, sizeof(struct fd6_texture_key)) == 0;
}

struct fd6_texture_state *
fd6_texture_state(struct fd_context *ctx, enum pipe_shader_type type,
                  struct fd_texture_stateobj *tex)
{
   struct fd6_context *fd6_ctx = fd6_context(ctx);
   struct fd6_texture_state *state = NULL;
   struct fd6_texture_key key;
   bool needs_border = false;

   memset(&key, 0, sizeof(key));

   for (unsigned i = 0; i < tex->num_textures; i++) {
      if (!tex->textures[i])
         continue;

      struct fd6_pipe_sampler_view *view =
         fd6_pipe_sampler_view(tex->textures[i]);

      /* NOTE that if the backing rsc was uncompressed between the
       * time that the CSO was originally created and now, the rsc
       * seqno would have changed, so we don't have to worry about
       * getting a bogus cache hit.
       */
      key.view[i].rsc_seqno = fd_resource(view->base.texture)->seqno;
      key.view[i].seqno = view->seqno;
   }

   for (unsigned i = 0; i < tex->num_samplers; i++) {
      if (!tex->samplers[i])
         continue;

      struct fd6_sampler_stateobj *sampler =
         fd6_sampler_stateobj(tex->samplers[i]);

      key.samp[i].seqno = sampler->seqno;

      needs_border |= sampler->needs_border;
   }

   key.type = type;
   key.bcolor_offset = fd6_border_color_offset(ctx, type, tex);

   uint32_t hash = key_hash(&key);
   fd_screen_lock(ctx->screen);
   struct hash_entry *entry =
      _mesa_hash_table_search_pre_hashed(fd6_ctx->tex_cache, hash, &key);

   if (entry) {
      fd6_texture_state_reference(&state, entry->data);
      goto out_unlock;
   }

   state = CALLOC_STRUCT(fd6_texture_state);

   /* NOTE: one ref for tex_cache, and second ref for returned state: */
   pipe_reference_init(&state->reference, 2);
   state->key = key;
   state->stateobj = fd_ringbuffer_new_object(ctx->pipe, 32 * 4);
   state->needs_border = needs_border;

   fd6_emit_textures(ctx, state->stateobj, type, tex, key.bcolor_offset, NULL);

   /* NOTE: uses copy of key in state obj, because pointer passed by caller
    * is probably on the stack
    */
   _mesa_hash_table_insert_pre_hashed(fd6_ctx->tex_cache, hash, &state->key,
                                      state);

out_unlock:
   fd_screen_unlock(ctx->screen);
   return state;
}

void
__fd6_texture_state_describe(char *buf, const struct fd6_texture_state *tex)
{
   sprintf(buf, "fd6_texture_state<%p>", tex);
}

void
__fd6_texture_state_destroy(struct fd6_texture_state *state)
{
   fd_ringbuffer_del(state->stateobj);
   free(state);
}

static void
fd6_rebind_resource(struct fd_context *ctx, struct fd_resource *rsc) assert_dt
{
   fd_screen_assert_locked(ctx->screen);

   if (!(rsc->dirty & FD_DIRTY_TEX))
      return;

   struct fd6_context *fd6_ctx = fd6_context(ctx);

   hash_table_foreach (fd6_ctx->tex_cache, entry) {
      struct fd6_texture_state *state = entry->data;

      for (unsigned i = 0; i < ARRAY_SIZE(state->key.view); i++) {
         if (rsc->seqno == state->key.view[i].rsc_seqno) {
            remove_tex_entry(fd6_ctx, entry);
            break;
         }
      }
   }
}

void
fd6_texture_init(struct pipe_context *pctx) disable_thread_safety_analysis
{
   struct fd_context *ctx = fd_context(pctx);
   struct fd6_context *fd6_ctx = fd6_context(ctx);

   pctx->create_sampler_state = fd6_sampler_state_create;
   pctx->delete_sampler_state = fd6_sampler_state_delete;
   pctx->bind_sampler_states = fd_sampler_states_bind;

   pctx->create_sampler_view = fd6_sampler_view_create;
   pctx->sampler_view_destroy = fd6_sampler_view_destroy;
   pctx->set_sampler_views = fd6_set_sampler_views;

   ctx->rebind_resource = fd6_rebind_resource;

   fd6_ctx->tex_cache = _mesa_hash_table_create(NULL, key_hash, key_equals);
}

void
fd6_texture_fini(struct pipe_context *pctx)
{
   struct fd_context *ctx = fd_context(pctx);
   struct fd6_context *fd6_ctx = fd6_context(ctx);

   fd_screen_lock(ctx->screen);

   hash_table_foreach (fd6_ctx->tex_cache, entry) {
      remove_tex_entry(fd6_ctx, entry);
   }

   fd_screen_unlock(ctx->screen);

   ralloc_free(fd6_ctx->tex_cache);
}