drivers/freedreno/freedreno_batch_cache.c

/*
 * Copyright (C) 2016 Rob Clark <robclark@freedesktop.org>
 *
 * 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 "util/hash_table.h"
#include "util/list.h"
#include "util/set.h"
#include "util/u_string.h"
#define XXH_INLINE_ALL
#include "util/xxhash.h"

#include "freedreno_batch.h"
#include "freedreno_batch_cache.h"
#include "freedreno_context.h"
#include "freedreno_resource.h"

/* Overview:
 *
 *   The batch cache provides lookup for mapping pipe_framebuffer_state
 *   to a batch.
 *
 *   It does this via hashtable, with key that roughly matches the
 *   pipe_framebuffer_state, as described below.
 *
 * Batch Cache hashtable key:
 *
 *   To serialize the key, and to avoid dealing with holding a reference to
 *   pipe_surface's (which hold a reference to pipe_resource and complicate
 *   the whole refcnting thing), the key is variable length and inline's the
 *   pertinent details of the pipe_surface.
 *
 * Batch:
 *
 *   Each batch needs to hold a reference to each resource it depends on (ie.
 *   anything that needs a mem2gmem).  And a weak reference to resources it
 *   renders to.  (If both src[n] and dst[n] are not NULL then they are the
 *   same.)
 *
 *   When a resource is destroyed, we need to remove entries in the batch
 *   cache that reference the resource, to avoid dangling pointer issues.
 *   So each resource holds a hashset of batches which have reference them
 *   in their hashtable key.
 *
 *   When a batch has weak reference to no more resources (ie. all the
 *   surfaces it rendered to are destroyed) the batch can be destroyed.
 *   Could happen in an app that renders and never uses the result.  More
 *   common scenario, I think, will be that some, but not all, of the
 *   surfaces are destroyed before the batch is submitted.
 *
 *   If (for example), batch writes to zsbuf but that surface is destroyed
 *   before batch is submitted, we can skip gmem2mem (but still need to
 *   alloc gmem space as before.  If the batch depended on previous contents
 *   of that surface, it would be holding a reference so the surface would
 *   not have been destroyed.
 */

struct fd_batch_key {
   uint32_t width;
   uint32_t height;
   uint16_t layers;
   uint16_t samples;
   uint16_t num_surfs;
   uint16_t ctx_seqno;
   struct {
      struct pipe_resource *texture;
      union pipe_surface_desc u;
      uint8_t pos, samples;
      uint16_t format;
   } surf[0];
};

static struct fd_batch_key *
key_alloc(unsigned num_surfs)
{
   struct fd_batch_key *key = CALLOC_VARIANT_LENGTH_STRUCT(
      fd_batch_key, sizeof(key->surf[0]) * num_surfs);
   return key;
}

uint32_t
fd_batch_key_hash(const void *_key)
{
   const struct fd_batch_key *key = _key;
   uint32_t hash = 0;
   hash = XXH32(key, offsetof(struct fd_batch_key, surf[0]), hash);
   hash = XXH32(key->surf, sizeof(key->surf[0]) * key->num_surfs, hash);
   return hash;
}

bool
fd_batch_key_equals(const void *_a, const void *_b)
{
   const struct fd_batch_key *a = _a;
   const struct fd_batch_key *b = _b;
   return (memcmp(a, b, offsetof(struct fd_batch_key, surf[0])) == 0) &&
          (memcmp(a->surf, b->surf, sizeof(a->surf[0]) * a->num_surfs) == 0);
}

struct fd_batch_key *
fd_batch_key_clone(void *mem_ctx, const struct fd_batch_key *key)
{
   unsigned sz =
      sizeof(struct fd_batch_key) + (sizeof(key->surf[0]) * key->num_surfs);
   struct fd_batch_key *new_key = rzalloc_size(mem_ctx, sz);
   memcpy(new_key, key, sz);
   return new_key;
}

void
fd_bc_init(struct fd_batch_cache *cache)
{
   cache->ht =
      _mesa_hash_table_create(NULL, fd_batch_key_hash, fd_batch_key_equals);
}

void
fd_bc_fini(struct fd_batch_cache *cache)
{
   _mesa_hash_table_destroy(cache->ht, NULL);
}

/* Flushes all batches in the batch cache.  Used at glFlush() and similar times. */
void
fd_bc_flush(struct fd_context *ctx, bool deferred) assert_dt
{
   struct fd_batch_cache *cache = &ctx->screen->batch_cache;

   /* fd_batch_flush() (and fd_batch_add_dep() which calls it indirectly)
    * can cause batches to be unref'd and freed under our feet, so grab
    * a reference to all the batches we need up-front.
    */
   struct fd_batch *batches[ARRAY_SIZE(cache->batches)] = {0};
   struct fd_batch *batch;
   unsigned n = 0;

   fd_screen_lock(ctx->screen);

   foreach_batch (batch, cache, cache->batch_mask) {
      if (batch->ctx == ctx) {
         fd_batch_reference_locked(&batches[n++], batch);
      }
   }

   /* deferred flush doesn't actually flush, but it marks every other
    * batch associated with the context as dependent on the current
    * batch.  So when the current batch gets flushed, all other batches
    * that came before also get flushed.
    */
   if (deferred) {
      struct fd_batch *current_batch = fd_context_batch(ctx);
      struct fd_batch *deps[ARRAY_SIZE(cache->batches)] = {0};
      unsigned ndeps = 0;

      /* To avoid a dependency loop, pull out any batches that already
       * have a dependency on the current batch.  This ensures the
       * following loop adding a dependency to the current_batch, all
       * remaining batches do not have a direct or indirect dependency
       * on the current_batch.
       *
       * The batches that have a dependency on the current batch will
       * be flushed immediately (after dropping screen lock) instead
       */
      for (unsigned i = 0; i < n; i++) {
         if ((batches[i] != current_batch) &&
             fd_batch_has_dep(batches[i], current_batch)) {
            /* We can't immediately flush while we hold the screen lock,
             * but that doesn't matter.  We just want to skip adding any
             * deps that would result in a loop, we can flush after we've
             * updated the dependency graph and dropped the lock.
             */
            fd_batch_reference_locked(&deps[ndeps++], batches[i]);
            fd_batch_reference_locked(&batches[i], NULL);
         }
      }

      for (unsigned i = 0; i < n; i++) {
         if (batches[i] && (batches[i] != current_batch) &&
               (batches[i]->ctx == current_batch->ctx)) {
            fd_batch_add_dep(current_batch, batches[i]);
         }
      }

      fd_batch_reference_locked(&current_batch, NULL);

      fd_screen_unlock(ctx->screen);

      /* If we have any batches that we could add a dependency on (unlikely)
       * flush them immediately.
       */
      for (unsigned i = 0; i < ndeps; i++) {
         fd_batch_flush(deps[i]);
         fd_batch_reference(&deps[i], NULL);
      }
   } else {
      fd_screen_unlock(ctx->screen);

      for (unsigned i = 0; i < n; i++) {
         fd_batch_flush(batches[i]);
      }
   }

   for (unsigned i = 0; i < n; i++) {
      fd_batch_reference(&batches[i], NULL);
   }
}

/**
 * Flushes the batch (if any) writing this resource.  Must not hold the screen
 * lock.
 */
void
fd_bc_flush_writer(struct fd_context *ctx, struct fd_resource *rsc) assert_dt
{
   fd_screen_lock(ctx->screen);
   struct fd_batch *write_batch = NULL;
   fd_batch_reference_locked(&write_batch, rsc->track->write_batch);
   fd_screen_unlock(ctx->screen);

   if (write_batch) {
      if (write_batch->ctx == ctx)
         fd_batch_flush(write_batch);
      fd_batch_reference(&write_batch, NULL);
   }
}

/**
 * Flushes any batches reading this resource.  Must not hold the screen lock.
 */
void
fd_bc_flush_readers(struct fd_context *ctx, struct fd_resource *rsc) assert_dt
{
   struct fd_batch *batch, *batches[32] = {};
   uint32_t batch_count = 0;

   /* This is a bit awkward, probably a fd_batch_flush_locked()
    * would make things simpler.. but we need to hold the lock
    * to iterate the batches which reference this resource.  So
    * we must first grab references under a lock, then flush.
    */
   fd_screen_lock(ctx->screen);
   foreach_batch (batch, &ctx->screen->batch_cache, rsc->track->batch_mask)
      fd_batch_reference_locked(&batches[batch_count++], batch);
   fd_screen_unlock(ctx->screen);

   for (int i = 0; i < batch_count; i++) {
      if (batches[i]->ctx == ctx)
         fd_batch_flush(batches[i]);
      fd_batch_reference(&batches[i], NULL);
   }
}

void
fd_bc_dump(struct fd_context *ctx, const char *fmt, ...)
{
   struct fd_batch_cache *cache = &ctx->screen->batch_cache;

   if (!FD_DBG(MSGS))
      return;

   fd_screen_lock(ctx->screen);

   va_list ap;
   va_start(ap, fmt);
   vprintf(fmt, ap);
   va_end(ap);

   for (int i = 0; i < ARRAY_SIZE(cache->batches); i++) {
      struct fd_batch *batch = cache->batches[i];
      if (batch) {
         printf("  %p<%u>%s\n", batch, batch->seqno,
                batch->needs_flush ? ", NEEDS FLUSH" : "");
      }
   }

   printf("----\n");

   fd_screen_unlock(ctx->screen);
}

/**
 * Note that when batch is flushed, it needs to remain in the cache so
 * that fd_bc_invalidate_resource() can work.. otherwise we can have
 * the case where a rsc is destroyed while a batch still has a dangling
 * reference to it.
 *
 * Note that the cmdstream (or, after the SUBMIT ioctl, the kernel)
 * would have a reference to the underlying bo, so it is ok for the
 * rsc to be destroyed before the batch.
 */
void
fd_bc_invalidate_batch(struct fd_batch *batch, bool remove)
{
   if (!batch)
      return;

   struct fd_batch_cache *cache = &batch->ctx->screen->batch_cache;
   struct fd_batch_key *key = batch->key;

   fd_screen_assert_locked(batch->ctx->screen);

   if (remove) {
      cache->batches[batch->idx] = NULL;
      cache->batch_mask &= ~(1 << batch->idx);
   }

   if (!key)
      return;

   DBG("%p: key=%p", batch, batch->key);
   for (unsigned idx = 0; idx < key->num_surfs; idx++) {
      struct fd_resource *rsc = fd_resource(key->surf[idx].texture);
      rsc->track->bc_batch_mask &= ~(1 << batch->idx);
   }

   struct hash_entry *entry =
      _mesa_hash_table_search_pre_hashed(cache->ht, batch->hash, key);
   _mesa_hash_table_remove(cache->ht, entry);
}

void
fd_bc_invalidate_resource(struct fd_resource *rsc, bool destroy)
{
   struct fd_screen *screen = fd_screen(rsc->b.b.screen);
   struct fd_batch *batch;

   fd_screen_lock(screen);

   if (destroy) {
      foreach_batch (batch, &screen->batch_cache, rsc->track->batch_mask) {
         struct set_entry *entry = _mesa_set_search_pre_hashed(batch->resources, rsc->hash, rsc);
         _mesa_set_remove(batch->resources, entry);
      }
      rsc->track->batch_mask = 0;

      fd_batch_reference_locked(&rsc->track->write_batch, NULL);
   }

   foreach_batch (batch, &screen->batch_cache, rsc->track->bc_batch_mask)
      fd_bc_invalidate_batch(batch, false);

   rsc->track->bc_batch_mask = 0;

   fd_screen_unlock(screen);
}

static struct fd_batch *
alloc_batch_locked(struct fd_batch_cache *cache, struct fd_context *ctx,
                   bool nondraw) assert_dt
{
   struct fd_batch *batch;
   uint32_t idx;

   fd_screen_assert_locked(ctx->screen);

   while ((idx = ffs(~cache->batch_mask)) == 0) {
#if 0
      for (unsigned i = 0; i < ARRAY_SIZE(cache->batches); i++) {
         batch = cache->batches[i];
         debug_printf("%d: needs_flush=%d, depends:", batch->idx, batch->needs_flush);
         set_foreach (batch->dependencies, entry) {
            struct fd_batch *dep = (struct fd_batch *)entry->key;
            debug_printf(" %d", dep->idx);
         }
         debug_printf("\n");
      }
#endif
      /* TODO: is LRU the better policy?  Or perhaps the batch that
       * depends on the fewest other batches?
       */
      struct fd_batch *flush_batch = NULL;
      for (unsigned i = 0; i < ARRAY_SIZE(cache->batches); i++) {
         if (!flush_batch || (cache->batches[i]->seqno < flush_batch->seqno))
            fd_batch_reference_locked(&flush_batch, cache->batches[i]);
      }

      /* we can drop lock temporarily here, since we hold a ref,
       * flush_batch won't disappear under us.
       */
      fd_screen_unlock(ctx->screen);
      DBG("%p: too many batches!  flush forced!", flush_batch);
      fd_batch_flush(flush_batch);
      fd_screen_lock(ctx->screen);

      /* While the resources get cleaned up automatically, the flush_batch
       * doesn't get removed from the dependencies of other batches, so
       * it won't be unref'd and will remain in the table.
       *
       * TODO maybe keep a bitmask of batches that depend on me, to make
       * this easier:
       */
      for (unsigned i = 0; i < ARRAY_SIZE(cache->batches); i++) {
         struct fd_batch *other = cache->batches[i];
         if (!other)
            continue;
         if (fd_batch_has_dep(other, flush_batch)) {
            other->dependents_mask &= ~(1 << flush_batch->idx);
            struct fd_batch *ref = flush_batch;
            fd_batch_reference_locked(&ref, NULL);
         }
      }

      fd_batch_reference_locked(&flush_batch, NULL);
   }

   idx--; /* bit zero returns 1 for ffs() */

   batch = fd_batch_create(ctx, nondraw);
   if (!batch)
      return NULL;

   batch->seqno = seqno_next(&cache->cnt);
   batch->idx = idx;
   cache->batch_mask |= (1 << idx);

   assert(cache->batches[idx] == NULL);
   cache->batches[idx] = batch;

   return batch;
}

static void
alloc_query_buf(struct fd_context *ctx, struct fd_batch *batch)
{
   if (batch->query_buf)
      return;

   if ((ctx->screen->gen < 3) || (ctx->screen->gen > 4))
      return;

   /* For gens that use fd_hw_query, pre-allocate an initially zero-sized
    * (unbacked) query buffer.  This simplifies draw/grid/etc-time resource
    * tracking.
    */
   struct pipe_screen *pscreen = &ctx->screen->base;
   struct pipe_resource templ = {
      .target = PIPE_BUFFER,
      .format = PIPE_FORMAT_R8_UNORM,
      .bind = PIPE_BIND_QUERY_BUFFER,
      .width0 = 0, /* create initially zero size buffer */
      .height0 = 1,
      .depth0 = 1,
      .array_size = 1,
      .last_level = 0,
      .nr_samples = 1,
   };
   batch->query_buf = pscreen->resource_create(pscreen, &templ);
}

struct fd_batch *
fd_bc_alloc_batch(struct fd_context *ctx, bool nondraw)
{
   struct fd_batch_cache *cache = &ctx->screen->batch_cache;
   struct fd_batch *batch;

   /* For normal draw batches, pctx->set_framebuffer_state() handles
    * this, but for nondraw batches, this is a nice central location
    * to handle them all.
    */
   if (nondraw)
      fd_context_switch_from(ctx);

   fd_screen_lock(ctx->screen);
   batch = alloc_batch_locked(cache, ctx, nondraw);
   fd_screen_unlock(ctx->screen);

   alloc_query_buf(ctx, batch);

   if (batch && nondraw)
      fd_context_switch_to(ctx, batch);

   return batch;
}

static struct fd_batch *
batch_from_key(struct fd_context *ctx, struct fd_batch_key *key) assert_dt
{
   struct fd_batch_cache *cache = &ctx->screen->batch_cache;
   struct fd_batch *batch = NULL;
   uint32_t hash = fd_batch_key_hash(key);
   struct hash_entry *entry =
      _mesa_hash_table_search_pre_hashed(cache->ht, hash, key);

   if (entry) {
      free(key);
      fd_batch_reference_locked(&batch, (struct fd_batch *)entry->data);
      assert(!batch->flushed);
      return batch;
   }

   batch = alloc_batch_locked(cache, ctx, false);
#ifdef DEBUG
   DBG("%p: hash=0x%08x, %ux%u, %u layers, %u samples", batch, hash, key->width,
       key->height, key->layers, key->samples);
   for (unsigned idx = 0; idx < key->num_surfs; idx++) {
      DBG("%p:  surf[%u]: %p (%s) (%u,%u / %u,%u,%u)", batch,
          key->surf[idx].pos, key->surf[idx].texture,
          util_format_name(key->surf[idx].format),
          key->surf[idx].u.buf.first_element, key->surf[idx].u.buf.last_element,
          key->surf[idx].u.tex.first_layer, key->surf[idx].u.tex.last_layer,
          key->surf[idx].u.tex.level);
   }
#endif
   if (!batch)
      return NULL;

   /* reset max_scissor, which will be adjusted on draws
    * according to the actual scissor.
    */
   batch->max_scissor.minx = ~0;
   batch->max_scissor.miny = ~0;
   batch->max_scissor.maxx = 0;
   batch->max_scissor.maxy = 0;

   _mesa_hash_table_insert_pre_hashed(cache->ht, hash, key, batch);
   batch->key = key;
   batch->hash = hash;

   for (unsigned idx = 0; idx < key->num_surfs; idx++) {
      struct fd_resource *rsc = fd_resource(key->surf[idx].texture);
      rsc->track->bc_batch_mask = (1 << batch->idx);
   }

   return batch;
}

static void
key_surf(struct fd_batch_key *key, unsigned idx, unsigned pos,
         struct pipe_surface *psurf)
{
   key->surf[idx].texture = psurf->texture;
   key->surf[idx].u = psurf->u;
   key->surf[idx].pos = pos;
   key->surf[idx].samples = MAX2(1, psurf->nr_samples);
   key->surf[idx].format = psurf->format;
}

struct fd_batch *
fd_batch_from_fb(struct fd_context *ctx,
                 const struct pipe_framebuffer_state *pfb)
{
   unsigned idx = 0, n = pfb->nr_cbufs + (pfb->zsbuf ? 1 : 0);
   struct fd_batch_key *key = key_alloc(n);

   key->width = pfb->width;
   key->height = pfb->height;
   key->layers = pfb->layers;
   key->samples = util_framebuffer_get_num_samples(pfb);
   key->ctx_seqno = ctx->seqno;

   if (pfb->zsbuf)
      key_surf(key, idx++, 0, pfb->zsbuf);

   for (unsigned i = 0; i < pfb->nr_cbufs; i++)
      if (pfb->cbufs[i])
         key_surf(key, idx++, i + 1, pfb->cbufs[i]);

   key->num_surfs = idx;

   fd_screen_lock(ctx->screen);
   struct fd_batch *batch = batch_from_key(ctx, key);
   fd_screen_unlock(ctx->screen);

   alloc_query_buf(ctx, batch);

   fd_batch_set_fb(batch, pfb);

   return batch;
}