]> www.pilppa.org Git - linux-2.6-omap-h63xx.git/commitdiff
Merge branch 'drm-gem-update' of git://git.kernel.org/pub/scm/linux/kernel/git/airlie...
authorLinus Torvalds <torvalds@linux-foundation.org>
Fri, 5 Dec 2008 05:44:40 +0000 (21:44 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 5 Dec 2008 05:44:40 +0000 (21:44 -0800)
* 'drm-gem-update' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/i915: Return error in i915_gem_set_to_gtt_domain if we're not in the GTT.
  drm/i915: Retry execbuffer pinning after clearing the GTT
  drm/i915: Move the execbuffer domain computations together
  drm/i915: Rename object_set_domain to object_set_to_gpu_domain
  drm/i915: Make a single set-to-cpu-domain path and use it wherever needed.
  drm/i915: Make a single set-to-gtt-domain path.
  drm/i915: If interrupted while setting object domains, still emit the flush.
  drm/i915: Move flushing list cleanup from flush request retire to request emit.
  drm/i915: Respect GM965/GM45 bit-17-instead-of-bit-11 option for swizzling.

drivers/gpu/drm/i915/i915_drv.h
drivers/gpu/drm/i915/i915_gem.c
drivers/gpu/drm/i915/i915_gem_proc.c
drivers/gpu/drm/i915/i915_gem_tiling.c
drivers/gpu/drm/i915/i915_reg.h

index 0a4f39b9a0ec1623b79fa4073b8580efb203caa0..adc972cc6bfc03a28038841b2f6a43ff578e0fb3 100644 (file)
@@ -244,6 +244,10 @@ typedef struct drm_i915_private {
                 * List of objects currently involved in rendering from the
                 * ringbuffer.
                 *
+                * Includes buffers having the contents of their GPU caches
+                * flushed, not necessarily primitives.  last_rendering_seqno
+                * represents when the rendering involved will be completed.
+                *
                 * A reference is held on the buffer while on this list.
                 */
                struct list_head active_list;
@@ -253,6 +257,8 @@ typedef struct drm_i915_private {
                 * still have a write_domain which needs to be flushed before
                 * unbinding.
                 *
+                * last_rendering_seqno is 0 while an object is in this list.
+                *
                 * A reference is held on the buffer while on this list.
                 */
                struct list_head flushing_list;
@@ -261,6 +267,8 @@ typedef struct drm_i915_private {
                 * LRU list of objects which are not in the ringbuffer and
                 * are ready to unbind, but are still in the GTT.
                 *
+                * last_rendering_seqno is 0 while an object is in this list.
+                *
                 * A reference is not held on the buffer while on this list,
                 * as merely being GTT-bound shouldn't prevent its being
                 * freed, and we'll pull it off the list in the free path.
@@ -371,8 +379,8 @@ struct drm_i915_gem_object {
        uint32_t agp_type;
 
        /**
-        * Flagging of which individual pages are valid in GEM_DOMAIN_CPU when
-        * GEM_DOMAIN_CPU is not in the object's read domain.
+        * If present, while GEM_DOMAIN_CPU is in the read domain this array
+        * flags which individual pages are valid.
         */
        uint8_t *page_cpu_valid;
 };
@@ -394,9 +402,6 @@ struct drm_i915_gem_request {
        /** Time at which this request was emitted, in jiffies. */
        unsigned long emitted_jiffies;
 
-       /** Cache domains that were flushed at the start of the request. */
-       uint32_t flush_domains;
-
        struct list_head list;
 };
 
index d58ddef468f8124c29f6af32412ee2ba31f9873e..3fde82be014f3a4b2e4046826b076d7540e17afc 100644 (file)
 
 #define I915_GEM_GPU_DOMAINS   (~(I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT))
 
-static int
-i915_gem_object_set_domain(struct drm_gem_object *obj,
-                           uint32_t read_domains,
-                           uint32_t write_domain);
-static int
-i915_gem_object_set_domain_range(struct drm_gem_object *obj,
-                                uint64_t offset,
-                                uint64_t size,
-                                uint32_t read_domains,
-                                uint32_t write_domain);
-static int
-i915_gem_set_domain(struct drm_gem_object *obj,
-                   struct drm_file *file_priv,
-                   uint32_t read_domains,
-                   uint32_t write_domain);
+static void
+i915_gem_object_set_to_gpu_domain(struct drm_gem_object *obj,
+                                 uint32_t read_domains,
+                                 uint32_t write_domain);
+static void i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj);
+static void i915_gem_object_flush_gtt_write_domain(struct drm_gem_object *obj);
+static void i915_gem_object_flush_cpu_write_domain(struct drm_gem_object *obj);
+static int i915_gem_object_set_to_gtt_domain(struct drm_gem_object *obj,
+                                            int write);
+static int i915_gem_object_set_to_cpu_domain(struct drm_gem_object *obj,
+                                            int write);
+static int i915_gem_object_set_cpu_read_domain_range(struct drm_gem_object *obj,
+                                                    uint64_t offset,
+                                                    uint64_t size);
+static void i915_gem_object_set_to_full_cpu_read_domain(struct drm_gem_object *obj);
 static int i915_gem_object_get_page_list(struct drm_gem_object *obj);
 static void i915_gem_object_free_page_list(struct drm_gem_object *obj);
 static int i915_gem_object_wait_rendering(struct drm_gem_object *obj);
@@ -162,8 +162,8 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data,
 
        mutex_lock(&dev->struct_mutex);
 
-       ret = i915_gem_object_set_domain_range(obj, args->offset, args->size,
-                                              I915_GEM_DOMAIN_CPU, 0);
+       ret = i915_gem_object_set_cpu_read_domain_range(obj, args->offset,
+                                                       args->size);
        if (ret != 0) {
                drm_gem_object_unreference(obj);
                mutex_unlock(&dev->struct_mutex);
@@ -260,8 +260,7 @@ i915_gem_gtt_pwrite(struct drm_device *dev, struct drm_gem_object *obj,
                mutex_unlock(&dev->struct_mutex);
                return ret;
        }
-       ret = i915_gem_set_domain(obj, file_priv,
-                                 I915_GEM_DOMAIN_GTT, I915_GEM_DOMAIN_GTT);
+       ret = i915_gem_object_set_to_gtt_domain(obj, 1);
        if (ret)
                goto fail;
 
@@ -320,8 +319,7 @@ i915_gem_shmem_pwrite(struct drm_device *dev, struct drm_gem_object *obj,
 
        mutex_lock(&dev->struct_mutex);
 
-       ret = i915_gem_set_domain(obj, file_priv,
-                                 I915_GEM_DOMAIN_CPU, I915_GEM_DOMAIN_CPU);
+       ret = i915_gem_object_set_to_cpu_domain(obj, 1);
        if (ret) {
                mutex_unlock(&dev->struct_mutex);
                return ret;
@@ -397,7 +395,8 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 }
 
 /**
- * Called when user space prepares to use an object
+ * Called when user space prepares to use an object with the CPU, either
+ * through the mmap ioctl's mapping or a GTT mapping.
  */
 int
 i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
@@ -405,11 +404,26 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
 {
        struct drm_i915_gem_set_domain *args = data;
        struct drm_gem_object *obj;
+       uint32_t read_domains = args->read_domains;
+       uint32_t write_domain = args->write_domain;
        int ret;
 
        if (!(dev->driver->driver_features & DRIVER_GEM))
                return -ENODEV;
 
+       /* Only handle setting domains to types used by the CPU. */
+       if (write_domain & ~(I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT))
+               return -EINVAL;
+
+       if (read_domains & ~(I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT))
+               return -EINVAL;
+
+       /* Having something in the write domain implies it's in the read
+        * domain, and only that read domain.  Enforce that in the request.
+        */
+       if (write_domain != 0 && read_domains != write_domain)
+               return -EINVAL;
+
        obj = drm_gem_object_lookup(dev, file_priv, args->handle);
        if (obj == NULL)
                return -EBADF;
@@ -417,10 +431,21 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
        mutex_lock(&dev->struct_mutex);
 #if WATCH_BUF
        DRM_INFO("set_domain_ioctl %p(%d), %08x %08x\n",
-                obj, obj->size, args->read_domains, args->write_domain);
+                obj, obj->size, read_domains, write_domain);
 #endif
-       ret = i915_gem_set_domain(obj, file_priv,
-                                 args->read_domains, args->write_domain);
+       if (read_domains & I915_GEM_DOMAIN_GTT) {
+               ret = i915_gem_object_set_to_gtt_domain(obj, write_domain != 0);
+
+               /* Silently promote "you're not bound, there was nothing to do"
+                * to success, since the client was just asking us to
+                * make sure everything was done.
+                */
+               if (ret == -EINVAL)
+                       ret = 0;
+       } else {
+               ret = i915_gem_object_set_to_cpu_domain(obj, write_domain != 0);
+       }
+
        drm_gem_object_unreference(obj);
        mutex_unlock(&dev->struct_mutex);
        return ret;
@@ -455,10 +480,9 @@ i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data,
        obj_priv = obj->driver_private;
 
        /* Pinned buffers may be scanout, so flush the cache */
-       if ((obj->write_domain & I915_GEM_DOMAIN_CPU) && obj_priv->pin_count) {
-               i915_gem_clflush_object(obj);
-               drm_agp_chipset_flush(dev);
-       }
+       if (obj_priv->pin_count)
+               i915_gem_object_flush_cpu_write_domain(obj);
+
        drm_gem_object_unreference(obj);
        mutex_unlock(&dev->struct_mutex);
        return ret;
@@ -532,7 +556,7 @@ i915_gem_object_free_page_list(struct drm_gem_object *obj)
 }
 
 static void
-i915_gem_object_move_to_active(struct drm_gem_object *obj)
+i915_gem_object_move_to_active(struct drm_gem_object *obj, uint32_t seqno)
 {
        struct drm_device *dev = obj->dev;
        drm_i915_private_t *dev_priv = dev->dev_private;
@@ -546,8 +570,20 @@ i915_gem_object_move_to_active(struct drm_gem_object *obj)
        /* Move from whatever list we were on to the tail of execution. */
        list_move_tail(&obj_priv->list,
                       &dev_priv->mm.active_list);
+       obj_priv->last_rendering_seqno = seqno;
 }
 
+static void
+i915_gem_object_move_to_flushing(struct drm_gem_object *obj)
+{
+       struct drm_device *dev = obj->dev;
+       drm_i915_private_t *dev_priv = dev->dev_private;
+       struct drm_i915_gem_object *obj_priv = obj->driver_private;
+
+       BUG_ON(!obj_priv->active);
+       list_move_tail(&obj_priv->list, &dev_priv->mm.flushing_list);
+       obj_priv->last_rendering_seqno = 0;
+}
 
 static void
 i915_gem_object_move_to_inactive(struct drm_gem_object *obj)
@@ -562,6 +598,7 @@ i915_gem_object_move_to_inactive(struct drm_gem_object *obj)
        else
                list_move_tail(&obj_priv->list, &dev_priv->mm.inactive_list);
 
+       obj_priv->last_rendering_seqno = 0;
        if (obj_priv->active) {
                obj_priv->active = 0;
                drm_gem_object_unreference(obj);
@@ -610,10 +647,28 @@ i915_add_request(struct drm_device *dev, uint32_t flush_domains)
 
        request->seqno = seqno;
        request->emitted_jiffies = jiffies;
-       request->flush_domains = flush_domains;
        was_empty = list_empty(&dev_priv->mm.request_list);
        list_add_tail(&request->list, &dev_priv->mm.request_list);
 
+       /* Associate any objects on the flushing list matching the write
+        * domain we're flushing with our flush.
+        */
+       if (flush_domains != 0) {
+               struct drm_i915_gem_object *obj_priv, *next;
+
+               list_for_each_entry_safe(obj_priv, next,
+                                        &dev_priv->mm.flushing_list, list) {
+                       struct drm_gem_object *obj = obj_priv->obj;
+
+                       if ((obj->write_domain & flush_domains) ==
+                           obj->write_domain) {
+                               obj->write_domain = 0;
+                               i915_gem_object_move_to_active(obj, seqno);
+                       }
+               }
+
+       }
+
        if (was_empty && !dev_priv->mm.suspended)
                schedule_delayed_work(&dev_priv->mm.retire_work, HZ);
        return seqno;
@@ -676,30 +731,10 @@ i915_gem_retire_request(struct drm_device *dev,
                         __func__, request->seqno, obj);
 #endif
 
-               if (obj->write_domain != 0) {
-                       list_move_tail(&obj_priv->list,
-                                      &dev_priv->mm.flushing_list);
-               } else {
+               if (obj->write_domain != 0)
+                       i915_gem_object_move_to_flushing(obj);
+               else
                        i915_gem_object_move_to_inactive(obj);
-               }
-       }
-
-       if (request->flush_domains != 0) {
-               struct drm_i915_gem_object *obj_priv, *next;
-
-               /* Clear the write domain and activity from any buffers
-                * that are just waiting for a flush matching the one retired.
-                */
-               list_for_each_entry_safe(obj_priv, next,
-                                        &dev_priv->mm.flushing_list, list) {
-                       struct drm_gem_object *obj = obj_priv->obj;
-
-                       if (obj->write_domain & request->flush_domains) {
-                               obj->write_domain = 0;
-                               i915_gem_object_move_to_inactive(obj);
-                       }
-               }
-
        }
 }
 
@@ -892,25 +927,10 @@ i915_gem_object_wait_rendering(struct drm_gem_object *obj)
        struct drm_i915_gem_object *obj_priv = obj->driver_private;
        int ret;
 
-       /* If there are writes queued to the buffer, flush and
-        * create a new seqno to wait for.
+       /* This function only exists to support waiting for existing rendering,
+        * not for emitting required flushes.
         */
-       if (obj->write_domain & ~(I915_GEM_DOMAIN_CPU|I915_GEM_DOMAIN_GTT)) {
-               uint32_t write_domain = obj->write_domain;
-#if WATCH_BUF
-               DRM_INFO("%s: flushing object %p from write domain %08x\n",
-                         __func__, obj, write_domain);
-#endif
-               i915_gem_flush(dev, 0, write_domain);
-
-               i915_gem_object_move_to_active(obj);
-               obj_priv->last_rendering_seqno = i915_add_request(dev,
-                                                                 write_domain);
-               BUG_ON(obj_priv->last_rendering_seqno == 0);
-#if WATCH_LRU
-               DRM_INFO("%s: flush moves to exec list %p\n", __func__, obj);
-#endif
-       }
+       BUG_ON((obj->write_domain & I915_GEM_GPU_DOMAINS) != 0);
 
        /* If there is rendering queued on the buffer being evicted, wait for
         * it.
@@ -950,24 +970,16 @@ i915_gem_object_unbind(struct drm_gem_object *obj)
                return -EINVAL;
        }
 
-       /* Wait for any rendering to complete
-        */
-       ret = i915_gem_object_wait_rendering(obj);
-       if (ret) {
-               DRM_ERROR("wait_rendering failed: %d\n", ret);
-               return ret;
-       }
-
        /* Move the object to the CPU domain to ensure that
         * any possible CPU writes while it's not in the GTT
         * are flushed when we go to remap it. This will
         * also ensure that all pending GPU writes are finished
         * before we unbind.
         */
-       ret = i915_gem_object_set_domain(obj, I915_GEM_DOMAIN_CPU,
-                                        I915_GEM_DOMAIN_CPU);
+       ret = i915_gem_object_set_to_cpu_domain(obj, 1);
        if (ret) {
-               DRM_ERROR("set_domain failed: %d\n", ret);
+               if (ret != -ERESTARTSYS)
+                       DRM_ERROR("set_domain failed: %d\n", ret);
                return ret;
        }
 
@@ -1082,6 +1094,19 @@ i915_gem_evict_something(struct drm_device *dev)
        return ret;
 }
 
+static int
+i915_gem_evict_everything(struct drm_device *dev)
+{
+       int ret;
+
+       for (;;) {
+               ret = i915_gem_evict_something(dev);
+               if (ret != 0)
+                       break;
+       }
+       return ret;
+}
+
 static int
 i915_gem_object_get_page_list(struct drm_gem_object *obj)
 {
@@ -1168,7 +1193,8 @@ i915_gem_object_bind_to_gtt(struct drm_gem_object *obj, unsigned alignment)
 
                ret = i915_gem_evict_something(dev);
                if (ret != 0) {
-                       DRM_ERROR("Failed to evict a buffer %d\n", ret);
+                       if (ret != -ERESTARTSYS)
+                               DRM_ERROR("Failed to evict a buffer %d\n", ret);
                        return ret;
                }
                goto search_free;
@@ -1228,6 +1254,143 @@ i915_gem_clflush_object(struct drm_gem_object *obj)
        drm_clflush_pages(obj_priv->page_list, obj->size / PAGE_SIZE);
 }
 
+/** Flushes any GPU write domain for the object if it's dirty. */
+static void
+i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj)
+{
+       struct drm_device *dev = obj->dev;
+       uint32_t seqno;
+
+       if ((obj->write_domain & I915_GEM_GPU_DOMAINS) == 0)
+               return;
+
+       /* Queue the GPU write cache flushing we need. */
+       i915_gem_flush(dev, 0, obj->write_domain);
+       seqno = i915_add_request(dev, obj->write_domain);
+       obj->write_domain = 0;
+       i915_gem_object_move_to_active(obj, seqno);
+}
+
+/** Flushes the GTT write domain for the object if it's dirty. */
+static void
+i915_gem_object_flush_gtt_write_domain(struct drm_gem_object *obj)
+{
+       if (obj->write_domain != I915_GEM_DOMAIN_GTT)
+               return;
+
+       /* No actual flushing is required for the GTT write domain.   Writes
+        * to it immediately go to main memory as far as we know, so there's
+        * no chipset flush.  It also doesn't land in render cache.
+        */
+       obj->write_domain = 0;
+}
+
+/** Flushes the CPU write domain for the object if it's dirty. */
+static void
+i915_gem_object_flush_cpu_write_domain(struct drm_gem_object *obj)
+{
+       struct drm_device *dev = obj->dev;
+
+       if (obj->write_domain != I915_GEM_DOMAIN_CPU)
+               return;
+
+       i915_gem_clflush_object(obj);
+       drm_agp_chipset_flush(dev);
+       obj->write_domain = 0;
+}
+
+/**
+ * Moves a single object to the GTT read, and possibly write domain.
+ *
+ * This function returns when the move is complete, including waiting on
+ * flushes to occur.
+ */
+static int
+i915_gem_object_set_to_gtt_domain(struct drm_gem_object *obj, int write)
+{
+       struct drm_i915_gem_object *obj_priv = obj->driver_private;
+       int ret;
+
+       /* Not valid to be called on unbound objects. */
+       if (obj_priv->gtt_space == NULL)
+               return -EINVAL;
+
+       i915_gem_object_flush_gpu_write_domain(obj);
+       /* Wait on any GPU rendering and flushing to occur. */
+       ret = i915_gem_object_wait_rendering(obj);
+       if (ret != 0)
+               return ret;
+
+       /* If we're writing through the GTT domain, then CPU and GPU caches
+        * will need to be invalidated at next use.
+        */
+       if (write)
+               obj->read_domains &= I915_GEM_DOMAIN_GTT;
+
+       i915_gem_object_flush_cpu_write_domain(obj);
+
+       /* It should now be out of any other write domains, and we can update
+        * the domain values for our changes.
+        */
+       BUG_ON((obj->write_domain & ~I915_GEM_DOMAIN_GTT) != 0);
+       obj->read_domains |= I915_GEM_DOMAIN_GTT;
+       if (write) {
+               obj->write_domain = I915_GEM_DOMAIN_GTT;
+               obj_priv->dirty = 1;
+       }
+
+       return 0;
+}
+
+/**
+ * Moves a single object to the CPU read, and possibly write domain.
+ *
+ * This function returns when the move is complete, including waiting on
+ * flushes to occur.
+ */
+static int
+i915_gem_object_set_to_cpu_domain(struct drm_gem_object *obj, int write)
+{
+       struct drm_device *dev = obj->dev;
+       int ret;
+
+       i915_gem_object_flush_gpu_write_domain(obj);
+       /* Wait on any GPU rendering and flushing to occur. */
+       ret = i915_gem_object_wait_rendering(obj);
+       if (ret != 0)
+               return ret;
+
+       i915_gem_object_flush_gtt_write_domain(obj);
+
+       /* If we have a partially-valid cache of the object in the CPU,
+        * finish invalidating it and free the per-page flags.
+        */
+       i915_gem_object_set_to_full_cpu_read_domain(obj);
+
+       /* Flush the CPU cache if it's still invalid. */
+       if ((obj->read_domains & I915_GEM_DOMAIN_CPU) == 0) {
+               i915_gem_clflush_object(obj);
+               drm_agp_chipset_flush(dev);
+
+               obj->read_domains |= I915_GEM_DOMAIN_CPU;
+       }
+
+       /* It should now be out of any other write domains, and we can update
+        * the domain values for our changes.
+        */
+       BUG_ON((obj->write_domain & ~I915_GEM_DOMAIN_CPU) != 0);
+
+       /* If we're writing through the CPU, then the GPU read domains will
+        * need to be invalidated at next use.
+        */
+       if (write) {
+               obj->read_domains &= I915_GEM_DOMAIN_CPU;
+               obj->write_domain = I915_GEM_DOMAIN_CPU;
+       }
+
+       return 0;
+}
+
 /*
  * Set the next domain for the specified object. This
  * may not actually perform the necessary flushing/invaliding though,
@@ -1339,16 +1502,18 @@ i915_gem_clflush_object(struct drm_gem_object *obj)
  *             MI_FLUSH
  *             drm_agp_chipset_flush
  */
-static int
-i915_gem_object_set_domain(struct drm_gem_object *obj,
-                           uint32_t read_domains,
-                           uint32_t write_domain)
+static void
+i915_gem_object_set_to_gpu_domain(struct drm_gem_object *obj,
+                                 uint32_t read_domains,
+                                 uint32_t write_domain)
 {
        struct drm_device               *dev = obj->dev;
        struct drm_i915_gem_object      *obj_priv = obj->driver_private;
        uint32_t                        invalidate_domains = 0;
        uint32_t                        flush_domains = 0;
-       int                             ret;
+
+       BUG_ON(read_domains & I915_GEM_DOMAIN_CPU);
+       BUG_ON(write_domain == I915_GEM_DOMAIN_CPU);
 
 #if WATCH_BUF
        DRM_INFO("%s: object %p read %08x -> %08x write %08x -> %08x\n",
@@ -1385,34 +1550,11 @@ i915_gem_object_set_domain(struct drm_gem_object *obj,
                DRM_INFO("%s: CPU domain flush %08x invalidate %08x\n",
                         __func__, flush_domains, invalidate_domains);
 #endif
-               /*
-                * If we're invaliding the CPU cache and flushing a GPU cache,
-                * then pause for rendering so that the GPU caches will be
-                * flushed before the cpu cache is invalidated
-                */
-               if ((invalidate_domains & I915_GEM_DOMAIN_CPU) &&
-                   (flush_domains & ~(I915_GEM_DOMAIN_CPU |
-                                      I915_GEM_DOMAIN_GTT))) {
-                       ret = i915_gem_object_wait_rendering(obj);
-                       if (ret)
-                               return ret;
-               }
                i915_gem_clflush_object(obj);
        }
 
        if ((write_domain | flush_domains) != 0)
                obj->write_domain = write_domain;
-
-       /* If we're invalidating the CPU domain, clear the per-page CPU
-        * domain list as well.
-        */
-       if (obj_priv->page_cpu_valid != NULL &&
-           (write_domain != 0 ||
-            read_domains & I915_GEM_DOMAIN_CPU)) {
-               drm_free(obj_priv->page_cpu_valid, obj->size / PAGE_SIZE,
-                        DRM_MEM_DRIVER);
-               obj_priv->page_cpu_valid = NULL;
-       }
        obj->read_domains = read_domains;
 
        dev->invalidate_domains |= invalidate_domains;
@@ -1423,47 +1565,94 @@ i915_gem_object_set_domain(struct drm_gem_object *obj,
                 obj->read_domains, obj->write_domain,
                 dev->invalidate_domains, dev->flush_domains);
 #endif
-       return 0;
 }
 
 /**
- * Set the read/write domain on a range of the object.
+ * Moves the object from a partially CPU read to a full one.
  *
- * Currently only implemented for CPU reads, otherwise drops to normal
- * i915_gem_object_set_domain().
+ * Note that this only resolves i915_gem_object_set_cpu_read_domain_range(),
+ * and doesn't handle transitioning from !(read_domains & I915_GEM_DOMAIN_CPU).
  */
-static int
-i915_gem_object_set_domain_range(struct drm_gem_object *obj,
-                                uint64_t offset,
-                                uint64_t size,
-                                uint32_t read_domains,
-                                uint32_t write_domain)
+static void
+i915_gem_object_set_to_full_cpu_read_domain(struct drm_gem_object *obj)
 {
+       struct drm_device *dev = obj->dev;
        struct drm_i915_gem_object *obj_priv = obj->driver_private;
-       int ret, i;
 
-       if (obj->read_domains & I915_GEM_DOMAIN_CPU)
-               return 0;
+       if (!obj_priv->page_cpu_valid)
+               return;
 
-       if (read_domains != I915_GEM_DOMAIN_CPU ||
-           write_domain != 0)
-               return i915_gem_object_set_domain(obj,
-                                                 read_domains, write_domain);
+       /* If we're partially in the CPU read domain, finish moving it in.
+        */
+       if (obj->read_domains & I915_GEM_DOMAIN_CPU) {
+               int i;
 
-       /* Wait on any GPU rendering to the object to be flushed. */
+               for (i = 0; i <= (obj->size - 1) / PAGE_SIZE; i++) {
+                       if (obj_priv->page_cpu_valid[i])
+                               continue;
+                       drm_clflush_pages(obj_priv->page_list + i, 1);
+               }
+               drm_agp_chipset_flush(dev);
+       }
+
+       /* Free the page_cpu_valid mappings which are now stale, whether
+        * or not we've got I915_GEM_DOMAIN_CPU.
+        */
+       drm_free(obj_priv->page_cpu_valid, obj->size / PAGE_SIZE,
+                DRM_MEM_DRIVER);
+       obj_priv->page_cpu_valid = NULL;
+}
+
+/**
+ * Set the CPU read domain on a range of the object.
+ *
+ * The object ends up with I915_GEM_DOMAIN_CPU in its read flags although it's
+ * not entirely valid.  The page_cpu_valid member of the object flags which
+ * pages have been flushed, and will be respected by
+ * i915_gem_object_set_to_cpu_domain() if it's called on to get a valid mapping
+ * of the whole object.
+ *
+ * This function returns when the move is complete, including waiting on
+ * flushes to occur.
+ */
+static int
+i915_gem_object_set_cpu_read_domain_range(struct drm_gem_object *obj,
+                                         uint64_t offset, uint64_t size)
+{
+       struct drm_i915_gem_object *obj_priv = obj->driver_private;
+       int i, ret;
+
+       if (offset == 0 && size == obj->size)
+               return i915_gem_object_set_to_cpu_domain(obj, 0);
+
+       i915_gem_object_flush_gpu_write_domain(obj);
+       /* Wait on any GPU rendering and flushing to occur. */
        ret = i915_gem_object_wait_rendering(obj);
-       if (ret)
+       if (ret != 0)
                return ret;
+       i915_gem_object_flush_gtt_write_domain(obj);
 
+       /* If we're already fully in the CPU read domain, we're done. */
+       if (obj_priv->page_cpu_valid == NULL &&
+           (obj->read_domains & I915_GEM_DOMAIN_CPU) != 0)
+               return 0;
+
+       /* Otherwise, create/clear the per-page CPU read domain flag if we're
+        * newly adding I915_GEM_DOMAIN_CPU
+        */
        if (obj_priv->page_cpu_valid == NULL) {
                obj_priv->page_cpu_valid = drm_calloc(1, obj->size / PAGE_SIZE,
                                                      DRM_MEM_DRIVER);
-       }
+               if (obj_priv->page_cpu_valid == NULL)
+                       return -ENOMEM;
+       } else if ((obj->read_domains & I915_GEM_DOMAIN_CPU) == 0)
+               memset(obj_priv->page_cpu_valid, 0, obj->size / PAGE_SIZE);
 
        /* Flush the cache on any pages that are still invalid from the CPU's
         * perspective.
         */
-       for (i = offset / PAGE_SIZE; i <= (offset + size - 1) / PAGE_SIZE; i++) {
+       for (i = offset / PAGE_SIZE; i <= (offset + size - 1) / PAGE_SIZE;
+            i++) {
                if (obj_priv->page_cpu_valid[i])
                        continue;
 
@@ -1472,39 +1661,14 @@ i915_gem_object_set_domain_range(struct drm_gem_object *obj,
                obj_priv->page_cpu_valid[i] = 1;
        }
 
-       return 0;
-}
-
-/**
- * Once all of the objects have been set in the proper domain,
- * perform the necessary flush and invalidate operations.
- *
- * Returns the write domains flushed, for use in flush tracking.
- */
-static uint32_t
-i915_gem_dev_set_domain(struct drm_device *dev)
-{
-       uint32_t flush_domains = dev->flush_domains;
-
-       /*
-        * Now that all the buffers are synced to the proper domains,
-        * flush and invalidate the collected domains
+       /* It should now be out of any other write domains, and we can update
+        * the domain values for our changes.
         */
-       if (dev->invalidate_domains | dev->flush_domains) {
-#if WATCH_EXEC
-               DRM_INFO("%s: invalidate_domains %08x flush_domains %08x\n",
-                         __func__,
-                        dev->invalidate_domains,
-                        dev->flush_domains);
-#endif
-               i915_gem_flush(dev,
-                              dev->invalidate_domains,
-                              dev->flush_domains);
-               dev->invalidate_domains = 0;
-               dev->flush_domains = 0;
-       }
+       BUG_ON((obj->write_domain & ~I915_GEM_DOMAIN_CPU) != 0);
 
-       return flush_domains;
+       obj->read_domains |= I915_GEM_DOMAIN_CPU;
+
+       return 0;
 }
 
 /**
@@ -1585,6 +1749,18 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj,
                        return -EINVAL;
                }
 
+               if (reloc.write_domain & I915_GEM_DOMAIN_CPU ||
+                   reloc.read_domains & I915_GEM_DOMAIN_CPU) {
+                       DRM_ERROR("reloc with read/write CPU domains: "
+                                 "obj %p target %d offset %d "
+                                 "read %08x write %08x",
+                                 obj, reloc.target_handle,
+                                 (int) reloc.offset,
+                                 reloc.read_domains,
+                                 reloc.write_domain);
+                       return -EINVAL;
+               }
+
                if (reloc.write_domain && target_obj->pending_write_domain &&
                    reloc.write_domain != target_obj->pending_write_domain) {
                        DRM_ERROR("Write domain conflict: "
@@ -1625,19 +1801,11 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj,
                        continue;
                }
 
-               /* Now that we're going to actually write some data in,
-                * make sure that any rendering using this buffer's contents
-                * is completed.
-                */
-               i915_gem_object_wait_rendering(obj);
-
-               /* As we're writing through the gtt, flush
-                * any CPU writes before we write the relocations
-                */
-               if (obj->write_domain & I915_GEM_DOMAIN_CPU) {
-                       i915_gem_clflush_object(obj);
-                       drm_agp_chipset_flush(dev);
-                       obj->write_domain = 0;
+               ret = i915_gem_object_set_to_gtt_domain(obj, 1);
+               if (ret != 0) {
+                       drm_gem_object_unreference(target_obj);
+                       i915_gem_object_unpin(obj);
+                       return -EINVAL;
                }
 
                /* Map the page containing the relocation we're going to
@@ -1779,6 +1947,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
        int ret, i, pinned = 0;
        uint64_t exec_offset;
        uint32_t seqno, flush_domains;
+       int pin_tries;
 
 #if WATCH_EXEC
        DRM_INFO("buffers_ptr %d buffer_count %d len %08x\n",
@@ -1827,14 +1996,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
                return -EBUSY;
        }
 
-       /* Zero the gloabl flush/invalidate flags. These
-        * will be modified as each object is bound to the
-        * gtt
-        */
-       dev->invalidate_domains = 0;
-       dev->flush_domains = 0;
-
-       /* Look up object handles and perform the relocations */
+       /* Look up object handles */
        for (i = 0; i < args->buffer_count; i++) {
                object_list[i] = drm_gem_object_lookup(dev, file_priv,
                                                       exec_list[i].handle);
@@ -1844,17 +2006,39 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
                        ret = -EBADF;
                        goto err;
                }
+       }
 
-               object_list[i]->pending_read_domains = 0;
-               object_list[i]->pending_write_domain = 0;
-               ret = i915_gem_object_pin_and_relocate(object_list[i],
-                                                      file_priv,
-                                                      &exec_list[i]);
-               if (ret) {
-                       DRM_ERROR("object bind and relocate failed %d\n", ret);
+       /* Pin and relocate */
+       for (pin_tries = 0; ; pin_tries++) {
+               ret = 0;
+               for (i = 0; i < args->buffer_count; i++) {
+                       object_list[i]->pending_read_domains = 0;
+                       object_list[i]->pending_write_domain = 0;
+                       ret = i915_gem_object_pin_and_relocate(object_list[i],
+                                                              file_priv,
+                                                              &exec_list[i]);
+                       if (ret)
+                               break;
+                       pinned = i + 1;
+               }
+               /* success */
+               if (ret == 0)
+                       break;
+
+               /* error other than GTT full, or we've already tried again */
+               if (ret != -ENOMEM || pin_tries >= 1) {
+                       DRM_ERROR("Failed to pin buffers %d\n", ret);
                        goto err;
                }
-               pinned = i + 1;
+
+               /* unpin all of our buffers */
+               for (i = 0; i < pinned; i++)
+                       i915_gem_object_unpin(object_list[i]);
+
+               /* evict everyone we can from the aperture */
+               ret = i915_gem_evict_everything(dev);
+               if (ret)
+                       goto err;
        }
 
        /* Set the pending read domains for the batch buffer to COMMAND */
@@ -1864,21 +2048,37 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 
        i915_verify_inactive(dev, __FILE__, __LINE__);
 
+       /* Zero the global flush/invalidate flags. These
+        * will be modified as new domains are computed
+        * for each object
+        */
+       dev->invalidate_domains = 0;
+       dev->flush_domains = 0;
+
        for (i = 0; i < args->buffer_count; i++) {
                struct drm_gem_object *obj = object_list[i];
 
-               /* make sure all previous memory operations have passed */
-               ret = i915_gem_object_set_domain(obj,
-                                                obj->pending_read_domains,
-                                                obj->pending_write_domain);
-               if (ret)
-                       goto err;
+               /* Compute new gpu domains and update invalidate/flush */
+               i915_gem_object_set_to_gpu_domain(obj,
+                                                 obj->pending_read_domains,
+                                                 obj->pending_write_domain);
        }
 
        i915_verify_inactive(dev, __FILE__, __LINE__);
 
-       /* Flush/invalidate caches and chipset buffer */
-       flush_domains = i915_gem_dev_set_domain(dev);
+       if (dev->invalidate_domains | dev->flush_domains) {
+#if WATCH_EXEC
+               DRM_INFO("%s: invalidate_domains %08x flush_domains %08x\n",
+                         __func__,
+                        dev->invalidate_domains,
+                        dev->flush_domains);
+#endif
+               i915_gem_flush(dev,
+                              dev->invalidate_domains,
+                              dev->flush_domains);
+               if (dev->flush_domains)
+                       (void)i915_add_request(dev, dev->flush_domains);
+       }
 
        i915_verify_inactive(dev, __FILE__, __LINE__);
 
@@ -1898,8 +2098,6 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
                              ~0);
 #endif
 
-       (void)i915_add_request(dev, flush_domains);
-
        /* Exec the batchbuffer */
        ret = i915_dispatch_gem_execbuffer(dev, args, exec_offset);
        if (ret) {
@@ -1927,10 +2125,8 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
        i915_file_priv->mm.last_gem_seqno = seqno;
        for (i = 0; i < args->buffer_count; i++) {
                struct drm_gem_object *obj = object_list[i];
-               struct drm_i915_gem_object *obj_priv = obj->driver_private;
 
-               i915_gem_object_move_to_active(obj);
-               obj_priv->last_rendering_seqno = seqno;
+               i915_gem_object_move_to_active(obj, seqno);
 #if WATCH_LRU
                DRM_INFO("%s: move to exec list %p\n", __func__, obj);
 #endif
@@ -2061,11 +2257,7 @@ i915_gem_pin_ioctl(struct drm_device *dev, void *data,
        /* XXX - flush the CPU caches for pinned objects
         * as the X server doesn't manage domains yet
         */
-       if (obj->write_domain & I915_GEM_DOMAIN_CPU) {
-               i915_gem_clflush_object(obj);
-               drm_agp_chipset_flush(dev);
-               obj->write_domain = 0;
-       }
+       i915_gem_object_flush_cpu_write_domain(obj);
        args->offset = obj_priv->gtt_offset;
        drm_gem_object_unreference(obj);
        mutex_unlock(&dev->struct_mutex);
@@ -2167,29 +2359,6 @@ void i915_gem_free_object(struct drm_gem_object *obj)
        drm_free(obj->driver_private, 1, DRM_MEM_DRIVER);
 }
 
-static int
-i915_gem_set_domain(struct drm_gem_object *obj,
-                   struct drm_file *file_priv,
-                   uint32_t read_domains,
-                   uint32_t write_domain)
-{
-       struct drm_device *dev = obj->dev;
-       int ret;
-       uint32_t flush_domains;
-
-       BUG_ON(!mutex_is_locked(&dev->struct_mutex));
-
-       ret = i915_gem_object_set_domain(obj, read_domains, write_domain);
-       if (ret)
-               return ret;
-       flush_domains = i915_gem_dev_set_domain(obj->dev);
-
-       if (flush_domains & ~(I915_GEM_DOMAIN_CPU|I915_GEM_DOMAIN_GTT))
-               (void) i915_add_request(dev, flush_domains);
-
-       return 0;
-}
-
 /** Unbinds all objects that are on the given buffer list. */
 static int
 i915_gem_evict_from_list(struct drm_device *dev, struct list_head *head)
index 93de15b4c9a722749d69d417d88cc1aba49d8280..e8d5abe1250ee14404bf8f9da972e34fd74ff6f0 100644 (file)
@@ -166,10 +166,9 @@ static int i915_gem_request_info(char *buf, char **start, off_t offset,
        list_for_each_entry(gem_request, &dev_priv->mm.request_list,
                            list)
        {
-               DRM_PROC_PRINT("    %d @ %d %08x\n",
+               DRM_PROC_PRINT("    %d @ %d\n",
                               gem_request->seqno,
-                              (int) (jiffies - gem_request->emitted_jiffies),
-                              gem_request->flush_domains);
+                              (int) (jiffies - gem_request->emitted_jiffies));
        }
        if (len > request + offset)
                return request;
index e8b85ac4ca041159553f5211e0bc3cbe58f03cc8..a8cb69469c641ff28f4d57df20b51d0dd5772f7c 100644 (file)
@@ -119,9 +119,10 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
                            dcc & DCC_CHANNEL_XOR_DISABLE) {
                                swizzle_x = I915_BIT_6_SWIZZLE_9_10;
                                swizzle_y = I915_BIT_6_SWIZZLE_9;
-                       } else if (IS_I965GM(dev) || IS_GM45(dev)) {
-                               /* GM965 only does bit 11-based channel
-                                * randomization
+                       } else if ((IS_I965GM(dev) || IS_GM45(dev)) &&
+                                  (dcc & DCC_CHANNEL_XOR_BIT_17) == 0) {
+                               /* GM965/GM45 does either bit 11 or bit 17
+                                * swizzling.
                                 */
                                swizzle_x = I915_BIT_6_SWIZZLE_9_10_11;
                                swizzle_y = I915_BIT_6_SWIZZLE_9_11;
index 0e476eba36e662b75ce24db2b7bb52162699f9e6..9d24aaeb8a453d76e2d23e2223af3bfc527a74fa 100644 (file)
 #define DCC_ADDRESSING_MODE_DUAL_CHANNEL_INTERLEAVED   (2 << 0)
 #define DCC_ADDRESSING_MODE_MASK                       (3 << 0)
 #define DCC_CHANNEL_XOR_DISABLE                                (1 << 10)
+#define DCC_CHANNEL_XOR_BIT_17                         (1 << 9)
 
 /** 965 MCH register controlling DRAM channel configuration */
 #define C0DRB3                 0x10206