diff options
Diffstat (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c')
| -rw-r--r-- | drivers/gpu/drm/i915/intel_ringbuffer.c | 213 |
1 files changed, 176 insertions, 37 deletions
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 2346b920bd86..1d5d613eb6be 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -318,6 +318,7 @@ gen7_render_ring_flush(struct intel_ring_buffer *ring, * TLB invalidate requires a post-sync write. */ flags |= PIPE_CONTROL_QW_WRITE; + flags |= PIPE_CONTROL_GLOBAL_GTT_IVB; /* Workaround: we must issue a pipe_control with CS-stall bit * set before a pipe_control command that has the state cache @@ -331,7 +332,7 @@ gen7_render_ring_flush(struct intel_ring_buffer *ring, intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4)); intel_ring_emit(ring, flags); - intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT); + intel_ring_emit(ring, scratch_addr); intel_ring_emit(ring, 0); intel_ring_advance(ring); @@ -467,6 +468,9 @@ init_pipe_control(struct intel_ring_buffer *ring) if (pc->cpu_page == NULL) goto err_unpin; + DRM_DEBUG_DRIVER("%s pipe control offset: 0x%08x\n", + ring->name, pc->gtt_offset); + pc->obj = obj; ring->private = pc; return 0; @@ -505,13 +509,25 @@ static int init_render_ring(struct intel_ring_buffer *ring) struct drm_i915_private *dev_priv = dev->dev_private; int ret = init_ring_common(ring); - if (INTEL_INFO(dev)->gen > 3) { + if (INTEL_INFO(dev)->gen > 3) I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(VS_TIMER_DISPATCH)); - if (IS_GEN7(dev)) - I915_WRITE(GFX_MODE_GEN7, - _MASKED_BIT_DISABLE(GFX_TLB_INVALIDATE_ALWAYS) | - _MASKED_BIT_ENABLE(GFX_REPLAY_MODE)); - } + + /* We need to disable the AsyncFlip performance optimisations in order + * to use MI_WAIT_FOR_EVENT within the CS. It should already be + * programmed to '1' on all products. + */ + if (INTEL_INFO(dev)->gen >= 6) + I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE)); + + /* Required for the hardware to program scanline values for waiting */ + if (INTEL_INFO(dev)->gen == 6) + I915_WRITE(GFX_MODE, + _MASKED_BIT_ENABLE(GFX_TLB_INVALIDATE_ALWAYS)); + + if (IS_GEN7(dev)) + I915_WRITE(GFX_MODE_GEN7, + _MASKED_BIT_DISABLE(GFX_TLB_INVALIDATE_ALWAYS) | + _MASKED_BIT_ENABLE(GFX_REPLAY_MODE)); if (INTEL_INFO(dev)->gen >= 5) { ret = init_pipe_control(ring); @@ -547,9 +563,14 @@ static int init_render_ring(struct intel_ring_buffer *ring) static void render_ring_cleanup(struct intel_ring_buffer *ring) { + struct drm_device *dev = ring->dev; + if (!ring->private) return; + if (HAS_BROKEN_CS_TLB(dev)) + drm_gem_object_unreference(to_gem_object(ring->private)); + cleanup_pipe_control(ring); } @@ -596,6 +617,13 @@ gen6_add_request(struct intel_ring_buffer *ring) return 0; } +static inline bool i915_gem_has_seqno_wrapped(struct drm_device *dev, + u32 seqno) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + return dev_priv->last_seqno < seqno; +} + /** * intel_ring_sync - sync the waiter to the signaller on seqno * @@ -626,11 +654,20 @@ gen6_ring_sync(struct intel_ring_buffer *waiter, if (ret) return ret; - intel_ring_emit(waiter, - dw1 | signaller->semaphore_register[waiter->id]); - intel_ring_emit(waiter, seqno); - intel_ring_emit(waiter, 0); - intel_ring_emit(waiter, MI_NOOP); + /* If seqno wrap happened, omit the wait with no-ops */ + if (likely(!i915_gem_has_seqno_wrapped(waiter->dev, seqno))) { + intel_ring_emit(waiter, + dw1 | + signaller->semaphore_register[waiter->id]); + intel_ring_emit(waiter, seqno); + intel_ring_emit(waiter, 0); + intel_ring_emit(waiter, MI_NOOP); + } else { + intel_ring_emit(waiter, MI_NOOP); + intel_ring_emit(waiter, MI_NOOP); + intel_ring_emit(waiter, MI_NOOP); + intel_ring_emit(waiter, MI_NOOP); + } intel_ring_advance(waiter); return 0; @@ -711,6 +748,12 @@ ring_get_seqno(struct intel_ring_buffer *ring, bool lazy_coherency) return intel_read_status_page(ring, I915_GEM_HWS_INDEX); } +static void +ring_set_seqno(struct intel_ring_buffer *ring, u32 seqno) +{ + intel_write_status_page(ring, I915_GEM_HWS_INDEX, seqno); +} + static u32 pc_render_get_seqno(struct intel_ring_buffer *ring, bool lazy_coherency) { @@ -718,6 +761,13 @@ pc_render_get_seqno(struct intel_ring_buffer *ring, bool lazy_coherency) return pc->cpu_page[0]; } +static void +pc_render_set_seqno(struct intel_ring_buffer *ring, u32 seqno) +{ + struct pipe_control *pc = ring->private; + pc->cpu_page[0] = seqno; +} + static bool gen5_ring_get_irq(struct intel_ring_buffer *ring) { @@ -969,6 +1019,8 @@ i965_dispatch_execbuffer(struct intel_ring_buffer *ring, return 0; } +/* Just userspace ABI convention to limit the wa batch bo to a resonable size */ +#define I830_BATCH_LIMIT (256*1024) static int i830_dispatch_execbuffer(struct intel_ring_buffer *ring, u32 offset, u32 len, @@ -976,15 +1028,47 @@ i830_dispatch_execbuffer(struct intel_ring_buffer *ring, { int ret; - ret = intel_ring_begin(ring, 4); - if (ret) - return ret; + if (flags & I915_DISPATCH_PINNED) { + ret = intel_ring_begin(ring, 4); + if (ret) + return ret; - intel_ring_emit(ring, MI_BATCH_BUFFER); - intel_ring_emit(ring, offset | (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE)); - intel_ring_emit(ring, offset + len - 8); - intel_ring_emit(ring, 0); - intel_ring_advance(ring); + intel_ring_emit(ring, MI_BATCH_BUFFER); + intel_ring_emit(ring, offset | (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE)); + intel_ring_emit(ring, offset + len - 8); + intel_ring_emit(ring, MI_NOOP); + intel_ring_advance(ring); + } else { + struct drm_i915_gem_object *obj = ring->private; + u32 cs_offset = obj->gtt_offset; + + if (len > I830_BATCH_LIMIT) + return -ENOSPC; + + ret = intel_ring_begin(ring, 9+3); + if (ret) + return ret; + /* Blit the batch (which has now all relocs applied) to the stable batch + * scratch bo area (so that the CS never stumbles over its tlb + * invalidation bug) ... */ + intel_ring_emit(ring, XY_SRC_COPY_BLT_CMD | + XY_SRC_COPY_BLT_WRITE_ALPHA | + XY_SRC_COPY_BLT_WRITE_RGB); + intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_GXCOPY | 4096); + intel_ring_emit(ring, 0); + intel_ring_emit(ring, (DIV_ROUND_UP(len, 4096) << 16) | 1024); + intel_ring_emit(ring, cs_offset); + intel_ring_emit(ring, 0); + intel_ring_emit(ring, 4096); + intel_ring_emit(ring, offset); + intel_ring_emit(ring, MI_FLUSH); + + /* ... and execute it. */ + intel_ring_emit(ring, MI_BATCH_BUFFER); + intel_ring_emit(ring, cs_offset | (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE)); + intel_ring_emit(ring, cs_offset + len - 8); + intel_ring_advance(ring); + } return 0; } @@ -1113,7 +1197,11 @@ static int intel_init_ring_buffer(struct drm_device *dev, return ret; } - obj = i915_gem_alloc_object(dev, ring->size); + obj = NULL; + if (!HAS_LLC(dev)) + obj = i915_gem_object_create_stolen(dev, ring->size); + if (obj == NULL) + obj = i915_gem_alloc_object(dev, ring->size); if (obj == NULL) { DRM_ERROR("Failed to allocate ringbuffer\n"); ret = -ENOMEM; @@ -1131,7 +1219,7 @@ static int intel_init_ring_buffer(struct drm_device *dev, goto err_unpin; ring->virtual_start = - ioremap_wc(dev_priv->mm.gtt->gma_bus_addr + obj->gtt_offset, + ioremap_wc(dev_priv->gtt.mappable_base + obj->gtt_offset, ring->size); if (ring->virtual_start == NULL) { DRM_ERROR("Failed to map ringbuffer.\n"); @@ -1297,7 +1385,8 @@ static int ring_wait_for_space(struct intel_ring_buffer *ring, int n) msleep(1); - ret = i915_gem_check_wedge(dev_priv, dev_priv->mm.interruptible); + ret = i915_gem_check_wedge(&dev_priv->gpu_error, + dev_priv->mm.interruptible); if (ret) return ret; } while (!time_after(jiffies, end)); @@ -1359,14 +1448,35 @@ intel_ring_alloc_seqno(struct intel_ring_buffer *ring) return i915_gem_get_seqno(ring->dev, &ring->outstanding_lazy_request); } +static int __intel_ring_begin(struct intel_ring_buffer *ring, + int bytes) +{ + int ret; + + if (unlikely(ring->tail + bytes > ring->effective_size)) { + ret = intel_wrap_ring_buffer(ring); + if (unlikely(ret)) + return ret; + } + + if (unlikely(ring->space < bytes)) { + ret = ring_wait_for_space(ring, bytes); + if (unlikely(ret)) + return ret; + } + + ring->space -= bytes; + return 0; +} + int intel_ring_begin(struct intel_ring_buffer *ring, int num_dwords) { drm_i915_private_t *dev_priv = ring->dev->dev_private; - int n = 4*num_dwords; int ret; - ret = i915_gem_check_wedge(dev_priv, dev_priv->mm.interruptible); + ret = i915_gem_check_wedge(&dev_priv->gpu_error, + dev_priv->mm.interruptible); if (ret) return ret; @@ -1375,20 +1485,21 @@ int intel_ring_begin(struct intel_ring_buffer *ring, if (ret) return ret; - if (unlikely(ring->tail + n > ring->effective_size)) { - ret = intel_wrap_ring_buffer(ring); - if (unlikely(ret)) - return ret; - } + return __intel_ring_begin(ring, num_dwords * sizeof(uint32_t)); +} - if (unlikely(ring->space < n)) { - ret = ring_wait_for_space(ring, n); - if (unlikely(ret)) - return ret; +void intel_ring_init_seqno(struct intel_ring_buffer *ring, u32 seqno) +{ + struct drm_i915_private *dev_priv = ring->dev->dev_private; + + BUG_ON(ring->outstanding_lazy_request); + + if (INTEL_INFO(ring->dev)->gen >= 6) { + I915_WRITE(RING_SYNC_0(ring->mmio_base), 0); + I915_WRITE(RING_SYNC_1(ring->mmio_base), 0); } - ring->space -= n; - return 0; + ring->set_seqno(ring, seqno); } void intel_ring_advance(struct intel_ring_buffer *ring) @@ -1396,7 +1507,7 @@ void intel_ring_advance(struct intel_ring_buffer *ring) struct drm_i915_private *dev_priv = ring->dev->dev_private; ring->tail &= ring->size - 1; - if (dev_priv->stop_rings & intel_ring_flag(ring)) + if (dev_priv->gpu_error.stop_rings & intel_ring_flag(ring)) return; ring->write_tail(ring, ring->tail); } @@ -1553,6 +1664,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->irq_put = gen6_ring_put_irq; ring->irq_enable_mask = GT_USER_INTERRUPT; ring->get_seqno = gen6_ring_get_seqno; + ring->set_seqno = ring_set_seqno; ring->sync_to = gen6_ring_sync; ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_INVALID; ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_RV; @@ -1563,6 +1675,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->add_request = pc_render_add_request; ring->flush = gen4_render_ring_flush; ring->get_seqno = pc_render_get_seqno; + ring->set_seqno = pc_render_set_seqno; ring->irq_get = gen5_ring_get_irq; ring->irq_put = gen5_ring_put_irq; ring->irq_enable_mask = GT_USER_INTERRUPT | GT_PIPE_NOTIFY; @@ -1573,6 +1686,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) else ring->flush = gen4_render_ring_flush; ring->get_seqno = ring_get_seqno; + ring->set_seqno = ring_set_seqno; if (IS_GEN2(dev)) { ring->irq_get = i8xx_ring_get_irq; ring->irq_put = i8xx_ring_put_irq; @@ -1596,6 +1710,27 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->init = init_render_ring; ring->cleanup = render_ring_cleanup; + /* Workaround batchbuffer to combat CS tlb bug. */ + if (HAS_BROKEN_CS_TLB(dev)) { + struct drm_i915_gem_object *obj; + int ret; + + obj = i915_gem_alloc_object(dev, I830_BATCH_LIMIT); + if (obj == NULL) { + DRM_ERROR("Failed to allocate batch bo\n"); + return -ENOMEM; + } + + ret = i915_gem_object_pin(obj, 0, true, false); + if (ret != 0) { + drm_gem_object_unreference(&obj->base); + DRM_ERROR("Failed to ping batch bo\n"); + return ret; + } + + ring->private = obj; + } + return intel_init_ring_buffer(dev, ring); } @@ -1623,6 +1758,7 @@ int intel_render_ring_init_dri(struct drm_device *dev, u64 start, u32 size) else ring->flush = gen4_render_ring_flush; ring->get_seqno = ring_get_seqno; + ring->set_seqno = ring_set_seqno; if (IS_GEN2(dev)) { ring->irq_get = i8xx_ring_get_irq; ring->irq_put = i8xx_ring_put_irq; @@ -1683,6 +1819,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev) ring->flush = gen6_ring_flush; ring->add_request = gen6_add_request; ring->get_seqno = gen6_ring_get_seqno; + ring->set_seqno = ring_set_seqno; ring->irq_enable_mask = GEN6_BSD_USER_INTERRUPT; ring->irq_get = gen6_ring_get_irq; ring->irq_put = gen6_ring_put_irq; @@ -1698,6 +1835,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev) ring->flush = bsd_ring_flush; ring->add_request = i9xx_add_request; ring->get_seqno = ring_get_seqno; + ring->set_seqno = ring_set_seqno; if (IS_GEN5(dev)) { ring->irq_enable_mask = GT_BSD_USER_INTERRUPT; ring->irq_get = gen5_ring_get_irq; @@ -1727,6 +1865,7 @@ int intel_init_blt_ring_buffer(struct drm_device *dev) ring->flush = blt_ring_flush; ring->add_request = gen6_add_request; ring->get_seqno = gen6_ring_get_seqno; + ring->set_seqno = ring_set_seqno; ring->irq_enable_mask = GEN6_BLITTER_USER_INTERRUPT; ring->irq_get = gen6_ring_get_irq; ring->irq_put = gen6_ring_put_irq; |