From 3b2cc8ab64bf5d643ac9a917653e0cb6fce32ffc Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Wed, 11 Jun 2014 16:17:16 +0100 Subject: drm/i915/bdw: Do not write the Semaphore Sync Registers in GEN8+ These do not exist anymore. Spotted while reading through intel_ringbuffer.c Signed-off-by: Oscar Mateo Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 279488addf3..0eaaaec78ba 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1746,14 +1746,15 @@ int intel_ring_cacheline_align(struct intel_engine_cs *ring) void intel_ring_init_seqno(struct intel_engine_cs *ring, u32 seqno) { - struct drm_i915_private *dev_priv = ring->dev->dev_private; + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; BUG_ON(ring->outstanding_lazy_seqno); - if (INTEL_INFO(ring->dev)->gen >= 6) { + if (INTEL_INFO(dev)->gen == 6 || INTEL_INFO(dev)->gen == 7) { I915_WRITE(RING_SYNC_0(ring->mmio_base), 0); I915_WRITE(RING_SYNC_1(ring->mmio_base), 0); - if (HAS_VEBOX(ring->dev)) + if (HAS_VEBOX(dev)) I915_WRITE(RING_SYNC_2(ring->mmio_base), 0); } -- cgit v1.2.3-70-g09d2 From 24f3a8cf7766e52a087904b4346794c7b410f957 Mon Sep 17 00:00:00 2001 From: Akash Goel Date: Tue, 17 Jun 2014 10:59:42 +0530 Subject: drm/i915: Added write-enable pte bit supportt This adds support for a write-enable bit in the entry of GTT. This is handled via a read-only flag in the GEM buffer object which is then used to see how to set the bit when writing the GTT entries. Currently by default the Batch buffer & Ring buffers are marked as read only. v2: Moved the pte override code for read-only bit to 'byt_pte_encode'. (Chris) Fixed the issue of leaving 'gt_old_ro' as unused. (Chris) v3: Removed the 'gt_old_ro' field, now setting RO bit only for Ring Buffers(Daniel). v4: Added a new 'flags' parameter to all the pte(gen6) encode & insert_entries functions, in lieu of overloading the cache_level enum (Daniel). v5: Removed the superfluous VLV check & changed the definition location of PTE_READ_ONLY flag (Imre) Reviewed-by: Imre Deak Signed-off-by: Akash Goel Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_drv.h | 6 +++++ drivers/gpu/drm/i915/i915_gem_gtt.c | 48 ++++++++++++++++++++------------- drivers/gpu/drm/i915/i915_gem_gtt.h | 5 ++-- drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++ 4 files changed, 41 insertions(+), 21 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index faef035f5a4..c5b02159898 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1660,6 +1660,12 @@ struct drm_i915_gem_object { unsigned int pin_mappable:1; unsigned int pin_display:1; + /* + * Is the object to be mapped as read-only to the GPU + * Only honoured if hardware has relevant pte bit + */ + unsigned long gt_ro:1; + /* * Is the GPU currently using a fence to access this buffer, */ diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 6f5d5a1e7c3..743ddfcf269 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -116,7 +116,7 @@ static inline gen8_ppgtt_pde_t gen8_pde_encode(struct drm_device *dev, static gen6_gtt_pte_t snb_pte_encode(dma_addr_t addr, enum i915_cache_level level, - bool valid) + bool valid, u32 unused) { gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0; pte |= GEN6_PTE_ADDR_ENCODE(addr); @@ -138,7 +138,7 @@ static gen6_gtt_pte_t snb_pte_encode(dma_addr_t addr, static gen6_gtt_pte_t ivb_pte_encode(dma_addr_t addr, enum i915_cache_level level, - bool valid) + bool valid, u32 unused) { gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0; pte |= GEN6_PTE_ADDR_ENCODE(addr); @@ -162,7 +162,7 @@ static gen6_gtt_pte_t ivb_pte_encode(dma_addr_t addr, static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr, enum i915_cache_level level, - bool valid) + bool valid, u32 flags) { gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0; pte |= GEN6_PTE_ADDR_ENCODE(addr); @@ -170,7 +170,8 @@ static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr, /* Mark the page as writeable. Other platforms don't have a * setting for read-only/writable, so this matches that behavior. */ - pte |= BYT_PTE_WRITEABLE; + if (!(flags & PTE_READ_ONLY)) + pte |= BYT_PTE_WRITEABLE; if (level != I915_CACHE_NONE) pte |= BYT_PTE_SNOOPED_BY_CPU_CACHES; @@ -180,7 +181,7 @@ static gen6_gtt_pte_t byt_pte_encode(dma_addr_t addr, static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr, enum i915_cache_level level, - bool valid) + bool valid, u32 unused) { gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0; pte |= HSW_PTE_ADDR_ENCODE(addr); @@ -193,7 +194,7 @@ static gen6_gtt_pte_t hsw_pte_encode(dma_addr_t addr, static gen6_gtt_pte_t iris_pte_encode(dma_addr_t addr, enum i915_cache_level level, - bool valid) + bool valid, u32 unused) { gen6_gtt_pte_t pte = valid ? GEN6_PTE_VALID : 0; pte |= HSW_PTE_ADDR_ENCODE(addr); @@ -307,7 +308,7 @@ static void gen8_ppgtt_clear_range(struct i915_address_space *vm, static void gen8_ppgtt_insert_entries(struct i915_address_space *vm, struct sg_table *pages, uint64_t start, - enum i915_cache_level cache_level) + enum i915_cache_level cache_level, u32 unused) { struct i915_hw_ppgtt *ppgtt = container_of(vm, struct i915_hw_ppgtt, base); @@ -645,7 +646,7 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m) uint32_t pd_entry; int pte, pde; - scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true); + scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true, 0); pd_addr = (gen6_gtt_pte_t __iomem *)dev_priv->gtt.gsm + ppgtt->pd_offset / sizeof(gen6_gtt_pte_t); @@ -947,7 +948,7 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm, unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES; unsigned last_pte, i; - scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true); + scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, true, 0); while (num_entries) { last_pte = first_pte + num_entries; @@ -970,7 +971,7 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm, static void gen6_ppgtt_insert_entries(struct i915_address_space *vm, struct sg_table *pages, uint64_t start, - enum i915_cache_level cache_level) + enum i915_cache_level cache_level, u32 flags) { struct i915_hw_ppgtt *ppgtt = container_of(vm, struct i915_hw_ppgtt, base); @@ -987,7 +988,8 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm, pt_vaddr[act_pte] = vm->pte_encode(sg_page_iter_dma_address(&sg_iter), - cache_level, true); + cache_level, true, flags); + if (++act_pte == I915_PPGTT_PT_ENTRIES) { kunmap_atomic(pt_vaddr); pt_vaddr = NULL; @@ -1224,8 +1226,12 @@ ppgtt_bind_vma(struct i915_vma *vma, enum i915_cache_level cache_level, u32 flags) { + /* Currently applicable only to VLV */ + if (vma->obj->gt_ro) + flags |= PTE_READ_ONLY; + vma->vm->insert_entries(vma->vm, vma->obj->pages, vma->node.start, - cache_level); + cache_level, flags); } static void ppgtt_unbind_vma(struct i915_vma *vma) @@ -1400,7 +1406,7 @@ static inline void gen8_set_pte(void __iomem *addr, gen8_gtt_pte_t pte) static void gen8_ggtt_insert_entries(struct i915_address_space *vm, struct sg_table *st, uint64_t start, - enum i915_cache_level level) + enum i915_cache_level level, u32 unused) { struct drm_i915_private *dev_priv = vm->dev->dev_private; unsigned first_entry = start >> PAGE_SHIFT; @@ -1446,7 +1452,7 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, static void gen6_ggtt_insert_entries(struct i915_address_space *vm, struct sg_table *st, uint64_t start, - enum i915_cache_level level) + enum i915_cache_level level, u32 flags) { struct drm_i915_private *dev_priv = vm->dev->dev_private; unsigned first_entry = start >> PAGE_SHIFT; @@ -1458,7 +1464,7 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) { addr = sg_page_iter_dma_address(&sg_iter); - iowrite32(vm->pte_encode(addr, level, true), >t_entries[i]); + iowrite32(vm->pte_encode(addr, level, true, flags), >t_entries[i]); i++; } @@ -1470,7 +1476,7 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, */ if (i != 0) WARN_ON(readl(>t_entries[i-1]) != - vm->pte_encode(addr, level, true)); + vm->pte_encode(addr, level, true, flags)); /* This next bit makes the above posting read even more important. We * want to flush the TLBs only after we're certain all the PTE updates @@ -1524,7 +1530,7 @@ static void gen6_ggtt_clear_range(struct i915_address_space *vm, first_entry, num_entries, max_entries)) num_entries = max_entries; - scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, use_scratch); + scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC, use_scratch, 0); for (i = 0; i < num_entries; i++) iowrite32(scratch_pte, >t_base[i]); @@ -1573,6 +1579,10 @@ static void ggtt_bind_vma(struct i915_vma *vma, struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj = vma->obj; + /* Currently applicable only to VLV */ + if (obj->gt_ro) + flags |= PTE_READ_ONLY; + /* If there is no aliasing PPGTT, or the caller needs a global mapping, * or we have a global mapping already but the cacheability flags have * changed, set the global PTEs. @@ -1589,7 +1599,7 @@ static void ggtt_bind_vma(struct i915_vma *vma, (cache_level != obj->cache_level)) { vma->vm->insert_entries(vma->vm, obj->pages, vma->node.start, - cache_level); + cache_level, flags); obj->has_global_gtt_mapping = 1; } } @@ -1601,7 +1611,7 @@ static void ggtt_bind_vma(struct i915_vma *vma, appgtt->base.insert_entries(&appgtt->base, vma->obj->pages, vma->node.start, - cache_level); + cache_level, flags); vma->obj->has_aliasing_ppgtt_mapping = 1; } } diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index 1b96a06be3c..8d6f7c18c40 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -154,6 +154,7 @@ struct i915_vma { void (*unbind_vma)(struct i915_vma *vma); /* Map an object into an address space with the given cache flags. */ #define GLOBAL_BIND (1<<0) +#define PTE_READ_ONLY (1<<1) void (*bind_vma)(struct i915_vma *vma, enum i915_cache_level cache_level, u32 flags); @@ -197,7 +198,7 @@ struct i915_address_space { /* FIXME: Need a more generic return type */ gen6_gtt_pte_t (*pte_encode)(dma_addr_t addr, enum i915_cache_level level, - bool valid); /* Create a valid PTE */ + bool valid, u32 flags); /* Create a valid PTE */ void (*clear_range)(struct i915_address_space *vm, uint64_t start, uint64_t length, @@ -205,7 +206,7 @@ struct i915_address_space { void (*insert_entries)(struct i915_address_space *vm, struct sg_table *st, uint64_t start, - enum i915_cache_level cache_level); + enum i915_cache_level cache_level, u32 flags); void (*cleanup)(struct i915_address_space *vm); }; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 0eaaaec78ba..b96edaf905d 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1397,6 +1397,9 @@ static int allocate_ring_buffer(struct intel_engine_cs *ring) if (obj == NULL) return -ENOMEM; + /* mark ring buffers as read-only from GPU side by default */ + obj->gt_ro = 1; + ret = i915_gem_obj_ggtt_pin(obj, PAGE_SIZE, PIN_MAPPABLE); if (ret) goto err_unref; -- cgit v1.2.3-70-g09d2 From 9c33baa6b3bbb01c1a88dceba986b20e6642cf31 Mon Sep 17 00:00:00 2001 From: Konrad Zapalowicz Date: Thu, 19 Jun 2014 19:07:15 +0200 Subject: drivers/i915: Fix unnoticed failure of init_ring_common() This commit add check for return value of init_ring_common() in the init_render_ring(). Now, when failure is detected the error code is propagated to the caller instead of being ignored. Signed-off-by: Konrad Zapalowicz Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index b96edaf905d..2faef2605e9 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -604,6 +604,8 @@ static int init_render_ring(struct intel_engine_cs *ring) struct drm_device *dev = ring->dev; struct drm_i915_private *dev_priv = dev->dev_private; int ret = init_ring_common(ring); + if (ret) + return ret; /* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */ if (INTEL_INFO(dev)->gen >= 4 && INTEL_INFO(dev)->gen < 7) -- cgit v1.2.3-70-g09d2 From f7b64236854b0862e9967090f227cf08d81c71f4 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 1 Jul 2014 02:41:36 -0700 Subject: drm/i915: Fix VCS2's ring name. It just fix a typo. v2: removing underscore to let this like all other ring names (Oscar) Cc: Oscar Mateo Reviewed-by (v1): Ben Widawsky Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 2faef2605e9..22c2b9a217b 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -2224,7 +2224,7 @@ int intel_init_bsd2_ring_buffer(struct drm_device *dev) return -EINVAL; } - ring->name = "bds2_ring"; + ring->name = "bsd2 ring"; ring->id = VCS2; ring->write_tail = ring_write_tail; -- cgit v1.2.3-70-g09d2 From 707d9cf9935cfba2d62dd80dc01dc5dc4530d4ca Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 30 Jun 2014 09:53:36 -0700 Subject: drm/i915: gen specific ring init Gen8 has already had some differentiation with how it handles rings. Semaphores bring yet more differences, and now is as good a time as any to do the split. Also, since gen8 doesn't actually use semaphores up until this point, put the proper "NULL" values in for the mbox info. v2: v1 had a stale commit message v3: Move everything in the is_semaphore_enabled() check v4: VCS2 rebase Remove double assignment of signal in render ring (Ville) v5: Adding missed VCS2 signal init on gen8+ (Rodrigo) Reviewed-by: Rodrigo Vivi Signed-off-by: Ben Widawsky Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 227 +++++++++++++++++++++----------- 1 file changed, 151 insertions(+), 76 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 22c2b9a217b..7ca1b946d8d 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -729,7 +729,11 @@ gen6_add_request(struct intel_engine_cs *ring) { int ret; - ret = ring->semaphore.signal(ring, 4); + if (ring->semaphore.signal) + ret = ring->semaphore.signal(ring, 4); + else + ret = intel_ring_begin(ring, 4); + if (ret) return ret; @@ -1952,40 +1956,59 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->id = RCS; ring->mmio_base = RENDER_RING_BASE; - if (INTEL_INFO(dev)->gen >= 6) { + if (INTEL_INFO(dev)->gen >= 8) { + ring->add_request = gen6_add_request; + ring->flush = gen8_render_ring_flush; + ring->irq_get = gen8_ring_get_irq; + ring->irq_put = gen8_ring_put_irq; + ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT; + ring->get_seqno = gen6_ring_get_seqno; + ring->set_seqno = ring_set_seqno; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } + } else if (INTEL_INFO(dev)->gen >= 6) { ring->add_request = gen6_add_request; ring->flush = gen7_render_ring_flush; if (INTEL_INFO(dev)->gen == 6) ring->flush = gen6_render_ring_flush; - if (INTEL_INFO(dev)->gen >= 8) { - ring->flush = gen8_render_ring_flush; - ring->irq_get = gen8_ring_get_irq; - ring->irq_put = gen8_ring_put_irq; - } else { - ring->irq_get = gen6_ring_get_irq; - ring->irq_put = gen6_ring_put_irq; - } + ring->irq_get = gen6_ring_get_irq; + ring->irq_put = gen6_ring_put_irq; ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT; ring->get_seqno = gen6_ring_get_seqno; ring->set_seqno = ring_set_seqno; - ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - /* - * The current semaphore is only applied on pre-gen8 platform. - * And there is no VCS2 ring on the pre-gen8 platform. So the - * semaphore between RCS and VCS2 is initialized as INVALID. - * Gen8 will initialize the sema between VCS2 and RCS later. - */ - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_RV; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_RB; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_RVE; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_VRSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_BRSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_VERSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + /* + * The current semaphore is only applied on pre-gen8 + * platform. And there is no VCS2 ring on the pre-gen8 + * platform. So the semaphore between RCS and VCS2 is + * initialized as INVALID. Gen8 will initialize the + * sema between VCS2 and RCS later. + */ + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_RV; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_RB; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_RVE; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_VRSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_BRSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_VERSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } else if (IS_GEN5(dev)) { ring->add_request = pc_render_add_request; ring->flush = gen4_render_ring_flush; @@ -2013,6 +2036,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->irq_enable_mask = I915_USER_INTERRUPT; } ring->write_tail = ring_write_tail; + if (IS_HASWELL(dev)) ring->dispatch_execbuffer = hsw_ring_dispatch_execbuffer; else if (IS_GEN8(dev)) @@ -2163,31 +2187,49 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev) ring->irq_put = gen8_ring_put_irq; ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + /* + * The current semaphore is only applied on + * pre-gen8 platform. And there is no VCS2 ring + * on the pre-gen8 platform. So the semaphore + * between VCS and VCS2 is initialized as + * INVALID. Gen8 will initialize the sema + * between VCS2 and VCS later. + */ + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } else { ring->irq_enable_mask = GT_BSD_USER_INTERRUPT; ring->irq_get = gen6_ring_get_irq; ring->irq_put = gen6_ring_put_irq; ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VR; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VB; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_VVE; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_RVSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_BVSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_VEVSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } - ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - /* - * The current semaphore is only applied on pre-gen8 platform. - * And there is no VCS2 ring on the pre-gen8 platform. So the - * semaphore between VCS and VCS2 is initialized as INVALID. - * Gen8 will initialize the sema between VCS2 and VCS later. - */ - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VR; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VB; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_VVE; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_RVSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_BVSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_VEVSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; } else { ring->mmio_base = BSD_RING_BASE; ring->flush = bsd_ring_flush; @@ -2283,30 +2325,47 @@ int intel_init_blt_ring_buffer(struct drm_device *dev) ring->irq_get = gen8_ring_get_irq; ring->irq_put = gen8_ring_put_irq; ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } else { ring->irq_enable_mask = GT_BLT_USER_INTERRUPT; ring->irq_get = gen6_ring_get_irq; ring->irq_put = gen6_ring_put_irq; ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.signal = gen6_signal; + ring->semaphore.sync_to = gen6_ring_sync; + /* + * The current semaphore is only applied on pre-gen8 + * platform. And there is no VCS2 ring on the pre-gen8 + * platform. So the semaphore between BCS and VCS2 is + * initialized as INVALID. Gen8 will initialize the + * sema between BCS and VCS2 later. + */ + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_BR; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_BV; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_BVE; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_RBSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_VBSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_VEBSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } - ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - /* - * The current semaphore is only applied on pre-gen8 platform. And - * there is no VCS2 ring on the pre-gen8 platform. So the semaphore - * between BCS and VCS2 is initialized as INVALID. - * Gen8 will initialize the sema between BCS and VCS2 later. - */ - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_BR; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_BV; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_BVE; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_RBSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_VBSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_VEBSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; ring->init = init_ring_common; return intel_init_ring_buffer(dev, ring); @@ -2333,24 +2392,40 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev) ring->irq_get = gen8_ring_get_irq; ring->irq_put = gen8_ring_put_irq; ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } else { ring->irq_enable_mask = PM_VEBOX_USER_INTERRUPT; ring->irq_get = hsw_vebox_get_irq; ring->irq_put = hsw_vebox_put_irq; ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer; + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.signal = gen6_signal; + ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VER; + ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_VEV; + ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VEB; + ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; + ring->semaphore.mbox.signal[RCS] = GEN6_RVESYNC; + ring->semaphore.mbox.signal[VCS] = GEN6_VVESYNC; + ring->semaphore.mbox.signal[BCS] = GEN6_BVESYNC; + ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; + ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + } } - ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_VER; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_VEV; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_VEB; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_RVESYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_VVESYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_BVESYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; ring->init = init_ring_common; return intel_init_ring_buffer(dev, ring); -- cgit v1.2.3-70-g09d2 From a1444b79fec2e95b705c5505c1f609306538dc49 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 30 Jun 2014 09:53:35 -0700 Subject: drm/i915: Make semaphore updates more precise MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With the ring mask we now have an easy way to know the number of rings in the system, and therefore can accurately predict the number of dwords to emit for semaphore signalling. This was not possible (easily) previously. There should be no functional impact, simply fewer instructions emitted. While we're here, simply do the round up to 2 instead of the fancier rounding we did before, which rounding up per mbox, ie 4. This also allows us to drop the unnecessary MI_NOOP, so not really 4, 3. v2: Use 3 dwords instead of 4 (Ville) Do the proper calculation to get the number of dwords to emit (Ville) Conditionally set .sync_to when semaphores are enabled (Ville) v3: Rebased on VCS2 Replace hweight_long with hweight32 (Ville) v4: Pull out the accidentally squashed hunk from the next patch after rebase (Daniel). v5: Fix conflict after rebase (Rodrigo) Reviewed-by: Rodrigo Vivi Reviewed-by: Ville Syrjälä (v1) Signed-off-by: Ben Widawsky Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 27 +++++++++------------------ 1 file changed, 9 insertions(+), 18 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 7ca1b946d8d..e9c8814bed3 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -679,23 +679,16 @@ static int gen6_signal(struct intel_engine_cs *signaller, struct drm_device *dev = signaller->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_engine_cs *useless; - int i, ret; + int i, ret, num_rings; - /* NB: In order to be able to do semaphore MBOX updates for varying - * number of rings, it's easiest if we round up each individual update - * to a multiple of 2 (since ring updates must always be a multiple of - * 2) even though the actual update only requires 3 dwords. - */ -#define MBOX_UPDATE_DWORDS 4 - if (i915_semaphore_is_enabled(dev)) - num_dwords += ((I915_NUM_RINGS-1) * MBOX_UPDATE_DWORDS); - else - return intel_ring_begin(signaller, num_dwords); +#define MBOX_UPDATE_DWORDS 3 + num_rings = hweight32(INTEL_INFO(dev)->ring_mask); + num_dwords += round_up((num_rings-1) * MBOX_UPDATE_DWORDS, 2); +#undef MBOX_UPDATE_DWORDS ret = intel_ring_begin(signaller, num_dwords); if (ret) return ret; -#undef MBOX_UPDATE_DWORDS for_each_ring(useless, dev_priv, i) { u32 mbox_reg = signaller->semaphore.mbox.signal[i]; @@ -703,15 +696,13 @@ static int gen6_signal(struct intel_engine_cs *signaller, intel_ring_emit(signaller, MI_LOAD_REGISTER_IMM(1)); intel_ring_emit(signaller, mbox_reg); intel_ring_emit(signaller, signaller->outstanding_lazy_seqno); - intel_ring_emit(signaller, MI_NOOP); - } else { - intel_ring_emit(signaller, MI_NOOP); - intel_ring_emit(signaller, MI_NOOP); - intel_ring_emit(signaller, MI_NOOP); - intel_ring_emit(signaller, MI_NOOP); } } + /* If num_dwords was rounded, make sure the tail pointer is correct */ + if (num_rings % 2 == 0) + intel_ring_emit(signaller, MI_NOOP); + return 0; } -- cgit v1.2.3-70-g09d2 From 3e78998a588c64d6447849deabf630e837505a24 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 30 Jun 2014 09:53:37 -0700 Subject: drm/i915/bdw: implement semaphore signal Semaphore signalling works similarly to previous GENs with the exception that the per ring mailboxes no longer exist. Instead you must define your own space, somewhere in the GTT. The comments in the code define the layout I've opted for, which should be fairly future proof. Ie. I tried to define offsets in abstract terms (NUM_RINGS, seqno size, etc). NOTE: If one wanted to move this to the HWSP they could. I've decided one 4k object would be easier to deal with, and provide potential wins with cache locality, but that's all speculative. v2: Update the macro to not need the other ring's ring->id (Chris) Update the comment to use the correct formula (Chris) v3: Move the macros the ringbuffer.h to prevent churn in next patch (Ville) v4: Fixed compilation rebase conflict commit 1ec9e26ddab06459e89a890431b2de064c5d1056 Author: Daniel Vetter Date: Fri Feb 14 14:01:11 2014 +0100 drm/i915: Consolidate binding parameters into flags v5: VCS2 rebase Replace hweight_long with hweight32 v6 (Rodrigo): * Add missed VC2 gen8 ring signal init * fixing conflicst on rebase * minor fixes on address table * remove WARN_ON Reviewed-by: Rodrigo Vivi Signed-off-by: Ben Widawsky Signed-off-by: Rodrigo Vivi [danvet: s/BUG_ON/WARN_ON/] Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/i915_reg.h | 5 +- drivers/gpu/drm/i915/intel_ringbuffer.c | 185 +++++++++++++++++++------------- drivers/gpu/drm/i915/intel_ringbuffer.h | 78 ++++++++++++-- 4 files changed, 189 insertions(+), 80 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 204de4032a4..7ef67b656e2 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1375,6 +1375,7 @@ struct drm_i915_private { struct pci_dev *bridge_dev; struct intel_engine_cs ring[I915_NUM_RINGS]; + struct drm_i915_gem_object *semaphore_obj; uint32_t last_seqno, next_seqno; drm_dma_handle_t *status_page_dmah; diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 66151ff1535..218ca7b7bb7 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -240,7 +240,7 @@ #define MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19) #define MI_DISPLAY_FLIP_IVB_PLANE_C (4 << 19) #define MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19) -#define MI_SEMAPHORE_MBOX MI_INSTR(0x16, 1) /* gen6+ */ +#define MI_SEMAPHORE_MBOX MI_INSTR(0x16, 1) /* gen6, gen7 */ #define MI_SEMAPHORE_GLOBAL_GTT (1<<22) #define MI_SEMAPHORE_UPDATE (1<<21) #define MI_SEMAPHORE_COMPARE (1<<20) @@ -266,6 +266,8 @@ #define MI_RESTORE_EXT_STATE_EN (1<<2) #define MI_FORCE_RESTORE (1<<1) #define MI_RESTORE_INHIBIT (1<<0) +#define MI_SEMAPHORE_SIGNAL MI_INSTR(0x1b, 0) /* GEN8+ */ +#define MI_SEMAPHORE_TARGET(engine) ((engine)<<15) #define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) #define MI_MEM_VIRTUAL (1 << 22) /* 965+ only */ #define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) @@ -360,6 +362,7 @@ #define PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE (1<<10) /* GM45+ only */ #define PIPE_CONTROL_INDIRECT_STATE_DISABLE (1<<9) #define PIPE_CONTROL_NOTIFY (1<<8) +#define PIPE_CONTROL_FLUSH_ENABLE (1<<7) /* gen7+ */ #define PIPE_CONTROL_VF_CACHE_INVALIDATE (1<<4) #define PIPE_CONTROL_CONST_CACHE_INVALIDATE (1<<3) #define PIPE_CONTROL_STATE_CACHE_INVALIDATE (1<<2) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index e9c8814bed3..e1aac25fc84 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -660,6 +660,13 @@ static int init_render_ring(struct intel_engine_cs *ring) static void render_ring_cleanup(struct intel_engine_cs *ring) { struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + + if (dev_priv->semaphore_obj) { + i915_gem_object_ggtt_unpin(dev_priv->semaphore_obj); + drm_gem_object_unreference(&dev_priv->semaphore_obj->base); + dev_priv->semaphore_obj = NULL; + } if (ring->scratch.obj == NULL) return; @@ -673,6 +680,80 @@ static void render_ring_cleanup(struct intel_engine_cs *ring) ring->scratch.obj = NULL; } +static int gen8_rcs_signal(struct intel_engine_cs *signaller, + unsigned int num_dwords) +{ +#define MBOX_UPDATE_DWORDS 8 + struct drm_device *dev = signaller->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_engine_cs *waiter; + int i, ret, num_rings; + + num_rings = hweight32(INTEL_INFO(dev)->ring_mask); + num_dwords += (num_rings-1) * MBOX_UPDATE_DWORDS; +#undef MBOX_UPDATE_DWORDS + + ret = intel_ring_begin(signaller, num_dwords); + if (ret) + return ret; + + for_each_ring(waiter, dev_priv, i) { + u64 gtt_offset = signaller->semaphore.signal_ggtt[i]; + if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID) + continue; + + intel_ring_emit(signaller, GFX_OP_PIPE_CONTROL(6)); + intel_ring_emit(signaller, PIPE_CONTROL_GLOBAL_GTT_IVB | + PIPE_CONTROL_QW_WRITE | + PIPE_CONTROL_FLUSH_ENABLE); + intel_ring_emit(signaller, lower_32_bits(gtt_offset)); + intel_ring_emit(signaller, upper_32_bits(gtt_offset)); + intel_ring_emit(signaller, signaller->outstanding_lazy_seqno); + intel_ring_emit(signaller, 0); + intel_ring_emit(signaller, MI_SEMAPHORE_SIGNAL | + MI_SEMAPHORE_TARGET(waiter->id)); + intel_ring_emit(signaller, 0); + } + + return 0; +} + +static int gen8_xcs_signal(struct intel_engine_cs *signaller, + unsigned int num_dwords) +{ +#define MBOX_UPDATE_DWORDS 6 + struct drm_device *dev = signaller->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_engine_cs *waiter; + int i, ret, num_rings; + + num_rings = hweight32(INTEL_INFO(dev)->ring_mask); + num_dwords += (num_rings-1) * MBOX_UPDATE_DWORDS; +#undef MBOX_UPDATE_DWORDS + + ret = intel_ring_begin(signaller, num_dwords); + if (ret) + return ret; + + for_each_ring(waiter, dev_priv, i) { + u64 gtt_offset = signaller->semaphore.signal_ggtt[i]; + if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID) + continue; + + intel_ring_emit(signaller, (MI_FLUSH_DW + 1) | + MI_FLUSH_DW_OP_STOREDW); + intel_ring_emit(signaller, lower_32_bits(gtt_offset) | + MI_FLUSH_DW_USE_GTT); + intel_ring_emit(signaller, upper_32_bits(gtt_offset)); + intel_ring_emit(signaller, signaller->outstanding_lazy_seqno); + intel_ring_emit(signaller, MI_SEMAPHORE_SIGNAL | + MI_SEMAPHORE_TARGET(waiter->id)); + intel_ring_emit(signaller, 0); + } + + return 0; +} + static int gen6_signal(struct intel_engine_cs *signaller, unsigned int num_dwords) { @@ -1942,12 +2023,30 @@ int intel_init_render_ring_buffer(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; struct intel_engine_cs *ring = &dev_priv->ring[RCS]; + struct drm_i915_gem_object *obj; + int ret; ring->name = "render ring"; ring->id = RCS; ring->mmio_base = RENDER_RING_BASE; if (INTEL_INFO(dev)->gen >= 8) { + if (i915_semaphore_is_enabled(dev)) { + obj = i915_gem_alloc_object(dev, 4096); + if (obj == NULL) { + DRM_ERROR("Failed to allocate semaphore bo. Disabling semaphores\n"); + i915.semaphores = 0; + } else { + i915_gem_object_set_cache_level(obj, I915_CACHE_LLC); + ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_NONBLOCK); + if (ret != 0) { + drm_gem_object_unreference(&obj->base); + DRM_ERROR("Failed to pin semaphore bo. Disabling semaphores\n"); + i915.semaphores = 0; + } else + dev_priv->semaphore_obj = obj; + } + } ring->add_request = gen6_add_request; ring->flush = gen8_render_ring_flush; ring->irq_get = gen8_ring_get_irq; @@ -1956,18 +2055,10 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->get_seqno = gen6_ring_get_seqno; ring->set_seqno = ring_set_seqno; if (i915_semaphore_is_enabled(dev)) { + WARN_ON(!dev_priv->semaphore_obj); ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + ring->semaphore.signal = gen8_rcs_signal; + GEN8_RING_SEMAPHORE_INIT; } } else if (INTEL_INFO(dev)->gen >= 6) { ring->add_request = gen6_add_request; @@ -2045,9 +2136,6 @@ int intel_init_render_ring_buffer(struct drm_device *dev) /* 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"); @@ -2180,25 +2268,8 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev) gen8_ring_dispatch_execbuffer; if (i915_semaphore_is_enabled(dev)) { ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - /* - * The current semaphore is only applied on - * pre-gen8 platform. And there is no VCS2 ring - * on the pre-gen8 platform. So the semaphore - * between VCS and VCS2 is initialized as - * INVALID. Gen8 will initialize the sema - * between VCS2 and VCS later. - */ - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + ring->semaphore.signal = gen8_xcs_signal; + GEN8_RING_SEMAPHORE_INIT; } } else { ring->irq_enable_mask = GT_BSD_USER_INTERRUPT; @@ -2273,24 +2344,10 @@ int intel_init_bsd2_ring_buffer(struct drm_device *dev) ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - /* - * The current semaphore is only applied on the pre-gen8. And there - * is no bsd2 ring on the pre-gen8. So now the semaphore_register - * between VCS2 and other ring is initialized as invalid. - * Gen8 will initialize the sema between VCS2 and other ring later. - */ - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; - + if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.signal = gen8_xcs_signal; + GEN8_RING_SEMAPHORE_INIT; + } ring->init = init_ring_common; return intel_init_ring_buffer(dev, ring); @@ -2318,17 +2375,8 @@ int intel_init_blt_ring_buffer(struct drm_device *dev) ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; if (i915_semaphore_is_enabled(dev)) { ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + ring->semaphore.signal = gen8_xcs_signal; + GEN8_RING_SEMAPHORE_INIT; } } else { ring->irq_enable_mask = GT_BLT_USER_INTERRUPT; @@ -2385,17 +2433,8 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev) ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; if (i915_semaphore_is_enabled(dev)) { ring->semaphore.sync_to = gen6_ring_sync; - ring->semaphore.signal = gen6_signal; - ring->semaphore.mbox.wait[RCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[BCS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VECS] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.wait[VCS2] = MI_SEMAPHORE_SYNC_INVALID; - ring->semaphore.mbox.signal[RCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[BCS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VECS] = GEN6_NOSYNC; - ring->semaphore.mbox.signal[VCS2] = GEN6_NOSYNC; + ring->semaphore.signal = gen8_xcs_signal; + GEN8_RING_SEMAPHORE_INIT; } } else { ring->irq_enable_mask = PM_VEBOX_USER_INTERRUPT; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 2e8b51645bc..17ef9cdf1b8 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -40,6 +40,32 @@ struct intel_hw_status_page { #define I915_READ_MODE(ring) I915_READ(RING_MI_MODE((ring)->mmio_base)) #define I915_WRITE_MODE(ring, val) I915_WRITE(RING_MI_MODE((ring)->mmio_base), val) +/* seqno size is actually only a uint32, but since we plan to use MI_FLUSH_DW to + * do the writes, and that must have qw aligned offsets, simply pretend it's 8b. + */ +#define i915_semaphore_seqno_size sizeof(uint64_t) +#define GEN8_SIGNAL_OFFSET(__ring, to) \ + (i915_gem_obj_ggtt_offset(dev_priv->semaphore_obj) + \ + ((__ring)->id * I915_NUM_RINGS * i915_semaphore_seqno_size) + \ + (i915_semaphore_seqno_size * (to))) + +#define GEN8_WAIT_OFFSET(__ring, from) \ + (i915_gem_obj_ggtt_offset(dev_priv->semaphore_obj) + \ + ((from) * I915_NUM_RINGS * i915_semaphore_seqno_size) + \ + (i915_semaphore_seqno_size * (__ring)->id)) + +#define GEN8_RING_SEMAPHORE_INIT do { \ + if (!dev_priv->semaphore_obj) { \ + break; \ + } \ + ring->semaphore.signal_ggtt[RCS] = GEN8_SIGNAL_OFFSET(ring, RCS); \ + ring->semaphore.signal_ggtt[VCS] = GEN8_SIGNAL_OFFSET(ring, VCS); \ + ring->semaphore.signal_ggtt[BCS] = GEN8_SIGNAL_OFFSET(ring, BCS); \ + ring->semaphore.signal_ggtt[VECS] = GEN8_SIGNAL_OFFSET(ring, VECS); \ + ring->semaphore.signal_ggtt[VCS2] = GEN8_SIGNAL_OFFSET(ring, VCS2); \ + ring->semaphore.signal_ggtt[ring->id] = MI_SEMAPHORE_SYNC_INVALID; \ + } while(0) + enum intel_ring_hangcheck_action { HANGCHECK_IDLE = 0, HANGCHECK_WAIT, @@ -127,15 +153,55 @@ struct intel_engine_cs { #define I915_DISPATCH_PINNED 0x2 void (*cleanup)(struct intel_engine_cs *ring); + /* GEN8 signal/wait table - never trust comments! + * signal to signal to signal to signal to signal to + * RCS VCS BCS VECS VCS2 + * -------------------------------------------------------------------- + * RCS | NOP (0x00) | VCS (0x08) | BCS (0x10) | VECS (0x18) | VCS2 (0x20) | + * |------------------------------------------------------------------- + * VCS | RCS (0x28) | NOP (0x30) | BCS (0x38) | VECS (0x40) | VCS2 (0x48) | + * |------------------------------------------------------------------- + * BCS | RCS (0x50) | VCS (0x58) | NOP (0x60) | VECS (0x68) | VCS2 (0x70) | + * |------------------------------------------------------------------- + * VECS | RCS (0x78) | VCS (0x80) | BCS (0x88) | NOP (0x90) | VCS2 (0x98) | + * |------------------------------------------------------------------- + * VCS2 | RCS (0xa0) | VCS (0xa8) | BCS (0xb0) | VECS (0xb8) | NOP (0xc0) | + * |------------------------------------------------------------------- + * + * Generalization: + * f(x, y) := (x->id * NUM_RINGS * seqno_size) + (seqno_size * y->id) + * ie. transpose of g(x, y) + * + * sync from sync from sync from sync from sync from + * RCS VCS BCS VECS VCS2 + * -------------------------------------------------------------------- + * RCS | NOP (0x00) | VCS (0x28) | BCS (0x50) | VECS (0x78) | VCS2 (0xa0) | + * |------------------------------------------------------------------- + * VCS | RCS (0x08) | NOP (0x30) | BCS (0x58) | VECS (0x80) | VCS2 (0xa8) | + * |------------------------------------------------------------------- + * BCS | RCS (0x10) | VCS (0x38) | NOP (0x60) | VECS (0x88) | VCS2 (0xb0) | + * |------------------------------------------------------------------- + * VECS | RCS (0x18) | VCS (0x40) | BCS (0x68) | NOP (0x90) | VCS2 (0xb8) | + * |------------------------------------------------------------------- + * VCS2 | RCS (0x20) | VCS (0x48) | BCS (0x70) | VECS (0x98) | NOP (0xc0) | + * |------------------------------------------------------------------- + * + * Generalization: + * g(x, y) := (y->id * NUM_RINGS * seqno_size) + (seqno_size * x->id) + * ie. transpose of f(x, y) + */ struct { u32 sync_seqno[I915_NUM_RINGS-1]; - struct { - /* our mbox written by others */ - u32 wait[I915_NUM_RINGS]; - /* mboxes this ring signals to */ - u32 signal[I915_NUM_RINGS]; - } mbox; + union { + struct { + /* our mbox written by others */ + u32 wait[I915_NUM_RINGS]; + /* mboxes this ring signals to */ + u32 signal[I915_NUM_RINGS]; + } mbox; + u64 signal_ggtt[I915_NUM_RINGS]; + }; /* AKA wait() */ int (*sync_to)(struct intel_engine_cs *ring, -- cgit v1.2.3-70-g09d2 From 5ee426ca135c91e071d23853e876b957526841c5 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 30 Jun 2014 09:53:38 -0700 Subject: drm/i915/bdw: implement semaphore wait Semaphore waits use a new instruction, MI_SEMAPHORE_WAIT. The seqno to wait on is all well defined by the table in the previous patch. There is nothing else different from previous GEN's semaphore synchronization code. v2: Update macros to not require the other ring's ring->id (Chris) v3: Add missing VCS2 gen8_ring_wait init besides s/ring_buffer/engine_cs (Rodrigo) Reviewed-by: Rodrigo Vivi Signed-off-by: Ben Widawsky Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_reg.h | 3 +++ drivers/gpu/drm/i915/intel_ringbuffer.c | 35 ++++++++++++++++++++++++++++----- 2 files changed, 33 insertions(+), 5 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 218ca7b7bb7..8c0f70de9fd 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -268,6 +268,9 @@ #define MI_RESTORE_INHIBIT (1<<0) #define MI_SEMAPHORE_SIGNAL MI_INSTR(0x1b, 0) /* GEN8+ */ #define MI_SEMAPHORE_TARGET(engine) ((engine)<<15) +#define MI_SEMAPHORE_WAIT MI_INSTR(0x1c, 2) /* GEN8+ */ +#define MI_SEMAPHORE_POLL (1<<15) +#define MI_SEMAPHORE_SAD_GTE_SDD (1<<12) #define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) #define MI_MEM_VIRTUAL (1 << 22) /* 965+ only */ #define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index e1aac25fc84..88196808cc4 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -832,6 +832,31 @@ static inline bool i915_gem_has_seqno_wrapped(struct drm_device *dev, * @signaller - ring which has, or will signal * @seqno - seqno which the waiter will block on */ + +static int +gen8_ring_sync(struct intel_engine_cs *waiter, + struct intel_engine_cs *signaller, + u32 seqno) +{ + struct drm_i915_private *dev_priv = waiter->dev->dev_private; + int ret; + + ret = intel_ring_begin(waiter, 4); + if (ret) + return ret; + + intel_ring_emit(waiter, MI_SEMAPHORE_WAIT | + MI_SEMAPHORE_GLOBAL_GTT | + MI_SEMAPHORE_SAD_GTE_SDD); + intel_ring_emit(waiter, seqno); + intel_ring_emit(waiter, + lower_32_bits(GEN8_WAIT_OFFSET(waiter, signaller->id))); + intel_ring_emit(waiter, + upper_32_bits(GEN8_WAIT_OFFSET(waiter, signaller->id))); + intel_ring_advance(waiter); + return 0; +} + static int gen6_ring_sync(struct intel_engine_cs *waiter, struct intel_engine_cs *signaller, @@ -2056,7 +2081,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->set_seqno = ring_set_seqno; if (i915_semaphore_is_enabled(dev)) { WARN_ON(!dev_priv->semaphore_obj); - ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.sync_to = gen8_ring_sync; ring->semaphore.signal = gen8_rcs_signal; GEN8_RING_SEMAPHORE_INIT; } @@ -2267,7 +2292,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev) ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; if (i915_semaphore_is_enabled(dev)) { - ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.sync_to = gen8_ring_sync; ring->semaphore.signal = gen8_xcs_signal; GEN8_RING_SEMAPHORE_INIT; } @@ -2343,8 +2368,8 @@ int intel_init_bsd2_ring_buffer(struct drm_device *dev) ring->irq_put = gen8_ring_put_irq; ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; - ring->semaphore.sync_to = gen6_ring_sync; if (i915_semaphore_is_enabled(dev)) { + ring->semaphore.sync_to = gen8_ring_sync; ring->semaphore.signal = gen8_xcs_signal; GEN8_RING_SEMAPHORE_INIT; } @@ -2374,7 +2399,7 @@ int intel_init_blt_ring_buffer(struct drm_device *dev) ring->irq_put = gen8_ring_put_irq; ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; if (i915_semaphore_is_enabled(dev)) { - ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.sync_to = gen8_ring_sync; ring->semaphore.signal = gen8_xcs_signal; GEN8_RING_SEMAPHORE_INIT; } @@ -2432,7 +2457,7 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev) ring->irq_put = gen8_ring_put_irq; ring->dispatch_execbuffer = gen8_ring_dispatch_execbuffer; if (i915_semaphore_is_enabled(dev)) { - ring->semaphore.sync_to = gen6_ring_sync; + ring->semaphore.sync_to = gen8_ring_sync; ring->semaphore.signal = gen8_xcs_signal; GEN8_RING_SEMAPHORE_INIT; } -- cgit v1.2.3-70-g09d2 From bae4fcd2c7b778a3f69e03e244221cd4932b8f67 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Mon, 30 Jun 2014 09:53:43 -0700 Subject: drm/i915/bdw: poll semaphores As Ville points out, it's possible/probable we don't actually need this. Potentially, this validates the letter of the spec, and not the spirit. Ville: > I discussed this on irc w/ Ben, and I was suggesting we don't need to > poll. Polling apparently can be used as a workaround for certain > hardware issues, but it looks like those issues shouldn't affect us, > for the momemnt at least. So my suggestion was to try w/o polling > first (since there could be some power cost to polling) and add the > poll bit if problems arise. Rodrigo: Spec suggests this as an W/A for GT3. However semaphores didn't worked in my BDW GT2 on Signal Mode. So pool mode is definitely needed. Reviewed-by: Rodrigo Vivi Tested-by: Rodrigo Vivi Signed-off-by: Ben Widawsky Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 88196808cc4..e18ed05dc0d 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -847,6 +847,7 @@ gen8_ring_sync(struct intel_engine_cs *waiter, intel_ring_emit(waiter, MI_SEMAPHORE_WAIT | MI_SEMAPHORE_GLOBAL_GTT | + MI_SEMAPHORE_POLL | MI_SEMAPHORE_SAD_GTE_SDD); intel_ring_emit(waiter, seqno); intel_ring_emit(waiter, -- cgit v1.2.3-70-g09d2 From 2919d2913c515fcfef6d50a46aae31128d454ce9 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 3 Jul 2014 16:28:02 +0100 Subject: drm/i915: Extract ringbuffer destroy & generalize alloc to take a ringbuf More prep work: with Execlists, we are going to start creating a lot of extra ringbuffers soon, so these functions are handy. No functional changes. v2: rename allocate/destroy_ring_buffer to alloc/destroy_ringbuffer_obj because the name is more meaningful and to mirror a similar function in the context world: i915_gem_alloc_context_obj(). Change suggested by Brad Volkin. Reviewed-by: Jesse Barnes Signed-off-by: Oscar Mateo Reviewed-by: Chris Wilson Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index e18ed05dc0d..40b17f246a6 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1482,15 +1482,25 @@ static int init_phys_status_page(struct intel_engine_cs *ring) return 0; } -static int allocate_ring_buffer(struct intel_engine_cs *ring) +static void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf) +{ + if (!ringbuf->obj) + return; + + iounmap(ringbuf->virtual_start); + i915_gem_object_ggtt_unpin(ringbuf->obj); + drm_gem_object_unreference(&ringbuf->obj->base); + ringbuf->obj = NULL; +} + +static int intel_alloc_ringbuffer_obj(struct drm_device *dev, + struct intel_ringbuffer *ringbuf) { - struct drm_device *dev = ring->dev; struct drm_i915_private *dev_priv = to_i915(dev); - struct intel_ringbuffer *ringbuf = ring->buffer; struct drm_i915_gem_object *obj; int ret; - if (intel_ring_initialized(ring)) + if (ringbuf->obj) return 0; obj = NULL; @@ -1562,7 +1572,7 @@ static int intel_init_ring_buffer(struct drm_device *dev, goto error; } - ret = allocate_ring_buffer(ring); + ret = intel_alloc_ringbuffer_obj(dev, ringbuf); if (ret) { DRM_ERROR("Failed to allocate ringbuffer %s: %d\n", ring->name, ret); goto error; @@ -1603,11 +1613,7 @@ void intel_cleanup_ring_buffer(struct intel_engine_cs *ring) intel_stop_ring_buffer(ring); WARN_ON(!IS_GEN2(ring->dev) && (I915_READ_MODE(ring) & MODE_IDLE) == 0); - iounmap(ringbuf->virtual_start); - - i915_gem_object_ggtt_unpin(ringbuf->obj); - drm_gem_object_unreference(&ringbuf->obj->base); - ringbuf->obj = NULL; + intel_destroy_ringbuffer_obj(ringbuf); ring->preallocated_lazy_request = NULL; ring->outstanding_lazy_seqno = 0; -- cgit v1.2.3-70-g09d2 From 64c58f2c48da85223f0db438fe4854bc3ef24353 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 3 Jul 2014 16:28:03 +0100 Subject: drm/i915: Generalize ring_space to take a ringbuf It's simple enough that it doesn't need to know anything about the engine. Trivial change. Reviewed-by: Jesse Barnes Signed-off-by: Oscar Mateo Reviewed-by: Chris Wilson Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 40b17f246a6..738140c0f5a 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -48,9 +48,8 @@ static inline int __ring_space(int head, int tail, int size) return space; } -static inline int ring_space(struct intel_engine_cs *ring) +static inline int ring_space(struct intel_ringbuffer *ringbuf) { - struct intel_ringbuffer *ringbuf = ring->buffer; return __ring_space(ringbuf->head & HEAD_ADDR, ringbuf->tail, ringbuf->size); } @@ -545,7 +544,7 @@ static int init_ring_common(struct intel_engine_cs *ring) else { ringbuf->head = I915_READ_HEAD(ring); ringbuf->tail = I915_READ_TAIL(ring) & TAIL_ADDR; - ringbuf->space = ring_space(ring); + ringbuf->space = ring_space(ringbuf); ringbuf->last_retired_head = -1; } @@ -1639,7 +1638,7 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n) ringbuf->head = ringbuf->last_retired_head; ringbuf->last_retired_head = -1; - ringbuf->space = ring_space(ring); + ringbuf->space = ring_space(ringbuf); if (ringbuf->space >= n) return 0; } @@ -1662,7 +1661,7 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n) ringbuf->head = ringbuf->last_retired_head; ringbuf->last_retired_head = -1; - ringbuf->space = ring_space(ring); + ringbuf->space = ring_space(ringbuf); return 0; } @@ -1691,7 +1690,7 @@ static int ring_wait_for_space(struct intel_engine_cs *ring, int n) trace_i915_ring_wait_begin(ring); do { ringbuf->head = I915_READ_HEAD(ring); - ringbuf->space = ring_space(ring); + ringbuf->space = ring_space(ringbuf); if (ringbuf->space >= n) { ret = 0; break; @@ -1743,7 +1742,7 @@ static int intel_wrap_ring_buffer(struct intel_engine_cs *ring) iowrite32(MI_NOOP, virt++); ringbuf->tail = 0; - ringbuf->space = ring_space(ring); + ringbuf->space = ring_space(ringbuf); return 0; } -- cgit v1.2.3-70-g09d2 From 1f767e02d69f6a37a86aa94fd2c7e2528ae1161e Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 3 Jul 2014 17:33:03 -0400 Subject: drm/i915: HWS must be in the mappable region for g33 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On g33, the documentation states "HWS_PGA: Format = Bits 28:12 of graphics memory address (bits 31:29 MBZ)." which translates to that the address of the HWS must be below 256MiB, which is conveniently the mappable aperture. This also appears to be true (but not documented as so) for gen4 and gen5. To generalise we force it into the low mappable region for all non-LLC platforms. If we locate the HWS at the top of the GTT the machine will hard hang during boot (fails on pnv, gm45, ilk and byt, but works on snb, ivb, hsw). v2: Add comments to explain why use PIN_MAPPABLE even though we have no intention of mapping the object. (Ville) Signed-off-by: Chris Wilson Cc: Ville Syrjälä Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 738140c0f5a..599709e80a1 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1432,6 +1432,7 @@ static int init_status_page(struct intel_engine_cs *ring) struct drm_i915_gem_object *obj; if ((obj = ring->status_page.obj) == NULL) { + unsigned flags; int ret; obj = i915_gem_alloc_object(ring->dev, 4096); @@ -1444,7 +1445,20 @@ static int init_status_page(struct intel_engine_cs *ring) if (ret) goto err_unref; - ret = i915_gem_obj_ggtt_pin(obj, 4096, 0); + flags = 0; + if (!HAS_LLC(ring->dev)) + /* On g33, we cannot place HWS above 256MiB, so + * restrict its pinning to the low mappable arena. + * Though this restriction is not documented for + * gen4, gen5, or byt, they also behave similarly + * and hang if the HWS is placed at the top of the + * GTT. To generalise, it appears that all !llc + * platforms have issues with us placing the HWS + * above the mappable region (even though we never + * actualy map it). + */ + flags |= PIN_MAPPABLE; + ret = i915_gem_obj_ggtt_pin(obj, 4096, flags); if (ret) { err_unref: drm_gem_object_unreference(&obj->base); -- cgit v1.2.3-70-g09d2 From 480c80338618867851659710d1a27c9cc85833d2 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 16 Jul 2014 09:49:40 +0200 Subject: drm/i915: Use genX_ prefix for gt irq enable/disable functions Traditionally we use genX_ for GT/render stuff and the codenames for display stuff. But the gt and pm interrupt handling functions on gen5/6+ stuck out as exceptions, so convert them. Looking at the diff this nicely realigns our ducks since almost all the callers are already platform-specific functions following the genX_ pattern. Spotted while reviewing some internal rps patches. No function change in this patch. Acked-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_irq.c | 24 ++++++++++++------------ drivers/gpu/drm/i915/intel_drv.h | 12 ++++++------ drivers/gpu/drm/i915/intel_pm.c | 4 ++-- drivers/gpu/drm/i915/intel_ringbuffer.c | 12 ++++++------ 4 files changed, 26 insertions(+), 26 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 7d61ca2a01d..dfe923a3cb9 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -182,12 +182,12 @@ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv, POSTING_READ(GTIMR); } -void ilk_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) { ilk_update_gt_irq(dev_priv, mask, mask); } -void ilk_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) { ilk_update_gt_irq(dev_priv, mask, 0); } @@ -220,12 +220,12 @@ static void snb_update_pm_irq(struct drm_i915_private *dev_priv, } } -void snb_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) { snb_update_pm_irq(dev_priv, mask, mask); } -void snb_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) { snb_update_pm_irq(dev_priv, mask, 0); } @@ -278,12 +278,12 @@ static void bdw_update_pm_irq(struct drm_i915_private *dev_priv, } } -void bdw_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen8_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) { bdw_update_pm_irq(dev_priv, mask, mask); } -void bdw_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen8_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask) { bdw_update_pm_irq(dev_priv, mask, 0); } @@ -1408,10 +1408,10 @@ static void gen6_pm_rps_work(struct work_struct *work) pm_iir = dev_priv->rps.pm_iir; dev_priv->rps.pm_iir = 0; if (INTEL_INFO(dev_priv->dev)->gen >= 8) - bdw_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); + gen8_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); else { /* Make sure not to corrupt PMIMR state used by ringbuffer */ - snb_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); + gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); } spin_unlock_irq(&dev_priv->irq_lock); @@ -1553,7 +1553,7 @@ static void ivybridge_parity_work(struct work_struct *work) out: WARN_ON(dev_priv->l3_parity.which_slice); spin_lock_irqsave(&dev_priv->irq_lock, flags); - ilk_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv->dev)); + gen5_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv->dev)); spin_unlock_irqrestore(&dev_priv->irq_lock, flags); mutex_unlock(&dev_priv->dev->struct_mutex); @@ -1567,7 +1567,7 @@ static void ivybridge_parity_error_irq_handler(struct drm_device *dev, u32 iir) return; spin_lock(&dev_priv->irq_lock); - ilk_disable_gt_irq(dev_priv, GT_PARITY_ERROR(dev)); + gen5_disable_gt_irq(dev_priv, GT_PARITY_ERROR(dev)); spin_unlock(&dev_priv->irq_lock); iir &= GT_PARITY_ERROR(dev); @@ -1622,7 +1622,7 @@ static void gen8_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir) spin_lock(&dev_priv->irq_lock); dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events; - bdw_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events); + gen8_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events); spin_unlock(&dev_priv->irq_lock); queue_work(dev_priv->wq, &dev_priv->rps.work); @@ -1969,7 +1969,7 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir) if (pm_iir & dev_priv->pm_rps_events) { spin_lock(&dev_priv->irq_lock); dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events; - snb_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events); + gen6_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events); spin_unlock(&dev_priv->irq_lock); queue_work(dev_priv->wq, &dev_priv->rps.work); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 6093ebdeb7c..9d97a50cae4 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -683,12 +683,12 @@ bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev, bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev, enum transcoder pch_transcoder, bool enable); -void ilk_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); -void ilk_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); -void snb_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); -void snb_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); -void bdw_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); -void bdw_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen8_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen8_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask); void intel_runtime_pm_disable_interrupts(struct drm_device *dev); void intel_runtime_pm_restore_interrupts(struct drm_device *dev); int intel_get_crtc_scanline(struct intel_crtc *crtc); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 6e03851a4fa..25ae4e6d3dd 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -3474,7 +3474,7 @@ static void gen8_enable_rps_interrupts(struct drm_device *dev) spin_lock_irq(&dev_priv->irq_lock); WARN_ON(dev_priv->rps.pm_iir); - bdw_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); + gen8_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); I915_WRITE(GEN8_GT_IIR(2), dev_priv->pm_rps_events); spin_unlock_irq(&dev_priv->irq_lock); } @@ -3485,7 +3485,7 @@ static void gen6_enable_rps_interrupts(struct drm_device *dev) spin_lock_irq(&dev_priv->irq_lock); WARN_ON(dev_priv->rps.pm_iir); - snb_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); + gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); I915_WRITE(GEN6_PMIIR, dev_priv->pm_rps_events); spin_unlock_irq(&dev_priv->irq_lock); } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 599709e80a1..b3d8f766fa7 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1004,7 +1004,7 @@ gen5_ring_get_irq(struct intel_engine_cs *ring) spin_lock_irqsave(&dev_priv->irq_lock, flags); if (ring->irq_refcount++ == 0) - ilk_enable_gt_irq(dev_priv, ring->irq_enable_mask); + gen5_enable_gt_irq(dev_priv, ring->irq_enable_mask); spin_unlock_irqrestore(&dev_priv->irq_lock, flags); return true; @@ -1019,7 +1019,7 @@ gen5_ring_put_irq(struct intel_engine_cs *ring) spin_lock_irqsave(&dev_priv->irq_lock, flags); if (--ring->irq_refcount == 0) - ilk_disable_gt_irq(dev_priv, ring->irq_enable_mask); + gen5_disable_gt_irq(dev_priv, ring->irq_enable_mask); spin_unlock_irqrestore(&dev_priv->irq_lock, flags); } @@ -1212,7 +1212,7 @@ gen6_ring_get_irq(struct intel_engine_cs *ring) GT_PARITY_ERROR(dev))); else I915_WRITE_IMR(ring, ~ring->irq_enable_mask); - ilk_enable_gt_irq(dev_priv, ring->irq_enable_mask); + gen5_enable_gt_irq(dev_priv, ring->irq_enable_mask); } spin_unlock_irqrestore(&dev_priv->irq_lock, flags); @@ -1232,7 +1232,7 @@ gen6_ring_put_irq(struct intel_engine_cs *ring) I915_WRITE_IMR(ring, ~GT_PARITY_ERROR(dev)); else I915_WRITE_IMR(ring, ~0); - ilk_disable_gt_irq(dev_priv, ring->irq_enable_mask); + gen5_disable_gt_irq(dev_priv, ring->irq_enable_mask); } spin_unlock_irqrestore(&dev_priv->irq_lock, flags); } @@ -1250,7 +1250,7 @@ hsw_vebox_get_irq(struct intel_engine_cs *ring) spin_lock_irqsave(&dev_priv->irq_lock, flags); if (ring->irq_refcount++ == 0) { I915_WRITE_IMR(ring, ~ring->irq_enable_mask); - snb_enable_pm_irq(dev_priv, ring->irq_enable_mask); + gen6_enable_pm_irq(dev_priv, ring->irq_enable_mask); } spin_unlock_irqrestore(&dev_priv->irq_lock, flags); @@ -1270,7 +1270,7 @@ hsw_vebox_put_irq(struct intel_engine_cs *ring) spin_lock_irqsave(&dev_priv->irq_lock, flags); if (--ring->irq_refcount == 0) { I915_WRITE_IMR(ring, ~0); - snb_disable_pm_irq(dev_priv, ring->irq_enable_mask); + gen6_disable_pm_irq(dev_priv, ring->irq_enable_mask); } spin_unlock_irqrestore(&dev_priv->irq_lock, flags); } -- cgit v1.2.3-70-g09d2 From 884ceacee308f0e4616d0c933518af2639f7b1d8 Mon Sep 17 00:00:00 2001 From: Kenneth Graunke Date: Sat, 28 Jun 2014 02:04:20 +0300 Subject: drm/i915: Refactor Broadwell PIPE_CONTROL emission into a helper. We'll want to reuse this for a workaround. Signed-off-by: Kenneth Graunke Reviewed-by: Jesse Barnes [danvet: Rmove now unused int.] Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 37 ++++++++++++++++++++------------- 1 file changed, 22 insertions(+), 15 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index b3d8f766fa7..2908896334f 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -379,13 +379,33 @@ gen7_render_ring_flush(struct intel_engine_cs *ring, return 0; } +static int +gen8_emit_pipe_control(struct intel_engine_cs *ring, + u32 flags, u32 scratch_addr) +{ + int ret; + + ret = intel_ring_begin(ring, 6); + if (ret) + return ret; + + intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6)); + intel_ring_emit(ring, flags); + intel_ring_emit(ring, scratch_addr); + intel_ring_emit(ring, 0); + intel_ring_emit(ring, 0); + intel_ring_emit(ring, 0); + intel_ring_advance(ring); + + return 0; +} + static int gen8_render_ring_flush(struct intel_engine_cs *ring, u32 invalidate_domains, u32 flush_domains) { u32 flags = 0; u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES; - int ret; flags |= PIPE_CONTROL_CS_STALL; @@ -404,20 +424,7 @@ gen8_render_ring_flush(struct intel_engine_cs *ring, flags |= PIPE_CONTROL_GLOBAL_GTT_IVB; } - ret = intel_ring_begin(ring, 6); - if (ret) - return ret; - - intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6)); - intel_ring_emit(ring, flags); - intel_ring_emit(ring, scratch_addr); - intel_ring_emit(ring, 0); - intel_ring_emit(ring, 0); - intel_ring_emit(ring, 0); - intel_ring_advance(ring); - - return 0; - + return gen8_emit_pipe_control(ring, flags, scratch_addr); } static void ring_write_tail(struct intel_engine_cs *ring, -- cgit v1.2.3-70-g09d2 From 02c9f7e3cfe76a7f54ef03438c36aade86cc1c8b Mon Sep 17 00:00:00 2001 From: Kenneth Graunke Date: Mon, 27 Jan 2014 14:20:16 -0800 Subject: drm/i915: Add the WaCsStallBeforeStateCacheInvalidate:bdw workaround. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On Broadwell, any PIPE_CONTROL with the "State Cache Invalidate" bit set must be preceded by a PIPE_CONTROL with the "CS Stall" bit set. Documented on the BSpec 3D workarounds page. Reviewed-by: Rafael Barbalho Signed-off-by: Kenneth Graunke [vsyrjala: add chv w/a note too] Signed-off-by: Ville Syrjälä Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 2908896334f..05969f03c0c 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -406,6 +406,7 @@ gen8_render_ring_flush(struct intel_engine_cs *ring, { u32 flags = 0; u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES; + int ret; flags |= PIPE_CONTROL_CS_STALL; @@ -422,6 +423,14 @@ gen8_render_ring_flush(struct intel_engine_cs *ring, flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE; flags |= PIPE_CONTROL_QW_WRITE; flags |= PIPE_CONTROL_GLOBAL_GTT_IVB; + + /* WaCsStallBeforeStateCacheInvalidate:bdw,chv */ + ret = gen8_emit_pipe_control(ring, + PIPE_CONTROL_CS_STALL | + PIPE_CONTROL_STALL_AT_SCOREBOARD, + 0); + if (ret) + return ret; } return gen8_emit_pipe_control(ring, flags, scratch_addr); -- cgit v1.2.3-70-g09d2 From ece4a17d237a79f63fbfaf3f724a12b6d500555c Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Thu, 7 Aug 2014 16:29:53 +0200 Subject: drm/i915: read HEAD register back in init_ring_common() to enforce ordering Withtout this, ring initialization fails reliabily during resume with [drm:init_ring_common] *ERROR* render ring initialization failed ctl 0001f001 head ffffff8804 tail 00000000 start 000e4000 This is not a complete fix, but it is verified to make the ring initialization failures during resume much less likely. We were not able to root-cause this bug (likely HW-specific to Gen4 chips) yet. This is therefore used as a ducttape before problem is fully understood and proper fix created, so that people don't suffer from completely unusable systems in the meantime. The discussion and debugging is happening at https://bugs.freedesktop.org/show_bug.cgi?id=76554 Signed-off-by: Jiri Kosina Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 05969f03c0c..16371a44442 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -532,6 +532,9 @@ static int init_ring_common(struct intel_engine_cs *ring) else ring_setup_phys_status_page(ring); + /* Enforce ordering by reading HEAD register back */ + I915_READ_HEAD(ring); + /* Initialize the ring. This must happen _after_ we've cleared the ring * registers with the above sequence (the readback of the HEAD registers * also enforces ordering), otherwise the hw might lose the new ring -- cgit v1.2.3-70-g09d2 From 403bdd10c815029694046adf5ffde0577cbd2866 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 7 Aug 2014 16:05:39 +0200 Subject: drm/i915: No busy-loop wait_for in the ring init code Doing a 1s wait (tops) with the cpu is a bit excessive. Tune it down like everything else in that code. v2: Also insert the missing space Chris spotted. Cc: Naresh Kumar Kachhi Cc: Chris Wilson Acked-by: Chris Wilson Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 16371a44442..117543e58d4 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -476,8 +476,8 @@ static bool stop_ring(struct intel_engine_cs *ring) if (!IS_GEN2(ring->dev)) { I915_WRITE_MODE(ring, _MASKED_BIT_ENABLE(STOP_RING)); - if (wait_for_atomic((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) { - DRM_ERROR("%s :timed out trying to stop ring\n", ring->name); + if (wait_for((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) { + DRM_ERROR("%s : timed out trying to stop ring\n", ring->name); return false; } } -- cgit v1.2.3-70-g09d2 From 9bec9b1334d687c0a9fcf3d3a1987a61b4826a45 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 11 Aug 2014 09:21:35 +0100 Subject: drm/i915: Double check ring is idle before declaring the GPU wedged During ring initialisation, sometimes we observe, though not in production hardware, that the idle flag is not set even though the ring is empty. Double check before giving up. Signed-off-by: Chris Wilson Cc: Damien Lespiau Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 117543e58d4..a059b64a0fb 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -478,7 +478,12 @@ static bool stop_ring(struct intel_engine_cs *ring) I915_WRITE_MODE(ring, _MASKED_BIT_ENABLE(STOP_RING)); if (wait_for((I915_READ_MODE(ring) & MODE_IDLE) != 0, 1000)) { DRM_ERROR("%s : timed out trying to stop ring\n", ring->name); - return false; + /* Sometimes we observe that the idle flag is not + * set even though the ring is empty. So double + * check before giving up. + */ + if (I915_READ_HEAD(ring) != I915_READ_TAIL(ring)) + return false; } } -- cgit v1.2.3-70-g09d2 From 84c2377fcee7a43cd964b62143e9a3714130bb0c Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 24 Jul 2014 17:04:15 +0100 Subject: drm/i915/bdw: Allocate ringbuffers for Logical Ring Contexts As we have said a couple of times by now, logical ring contexts have their own ringbuffers: not only the backing pages, but the whole management struct. In a previous version of the series, this was achieved with two separate patches: drm/i915/bdw: Allocate ringbuffer backing objects for default global LRC drm/i915/bdw: Allocate ringbuffer for user-created LRCs Signed-off-by: Oscar Mateo Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_lrc.c | 38 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.c | 6 +++--- drivers/gpu/drm/i915/intel_ringbuffer.h | 4 ++++ 4 files changed, 46 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index cbae19bab4b..eccb8e406e9 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -635,6 +635,7 @@ struct intel_context { /* Execlists */ struct { struct drm_i915_gem_object *state; + struct intel_ringbuffer *ringbuf; } engine[I915_NUM_RINGS]; struct list_head link; diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 9f30ee80e48..0c80bb1f542 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -65,7 +65,11 @@ void intel_lr_context_free(struct intel_context *ctx) for (i = 0; i < I915_NUM_RINGS; i++) { struct drm_i915_gem_object *ctx_obj = ctx->engine[i].state; + struct intel_ringbuffer *ringbuf = ctx->engine[i].ringbuf; + if (ctx_obj) { + intel_destroy_ringbuffer_obj(ringbuf); + kfree(ringbuf); i915_gem_object_ggtt_unpin(ctx_obj); drm_gem_object_unreference(&ctx_obj->base); } @@ -99,6 +103,7 @@ int intel_lr_context_deferred_create(struct intel_context *ctx, struct drm_device *dev = ring->dev; struct drm_i915_gem_object *ctx_obj; uint32_t context_size; + struct intel_ringbuffer *ringbuf; int ret; WARN_ON(ctx->legacy_hw_ctx.rcs_state != NULL); @@ -119,6 +124,39 @@ int intel_lr_context_deferred_create(struct intel_context *ctx, return ret; } + ringbuf = kzalloc(sizeof(*ringbuf), GFP_KERNEL); + if (!ringbuf) { + DRM_DEBUG_DRIVER("Failed to allocate ringbuffer %s\n", + ring->name); + i915_gem_object_ggtt_unpin(ctx_obj); + drm_gem_object_unreference(&ctx_obj->base); + ret = -ENOMEM; + return ret; + } + + ringbuf->size = 32 * PAGE_SIZE; + ringbuf->effective_size = ringbuf->size; + ringbuf->head = 0; + ringbuf->tail = 0; + ringbuf->space = ringbuf->size; + ringbuf->last_retired_head = -1; + + /* TODO: For now we put this in the mappable region so that we can reuse + * the existing ringbuffer code which ioremaps it. When we start + * creating many contexts, this will no longer work and we must switch + * to a kmapish interface. + */ + ret = intel_alloc_ringbuffer_obj(dev, ringbuf); + if (ret) { + DRM_DEBUG_DRIVER("Failed to allocate ringbuffer obj %s: %d\n", + ring->name, ret); + kfree(ringbuf); + i915_gem_object_ggtt_unpin(ctx_obj); + drm_gem_object_unreference(&ctx_obj->base); + return ret; + } + + ctx->engine[ring->id].ringbuf = ringbuf; ctx->engine[ring->id].state = ctx_obj; return 0; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index a059b64a0fb..064652034d7 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1519,7 +1519,7 @@ static int init_phys_status_page(struct intel_engine_cs *ring) return 0; } -static void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf) +void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf) { if (!ringbuf->obj) return; @@ -1530,8 +1530,8 @@ static void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf) ringbuf->obj = NULL; } -static int intel_alloc_ringbuffer_obj(struct drm_device *dev, - struct intel_ringbuffer *ringbuf) +int intel_alloc_ringbuffer_obj(struct drm_device *dev, + struct intel_ringbuffer *ringbuf) { struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_gem_object *obj; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 70525d0c2c7..669cc7527f9 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -355,6 +355,10 @@ intel_write_status_page(struct intel_engine_cs *ring, #define I915_GEM_HWS_SCRATCH_INDEX 0x30 #define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT) +void intel_destroy_ringbuffer_obj(struct intel_ringbuffer *ringbuf); +int intel_alloc_ringbuffer_obj(struct drm_device *dev, + struct intel_ringbuffer *ringbuf); + void intel_stop_ring_buffer(struct intel_engine_cs *ring); void intel_cleanup_ring_buffer(struct intel_engine_cs *ring); -- cgit v1.2.3-70-g09d2 From 0c7dd53b84def4fbbba907bef3d32a5171b617a5 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Mon, 11 Aug 2014 16:17:44 +0200 Subject: drm/i915/bdw: Add a context and an engine pointers to the ringbuffer Any given ringbuffer is unequivocally tied to one context and one engine. By setting the appropriate pointers to them, the ringbuffer struct holds all the infromation you might need to submit a workload for processing, Execlists style. v2: Drop ring->ctx since that looks terribly ill-defined for legacy ringbuffer submission. Signed-off-by: Oscar Mateo (v1) Acked-by: Damien Lespiau (v2) Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_lrc.c | 1 + drivers/gpu/drm/i915/intel_ringbuffer.c | 1 + drivers/gpu/drm/i915/intel_ringbuffer.h | 2 ++ 3 files changed, 4 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 0c80bb1f542..8f2d14da622 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -134,6 +134,7 @@ int intel_lr_context_deferred_create(struct intel_context *ctx, return ret; } + ringbuf->ring = ring; ringbuf->size = 32 * PAGE_SIZE; ringbuf->effective_size = ringbuf->size; ringbuf->head = 0; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 064652034d7..c35f956ed6a 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1594,6 +1594,7 @@ static int intel_init_ring_buffer(struct drm_device *dev, INIT_LIST_HEAD(&ring->active_list); INIT_LIST_HEAD(&ring->request_list); ringbuf->size = 32 * PAGE_SIZE; + ringbuf->ring = ring; memset(ring->semaphore.sync_seqno, 0, sizeof(ring->semaphore.sync_seqno)); init_waitqueue_head(&ring->irq_queue); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 669cc7527f9..fe9d9d9d359 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -90,6 +90,8 @@ struct intel_ringbuffer { struct drm_i915_gem_object *obj; void __iomem *virtual_start; + struct intel_engine_cs *ring; + u32 head; u32 tail; int space; -- cgit v1.2.3-70-g09d2 From 48d823878d64f93163f5a949623346748bbce1b4 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 24 Jul 2014 17:04:23 +0100 Subject: drm/i915/bdw: Generic logical ring init and cleanup Allocate and populate the default LRC for every ring, call gen-specific init/cleanup, init/fini the command parser and set the status page (now inside the LRC object). These are things all engines/rings have in common. Stopping the ring before cleanup and initializing the seqnos is left as a TODO task (we need more infrastructure in place before we can achieve this). v2: Check the ringbuffer backing obj for ring_is_initialized, instead of the context backing obj (similar, but not exactly the same). Signed-off-by: Oscar Mateo Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_gem_context.c | 4 --- drivers/gpu/drm/i915/intel_lrc.c | 54 +++++++++++++++++++++++++++++++-- drivers/gpu/drm/i915/intel_ringbuffer.c | 17 +++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.h | 6 +--- 4 files changed, 70 insertions(+), 11 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index bcb41002aa1..7a08f3e9e1a 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -445,10 +445,6 @@ int i915_gem_context_init(struct drm_device *dev) /* NB: RCS will hold a ref for all rings */ ring->default_context = ctx; - - /* FIXME: we really only want to do this for initialized rings */ - if (i915.enable_execlists) - intel_lr_context_deferred_create(ctx, ring); } DRM_DEBUG_DRIVER("%s context support initialized\n", diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 9c2ff8f11c9..ed7a4ff3bbd 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -110,12 +110,60 @@ void intel_logical_ring_stop(struct intel_engine_cs *ring) void intel_logical_ring_cleanup(struct intel_engine_cs *ring) { - /* TODO */ + if (!intel_ring_initialized(ring)) + return; + + /* TODO: make sure the ring is stopped */ + ring->preallocated_lazy_request = NULL; + ring->outstanding_lazy_seqno = 0; + + if (ring->cleanup) + ring->cleanup(ring); + + i915_cmd_parser_fini_ring(ring); + + if (ring->status_page.obj) { + kunmap(sg_page(ring->status_page.obj->pages->sgl)); + ring->status_page.obj = NULL; + } } static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *ring) { - /* TODO */ + int ret; + struct intel_context *dctx = ring->default_context; + struct drm_i915_gem_object *dctx_obj; + + /* Intentionally left blank. */ + ring->buffer = NULL; + + ring->dev = dev; + INIT_LIST_HEAD(&ring->active_list); + INIT_LIST_HEAD(&ring->request_list); + init_waitqueue_head(&ring->irq_queue); + + ret = intel_lr_context_deferred_create(dctx, ring); + if (ret) + return ret; + + /* The status page is offset 0 from the context object in LRCs. */ + dctx_obj = dctx->engine[ring->id].state; + ring->status_page.gfx_addr = i915_gem_obj_ggtt_offset(dctx_obj); + ring->status_page.page_addr = kmap(sg_page(dctx_obj->pages->sgl)); + if (ring->status_page.page_addr == NULL) + return -ENOMEM; + ring->status_page.obj = dctx_obj; + + ret = i915_cmd_parser_init_ring(ring); + if (ret) + return ret; + + if (ring->init) { + ret = ring->init(ring); + if (ret) + return ret; + } + return 0; } @@ -399,6 +447,8 @@ int intel_lr_context_deferred_create(struct intel_context *ctx, int ret; WARN_ON(ctx->legacy_hw_ctx.rcs_state != NULL); + if (ctx->engine[ring->id].state) + return 0; context_size = round_up(get_lr_context_size(ring), 4096); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index c35f956ed6a..e4b97f5c579 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -40,6 +40,23 @@ */ #define CACHELINE_BYTES 64 +bool +intel_ring_initialized(struct intel_engine_cs *ring) +{ + struct drm_device *dev = ring->dev; + + if (!dev) + return false; + + if (i915.enable_execlists) { + struct intel_context *dctx = ring->default_context; + struct intel_ringbuffer *ringbuf = dctx->engine[ring->id].ringbuf; + + return ringbuf->obj; + } else + return ring->buffer && ring->buffer->obj; +} + static inline int __ring_space(int head, int tail, int size) { int space = head - (tail + I915_RING_FREE_SPACE); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index fe9d9d9d359..fbe54ef6a9a 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -289,11 +289,7 @@ struct intel_engine_cs { u32 (*get_cmd_length_mask)(u32 cmd_header); }; -static inline bool -intel_ring_initialized(struct intel_engine_cs *ring) -{ - return ring->buffer && ring->buffer->obj; -} +bool intel_ring_initialized(struct intel_engine_cs *ring); static inline unsigned intel_ring_flag(struct intel_engine_cs *ring) -- cgit v1.2.3-70-g09d2 From 9b1136d505b1de5478e11b59ca59cf8ce2a33217 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 24 Jul 2014 17:04:24 +0100 Subject: drm/i915/bdw: GEN-specific logical ring init Logical rings do not need most of the initialization their legacy ringbuffer counterparts do: we just need the pipe control object for the render ring, enable Execlists on the hardware and a few workarounds. v2: Squash with: "drm/i915: Extract pipe control fini & make init outside accesible". Signed-off-by: Oscar Mateo Reviewed-by: Damien Lespiau [danvet: Make checkpatch happy.] Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_lrc.c | 54 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.c | 34 +++++++++++++-------- drivers/gpu/drm/i915/intel_ringbuffer.h | 3 ++ 3 files changed, 78 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index ed7a4ff3bbd..1a1f5f98f05 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -108,6 +108,49 @@ void intel_logical_ring_stop(struct intel_engine_cs *ring) /* TODO */ } +static int gen8_init_common_ring(struct intel_engine_cs *ring) +{ + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + + I915_WRITE(RING_MODE_GEN7(ring), + _MASKED_BIT_DISABLE(GFX_REPLAY_MODE) | + _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE)); + POSTING_READ(RING_MODE_GEN7(ring)); + DRM_DEBUG_DRIVER("Execlists enabled for %s\n", ring->name); + + memset(&ring->hangcheck, 0, sizeof(ring->hangcheck)); + + return 0; +} + +static int gen8_init_render_ring(struct intel_engine_cs *ring) +{ + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + int ret; + + ret = gen8_init_common_ring(ring); + if (ret) + return ret; + + /* 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. + * + * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv,bdw,chv + */ + I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE)); + + ret = intel_init_pipe_control(ring); + if (ret) + return ret; + + I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING)); + + return ret; +} + void intel_logical_ring_cleanup(struct intel_engine_cs *ring) { if (!intel_ring_initialized(ring)) @@ -178,6 +221,9 @@ static int logical_render_ring_init(struct drm_device *dev) ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT; + ring->init = gen8_init_render_ring; + ring->cleanup = intel_fini_pipe_control; + return logical_ring_init(dev, ring); } @@ -192,6 +238,8 @@ static int logical_bsd_ring_init(struct drm_device *dev) ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT; + ring->init = gen8_init_common_ring; + return logical_ring_init(dev, ring); } @@ -206,6 +254,8 @@ static int logical_bsd2_ring_init(struct drm_device *dev) ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT; + ring->init = gen8_init_common_ring; + return logical_ring_init(dev, ring); } @@ -220,6 +270,8 @@ static int logical_blt_ring_init(struct drm_device *dev) ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT; + ring->init = gen8_init_common_ring; + return logical_ring_init(dev, ring); } @@ -234,6 +286,8 @@ static int logical_vebox_ring_init(struct drm_device *dev) ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT; + ring->init = gen8_init_common_ring; + return logical_ring_init(dev, ring); } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index e4b97f5c579..dab5e7c7903 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -597,8 +597,25 @@ out: return ret; } -static int -init_pipe_control(struct intel_engine_cs *ring) +void +intel_fini_pipe_control(struct intel_engine_cs *ring) +{ + struct drm_device *dev = ring->dev; + + if (ring->scratch.obj == NULL) + return; + + if (INTEL_INFO(dev)->gen >= 5) { + kunmap(sg_page(ring->scratch.obj->pages->sgl)); + i915_gem_object_ggtt_unpin(ring->scratch.obj); + } + + drm_gem_object_unreference(&ring->scratch.obj->base); + ring->scratch.obj = NULL; +} + +int +intel_init_pipe_control(struct intel_engine_cs *ring) { int ret; @@ -673,7 +690,7 @@ static int init_render_ring(struct intel_engine_cs *ring) _MASKED_BIT_ENABLE(GFX_REPLAY_MODE)); if (INTEL_INFO(dev)->gen >= 5) { - ret = init_pipe_control(ring); + ret = intel_init_pipe_control(ring); if (ret) return ret; } @@ -708,16 +725,7 @@ static void render_ring_cleanup(struct intel_engine_cs *ring) dev_priv->semaphore_obj = NULL; } - if (ring->scratch.obj == NULL) - return; - - if (INTEL_INFO(dev)->gen >= 5) { - kunmap(sg_page(ring->scratch.obj->pages->sgl)); - i915_gem_object_ggtt_unpin(ring->scratch.obj); - } - - drm_gem_object_unreference(&ring->scratch.obj->base); - ring->scratch.obj = NULL; + intel_fini_pipe_control(ring); } static int gen8_rcs_signal(struct intel_engine_cs *signaller, diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index fbe54ef6a9a..677df0d7be4 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -381,6 +381,9 @@ void intel_ring_init_seqno(struct intel_engine_cs *ring, u32 seqno); int intel_ring_flush_all_caches(struct intel_engine_cs *ring); int intel_ring_invalidate_all_caches(struct intel_engine_cs *ring); +void intel_fini_pipe_control(struct intel_engine_cs *ring); +int intel_init_pipe_control(struct intel_engine_cs *ring); + int intel_init_render_ring_buffer(struct drm_device *dev); int intel_init_bsd_ring_buffer(struct drm_device *dev); int intel_init_bsd2_ring_buffer(struct drm_device *dev); -- cgit v1.2.3-70-g09d2 From 82e104cc266c6da30a30fc5028b2f0236c669cd7 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 24 Jul 2014 17:04:26 +0100 Subject: drm/i915/bdw: New logical ring submission mechanism Well, new-ish: if all this code looks familiar, that's because it's a clone of the existing submission mechanism (with some modifications here and there to adapt it to LRCs and Execlists). And why did we do this instead of reusing code, one might wonder? Well, there are some fears that the differences are big enough that they will end up breaking all platforms. Also, Execlists offer several advantages, like control over when the GPU is done with a given workload, that can help simplify the submission mechanism, no doubt. I am interested in getting Execlists to work first and foremost, but in the future this parallel submission mechanism will help us to fine tune the mechanism without affecting old gens. v2: Pass the ringbuffer only (whenever possible). Signed-off-by: Oscar Mateo Reviewed-by: Damien Lespiau [danvet: Appease checkpatch. Again. And drop the legacy sarea gunk that somehow crept in.] Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_lrc.c | 189 ++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_lrc.h | 13 +++ drivers/gpu/drm/i915/intel_ringbuffer.c | 22 ++-- drivers/gpu/drm/i915/intel_ringbuffer.h | 3 + 4 files changed, 217 insertions(+), 10 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index c9518c6261d..31025847d68 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -108,6 +108,195 @@ void intel_logical_ring_stop(struct intel_engine_cs *ring) /* TODO */ } +void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf) +{ + intel_logical_ring_advance(ringbuf); + + if (intel_ring_stopped(ringbuf->ring)) + return; + + /* TODO: how to submit a context to the ELSP is not here yet */ +} + +static int logical_ring_alloc_seqno(struct intel_engine_cs *ring) +{ + if (ring->outstanding_lazy_seqno) + return 0; + + if (ring->preallocated_lazy_request == NULL) { + struct drm_i915_gem_request *request; + + request = kmalloc(sizeof(*request), GFP_KERNEL); + if (request == NULL) + return -ENOMEM; + + ring->preallocated_lazy_request = request; + } + + return i915_gem_get_seqno(ring->dev, &ring->outstanding_lazy_seqno); +} + +static int logical_ring_wait_request(struct intel_ringbuffer *ringbuf, + int bytes) +{ + struct intel_engine_cs *ring = ringbuf->ring; + struct drm_i915_gem_request *request; + u32 seqno = 0; + int ret; + + if (ringbuf->last_retired_head != -1) { + ringbuf->head = ringbuf->last_retired_head; + ringbuf->last_retired_head = -1; + + ringbuf->space = intel_ring_space(ringbuf); + if (ringbuf->space >= bytes) + return 0; + } + + list_for_each_entry(request, &ring->request_list, list) { + if (__intel_ring_space(request->tail, ringbuf->tail, + ringbuf->size) >= bytes) { + seqno = request->seqno; + break; + } + } + + if (seqno == 0) + return -ENOSPC; + + ret = i915_wait_seqno(ring, seqno); + if (ret) + return ret; + + /* TODO: make sure we update the right ringbuffer's last_retired_head + * when retiring requests */ + i915_gem_retire_requests_ring(ring); + ringbuf->head = ringbuf->last_retired_head; + ringbuf->last_retired_head = -1; + + ringbuf->space = intel_ring_space(ringbuf); + return 0; +} + +static int logical_ring_wait_for_space(struct intel_ringbuffer *ringbuf, + int bytes) +{ + struct intel_engine_cs *ring = ringbuf->ring; + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + unsigned long end; + int ret; + + ret = logical_ring_wait_request(ringbuf, bytes); + if (ret != -ENOSPC) + return ret; + + /* Force the context submission in case we have been skipping it */ + intel_logical_ring_advance_and_submit(ringbuf); + + /* With GEM the hangcheck timer should kick us out of the loop, + * leaving it early runs the risk of corrupting GEM state (due + * to running on almost untested codepaths). But on resume + * timers don't work yet, so prevent a complete hang in that + * case by choosing an insanely large timeout. */ + end = jiffies + 60 * HZ; + + do { + ringbuf->head = I915_READ_HEAD(ring); + ringbuf->space = intel_ring_space(ringbuf); + if (ringbuf->space >= bytes) { + ret = 0; + break; + } + + msleep(1); + + if (dev_priv->mm.interruptible && signal_pending(current)) { + ret = -ERESTARTSYS; + break; + } + + ret = i915_gem_check_wedge(&dev_priv->gpu_error, + dev_priv->mm.interruptible); + if (ret) + break; + + if (time_after(jiffies, end)) { + ret = -EBUSY; + break; + } + } while (1); + + return ret; +} + +static int logical_ring_wrap_buffer(struct intel_ringbuffer *ringbuf) +{ + uint32_t __iomem *virt; + int rem = ringbuf->size - ringbuf->tail; + + if (ringbuf->space < rem) { + int ret = logical_ring_wait_for_space(ringbuf, rem); + + if (ret) + return ret; + } + + virt = ringbuf->virtual_start + ringbuf->tail; + rem /= 4; + while (rem--) + iowrite32(MI_NOOP, virt++); + + ringbuf->tail = 0; + ringbuf->space = intel_ring_space(ringbuf); + + return 0; +} + +static int logical_ring_prepare(struct intel_ringbuffer *ringbuf, int bytes) +{ + int ret; + + if (unlikely(ringbuf->tail + bytes > ringbuf->effective_size)) { + ret = logical_ring_wrap_buffer(ringbuf); + if (unlikely(ret)) + return ret; + } + + if (unlikely(ringbuf->space < bytes)) { + ret = logical_ring_wait_for_space(ringbuf, bytes); + if (unlikely(ret)) + return ret; + } + + return 0; +} + +int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords) +{ + struct intel_engine_cs *ring = ringbuf->ring; + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + int ret; + + ret = i915_gem_check_wedge(&dev_priv->gpu_error, + dev_priv->mm.interruptible); + if (ret) + return ret; + + ret = logical_ring_prepare(ringbuf, num_dwords * sizeof(uint32_t)); + if (ret) + return ret; + + /* Preallocate the olr before touching the ring */ + ret = logical_ring_alloc_seqno(ring); + if (ret) + return ret; + + ringbuf->space -= num_dwords * sizeof(uint32_t); + return 0; +} + static int gen8_init_common_ring(struct intel_engine_cs *ring) { struct drm_device *dev = ring->dev; diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index bf0eff4e9f0..4e032875c1f 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -29,6 +29,19 @@ void intel_logical_ring_stop(struct intel_engine_cs *ring); void intel_logical_ring_cleanup(struct intel_engine_cs *ring); int intel_logical_rings_init(struct drm_device *dev); +void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf); +static inline void intel_logical_ring_advance(struct intel_ringbuffer *ringbuf) +{ + ringbuf->tail &= ringbuf->size - 1; +} +static inline void intel_logical_ring_emit(struct intel_ringbuffer *ringbuf, + u32 data) +{ + iowrite32(data, ringbuf->virtual_start + ringbuf->tail); + ringbuf->tail += 4; +} +int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords); + /* Logical Ring Contexts */ void intel_lr_context_free(struct intel_context *ctx); int intel_lr_context_deferred_create(struct intel_context *ctx, diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index dab5e7c7903..0bfa018fab2 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -57,7 +57,7 @@ intel_ring_initialized(struct intel_engine_cs *ring) return ring->buffer && ring->buffer->obj; } -static inline int __ring_space(int head, int tail, int size) +int __intel_ring_space(int head, int tail, int size) { int space = head - (tail + I915_RING_FREE_SPACE); if (space < 0) @@ -65,12 +65,13 @@ static inline int __ring_space(int head, int tail, int size) return space; } -static inline int ring_space(struct intel_ringbuffer *ringbuf) +int intel_ring_space(struct intel_ringbuffer *ringbuf) { - return __ring_space(ringbuf->head & HEAD_ADDR, ringbuf->tail, ringbuf->size); + return __intel_ring_space(ringbuf->head & HEAD_ADDR, + ringbuf->tail, ringbuf->size); } -static bool intel_ring_stopped(struct intel_engine_cs *ring) +bool intel_ring_stopped(struct intel_engine_cs *ring) { struct drm_i915_private *dev_priv = ring->dev->dev_private; return dev_priv->gpu_error.stop_rings & intel_ring_flag(ring); @@ -585,7 +586,7 @@ static int init_ring_common(struct intel_engine_cs *ring) else { ringbuf->head = I915_READ_HEAD(ring); ringbuf->tail = I915_READ_TAIL(ring) & TAIL_ADDR; - ringbuf->space = ring_space(ringbuf); + ringbuf->space = intel_ring_space(ringbuf); ringbuf->last_retired_head = -1; } @@ -1702,13 +1703,14 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n) ringbuf->head = ringbuf->last_retired_head; ringbuf->last_retired_head = -1; - ringbuf->space = ring_space(ringbuf); + ringbuf->space = intel_ring_space(ringbuf); if (ringbuf->space >= n) return 0; } list_for_each_entry(request, &ring->request_list, list) { - if (__ring_space(request->tail, ringbuf->tail, ringbuf->size) >= n) { + if (__intel_ring_space(request->tail, ringbuf->tail, + ringbuf->size) >= n) { seqno = request->seqno; break; } @@ -1725,7 +1727,7 @@ static int intel_ring_wait_request(struct intel_engine_cs *ring, int n) ringbuf->head = ringbuf->last_retired_head; ringbuf->last_retired_head = -1; - ringbuf->space = ring_space(ringbuf); + ringbuf->space = intel_ring_space(ringbuf); return 0; } @@ -1754,7 +1756,7 @@ static int ring_wait_for_space(struct intel_engine_cs *ring, int n) trace_i915_ring_wait_begin(ring); do { ringbuf->head = I915_READ_HEAD(ring); - ringbuf->space = ring_space(ringbuf); + ringbuf->space = intel_ring_space(ringbuf); if (ringbuf->space >= n) { ret = 0; break; @@ -1806,7 +1808,7 @@ static int intel_wrap_ring_buffer(struct intel_engine_cs *ring) iowrite32(MI_NOOP, virt++); ringbuf->tail = 0; - ringbuf->space = ring_space(ringbuf); + ringbuf->space = intel_ring_space(ringbuf); return 0; } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 677df0d7be4..81bad364e36 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -374,6 +374,9 @@ static inline void intel_ring_advance(struct intel_engine_cs *ring) struct intel_ringbuffer *ringbuf = ring->buffer; ringbuf->tail &= ringbuf->size - 1; } +int __intel_ring_space(int head, int tail, int size); +int intel_ring_space(struct intel_ringbuffer *ringbuf); +bool intel_ring_stopped(struct intel_engine_cs *ring); void __intel_ring_advance(struct intel_engine_cs *ring); int __must_check intel_ring_idle(struct intel_engine_cs *ring); -- cgit v1.2.3-70-g09d2 From 4712274c362b7730a1c6e01c9a51a6d46f5b7f43 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 24 Jul 2014 17:04:28 +0100 Subject: drm/i915/bdw: GEN-specific logical ring emit flush Same as the legacy-style ring->flush. v2: The BSD invalidate bit still exists in GEN8! Add it for the VCS rings (but still consolidate the blt and bsd ring flushes into one). This was noticed by Brad Volkin. v3: The command for BSD and for other rings is slightly different: get it exactly the same as in gen6_ring_flush + gen6_bsd_ring_flush Signed-off-by: Oscar Mateo Reviewed-by: Damien Lespiau [danvet: Checkpatch.] Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_lrc.c | 85 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.c | 7 --- drivers/gpu/drm/i915/intel_ringbuffer.h | 10 ++++ 3 files changed, 95 insertions(+), 7 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 94f8b408764..a88fa6e9360 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -340,6 +340,86 @@ static int gen8_init_render_ring(struct intel_engine_cs *ring) return ret; } +static int gen8_emit_flush(struct intel_ringbuffer *ringbuf, + u32 invalidate_domains, + u32 unused) +{ + struct intel_engine_cs *ring = ringbuf->ring; + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + uint32_t cmd; + int ret; + + ret = intel_logical_ring_begin(ringbuf, 4); + if (ret) + return ret; + + cmd = MI_FLUSH_DW + 1; + + if (ring == &dev_priv->ring[VCS]) { + if (invalidate_domains & I915_GEM_GPU_DOMAINS) + cmd |= MI_INVALIDATE_TLB | MI_INVALIDATE_BSD | + MI_FLUSH_DW_STORE_INDEX | + MI_FLUSH_DW_OP_STOREDW; + } else { + if (invalidate_domains & I915_GEM_DOMAIN_RENDER) + cmd |= MI_INVALIDATE_TLB | MI_FLUSH_DW_STORE_INDEX | + MI_FLUSH_DW_OP_STOREDW; + } + + intel_logical_ring_emit(ringbuf, cmd); + intel_logical_ring_emit(ringbuf, + I915_GEM_HWS_SCRATCH_ADDR | + MI_FLUSH_DW_USE_GTT); + intel_logical_ring_emit(ringbuf, 0); /* upper addr */ + intel_logical_ring_emit(ringbuf, 0); /* value */ + intel_logical_ring_advance(ringbuf); + + return 0; +} + +static int gen8_emit_flush_render(struct intel_ringbuffer *ringbuf, + u32 invalidate_domains, + u32 flush_domains) +{ + struct intel_engine_cs *ring = ringbuf->ring; + u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES; + u32 flags = 0; + int ret; + + flags |= PIPE_CONTROL_CS_STALL; + + if (flush_domains) { + flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH; + flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH; + } + + if (invalidate_domains) { + flags |= PIPE_CONTROL_TLB_INVALIDATE; + flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE; + flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE; + flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE; + flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE; + flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE; + flags |= PIPE_CONTROL_QW_WRITE; + flags |= PIPE_CONTROL_GLOBAL_GTT_IVB; + } + + ret = intel_logical_ring_begin(ringbuf, 6); + if (ret) + return ret; + + intel_logical_ring_emit(ringbuf, GFX_OP_PIPE_CONTROL(6)); + intel_logical_ring_emit(ringbuf, flags); + intel_logical_ring_emit(ringbuf, scratch_addr); + intel_logical_ring_emit(ringbuf, 0); + intel_logical_ring_emit(ringbuf, 0); + intel_logical_ring_emit(ringbuf, 0); + intel_logical_ring_advance(ringbuf); + + return 0; +} + static u32 gen8_get_seqno(struct intel_engine_cs *ring, bool lazy_coherency) { return intel_read_status_page(ring, I915_GEM_HWS_INDEX); @@ -451,6 +531,7 @@ static int logical_render_ring_init(struct drm_device *dev) ring->get_seqno = gen8_get_seqno; ring->set_seqno = gen8_set_seqno; ring->emit_request = gen8_emit_request; + ring->emit_flush = gen8_emit_flush_render; return logical_ring_init(dev, ring); } @@ -470,6 +551,7 @@ static int logical_bsd_ring_init(struct drm_device *dev) ring->get_seqno = gen8_get_seqno; ring->set_seqno = gen8_set_seqno; ring->emit_request = gen8_emit_request; + ring->emit_flush = gen8_emit_flush; return logical_ring_init(dev, ring); } @@ -489,6 +571,7 @@ static int logical_bsd2_ring_init(struct drm_device *dev) ring->get_seqno = gen8_get_seqno; ring->set_seqno = gen8_set_seqno; ring->emit_request = gen8_emit_request; + ring->emit_flush = gen8_emit_flush; return logical_ring_init(dev, ring); } @@ -508,6 +591,7 @@ static int logical_blt_ring_init(struct drm_device *dev) ring->get_seqno = gen8_get_seqno; ring->set_seqno = gen8_set_seqno; ring->emit_request = gen8_emit_request; + ring->emit_flush = gen8_emit_flush; return logical_ring_init(dev, ring); } @@ -527,6 +611,7 @@ static int logical_vebox_ring_init(struct drm_device *dev) ring->get_seqno = gen8_get_seqno; ring->set_seqno = gen8_set_seqno; ring->emit_request = gen8_emit_request; + ring->emit_flush = gen8_emit_flush; return logical_ring_init(dev, ring); } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 0bfa018fab2..4236014c1cd 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -33,13 +33,6 @@ #include "i915_trace.h" #include "intel_drv.h" -/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill, - * but keeps the logic simple. Indeed, the whole purpose of this macro is just - * to give some inclination as to some of the magic values used in the various - * workarounds! - */ -#define CACHELINE_BYTES 64 - bool intel_ring_initialized(struct intel_engine_cs *ring) { diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 467885159a8..e497837c772 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -5,6 +5,13 @@ #define I915_CMD_HASH_ORDER 9 +/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill, + * but keeps the logic simple. Indeed, the whole purpose of this macro is just + * to give some inclination as to some of the magic values used in the various + * workarounds! + */ +#define CACHELINE_BYTES 64 + /* * Gen2 BSpec "1. Programming Environment" / 1.4.4.6 "Ring Buffer Use" * Gen3 BSpec "vol1c Memory Interface Functions" / 2.3.4.5 "Ring Buffer Use" @@ -218,6 +225,9 @@ struct intel_engine_cs { /* Execlists */ int (*emit_request)(struct intel_ringbuffer *ringbuf); + int (*emit_flush)(struct intel_ringbuffer *ringbuf, + u32 invalidate_domains, + u32 flush_domains); /** * List of objects currently involved in rendering from the -- cgit v1.2.3-70-g09d2 From 896ab1a5d54269b463a24194c2e4a369103b46d8 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 6 Aug 2014 15:04:51 +0200 Subject: drm/i915: Fix up checks for aliasing ppgtt A subsequent patch will no longer initialize the aliasing ppgtt if we have full ppgtt enabled, since we simply don't need that any more. Unfortunately a few places check for the aliasing ppgtt instead of checking for ppgtt in general. Fix them up. One special case are the gtt offset and size macros, which have some code to remap the aliasing ppgtt to the global gtt. The aliasing ppgtt is _not_ a logical address space, so passing that in as the vm is plain and simple a bug. So just WARN about it and carry on - we have a gracefully fall-through anyway if we can't find the vma. Reviewed-by: Michel Thierry Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_cmd_parser.c | 4 +--- drivers/gpu/drm/i915/i915_dma.c | 2 +- drivers/gpu/drm/i915/i915_gem.c | 8 ++------ drivers/gpu/drm/i915/intel_ringbuffer.c | 4 +--- 4 files changed, 5 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c index dea99d92fb4..c45856bcc8b 100644 --- a/drivers/gpu/drm/i915/i915_cmd_parser.c +++ b/drivers/gpu/drm/i915/i915_cmd_parser.c @@ -842,8 +842,6 @@ finish: */ bool i915_needs_cmd_parser(struct intel_engine_cs *ring) { - struct drm_i915_private *dev_priv = ring->dev->dev_private; - if (!ring->needs_cmd_parser) return false; @@ -852,7 +850,7 @@ bool i915_needs_cmd_parser(struct intel_engine_cs *ring) * disabled. That will cause all of the parser's PPGTT checks to * fail. For now, disable parsing when PPGTT is off. */ - if (!dev_priv->mm.aliasing_ppgtt) + if (USES_PPGTT(ring->dev)) return false; return (i915.enable_cmd_parser == 1); diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 1763fbf34e1..895f9f2f35e 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -999,7 +999,7 @@ static int i915_getparam(struct drm_device *dev, void *data, value = HAS_WT(dev); break; case I915_PARAM_HAS_ALIASING_PPGTT: - value = dev_priv->mm.aliasing_ppgtt || USES_FULL_PPGTT(dev); + value = USES_PPGTT(dev); break; case I915_PARAM_HAS_WAIT_TIMEOUT: value = 1; diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index c8404a43950..6a779509701 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -5106,9 +5106,7 @@ unsigned long i915_gem_obj_offset(struct drm_i915_gem_object *o, struct drm_i915_private *dev_priv = o->base.dev->dev_private; struct i915_vma *vma; - if (!dev_priv->mm.aliasing_ppgtt || - vm == &dev_priv->mm.aliasing_ppgtt->base) - vm = &dev_priv->gtt.base; + WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base); list_for_each_entry(vma, &o->vma_list, vma_link) { if (vma->vm == vm) @@ -5149,9 +5147,7 @@ unsigned long i915_gem_obj_size(struct drm_i915_gem_object *o, struct drm_i915_private *dev_priv = o->base.dev->dev_private; struct i915_vma *vma; - if (!dev_priv->mm.aliasing_ppgtt || - vm == &dev_priv->mm.aliasing_ppgtt->base) - vm = &dev_priv->gtt.base; + WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base); BUG_ON(list_empty(&o->vma_list)); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 4236014c1cd..13543f8528c 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -2006,9 +2006,7 @@ gen8_ring_dispatch_execbuffer(struct intel_engine_cs *ring, u64 offset, u32 len, unsigned flags) { - struct drm_i915_private *dev_priv = ring->dev->dev_private; - bool ppgtt = dev_priv->mm.aliasing_ppgtt != NULL && - !(flags & I915_DISPATCH_SECURE); + bool ppgtt = USES_PPGTT(ring->dev) && !(flags & I915_DISPATCH_SECURE); int ret; ret = intel_ring_begin(ring, 4); -- cgit v1.2.3-70-g09d2 From cc9130be805d955f0e06642e57741dd9df1fbc86 Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Thu, 24 Jul 2014 17:04:42 +0100 Subject: drm/i915/bdw: Make sure gpu reset still works with Execlists If we reset a ring after a hang, we have to make sure that we clear out all queued Execlists requests. v2: The ring is, at this point, already being correctly re-programmed for Execlists, and the hangcheck counters cleared. v3: Daniel suggests to drop the "if (execlists)" because the Execlists queue should be empty in legacy mode (which is true, if we do the INIT_LIST_HEAD). v4: Do the pending intel_runtime_pm_put Signed-off-by: Oscar Mateo Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_gem.c | 12 ++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.c | 1 + 2 files changed, 13 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 32fa1e9eb84..aa4103bdd35 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2551,6 +2551,18 @@ static void i915_gem_reset_ring_cleanup(struct drm_i915_private *dev_priv, i915_gem_free_request(request); } + while (!list_empty(&ring->execlist_queue)) { + struct intel_ctx_submit_request *submit_req; + + submit_req = list_first_entry(&ring->execlist_queue, + struct intel_ctx_submit_request, + execlist_link); + list_del(&submit_req->execlist_link); + intel_runtime_pm_put(dev_priv); + i915_gem_context_unreference(submit_req->ctx); + kfree(submit_req); + } + /* These may not have been flush before the reset, do so now */ kfree(ring->preallocated_lazy_request); ring->preallocated_lazy_request = NULL; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 13543f8528c..4fb1ec95ec0 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1612,6 +1612,7 @@ static int intel_init_ring_buffer(struct drm_device *dev, ring->dev = dev; INIT_LIST_HEAD(&ring->active_list); INIT_LIST_HEAD(&ring->request_list); + INIT_LIST_HEAD(&ring->execlist_queue); ringbuf->size = 32 * PAGE_SIZE; ringbuf->ring = ring; memset(ring->semaphore.sync_seqno, 0, sizeof(ring->semaphore.sync_seqno)); -- cgit v1.2.3-70-g09d2 From c5ad011d7d256ecbe173324029e992817194d2b0 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Mon, 4 Aug 2014 03:51:38 -0700 Subject: drm/i915: FBC flush nuke for BDW According to spec FBC on BDW and HSW are identical without any gaps. So let's copy the nuke and let FBC really start compressing stuff. Without this patch we can verify with false color that nothing is being compressed. With the nuke in place and false color it is possible to see false color debugs. Unfortunatelly on some rings like BCS on BDW we have to avoid Bits 22:18 on LRIs due to a high risk of hung. So, when using Blt ring for frontbuffer rend cache would never been cleaned and FBC would stop compressing buffer. One alternative is to cache clean on software frontbuffer tracking. v2: Fix rebase conflict. v3: Do not clean cache on BCS ring. Instead use sw frontbuffer tracking. Signed-off-by: Rodrigo Vivi Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_display.c | 3 +++ drivers/gpu/drm/i915/intel_pm.c | 10 ++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.c | 9 ++++++++- 4 files changed, 22 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 5957db463e2..f13f30d6517 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -2772,6 +2772,7 @@ extern void intel_modeset_setup_hw_state(struct drm_device *dev, extern void i915_redisable_vga(struct drm_device *dev); extern void i915_redisable_vga_power_on(struct drm_device *dev); extern bool intel_fbc_enabled(struct drm_device *dev); +extern void gen8_fbc_sw_flush(struct drm_device *dev, u32 value); extern void intel_disable_fbc(struct drm_device *dev); extern bool ironlake_set_drps(struct drm_device *dev, u8 val); extern void intel_init_pch_refclk(struct drm_device *dev); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 794aa5e611e..cbd15916372 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -9082,6 +9082,9 @@ void intel_frontbuffer_flush(struct drm_device *dev, intel_mark_fb_busy(dev, frontbuffer_bits, NULL); intel_edp_psr_flush(dev, frontbuffer_bits); + + if (IS_GEN8(dev)) + gen8_fbc_sw_flush(dev, FBC_REND_CACHE_CLEAN); } /** diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 2253f878adf..b9edfd426a1 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -345,6 +345,16 @@ bool intel_fbc_enabled(struct drm_device *dev) return dev_priv->display.fbc_enabled(dev); } +void gen8_fbc_sw_flush(struct drm_device *dev, u32 value) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + + if (!IS_GEN8(dev)) + return; + + I915_WRITE(MSG_FBC_REND_STATE, value); +} + static void intel_fbc_work_fn(struct work_struct *__work) { struct intel_fbc_work *work = diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 4fb1ec95ec0..de7654623ac 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -444,7 +444,14 @@ gen8_render_ring_flush(struct intel_engine_cs *ring, return ret; } - return gen8_emit_pipe_control(ring, flags, scratch_addr); + ret = gen8_emit_pipe_control(ring, flags, scratch_addr); + if (ret) + return ret; + + if (!invalidate_domains && flush_domains) + return gen7_ring_fbc_flush(ring, FBC_REND_NUKE); + + return 0; } static void ring_write_tail(struct intel_engine_cs *ring, -- cgit v1.2.3-70-g09d2 From 86d7f23842f1bce3ab5e8c8d0c676112bbc4c99b Mon Sep 17 00:00:00 2001 From: Arun Siluvery Date: Tue, 26 Aug 2014 14:44:50 +0100 Subject: drm/i915/bdw: Apply workarounds in render ring init function MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For BDW workarounds are currently initialized in init_clock_gating() but they are lost during reset, suspend/resume etc; this patch moves the WAs that are part of register state context to render ring init fn otherwise default context ends up with incorrect values as they don't get initialized until init_clock_gating fn. v2: Add workarounds to golden render state This method has its own issues, first of all this is different for each gen and it is generated using a tool so adding new workaround and mainitaining them across gens is not a straightforward process. v3: Use LRIs to emit these workarounds (Ville) Instead of modifying the golden render state the same LRIs are emitted from within the driver. v4: Use abstract name when exporting gen specific routines (Chris) For: VIZ-4092 Signed-off-by: Arun Siluvery Reviewed-by: Ville Syrjälä Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_gem_context.c | 6 +++ drivers/gpu/drm/i915/intel_pm.c | 48 -------------------- drivers/gpu/drm/i915/intel_ringbuffer.c | 79 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.h | 2 + 4 files changed, 87 insertions(+), 48 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 62ee178b1ed..a5221d8f158 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -628,6 +628,12 @@ done: ring->last_context = to; if (uninitialized) { + if (ring->init_context) { + ret = ring->init_context(ring); + if (ret) + DRM_ERROR("ring init context: %d\n", ret); + } + ret = i915_gem_render_state_init(ring); if (ret) DRM_ERROR("init render state: %d\n", ret); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index b9edfd426a1..71802385968 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -5536,37 +5536,12 @@ static void broadwell_init_clock_gating(struct drm_device *dev) /* FIXME(BDW): Check all the w/a, some might only apply to * pre-production hw. */ - /* WaDisablePartialInstShootdown:bdw */ - I915_WRITE(GEN8_ROW_CHICKEN, - _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE)); - - /* WaDisableThreadStallDopClockGating:bdw */ - /* FIXME: Unclear whether we really need this on production bdw. */ - I915_WRITE(GEN8_ROW_CHICKEN, - _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE)); - /* - * This GEN8_CENTROID_PIXEL_OPT_DIS W/A is only needed for - * pre-production hardware - */ - I915_WRITE(HALF_SLICE_CHICKEN3, - _MASKED_BIT_ENABLE(GEN8_CENTROID_PIXEL_OPT_DIS)); - I915_WRITE(HALF_SLICE_CHICKEN3, - _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS)); I915_WRITE(GAMTARBMODE, _MASKED_BIT_ENABLE(ARB_MODE_BWGTLB_DISABLE)); I915_WRITE(_3D_CHICKEN3, _MASKED_BIT_ENABLE(_3D_CHICKEN_SDE_LIMIT_FIFO_POLY_DEPTH(2))); - I915_WRITE(COMMON_SLICE_CHICKEN2, - _MASKED_BIT_ENABLE(GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE)); - - I915_WRITE(GEN7_HALF_SLICE_CHICKEN1, - _MASKED_BIT_ENABLE(GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE)); - - /* WaDisableDopClockGating:bdw May not be needed for production */ - I915_WRITE(GEN7_ROW_CHICKEN2, - _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE)); /* WaSwitchSolVfFArbitrationPriority:bdw */ I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL); @@ -5582,31 +5557,12 @@ static void broadwell_init_clock_gating(struct drm_device *dev) BDW_DPRS_MASK_VBLANK_SRD); } - /* Use Force Non-Coherent whenever executing a 3D context. This is a - * workaround for for a possible hang in the unlikely event a TLB - * invalidation occurs during a PSD flush. - */ - I915_WRITE(HDC_CHICKEN0, - I915_READ(HDC_CHICKEN0) | - _MASKED_BIT_ENABLE(HDC_FORCE_NON_COHERENT)); - /* WaVSRefCountFullforceMissDisable:bdw */ /* WaDSRefCountFullforceMissDisable:bdw */ I915_WRITE(GEN7_FF_THREAD_MODE, I915_READ(GEN7_FF_THREAD_MODE) & ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME)); - /* - * BSpec recommends 8x4 when MSAA is used, - * however in practice 16x4 seems fastest. - * - * Note that PS/WM thread counts depend on the WIZ hashing - * disable bit, which we don't touch here, but it's good - * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM). - */ - I915_WRITE(GEN7_GT_MODE, - GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4); - I915_WRITE(GEN6_RC_SLEEP_PSMI_CONTROL, _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE)); @@ -5614,10 +5570,6 @@ static void broadwell_init_clock_gating(struct drm_device *dev) I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) | GEN8_SDEUNIT_CLOCK_GATE_DISABLE); - /* Wa4x4STCOptimizationDisable:bdw */ - I915_WRITE(CACHE_MODE_1, - _MASKED_BIT_ENABLE(GEN8_4x4_STC_OPTIMIZATION_DISABLE)); - lpt_init_clock_gating(dev); } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index de7654623ac..1d5bfdb4fe9 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -657,6 +657,84 @@ err: return ret; } +static inline void intel_ring_emit_wa(struct intel_engine_cs *ring, + u32 addr, u32 value) +{ + intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1)); + intel_ring_emit(ring, addr); + intel_ring_emit(ring, value); +} + +static int gen8_init_workarounds(struct intel_engine_cs *ring) +{ + int ret; + + /* + * workarounds applied in this fn are part of register state context, + * they need to be re-initialized followed by gpu reset, suspend/resume, + * module reload. + */ + + /* + * update the number of dwords required based on the + * actual number of workarounds applied + */ + ret = intel_ring_begin(ring, 24); + if (ret) + return ret; + + /* WaDisablePartialInstShootdown:bdw */ + /* WaDisableThreadStallDopClockGating:bdw */ + /* FIXME: Unclear whether we really need this on production bdw. */ + intel_ring_emit_wa(ring, GEN8_ROW_CHICKEN, + _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE + | STALL_DOP_GATING_DISABLE)); + + /* WaDisableDopClockGating:bdw May not be needed for production */ + intel_ring_emit_wa(ring, GEN7_ROW_CHICKEN2, + _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE)); + + /* + * This GEN8_CENTROID_PIXEL_OPT_DIS W/A is only needed for + * pre-production hardware + */ + intel_ring_emit_wa(ring, HALF_SLICE_CHICKEN3, + _MASKED_BIT_ENABLE(GEN8_CENTROID_PIXEL_OPT_DIS + | GEN8_SAMPLER_POWER_BYPASS_DIS)); + + intel_ring_emit_wa(ring, GEN7_HALF_SLICE_CHICKEN1, + _MASKED_BIT_ENABLE(GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE)); + + intel_ring_emit_wa(ring, COMMON_SLICE_CHICKEN2, + _MASKED_BIT_ENABLE(GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE)); + + /* Use Force Non-Coherent whenever executing a 3D context. This is a + * workaround for for a possible hang in the unlikely event a TLB + * invalidation occurs during a PSD flush. + */ + intel_ring_emit_wa(ring, HDC_CHICKEN0, + _MASKED_BIT_ENABLE(HDC_FORCE_NON_COHERENT)); + + /* Wa4x4STCOptimizationDisable:bdw */ + intel_ring_emit_wa(ring, CACHE_MODE_1, + _MASKED_BIT_ENABLE(GEN8_4x4_STC_OPTIMIZATION_DISABLE)); + + /* + * BSpec recommends 8x4 when MSAA is used, + * however in practice 16x4 seems fastest. + * + * Note that PS/WM thread counts depend on the WIZ hashing + * disable bit, which we don't touch here, but it's good + * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM). + */ + intel_ring_emit_wa(ring, GEN7_GT_MODE, + GEN6_WIZ_HASHING_MASK | GEN6_WIZ_HASHING_16x4); + + intel_ring_advance(ring); + + return 0; +} + static int init_render_ring(struct intel_engine_cs *ring) { struct drm_device *dev = ring->dev; @@ -2143,6 +2221,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) dev_priv->semaphore_obj = obj; } } + ring->init_context = gen8_init_workarounds; ring->add_request = gen6_add_request; ring->flush = gen8_render_ring_flush; ring->irq_get = gen8_ring_get_irq; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 9cbf7b0ebc9..96479c89f4b 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -148,6 +148,8 @@ struct intel_engine_cs { int (*init)(struct intel_engine_cs *ring); + int (*init_context)(struct intel_engine_cs *ring); + void (*write_tail)(struct intel_engine_cs *ring, u32 value); int __must_check (*flush)(struct intel_engine_cs *ring, -- cgit v1.2.3-70-g09d2 From 888b59951ed5ac450fe3ddd7b3937d905b9bafbc Mon Sep 17 00:00:00 2001 From: Arun Siluvery Date: Tue, 26 Aug 2014 14:44:51 +0100 Subject: drm/i915/bdw: Export workaround data to debugfs The workarounds that are applied are exported to a debugfs file; this is used to verify their state after the test case (reset or suspend/resume etc). This patch is only required to support i-g-t. Signed-off-by: Arun Siluvery Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_debugfs.c | 40 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/i915_drv.h | 14 ++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.c | 23 +++++++++++++++++++ 3 files changed, 77 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 9d93af126f9..1467cc1a47a 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2628,6 +2628,45 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused) return 0; } +static int intel_wa_registers(struct seq_file *m, void *unused) +{ + int i; + int ret; + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + + if (!IS_BROADWELL(dev)) { + DRM_DEBUG_DRIVER("Workaround table not available !!\n"); + return -EINVAL; + } + + ret = mutex_lock_interruptible(&dev->struct_mutex); + if (ret) + return ret; + + intel_runtime_pm_get(dev_priv); + + seq_printf(m, "Workarounds applied: %d\n", dev_priv->num_wa_regs); + for (i = 0; i < dev_priv->num_wa_regs; ++i) { + u32 addr, mask; + + addr = dev_priv->intel_wa_regs[i].addr; + mask = dev_priv->intel_wa_regs[i].mask; + dev_priv->intel_wa_regs[i].value = I915_READ(addr) | mask; + if (dev_priv->intel_wa_regs[i].addr) + seq_printf(m, "0x%X: 0x%08X, mask: 0x%08X\n", + dev_priv->intel_wa_regs[i].addr, + dev_priv->intel_wa_regs[i].value, + dev_priv->intel_wa_regs[i].mask); + } + + intel_runtime_pm_put(dev_priv); + mutex_unlock(&dev->struct_mutex); + + return 0; +} + struct pipe_crc_info { const char *name; struct drm_device *dev; @@ -4159,6 +4198,7 @@ static const struct drm_info_list i915_debugfs_list[] = { {"i915_semaphore_status", i915_semaphore_status, 0}, {"i915_shared_dplls_info", i915_shared_dplls_info, 0}, {"i915_dp_mst_info", i915_dp_mst_info, 0}, + {"intel_wa_registers", intel_wa_registers, 0} }; #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index f13f30d6517..95820652e15 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1556,6 +1556,20 @@ struct drm_i915_private { struct intel_shared_dpll shared_dplls[I915_NUM_PLLS]; int dpio_phy_iosf_port[I915_NUM_PHYS_VLV]; + /* + * workarounds are currently applied at different places and + * changes are being done to consolidate them so exact count is + * not clear at this point, use a max value for now. + */ +#define I915_MAX_WA_REGS 16 + struct { + u32 addr; + u32 value; + /* bitmask representing WA bits */ + u32 mask; + } intel_wa_regs[I915_MAX_WA_REGS]; + u32 num_wa_regs; + /* Reclocking support */ bool render_reclock_avail; bool lvds_downclock_avail; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 1d5bfdb4fe9..14b517d61b4 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -660,20 +660,40 @@ err: static inline void intel_ring_emit_wa(struct intel_engine_cs *ring, u32 addr, u32 value) { + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + + if (dev_priv->num_wa_regs > I915_MAX_WA_REGS) + return; + intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1)); intel_ring_emit(ring, addr); intel_ring_emit(ring, value); + + dev_priv->intel_wa_regs[dev_priv->num_wa_regs].addr = addr; + dev_priv->intel_wa_regs[dev_priv->num_wa_regs].mask = (value) & 0xFFFF; + /* value is updated with the status of remaining bits of this + * register when it is read from debugfs file + */ + dev_priv->intel_wa_regs[dev_priv->num_wa_regs].value = value; + dev_priv->num_wa_regs++; + + return; } static int gen8_init_workarounds(struct intel_engine_cs *ring) { int ret; + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; /* * workarounds applied in this fn are part of register state context, * they need to be re-initialized followed by gpu reset, suspend/resume, * module reload. */ + dev_priv->num_wa_regs = 0; + memset(dev_priv->intel_wa_regs, 0, sizeof(dev_priv->intel_wa_regs)); /* * update the number of dwords required based on the @@ -732,6 +752,9 @@ static int gen8_init_workarounds(struct intel_engine_cs *ring) intel_ring_advance(ring); + DRM_DEBUG_DRIVER("Number of Workarounds applied: %d\n", + dev_priv->num_wa_regs); + return 0; } -- cgit v1.2.3-70-g09d2 From 00e1e623e62cd8452e28633182b91ddcbb70cc7c Mon Sep 17 00:00:00 2001 From: Ville Syrjälä Date: Wed, 27 Aug 2014 17:33:12 +0300 Subject: drm/i915: Init some CHV workarounds via LRIs in ring->init_context() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Follow the BDW example and apply the workarounds touching registers which are saved in the context image through LRIs in the new ring->init_context() hook. This makes Mesa much happier and eg. glxgears doesn't hang after the first frame. Cc: Arun Siluvery Signed-off-by: Ville Syrjälä [danvet: Add missing wa table initialization to avoid a functional conflict with Arun's wa table debugfs support.] Reviewed-by: "Barbalho, Rafael" Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_pm.c | 14 ---------- drivers/gpu/drm/i915/intel_ringbuffer.c | 46 +++++++++++++++++++++++++++++++-- 2 files changed, 44 insertions(+), 16 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 6bf533100da..376cc2c8751 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6011,14 +6011,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev) I915_WRITE(MI_ARB_VLV, MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE); - /* WaDisablePartialInstShootdown:chv */ - I915_WRITE(GEN8_ROW_CHICKEN, - _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE)); - - /* WaDisableThreadStallDopClockGating:chv */ - I915_WRITE(GEN8_ROW_CHICKEN, - _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE)); - /* WaVSRefCountFullforceMissDisable:chv */ /* WaDSRefCountFullforceMissDisable:chv */ I915_WRITE(GEN7_FF_THREAD_MODE, @@ -6037,10 +6029,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev) I915_WRITE(GEN8_UCGCTL6, I915_READ(GEN8_UCGCTL6) | GEN8_SDEUNIT_CLOCK_GATE_DISABLE); - /* WaDisableSamplerPowerBypass:chv (pre-production hw) */ - I915_WRITE(HALF_SLICE_CHICKEN3, - _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS)); - /* WaDisableGunitClockGating:chv (pre-production hw) */ I915_WRITE(VLV_GUNIT_CLOCK_GATE, I915_READ(VLV_GUNIT_CLOCK_GATE) | GINT_DIS); @@ -6050,8 +6038,6 @@ static void cherryview_init_clock_gating(struct drm_device *dev) _MASKED_BIT_ENABLE(GEN8_FF_DOP_CLOCK_GATE_DISABLE)); /* WaDisableDopClockGating:chv (pre-production hw) */ - I915_WRITE(GEN7_ROW_CHICKEN2, - _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE)); I915_WRITE(GEN6_UCGCTL1, I915_READ(GEN6_UCGCTL1) | GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE); } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 14b517d61b4..12135ef3606 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -681,7 +681,7 @@ static inline void intel_ring_emit_wa(struct intel_engine_cs *ring, return; } -static int gen8_init_workarounds(struct intel_engine_cs *ring) +static int bdw_init_workarounds(struct intel_engine_cs *ring) { int ret; struct drm_device *dev = ring->dev; @@ -758,6 +758,45 @@ static int gen8_init_workarounds(struct intel_engine_cs *ring) return 0; } +static int chv_init_workarounds(struct intel_engine_cs *ring) +{ + int ret; + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + + /* + * workarounds applied in this fn are part of register state context, + * they need to be re-initialized followed by gpu reset, suspend/resume, + * module reload. + */ + dev_priv->num_wa_regs = 0; + memset(dev_priv->intel_wa_regs, 0, sizeof(dev_priv->intel_wa_regs)); + + ret = intel_ring_begin(ring, 12); + if (ret) + return ret; + + /* WaDisablePartialInstShootdown:chv */ + intel_ring_emit_wa(ring, GEN8_ROW_CHICKEN, + _MASKED_BIT_ENABLE(PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE)); + + /* WaDisableThreadStallDopClockGating:chv */ + intel_ring_emit_wa(ring, GEN8_ROW_CHICKEN, + _MASKED_BIT_ENABLE(STALL_DOP_GATING_DISABLE)); + + /* WaDisableDopClockGating:chv (pre-production hw) */ + intel_ring_emit_wa(ring, GEN7_ROW_CHICKEN2, + _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE)); + + /* WaDisableSamplerPowerBypass:chv (pre-production hw) */ + intel_ring_emit_wa(ring, HALF_SLICE_CHICKEN3, + _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS)); + + intel_ring_advance(ring); + + return 0; +} + static int init_render_ring(struct intel_engine_cs *ring) { struct drm_device *dev = ring->dev; @@ -2244,7 +2283,10 @@ int intel_init_render_ring_buffer(struct drm_device *dev) dev_priv->semaphore_obj = obj; } } - ring->init_context = gen8_init_workarounds; + if (IS_CHERRYVIEW(dev)) + ring->init_context = chv_init_workarounds; + else + ring->init_context = bdw_init_workarounds; ring->add_request = gen6_add_request; ring->flush = gen8_render_ring_flush; ring->irq_get = gen8_ring_get_irq; -- cgit v1.2.3-70-g09d2 From 55820e1e840def3802fc366607f2b25e31036ab1 Mon Sep 17 00:00:00 2001 From: Damien Lespiau Date: Sat, 30 Aug 2014 16:51:00 +0100 Subject: drm/i915: Don't overrun the intel_wa_regs array When entering intel_ring_emit_wa() with num_wa_regs equal to I915_MAX_WA_REGS, we end up indexing the intel_wa_regs array beyond its allocation. Fix the check then. Signed-off-by: Damien Lespiau Reviewed-by: Arun Siluvery Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 12135ef3606..96618c0c508 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -663,7 +663,7 @@ static inline void intel_ring_emit_wa(struct intel_engine_cs *ring, struct drm_device *dev = ring->dev; struct drm_i915_private *dev_priv = dev->dev_private; - if (dev_priv->num_wa_regs > I915_MAX_WA_REGS) + if (dev_priv->num_wa_regs >= I915_MAX_WA_REGS) return; intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1)); -- cgit v1.2.3-70-g09d2 From 04ad2dc7116347a4219b13935c7569ceaab95155 Mon Sep 17 00:00:00 2001 From: Damien Lespiau Date: Sat, 30 Aug 2014 16:51:01 +0100 Subject: drm/i915: Don't silently discard workarounds If we happen to emit more than I915_MAX_WA_REGS workarounds, we will currently discard them, not even emit the LRI. Not really what we want, so warn loudly. Signed-off-by: Damien Lespiau Reviewed-by: Arun Siluvery Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 96618c0c508..5ef0f994b0b 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -663,7 +663,7 @@ static inline void intel_ring_emit_wa(struct intel_engine_cs *ring, struct drm_device *dev = ring->dev; struct drm_i915_private *dev_priv = dev->dev_private; - if (dev_priv->num_wa_regs >= I915_MAX_WA_REGS) + if (WARN_ON(dev_priv->num_wa_regs >= I915_MAX_WA_REGS)) return; intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1)); -- cgit v1.2.3-70-g09d2 From b07ba1dc78a251fc02992a35b0fd8757029566e4 Mon Sep 17 00:00:00 2001 From: Damien Lespiau Date: Sat, 30 Aug 2014 16:51:02 +0100 Subject: drm/i915: Remove unneeded brackets Signed-off-by: Damien Lespiau Reviewed-by: Arun Siluvery Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 5ef0f994b0b..f8aadc3c212 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -671,7 +671,7 @@ static inline void intel_ring_emit_wa(struct intel_engine_cs *ring, intel_ring_emit(ring, value); dev_priv->intel_wa_regs[dev_priv->num_wa_regs].addr = addr; - dev_priv->intel_wa_regs[dev_priv->num_wa_regs].mask = (value) & 0xFFFF; + dev_priv->intel_wa_regs[dev_priv->num_wa_regs].mask = value & 0xFFFF; /* value is updated with the status of remaining bits of this * register when it is read from debugfs file */ -- cgit v1.2.3-70-g09d2 From 95468892fdfeef6d1004b524e35957629efdbe00 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 7 Aug 2014 15:39:54 +0100 Subject: drm/i915: Reset the HEAD pointer for the ring after writing START MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Ville found an old w/a documented for g4x that suggested that we need to reset the HEAD after writing START. This is a useful fixup for some of the g4x ring initialisation woes, but as usual, not all. v2: Do the rewrite unconditionally anyway References: https://bugs.freedesktop.org/show_bug.cgi?id=76554 Signed-off-by: Chris Wilson Cc: Ville Syrjälä Reviewed-by: Ville Syrjälä Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index f8aadc3c212..85fc2b1a124 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -563,6 +563,14 @@ static int init_ring_common(struct intel_engine_cs *ring) * also enforces ordering), otherwise the hw might lose the new ring * register values. */ I915_WRITE_START(ring, i915_gem_obj_ggtt_offset(obj)); + + /* WaClearRingBufHeadRegAtInit:ctg,elk */ + if (I915_READ_HEAD(ring)) + DRM_DEBUG("%s initialization failed [head=%08x], fudging\n", + ring->name, I915_READ_HEAD(ring)); + I915_WRITE_HEAD(ring, 0); + (void)I915_READ_HEAD(ring); + I915_WRITE_CTL(ring, ((ringbuf->size - PAGE_SIZE) & RING_NR_PAGES) | RING_VALID); -- cgit v1.2.3-70-g09d2 From c4d69da167fa967749aeb70bc0e94a457e5d00c1 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 8 Sep 2014 14:25:41 +0100 Subject: drm/i915: Evict CS TLBs between batches MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Running igt, I was encountering the invalid TLB bug on my 845g, despite that it was using the CS workaround. Examining the w/a buffer in the error state, showed that the copy from the user batch into the workaround itself was suffering from the invalid TLB bug (the first cacheline was broken with the first two words reversed). Time to try a fresh approach. This extends the workaround to write into each page of our scratch buffer in order to overflow the TLB and evict the invalid entries. This could be refined to only do so after we update the GTT, but for simplicity, we do it before each batch. I suspect this supersedes our current workaround, but for safety keep doing both. v2: The magic number shall be 2. This doesn't conclusively prove that it is the mythical TLB bug we've been trying to workaround for so long, that it requires touching a number of pages to prevent the corruption indicates to me that it is TLB related, but the corruption (the reversed cacheline) is more subtle than a TLB bug, where we would expect it to read the wrong page entirely. Oh well, it prevents a reliable hang for me and so probably for others as well. Signed-off-by: Chris Wilson Cc: Daniel Vetter Cc: Ville Syrjälä Cc: stable@vger.kernel.org Reviewed-by: Daniel Vetter Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_reg.h | 12 ++++-- drivers/gpu/drm/i915/intel_ringbuffer.c | 66 +++++++++++++++++++-------------- 2 files changed, 47 insertions(+), 31 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index e4d7607da2c..f29b44c86a2 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -334,16 +334,20 @@ #define GFX_OP_DESTBUFFER_INFO ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1) #define GFX_OP_DRAWRECT_INFO ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3)) #define GFX_OP_DRAWRECT_INFO_I965 ((0x7900<<16)|0x2) -#define SRC_COPY_BLT_CMD ((2<<29)|(0x43<<22)|4) + +#define COLOR_BLT_CMD (2<<29 | 0x40<<22 | (5-2)) +#define SRC_COPY_BLT_CMD ((2<<29)|(0x43<<22)|4) #define XY_SRC_COPY_BLT_CMD ((2<<29)|(0x53<<22)|6) #define XY_MONO_SRC_COPY_IMM_BLT ((2<<29)|(0x71<<22)|5) -#define XY_SRC_COPY_BLT_WRITE_ALPHA (1<<21) -#define XY_SRC_COPY_BLT_WRITE_RGB (1<<20) +#define BLT_WRITE_A (2<<20) +#define BLT_WRITE_RGB (1<<20) +#define BLT_WRITE_RGBA (BLT_WRITE_RGB | BLT_WRITE_A) #define BLT_DEPTH_8 (0<<24) #define BLT_DEPTH_16_565 (1<<24) #define BLT_DEPTH_16_1555 (2<<24) #define BLT_DEPTH_32 (3<<24) -#define BLT_ROP_GXCOPY (0xcc<<16) +#define BLT_ROP_SRC_COPY (0xcc<<16) +#define BLT_ROP_COLOR_COPY (0xf0<<16) #define XY_SRC_COPY_BLT_SRC_TILED (1<<15) /* 965+ only */ #define XY_SRC_COPY_BLT_DST_TILED (1<<11) /* 965+ only */ #define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 16371a44442..2d068edd1ad 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1363,54 +1363,66 @@ i965_dispatch_execbuffer(struct intel_engine_cs *ring, /* Just userspace ABI convention to limit the wa batch bo to a resonable size */ #define I830_BATCH_LIMIT (256*1024) +#define I830_TLB_ENTRIES (2) +#define I830_WA_SIZE max(I830_TLB_ENTRIES*4096, I830_BATCH_LIMIT) static int i830_dispatch_execbuffer(struct intel_engine_cs *ring, u64 offset, u32 len, unsigned flags) { + u32 cs_offset = ring->scratch.gtt_offset; int ret; - if (flags & I915_DISPATCH_PINNED) { - ret = intel_ring_begin(ring, 4); - if (ret) - return ret; + ret = intel_ring_begin(ring, 6); + 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, MI_NOOP); - intel_ring_advance(ring); - } else { - u32 cs_offset = ring->scratch.gtt_offset; + /* Evict the invalid PTE TLBs */ + intel_ring_emit(ring, COLOR_BLT_CMD | BLT_WRITE_RGBA); + intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | 4096); + intel_ring_emit(ring, I830_TLB_ENTRIES << 16 | 4); /* load each page */ + intel_ring_emit(ring, cs_offset); + intel_ring_emit(ring, 0xdeadbeef); + intel_ring_emit(ring, MI_NOOP); + intel_ring_advance(ring); + if ((flags & I915_DISPATCH_PINNED) == 0) { if (len > I830_BATCH_LIMIT) return -ENOSPC; - ret = intel_ring_begin(ring, 9+3); + ret = intel_ring_begin(ring, 6 + 2); 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); + + /* 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, SRC_COPY_BLT_CMD | BLT_WRITE_RGBA); + intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096); + 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); + intel_ring_emit(ring, MI_NOOP); + intel_ring_advance(ring); /* ... 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); + offset = cs_offset; } + 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, MI_NOOP); + intel_ring_advance(ring); + return 0; } @@ -2200,7 +2212,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) /* Workaround batchbuffer to combat CS tlb bug. */ if (HAS_BROKEN_CS_TLB(dev)) { - obj = i915_gem_alloc_object(dev, I830_BATCH_LIMIT); + obj = i915_gem_alloc_object(dev, I830_WA_SIZE); if (obj == NULL) { DRM_ERROR("Failed to allocate batch bo\n"); return -ENOMEM; -- cgit v1.2.3-70-g09d2 From 611a7a4fd8b5fb6b25ab1f8bdcde61800a7feacf Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 12 Sep 2014 07:37:42 +0100 Subject: drm/i915: Fix SRC_COPY width on 830/845g One small change I forgot to make in commit c4d69da167fa967749aeb70bc0e94a457e5d00c1 Author: Chris Wilson Date: Mon Sep 8 14:25:41 2014 +0100 drm/i915: Evict CS TLBs between batches was to update the copy width for the compact BLT copy instruction. Reported-by: Thomas Richter Signed-off-by: Chris Wilson Cc: Thomas Richter Cc: Jani Nikula Cc: stable@vger.kernel.org Tested-by: Thomas Richter Acked-by: Daniel Vetter Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/intel_ringbuffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 2d068edd1ad..47a126a0493 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1400,7 +1400,7 @@ i830_dispatch_execbuffer(struct intel_engine_cs *ring, */ intel_ring_emit(ring, SRC_COPY_BLT_CMD | BLT_WRITE_RGBA); intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096); - intel_ring_emit(ring, DIV_ROUND_UP(len, 4096) << 16 | 1024); + intel_ring_emit(ring, DIV_ROUND_UP(len, 4096) << 16 | 4096); intel_ring_emit(ring, cs_offset); intel_ring_emit(ring, 4096); intel_ring_emit(ring, offset); -- cgit v1.2.3-70-g09d2 From 770722585639bc2da683e72f610d5f614298e415 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 10 Sep 2014 12:18:27 +0100 Subject: drm/i915: HSW always use GGTT selector for secure batches MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit gen6 and earlier conflate address space selection (ppgtt vs ggtt) with the security bit (i.e. only privileged batches were allowed to run from ggtt). From Haswell only, you are able to select the security bit separate from the address space - and we always requested to use ppgtt. This breaks the golden render state batch execution with full-ppgtt as that is only present in the global GTT and more generally any secure batch that is not colocated in the ppgtt and ggtt. So we need to disable the use of the ppgtt selector bit for secure batches, or else we hang immediately upon boot and thence after every GPU reset... v2: Only HSW differentiates between secure dispatch and ggtt, so simply ignore the differentiation and always use secure==ggtt. Signed-off-by: Chris Wilson Cc: Ville Syrjälä Reviewed-by: Ville Syrjälä [danvet: Rectify commit message as noted by Chris.] Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 109de2eeb9a..25795f2efdc 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -2203,8 +2203,9 @@ hsw_ring_dispatch_execbuffer(struct intel_engine_cs *ring, return ret; intel_ring_emit(ring, - MI_BATCH_BUFFER_START | MI_BATCH_PPGTT_HSW | - (flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE_HSW)); + MI_BATCH_BUFFER_START | + (flags & I915_DISPATCH_SECURE ? + 0 : MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW)); /* bit0-7 is the length on GEN6+ */ intel_ring_emit(ring, offset); intel_ring_advance(ring); -- cgit v1.2.3-70-g09d2 From d37cf5f7e1b315585940a735a8508d955ffc0f16 Mon Sep 17 00:00:00 2001 From: Mika Kuoppala Date: Fri, 19 Sep 2014 20:05:26 +0300 Subject: drm/i915/bdw: Cleanup pre prod workarounds MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit as these have been fixed in production hw and hurt performance if applied. v2: adjust requested ring space (Ville) Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=83482 Tested-by: zhoujian Signed-off-by: Mika Kuoppala Reviewed-by: Ville Syrjälä Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/intel_ringbuffer.c | 15 ++------------- 1 file changed, 2 insertions(+), 13 deletions(-) (limited to 'drivers/gpu/drm/i915/intel_ringbuffer.c') diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 25795f2efdc..6dc981f0671 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -707,7 +707,7 @@ static int bdw_init_workarounds(struct intel_engine_cs *ring) * update the number of dwords required based on the * actual number of workarounds applied */ - ret = intel_ring_begin(ring, 24); + ret = intel_ring_begin(ring, 18); if (ret) return ret; @@ -722,19 +722,8 @@ static int bdw_init_workarounds(struct intel_engine_cs *ring) intel_ring_emit_wa(ring, GEN7_ROW_CHICKEN2, _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE)); - /* - * This GEN8_CENTROID_PIXEL_OPT_DIS W/A is only needed for - * pre-production hardware - */ intel_ring_emit_wa(ring, HALF_SLICE_CHICKEN3, - _MASKED_BIT_ENABLE(GEN8_CENTROID_PIXEL_OPT_DIS - | GEN8_SAMPLER_POWER_BYPASS_DIS)); - - intel_ring_emit_wa(ring, GEN7_HALF_SLICE_CHICKEN1, - _MASKED_BIT_ENABLE(GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE)); - - intel_ring_emit_wa(ring, COMMON_SLICE_CHICKEN2, - _MASKED_BIT_ENABLE(GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE)); + _MASKED_BIT_ENABLE(GEN8_SAMPLER_POWER_BYPASS_DIS)); /* Use Force Non-Coherent whenever executing a 3D context. This is a * workaround for for a possible hang in the unlikely event a TLB -- cgit v1.2.3-70-g09d2