From 9d87fa2138d06ff400551800d67d522625033e35 Mon Sep 17 00:00:00 2001 From: Jerome Glisse Date: Wed, 7 Apr 2010 10:21:19 +0000 Subject: drm/ttm: split no_wait argument in 2 GPU or reserve wait There is case where we want to be able to wait only for the GPU while not waiting for other buffer to be unreserved. This patch split the no_wait argument all the way down in the whole ttm path so that upper level can decide on what to wait on or not. [airlied: squashed these 4 for bisectability reasons.] drm/radeon/kms: update to TTM no_wait splitted argument drm/nouveau: update to TTM no_wait splitted argument drm/vmwgfx: update to TTM no_wait splitted argument [vmwgfx patch: Reviewed-by: Thomas Hellstrom ] Signed-off-by: Jerome Glisse Acked-by: Thomas Hellstrom Signed-off-by: Dave Airlie --- drivers/gpu/drm/nouveau/nouveau_bo.c | 45 ++++++++++++++++++++---------------- 1 file changed, 25 insertions(+), 20 deletions(-) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 026612471c92..5a167de895c6 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -219,7 +219,7 @@ nouveau_bo_pin(struct nouveau_bo *nvbo, uint32_t memtype) for (i = 0; i < nvbo->placement.num_placement; i++) nvbo->placements[i] |= TTM_PL_FLAG_NO_EVICT; - ret = ttm_bo_validate(bo, &nvbo->placement, false, false); + ret = ttm_bo_validate(bo, &nvbo->placement, false, false, false); if (ret == 0) { switch (bo->mem.mem_type) { case TTM_PL_VRAM: @@ -256,7 +256,7 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo) for (i = 0; i < nvbo->placement.num_placement; i++) nvbo->placements[i] &= ~TTM_PL_FLAG_NO_EVICT; - ret = ttm_bo_validate(bo, &nvbo->placement, false, false); + ret = ttm_bo_validate(bo, &nvbo->placement, false, false, false); if (ret == 0) { switch (bo->mem.mem_type) { case TTM_PL_VRAM: @@ -456,7 +456,8 @@ nouveau_bo_evict_flags(struct ttm_buffer_object *bo, struct ttm_placement *pl) static int nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan, - struct nouveau_bo *nvbo, bool evict, bool no_wait, + struct nouveau_bo *nvbo, bool evict, + bool no_wait_reserve, bool no_wait_gpu, struct ttm_mem_reg *new_mem) { struct nouveau_fence *fence = NULL; @@ -467,7 +468,7 @@ nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan, return ret; ret = ttm_bo_move_accel_cleanup(&nvbo->bo, fence, NULL, - evict, no_wait, new_mem); + evict, no_wait_reserve, no_wait_gpu, new_mem); if (nvbo->channel && nvbo->channel != chan) ret = nouveau_fence_wait(fence, NULL, false, false); nouveau_fence_unref((void *)&fence); @@ -491,7 +492,8 @@ nouveau_bo_mem_ctxdma(struct nouveau_bo *nvbo, struct nouveau_channel *chan, static int nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr, - int no_wait, struct ttm_mem_reg *new_mem) + bool no_wait_reserve, bool no_wait_gpu, + struct ttm_mem_reg *new_mem) { struct nouveau_bo *nvbo = nouveau_bo(bo); struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev); @@ -569,12 +571,13 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr, dst_offset += (PAGE_SIZE * line_count); } - return nouveau_bo_move_accel_cleanup(chan, nvbo, evict, no_wait, new_mem); + return nouveau_bo_move_accel_cleanup(chan, nvbo, evict, no_wait_reserve, no_wait_gpu, new_mem); } static int nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr, - bool no_wait, struct ttm_mem_reg *new_mem) + bool no_wait_reserve, bool no_wait_gpu, + struct ttm_mem_reg *new_mem) { u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING; struct ttm_placement placement; @@ -587,7 +590,7 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr, tmp_mem = *new_mem; tmp_mem.mm_node = NULL; - ret = ttm_bo_mem_space(bo, &placement, &tmp_mem, intr, no_wait); + ret = ttm_bo_mem_space(bo, &placement, &tmp_mem, intr, no_wait_reserve, no_wait_gpu); if (ret) return ret; @@ -595,11 +598,11 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr, if (ret) goto out; - ret = nouveau_bo_move_m2mf(bo, true, intr, no_wait, &tmp_mem); + ret = nouveau_bo_move_m2mf(bo, true, intr, no_wait_reserve, no_wait_gpu, &tmp_mem); if (ret) goto out; - ret = ttm_bo_move_ttm(bo, evict, no_wait, new_mem); + ret = ttm_bo_move_ttm(bo, evict, no_wait_reserve, no_wait_gpu, new_mem); out: if (tmp_mem.mm_node) { spin_lock(&bo->bdev->glob->lru_lock); @@ -612,7 +615,8 @@ out: static int nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr, - bool no_wait, struct ttm_mem_reg *new_mem) + bool no_wait_reserve, bool no_wait_gpu, + struct ttm_mem_reg *new_mem) { u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING; struct ttm_placement placement; @@ -625,15 +629,15 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr, tmp_mem = *new_mem; tmp_mem.mm_node = NULL; - ret = ttm_bo_mem_space(bo, &placement, &tmp_mem, intr, no_wait); + ret = ttm_bo_mem_space(bo, &placement, &tmp_mem, intr, no_wait_reserve, no_wait_gpu); if (ret) return ret; - ret = ttm_bo_move_ttm(bo, evict, no_wait, &tmp_mem); + ret = ttm_bo_move_ttm(bo, evict, no_wait_reserve, no_wait_gpu, &tmp_mem); if (ret) goto out; - ret = nouveau_bo_move_m2mf(bo, evict, intr, no_wait, new_mem); + ret = nouveau_bo_move_m2mf(bo, evict, intr, no_wait_reserve, no_wait_gpu, new_mem); if (ret) goto out; @@ -700,7 +704,8 @@ nouveau_bo_vm_cleanup(struct ttm_buffer_object *bo, static int nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr, - bool no_wait, struct ttm_mem_reg *new_mem) + bool no_wait_reserve, bool no_wait_gpu, + struct ttm_mem_reg *new_mem) { struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev); struct nouveau_bo *nvbo = nouveau_bo(bo); @@ -715,7 +720,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr, /* Software copy if the card isn't up and running yet. */ if (dev_priv->init_state != NOUVEAU_CARD_INIT_DONE || !dev_priv->channel) { - ret = ttm_bo_move_memcpy(bo, evict, no_wait, new_mem); + ret = ttm_bo_move_memcpy(bo, evict, no_wait_reserve, no_wait_gpu, new_mem); goto out; } @@ -729,17 +734,17 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr, /* Hardware assisted copy. */ if (new_mem->mem_type == TTM_PL_SYSTEM) - ret = nouveau_bo_move_flipd(bo, evict, intr, no_wait, new_mem); + ret = nouveau_bo_move_flipd(bo, evict, intr, no_wait_reserve, no_wait_gpu, new_mem); else if (old_mem->mem_type == TTM_PL_SYSTEM) - ret = nouveau_bo_move_flips(bo, evict, intr, no_wait, new_mem); + ret = nouveau_bo_move_flips(bo, evict, intr, no_wait_reserve, no_wait_gpu, new_mem); else - ret = nouveau_bo_move_m2mf(bo, evict, intr, no_wait, new_mem); + ret = nouveau_bo_move_m2mf(bo, evict, intr, no_wait_reserve, no_wait_gpu, new_mem); if (!ret) goto out; /* Fallback to software copy. */ - ret = ttm_bo_move_memcpy(bo, evict, no_wait, new_mem); + ret = ttm_bo_move_memcpy(bo, evict, no_wait_reserve, no_wait_gpu, new_mem); out: if (ret) -- cgit From f32f02fd81f3177cce0c16cc7d210fcc9cad953c Mon Sep 17 00:00:00 2001 From: Jerome Glisse Date: Fri, 9 Apr 2010 14:39:25 +0200 Subject: drm/nouveau/kms: add support for new TTM fault callback V5 This add the support for the new fault callback, does change anything from driver point of view, thought it should allow nouveau to add support for unmappable VRAM. Improvement: store the aperture base in a variable so that we don't call a function to get it on each fault. Patch hasn't been tested on any hw. V2 don't derefence bo->mem.mm_node as it's not NULL only for VRAM or GTT V3 update after io_mem_reserve/io_mem_free callback balancing V4 callback has to ioremap V5 ioremap is done by ttm Signed-off-by: Jerome Glisse Signed-off-by: Dave Airlie --- drivers/gpu/drm/nouveau/nouveau_bo.c | 58 +++++++++++++++++++++++++++++++++--- 1 file changed, 54 insertions(+), 4 deletions(-) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 5a167de895c6..288c2ecd937a 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -386,8 +386,7 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, break; case TTM_PL_VRAM: man->flags = TTM_MEMTYPE_FLAG_FIXED | - TTM_MEMTYPE_FLAG_MAPPABLE | - TTM_MEMTYPE_FLAG_NEEDS_IOREMAP; + TTM_MEMTYPE_FLAG_MAPPABLE; man->available_caching = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC; man->default_caching = TTM_PL_FLAG_WC; @@ -403,8 +402,7 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, case TTM_PL_TT: switch (dev_priv->gart_info.type) { case NOUVEAU_GART_AGP: - man->flags = TTM_MEMTYPE_FLAG_MAPPABLE | - TTM_MEMTYPE_FLAG_NEEDS_IOREMAP; + man->flags = TTM_MEMTYPE_FLAG_MAPPABLE; man->available_caching = TTM_PL_FLAG_UNCACHED; man->default_caching = TTM_PL_FLAG_UNCACHED; break; @@ -761,6 +759,55 @@ nouveau_bo_verify_access(struct ttm_buffer_object *bo, struct file *filp) return 0; } +static int +nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem) +{ + struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type]; + struct drm_nouveau_private *dev_priv = nouveau_bdev(bdev); + struct drm_device *dev = dev_priv->dev; + + mem->bus.addr = NULL; + mem->bus.offset = 0; + mem->bus.size = mem->num_pages << PAGE_SHIFT; + mem->bus.base = 0; + mem->bus.is_iomem = false; + if (!(man->flags & TTM_MEMTYPE_FLAG_MAPPABLE)) + return -EINVAL; + switch (mem->mem_type) { + case TTM_PL_SYSTEM: + /* System memory */ + return 0; + case TTM_PL_TT: +#if __OS_HAS_AGP + if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) { + mem->bus.offset = mem->mm_node->start << PAGE_SHIFT; + mem->bus.base = dev_priv->gart_info.aper_base; + mem->bus.is_iomem = true; + } +#endif + break; + case TTM_PL_VRAM: + mem->bus.offset = mem->mm_node->start << PAGE_SHIFT; + mem->bus.base = drm_get_resource_start(dev, 1); + mem->bus.is_iomem = true; + break; + default: + return -EINVAL; + } + return 0; +} + +static void +nouveau_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem) +{ +} + +static int +nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo) +{ + return 0; +} + struct ttm_bo_driver nouveau_bo_driver = { .create_ttm_backend_entry = nouveau_bo_create_ttm_backend_entry, .invalidate_caches = nouveau_bo_invalidate_caches, @@ -773,5 +820,8 @@ struct ttm_bo_driver nouveau_bo_driver = { .sync_obj_flush = nouveau_fence_flush, .sync_obj_unref = nouveau_fence_unref, .sync_obj_ref = nouveau_fence_ref, + .fault_reserve_notify = &nouveau_ttm_fault_reserve_notify, + .io_mem_reserve = &nouveau_ttm_io_mem_reserve, + .io_mem_free = &nouveau_ttm_io_mem_free, }; -- cgit From 0c321c79627189204d7d0bf65ab19f5ac419abed Mon Sep 17 00:00:00 2001 From: Jerome Glisse Date: Wed, 7 Apr 2010 10:21:27 +0000 Subject: drm/ttm: remove io_ field from TTM V6 All TTM driver have been converted to new io_mem_reserve/free interface which allow driver to choose and return proper io base, offset to core TTM for ioremapping if necessary. This patch remove what is now deadcode. V2 adapt to match with change in first patch of the patchset V3 update after io_mem_reserve/io_mem_free callback balancing V4 adjust to minor cleanup V5 remove the needs ioremap flag V6 keep the ioremapping facility in TTM [airlied- squashed driver removals in here also] Signed-off-by: Jerome Glisse Reviewed-by: Thomas Hellstrom Signed-off-by: Dave Airlie --- drivers/gpu/drm/nouveau/nouveau_bo.c | 11 ----------- 1 file changed, 11 deletions(-) (limited to 'drivers/gpu/drm/nouveau/nouveau_bo.c') diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 288c2ecd937a..34be1924218f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -390,13 +390,6 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, man->available_caching = TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_WC; man->default_caching = TTM_PL_FLAG_WC; - - man->io_addr = NULL; - man->io_offset = drm_get_resource_start(dev, 1); - man->io_size = drm_get_resource_len(dev, 1); - if (man->io_size > nouveau_mem_fb_amount(dev)) - man->io_size = nouveau_mem_fb_amount(dev); - man->gpu_offset = dev_priv->vm_vram_base; break; case TTM_PL_TT: @@ -417,10 +410,6 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, dev_priv->gart_info.type); return -EINVAL; } - - man->io_offset = dev_priv->gart_info.aper_base; - man->io_size = dev_priv->gart_info.aper_size; - man->io_addr = NULL; man->gpu_offset = dev_priv->vm_gart_base; break; default: -- cgit