diff --git a/Documentation/driver-api/dma-buf.rst b/Documentation/driver-api/dma-buf.rst index b541e97c7ab1aa1226f286d3686a0317f4132d1d..c78db28519f740127a02ae956085e7a4ca9113b1 100644 --- a/Documentation/driver-api/dma-buf.rst +++ b/Documentation/driver-api/dma-buf.rst @@ -118,13 +118,13 @@ Kernel Functions and Structures Reference Reservation Objects ------------------- -.. kernel-doc:: drivers/dma-buf/reservation.c +.. kernel-doc:: drivers/dma-buf/dma-resv.c :doc: Reservation Object Overview -.. kernel-doc:: drivers/dma-buf/reservation.c +.. kernel-doc:: drivers/dma-buf/dma-resv.c :export: -.. kernel-doc:: include/linux/reservation.h +.. kernel-doc:: include/linux/dma-resv.h :internal: DMA Fences diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst index 73c51b5a0997827d8cd9adca9fe5a2a3282be100..6792fa9b6b6bd1dd058ef6f76ba6b85428ee182b 100644 --- a/Documentation/gpu/todo.rst +++ b/Documentation/gpu/todo.rst @@ -206,10 +206,10 @@ Generic fbdev defio support --------------------------- The defio support code in the fbdev core has some very specific requirements, -which means drivers need to have a special framebuffer for fbdev. Which prevents -us from using the generic fbdev emulation code everywhere. The main issue is -that it uses some fields in struct page itself, which breaks shmem gem objects -(and other things). +which means drivers need to have a special framebuffer for fbdev. The main +issue is that it uses some fields in struct page itself, which breaks shmem +gem objects (and other things). To support defio, affected drivers require +the use of a shadow buffer, which may add CPU and memory overhead. Possible solution would be to write our own defio mmap code in the drm fbdev emulation. It would need to fully wrap the existing mmap ops, forwarding diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index 433d91d710e4e3f7158ef40950a64c57c951b672..d377b4ca66bfa61ea92706c74614e8cccac52242 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -45,10 +45,10 @@ static char *dmabuffs_dname(struct dentry *dentry, char *buffer, int buflen) size_t ret = 0; dmabuf = dentry->d_fsdata; - mutex_lock(&dmabuf->lock); + dma_resv_lock(dmabuf->resv, NULL); if (dmabuf->name) ret = strlcpy(name, dmabuf->name, DMA_BUF_NAME_LEN); - mutex_unlock(&dmabuf->lock); + dma_resv_unlock(dmabuf->resv); return dynamic_dname(dentry, buffer, buflen, "/%s:%s", dentry->d_name.name, ret > 0 ? name : ""); @@ -334,7 +334,7 @@ static long dma_buf_set_name(struct dma_buf *dmabuf, const char __user *buf) if (IS_ERR(name)) return PTR_ERR(name); - mutex_lock(&dmabuf->lock); + dma_resv_lock(dmabuf->resv, NULL); if (!list_empty(&dmabuf->attachments)) { ret = -EBUSY; kfree(name); @@ -344,7 +344,7 @@ static long dma_buf_set_name(struct dma_buf *dmabuf, const char __user *buf) dmabuf->name = name; out_unlock: - mutex_unlock(&dmabuf->lock); + dma_resv_unlock(dmabuf->resv); return ret; } @@ -403,10 +403,10 @@ static void dma_buf_show_fdinfo(struct seq_file *m, struct file *file) /* Don't count the temporary reference taken inside procfs seq_show */ seq_printf(m, "count:\t%ld\n", file_count(dmabuf->file) - 1); seq_printf(m, "exp_name:\t%s\n", dmabuf->exp_name); - mutex_lock(&dmabuf->lock); + dma_resv_lock(dmabuf->resv, NULL); if (dmabuf->name) seq_printf(m, "name:\t%s\n", dmabuf->name); - mutex_unlock(&dmabuf->lock); + dma_resv_unlock(dmabuf->resv); } static const struct file_operations dma_buf_fops = { @@ -525,6 +525,10 @@ struct dma_buf *dma_buf_export(const struct dma_buf_export_info *exp_info) return ERR_PTR(-EINVAL); } + if (WARN_ON(exp_info->ops->cache_sgt_mapping && + exp_info->ops->dynamic_mapping)) + return ERR_PTR(-EINVAL); + if (!try_module_get(exp_info->owner)) return ERR_PTR(-ENOENT); @@ -645,10 +649,11 @@ void dma_buf_put(struct dma_buf *dmabuf) EXPORT_SYMBOL_GPL(dma_buf_put); /** - * dma_buf_attach - Add the device to dma_buf's attachments list; optionally, + * dma_buf_dynamic_attach - Add the device to dma_buf's attachments list; optionally, * calls attach() of dma_buf_ops to allow device-specific attach functionality - * @dmabuf: [in] buffer to attach device to. - * @dev: [in] device to be attached. + * @dmabuf: [in] buffer to attach device to. + * @dev: [in] device to be attached. + * @dynamic_mapping: [in] calling convention for map/unmap * * Returns struct dma_buf_attachment pointer for this attachment. Attachments * must be cleaned up by calling dma_buf_detach(). @@ -662,8 +667,9 @@ EXPORT_SYMBOL_GPL(dma_buf_put); * accessible to @dev, and cannot be moved to a more suitable place. This is * indicated with the error code -EBUSY. */ -struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf, - struct device *dev) +struct dma_buf_attachment * +dma_buf_dynamic_attach(struct dma_buf *dmabuf, struct device *dev, + bool dynamic_mapping) { struct dma_buf_attachment *attach; int ret; @@ -677,25 +683,69 @@ struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf, attach->dev = dev; attach->dmabuf = dmabuf; - - mutex_lock(&dmabuf->lock); + attach->dynamic_mapping = dynamic_mapping; if (dmabuf->ops->attach) { ret = dmabuf->ops->attach(dmabuf, attach); if (ret) goto err_attach; } + dma_resv_lock(dmabuf->resv, NULL); list_add(&attach->node, &dmabuf->attachments); + dma_resv_unlock(dmabuf->resv); - mutex_unlock(&dmabuf->lock); + /* When either the importer or the exporter can't handle dynamic + * mappings we cache the mapping here to avoid issues with the + * reservation object lock. + */ + if (dma_buf_attachment_is_dynamic(attach) != + dma_buf_is_dynamic(dmabuf)) { + struct sg_table *sgt; + + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_lock(attach->dmabuf->resv, NULL); + + sgt = dmabuf->ops->map_dma_buf(attach, DMA_BIDIRECTIONAL); + if (!sgt) + sgt = ERR_PTR(-ENOMEM); + if (IS_ERR(sgt)) { + ret = PTR_ERR(sgt); + goto err_unlock; + } + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_unlock(attach->dmabuf->resv); + attach->sgt = sgt; + attach->dir = DMA_BIDIRECTIONAL; + } return attach; err_attach: kfree(attach); - mutex_unlock(&dmabuf->lock); + return ERR_PTR(ret); + +err_unlock: + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_unlock(attach->dmabuf->resv); + + dma_buf_detach(dmabuf, attach); return ERR_PTR(ret); } +EXPORT_SYMBOL_GPL(dma_buf_dynamic_attach); + +/** + * dma_buf_attach - Wrapper for dma_buf_dynamic_attach + * @dmabuf: [in] buffer to attach device to. + * @dev: [in] device to be attached. + * + * Wrapper to call dma_buf_dynamic_attach() for drivers which still use a static + * mapping. + */ +struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf, + struct device *dev) +{ + return dma_buf_dynamic_attach(dmabuf, dev, false); +} EXPORT_SYMBOL_GPL(dma_buf_attach); /** @@ -711,15 +761,22 @@ void dma_buf_detach(struct dma_buf *dmabuf, struct dma_buf_attachment *attach) if (WARN_ON(!dmabuf || !attach)) return; - if (attach->sgt) + if (attach->sgt) { + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_lock(attach->dmabuf->resv, NULL); + dmabuf->ops->unmap_dma_buf(attach, attach->sgt, attach->dir); - mutex_lock(&dmabuf->lock); + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_unlock(attach->dmabuf->resv); + } + + dma_resv_lock(dmabuf->resv, NULL); list_del(&attach->node); + dma_resv_unlock(dmabuf->resv); if (dmabuf->ops->detach) dmabuf->ops->detach(dmabuf, attach); - mutex_unlock(&dmabuf->lock); kfree(attach); } EXPORT_SYMBOL_GPL(dma_buf_detach); @@ -749,6 +806,9 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *attach, if (WARN_ON(!attach || !attach->dmabuf)) return ERR_PTR(-EINVAL); + if (dma_buf_attachment_is_dynamic(attach)) + dma_resv_assert_held(attach->dmabuf->resv); + if (attach->sgt) { /* * Two mappings with different directions for the same @@ -761,6 +821,9 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *attach, return attach->sgt; } + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_assert_held(attach->dmabuf->resv); + sg_table = attach->dmabuf->ops->map_dma_buf(attach, direction); if (!sg_table) sg_table = ERR_PTR(-ENOMEM); @@ -793,9 +856,15 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach, if (WARN_ON(!attach || !attach->dmabuf || !sg_table)) return; + if (dma_buf_attachment_is_dynamic(attach)) + dma_resv_assert_held(attach->dmabuf->resv); + if (attach->sgt == sg_table) return; + if (dma_buf_is_dynamic(attach->dmabuf)) + dma_resv_assert_held(attach->dmabuf->resv); + attach->dmabuf->ops->unmap_dma_buf(attach, sg_table, direction); } EXPORT_SYMBOL_GPL(dma_buf_unmap_attachment); @@ -1171,13 +1240,10 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused) "size", "flags", "mode", "count", "ino"); list_for_each_entry(buf_obj, &db_list.head, list_node) { - ret = mutex_lock_interruptible(&buf_obj->lock); - if (ret) { - seq_puts(s, - "\tERROR locking buffer object: skipping\n"); - continue; - } + ret = dma_resv_lock_interruptible(buf_obj->resv, NULL); + if (ret) + goto error_unlock; seq_printf(s, "%08zu\t%08x\t%08x\t%08ld\t%s\t%08lu\t%s\n", buf_obj->size, @@ -1223,19 +1289,23 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused) seq_printf(s, "\t%s\n", dev_name(attach_obj->dev)); attach_count++; } + dma_resv_unlock(buf_obj->resv); seq_printf(s, "Total %d devices attached\n\n", attach_count); count++; size += buf_obj->size; - mutex_unlock(&buf_obj->lock); } seq_printf(s, "\nTotal %d objects, %zu bytes\n", count, size); mutex_unlock(&db_list.lock); return 0; + +error_unlock: + mutex_unlock(&db_list.lock); + return ret; } DEFINE_SHOW_ATTRIBUTE(dma_buf_debug); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index a07897e21526145e6ee28fd9b9f38efaedb4d4e2..d2e7b4fe565992ad03ef49e9284f2053b73821cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -613,7 +613,7 @@ static int reserve_bo_and_vm(struct kgd_mem *mem, amdgpu_vm_get_pd_bo(vm, &ctx->list, &ctx->vm_pd[0]); ret = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->list, - false, &ctx->duplicates, true); + false, &ctx->duplicates); if (!ret) ctx->reserved = true; else { @@ -686,7 +686,7 @@ static int reserve_bo_and_cond_vms(struct kgd_mem *mem, } ret = ttm_eu_reserve_buffers(&ctx->ticket, &ctx->list, - false, &ctx->duplicates, true); + false, &ctx->duplicates); if (!ret) ctx->reserved = true; else @@ -1805,8 +1805,7 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) } /* Reserve all BOs and page tables for validation */ - ret = ttm_eu_reserve_buffers(&ticket, &resv_list, false, &duplicates, - true); + ret = ttm_eu_reserve_buffers(&ticket, &resv_list, false, &duplicates); WARN(!list_empty(&duplicates), "Duplicates should be empty"); if (ret) goto out_free; @@ -2004,7 +2003,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence **ef) } ret = ttm_eu_reserve_buffers(&ctx.ticket, &ctx.list, - false, &duplicate_save, true); + false, &duplicate_save); if (ret) { pr_debug("Memory eviction: TTM Reserve Failed. Try again\n"); goto ttm_reserve_fail; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 6cfaa984cc3d6f47171190c95bb018ca68e34e39..a169ff16277f2034d9f341d2ad76c38958ed128d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -581,7 +581,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, } r = ttm_eu_reserve_buffers(&p->ticket, &p->validated, true, - &duplicates, false); + &duplicates); if (unlikely(r != 0)) { if (r != -ERESTARTSYS) DRM_ERROR("ttm_eu_reserve_buffers failed.\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c index 35a8d3c96fc90149cb9ccce335739430d78df062..08047bc4d58860af370868a867b0cdaf5d5f5b8e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c @@ -80,7 +80,7 @@ int amdgpu_map_static_csa(struct amdgpu_device *adev, struct amdgpu_vm *vm, list_add(&csa_tv.head, &list); amdgpu_vm_get_pd_bo(vm, &list, &pd); - r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL, false); + r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL); if (r) { DRM_ERROR("failed to reserve CSA,PD BOs: err=%d\n", r); return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c index 4917b548b7f2d87efff68c234f62ad9991076f96..e2eec7b6633407916513249f7826d10e2a4eba65 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c @@ -34,26 +34,11 @@ #include "amdgpu.h" #include "amdgpu_display.h" #include "amdgpu_gem.h" +#include "amdgpu_dma_buf.h" #include <drm/amdgpu_drm.h> #include <linux/dma-buf.h> #include <linux/dma-fence-array.h> -/** - * amdgpu_gem_prime_get_sg_table - &drm_driver.gem_prime_get_sg_table - * implementation - * @obj: GEM buffer object (BO) - * - * Returns: - * A scatter/gather table for the pinned pages of the BO's memory. - */ -struct sg_table *amdgpu_gem_prime_get_sg_table(struct drm_gem_object *obj) -{ - struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); - int npages = bo->tbo.num_pages; - - return drm_prime_pages_to_sg(bo->tbo.ttm->pages, npages); -} - /** * amdgpu_gem_prime_vmap - &dma_buf_ops.vmap implementation * @obj: GEM BO @@ -179,92 +164,126 @@ __dma_resv_make_exclusive(struct dma_resv *obj) } /** - * amdgpu_dma_buf_map_attach - &dma_buf_ops.attach implementation - * @dma_buf: Shared DMA buffer + * amdgpu_dma_buf_attach - &dma_buf_ops.attach implementation + * + * @dmabuf: DMA-buf where we attach to + * @attach: attachment to add + * + * Add the attachment as user to the exported DMA-buf. + */ +static int amdgpu_dma_buf_attach(struct dma_buf *dmabuf, + struct dma_buf_attachment *attach) +{ + struct drm_gem_object *obj = dmabuf->priv; + struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); + struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + int r; + + if (attach->dev->driver == adev->dev->driver) + return 0; + + r = amdgpu_bo_reserve(bo, false); + if (unlikely(r != 0)) + return r; + + /* + * We only create shared fences for internal use, but importers + * of the dmabuf rely on exclusive fences for implicitly + * tracking write hazards. As any of the current fences may + * correspond to a write, we need to convert all existing + * fences on the reservation object into a single exclusive + * fence. + */ + r = __dma_resv_make_exclusive(bo->tbo.base.resv); + if (r) + return r; + + bo->prime_shared_count++; + amdgpu_bo_unreserve(bo); + return 0; +} + +/** + * amdgpu_dma_buf_detach - &dma_buf_ops.detach implementation + * + * @dmabuf: DMA-buf where we remove the attachment from + * @attach: the attachment to remove + * + * Called when an attachment is removed from the DMA-buf. + */ +static void amdgpu_dma_buf_detach(struct dma_buf *dmabuf, + struct dma_buf_attachment *attach) +{ + struct drm_gem_object *obj = dmabuf->priv; + struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); + struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + + if (attach->dev->driver != adev->dev->driver && bo->prime_shared_count) + bo->prime_shared_count--; +} + +/** + * amdgpu_dma_buf_map - &dma_buf_ops.map_dma_buf implementation * @attach: DMA-buf attachment + * @dir: DMA direction * * Makes sure that the shared DMA buffer can be accessed by the target device. * For now, simply pins it to the GTT domain, where it should be accessible by * all DMA devices. * * Returns: - * 0 on success or a negative error code on failure. + * sg_table filled with the DMA addresses to use or ERR_PRT with negative error + * code. */ -static int amdgpu_dma_buf_map_attach(struct dma_buf *dma_buf, - struct dma_buf_attachment *attach) +static struct sg_table *amdgpu_dma_buf_map(struct dma_buf_attachment *attach, + enum dma_data_direction dir) { + struct dma_buf *dma_buf = attach->dmabuf; struct drm_gem_object *obj = dma_buf->priv; struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); - struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + struct sg_table *sgt; long r; - r = drm_gem_map_attach(dma_buf, attach); - if (r) - return r; - - r = amdgpu_bo_reserve(bo, false); - if (unlikely(r != 0)) - goto error_detach; - - - if (attach->dev->driver != adev->dev->driver) { - /* - * We only create shared fences for internal use, but importers - * of the dmabuf rely on exclusive fences for implicitly - * tracking write hazards. As any of the current fences may - * correspond to a write, we need to convert all existing - * fences on the reservation object into a single exclusive - * fence. - */ - r = __dma_resv_make_exclusive(bo->tbo.base.resv); - if (r) - goto error_unreserve; - } - - /* pin buffer into GTT */ r = amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT); if (r) - goto error_unreserve; + return ERR_PTR(r); - if (attach->dev->driver != adev->dev->driver) - bo->prime_shared_count++; + sgt = drm_prime_pages_to_sg(bo->tbo.ttm->pages, bo->tbo.num_pages); + if (IS_ERR(sgt)) + return sgt; -error_unreserve: - amdgpu_bo_unreserve(bo); + if (!dma_map_sg_attrs(attach->dev, sgt->sgl, sgt->nents, dir, + DMA_ATTR_SKIP_CPU_SYNC)) + goto error_free; -error_detach: - if (r) - drm_gem_map_detach(dma_buf, attach); - return r; + return sgt; + +error_free: + sg_free_table(sgt); + kfree(sgt); + return ERR_PTR(-ENOMEM); } /** - * amdgpu_dma_buf_map_detach - &dma_buf_ops.detach implementation - * @dma_buf: Shared DMA buffer + * amdgpu_dma_buf_unmap - &dma_buf_ops.unmap_dma_buf implementation * @attach: DMA-buf attachment + * @sgt: sg_table to unmap + * @dir: DMA direction * * This is called when a shared DMA buffer no longer needs to be accessible by * another device. For now, simply unpins the buffer from GTT. */ -static void amdgpu_dma_buf_map_detach(struct dma_buf *dma_buf, - struct dma_buf_attachment *attach) +static void amdgpu_dma_buf_unmap(struct dma_buf_attachment *attach, + struct sg_table *sgt, + enum dma_data_direction dir) { - struct drm_gem_object *obj = dma_buf->priv; + struct drm_gem_object *obj = attach->dmabuf->priv; struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj); - struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); - int ret = 0; - - ret = amdgpu_bo_reserve(bo, true); - if (unlikely(ret != 0)) - goto error; + dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir); + sg_free_table(sgt); + kfree(sgt); amdgpu_bo_unpin(bo); - if (attach->dev->driver != adev->dev->driver && bo->prime_shared_count) - bo->prime_shared_count--; - amdgpu_bo_unreserve(bo); - -error: - drm_gem_map_detach(dma_buf, attach); } /** @@ -308,10 +327,11 @@ static int amdgpu_dma_buf_begin_cpu_access(struct dma_buf *dma_buf, } const struct dma_buf_ops amdgpu_dmabuf_ops = { - .attach = amdgpu_dma_buf_map_attach, - .detach = amdgpu_dma_buf_map_detach, - .map_dma_buf = drm_gem_map_dma_buf, - .unmap_dma_buf = drm_gem_unmap_dma_buf, + .dynamic_mapping = true, + .attach = amdgpu_dma_buf_attach, + .detach = amdgpu_dma_buf_detach, + .map_dma_buf = amdgpu_dma_buf_map, + .unmap_dma_buf = amdgpu_dma_buf_unmap, .release = drm_gem_dmabuf_release, .begin_cpu_access = amdgpu_dma_buf_begin_cpu_access, .mmap = drm_gem_dmabuf_mmap, @@ -349,31 +369,28 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_gem_object *gobj, } /** - * amdgpu_gem_prime_import_sg_table - &drm_driver.gem_prime_import_sg_table - * implementation + * amdgpu_dma_buf_create_obj - create BO for DMA-buf import + * * @dev: DRM device - * @attach: DMA-buf attachment - * @sg: Scatter/gather table + * @dma_buf: DMA-buf * - * Imports shared DMA buffer memory exported by another device. + * Creates an empty SG BO for DMA-buf import. * * Returns: * A new GEM BO of the given DRM device, representing the memory * described by the given DMA-buf attachment and scatter/gather table. */ -struct drm_gem_object * -amdgpu_gem_prime_import_sg_table(struct drm_device *dev, - struct dma_buf_attachment *attach, - struct sg_table *sg) +static struct drm_gem_object * +amdgpu_dma_buf_create_obj(struct drm_device *dev, struct dma_buf *dma_buf) { - struct dma_resv *resv = attach->dmabuf->resv; + struct dma_resv *resv = dma_buf->resv; struct amdgpu_device *adev = dev->dev_private; struct amdgpu_bo *bo; struct amdgpu_bo_param bp; int ret; memset(&bp, 0, sizeof(bp)); - bp.size = attach->dmabuf->size; + bp.size = dma_buf->size; bp.byte_align = PAGE_SIZE; bp.domain = AMDGPU_GEM_DOMAIN_CPU; bp.flags = 0; @@ -384,11 +401,9 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev, if (ret) goto error; - bo->tbo.sg = sg; - bo->tbo.ttm->sg = sg; bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT; bo->preferred_domains = AMDGPU_GEM_DOMAIN_GTT; - if (attach->dmabuf->ops != &amdgpu_dmabuf_ops) + if (dma_buf->ops != &amdgpu_dmabuf_ops) bo->prime_shared_count = 1; dma_resv_unlock(resv); @@ -404,15 +419,15 @@ amdgpu_gem_prime_import_sg_table(struct drm_device *dev, * @dev: DRM device * @dma_buf: Shared DMA buffer * - * The main work is done by the &drm_gem_prime_import helper, which in turn - * uses &amdgpu_gem_prime_import_sg_table. + * Import a dma_buf into a the driver and potentially create a new GEM object. * * Returns: * GEM BO representing the shared DMA buffer for the given device. */ struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev, - struct dma_buf *dma_buf) + struct dma_buf *dma_buf) { + struct dma_buf_attachment *attach; struct drm_gem_object *obj; if (dma_buf->ops == &amdgpu_dmabuf_ops) { @@ -427,5 +442,17 @@ struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev, } } - return drm_gem_prime_import(dev, dma_buf); + obj = amdgpu_dma_buf_create_obj(dev, dma_buf); + if (IS_ERR(obj)) + return obj; + + attach = dma_buf_dynamic_attach(dma_buf, dev->dev, true); + if (IS_ERR(attach)) { + drm_gem_object_put(obj); + return ERR_CAST(attach); + } + + get_dma_buf(dma_buf); + obj->import_attach = attach; + return obj; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.h index 5012e6ab58f10549e7c8c4fe07215e2e8baa066a..ec447a7b6b28394bcfe1184fe2ab45f0e5610419 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.h @@ -25,11 +25,6 @@ #include <drm/drm_gem.h> -struct sg_table *amdgpu_gem_prime_get_sg_table(struct drm_gem_object *obj); -struct drm_gem_object * -amdgpu_gem_prime_import_sg_table(struct drm_device *dev, - struct dma_buf_attachment *attach, - struct sg_table *sg); struct dma_buf *amdgpu_gem_prime_export(struct drm_gem_object *gobj, int flags); struct drm_gem_object *amdgpu_gem_prime_import(struct drm_device *dev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 01f6a3b9e0395ce8151ec59a32fb3ec798566600..7d16e6edda62f9e3d21b32456eec74e37aa5206d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -1381,8 +1381,6 @@ static struct drm_driver kms_driver = { .prime_fd_to_handle = drm_gem_prime_fd_to_handle, .gem_prime_export = amdgpu_gem_prime_export, .gem_prime_import = amdgpu_gem_prime_import, - .gem_prime_get_sg_table = amdgpu_gem_prime_get_sg_table, - .gem_prime_import_sg_table = amdgpu_gem_prime_import_sg_table, .gem_prime_vmap = amdgpu_gem_prime_vmap, .gem_prime_vunmap = amdgpu_gem_prime_vunmap, .gem_prime_mmap = amdgpu_gem_prime_mmap, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index 5e8bdded265faf6ec1dce44bff0ee8e885841d8a..19705e399905b5dacd1a8abeb64276f213d91054 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -71,7 +71,7 @@ */ static int amdgpu_gart_dummy_page_init(struct amdgpu_device *adev) { - struct page *dummy_page = adev->mman.bdev.glob->dummy_read_page; + struct page *dummy_page = ttm_bo_glob.dummy_read_page; if (adev->dummy_page_addr) return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 6cf8862ebba1df1866f06885aad60be94c0d5930..4277125a79ee45ef61a8f9ad42f903e9b0a0fe7f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -175,7 +175,7 @@ void amdgpu_gem_object_close(struct drm_gem_object *obj, amdgpu_vm_get_pd_bo(vm, &list, &vm_pd); - r = ttm_eu_reserve_buffers(&ticket, &list, false, &duplicates, false); + r = ttm_eu_reserve_buffers(&ticket, &list, false, &duplicates); if (r) { dev_err(adev->dev, "leaking bo va because " "we fail to reserve bo (%d)\n", r); @@ -641,7 +641,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data, amdgpu_vm_get_pd_bo(&fpriv->vm, &list, &vm_pd); - r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates, false); + r = ttm_eu_reserve_buffers(&ticket, &list, true, &duplicates); if (r) goto error_unref; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index a61b0d9ab9865749ff306547b764fa2e5c6d00b8..6d34e766cd54c031d059423e7569529f92f45858 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -39,6 +39,7 @@ #include <linux/slab.h> #include <linux/swap.h> #include <linux/swiotlb.h> +#include <linux/dma-buf.h> #include <drm/ttm/ttm_bo_api.h> #include <drm/ttm/ttm_bo_driver.h> @@ -764,6 +765,7 @@ static unsigned long amdgpu_ttm_io_mem_pfn(struct ttm_buffer_object *bo, */ struct amdgpu_ttm_tt { struct ttm_dma_tt ttm; + struct drm_gem_object *gobj; u64 offset; uint64_t userptr; struct task_struct *usertask; @@ -1228,6 +1230,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo, return NULL; } gtt->ttm.ttm.func = &amdgpu_backend_func; + gtt->gobj = &bo->base; /* allocate space for the uninitialized page entries */ if (ttm_sg_tt_init(>t->ttm, bo, page_flags)) { @@ -1248,7 +1251,6 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm, { struct amdgpu_device *adev = amdgpu_ttm_adev(ttm->bdev); struct amdgpu_ttm_tt *gtt = (void *)ttm; - bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG); /* user pages are bound by amdgpu_ttm_tt_pin_userptr() */ if (gtt && gtt->userptr) { @@ -1261,7 +1263,19 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm, return 0; } - if (slave && ttm->sg) { + if (ttm->page_flags & TTM_PAGE_FLAG_SG) { + if (!ttm->sg) { + struct dma_buf_attachment *attach; + struct sg_table *sgt; + + attach = gtt->gobj->import_attach; + sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL); + if (IS_ERR(sgt)) + return PTR_ERR(sgt); + + ttm->sg = sgt; + } + drm_prime_sg_to_page_addr_arrays(ttm->sg, ttm->pages, gtt->ttm.dma_address, ttm->num_pages); @@ -1288,9 +1302,8 @@ static int amdgpu_ttm_tt_populate(struct ttm_tt *ttm, */ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm) { - struct amdgpu_device *adev; struct amdgpu_ttm_tt *gtt = (void *)ttm; - bool slave = !!(ttm->page_flags & TTM_PAGE_FLAG_SG); + struct amdgpu_device *adev; if (gtt && gtt->userptr) { amdgpu_ttm_tt_set_user_pages(ttm, NULL); @@ -1299,7 +1312,16 @@ static void amdgpu_ttm_tt_unpopulate(struct ttm_tt *ttm) return; } - if (slave) + if (ttm->sg && gtt->gobj->import_attach) { + struct dma_buf_attachment *attach; + + attach = gtt->gobj->import_attach; + dma_buf_unmap_attachment(attach, ttm->sg, DMA_BIDIRECTIONAL); + ttm->sg = NULL; + return; + } + + if (ttm->page_flags & TTM_PAGE_FLAG_SG) return; adev = amdgpu_ttm_adev(ttm->bdev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 397bf5677ff251a86bb4c8522950d01c0b41b23f..e775f271f1edfbb60fff3f52cae9be81bdb9b6bd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -610,19 +610,18 @@ void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo) void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev, struct amdgpu_vm *vm) { - struct ttm_bo_global *glob = adev->mman.bdev.glob; struct amdgpu_vm_bo_base *bo_base; if (vm->bulk_moveable) { - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); ttm_bo_bulk_move_lru_tail(&vm->lru_bulk_move); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); return; } memset(&vm->lru_bulk_move, 0, sizeof(vm->lru_bulk_move)); - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); list_for_each_entry(bo_base, &vm->idle, vm_status) { struct amdgpu_bo *bo = bo_base->bo; @@ -634,7 +633,7 @@ void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev, ttm_bo_move_to_lru_tail(&bo->shadow->tbo, &vm->lru_bulk_move); } - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); vm->bulk_moveable = true; } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index b3d4403b7d9f81b3d8162fc1e8574588e7f1e2cb..8c250fd07015aa07ab0a918cb5a29b37bddb59bb 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -4596,7 +4596,7 @@ static int dm_plane_helper_prepare_fb(struct drm_plane *plane, tv.num_shared = 1; list_add(&tv.head, &list); - r = ttm_eu_reserve_buffers(&ticket, &list, false, NULL, true); + r = ttm_eu_reserve_buffers(&ticket, &list, false, NULL); if (r) { dev_err(adev->dev, "fail to reserve bo (%d)\n", r); return r; diff --git a/drivers/gpu/drm/bochs/bochs_kms.c b/drivers/gpu/drm/bochs/bochs_kms.c index 02a9c1ed165bba9871d3af1b6759fe8647da46e6..3f0006c2470d1a28d61fb03853eac3b3f6470c1f 100644 --- a/drivers/gpu/drm/bochs/bochs_kms.c +++ b/drivers/gpu/drm/bochs/bochs_kms.c @@ -69,33 +69,11 @@ static void bochs_pipe_update(struct drm_simple_display_pipe *pipe, } } -static int bochs_pipe_prepare_fb(struct drm_simple_display_pipe *pipe, - struct drm_plane_state *new_state) -{ - struct drm_gem_vram_object *gbo; - - if (!new_state->fb) - return 0; - gbo = drm_gem_vram_of_gem(new_state->fb->obj[0]); - return drm_gem_vram_pin(gbo, DRM_GEM_VRAM_PL_FLAG_VRAM); -} - -static void bochs_pipe_cleanup_fb(struct drm_simple_display_pipe *pipe, - struct drm_plane_state *old_state) -{ - struct drm_gem_vram_object *gbo; - - if (!old_state->fb) - return; - gbo = drm_gem_vram_of_gem(old_state->fb->obj[0]); - drm_gem_vram_unpin(gbo); -} - static const struct drm_simple_display_pipe_funcs bochs_pipe_funcs = { .enable = bochs_pipe_enable, .update = bochs_pipe_update, - .prepare_fb = bochs_pipe_prepare_fb, - .cleanup_fb = bochs_pipe_cleanup_fb, + .prepare_fb = drm_gem_vram_simple_display_pipe_prepare_fb, + .cleanup_fb = drm_gem_vram_simple_display_pipe_cleanup_fb, }; static int bochs_connector_get_modes(struct drm_connector *connector) diff --git a/drivers/gpu/drm/cirrus/cirrus.c b/drivers/gpu/drm/cirrus/cirrus.c index 7d08d067e1a4aa8def5a331c1c1e55302b4724df..248c9f765c45a0877422f2ec62ddd3900a4c70a2 100644 --- a/drivers/gpu/drm/cirrus/cirrus.c +++ b/drivers/gpu/drm/cirrus/cirrus.c @@ -390,7 +390,7 @@ static int cirrus_conn_init(struct cirrus_device *cirrus) /* ------------------------------------------------------------------ */ /* cirrus (simple) display pipe */ -static enum drm_mode_status cirrus_pipe_mode_valid(struct drm_crtc *crtc, +static enum drm_mode_status cirrus_pipe_mode_valid(struct drm_simple_display_pipe *pipe, const struct drm_display_mode *mode) { if (cirrus_check_size(mode->hdisplay, mode->vdisplay, NULL) < 0) diff --git a/drivers/gpu/drm/drm_cache.c b/drivers/gpu/drm/drm_cache.c index 3bd76e918b5d29ab5e37a733960925d765c080c9..03e01b000f7acb03197c2bd95564db642ee79f5f 100644 --- a/drivers/gpu/drm/drm_cache.c +++ b/drivers/gpu/drm/drm_cache.c @@ -62,10 +62,10 @@ static void drm_cache_flush_clflush(struct page *pages[], { unsigned long i; - mb(); + mb(); /*Full memory barrier used before so that CLFLUSH is ordered*/ for (i = 0; i < num_pages; i++) drm_clflush_page(*pages++); - mb(); + mb(); /*Also used after CLFLUSH so that all cache is flushed*/ } #endif @@ -92,6 +92,7 @@ drm_clflush_pages(struct page *pages[], unsigned long num_pages) #elif defined(__powerpc__) unsigned long i; + for (i = 0; i < num_pages; i++) { struct page *page = pages[i]; void *page_virtual; @@ -125,10 +126,10 @@ drm_clflush_sg(struct sg_table *st) if (static_cpu_has(X86_FEATURE_CLFLUSH)) { struct sg_page_iter sg_iter; - mb(); + mb(); /*CLFLUSH is ordered only by using memory barriers*/ for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) drm_clflush_page(sg_page_iter_page(&sg_iter)); - mb(); + mb(); /*Make sure that all cache line entry is flushed*/ return; } @@ -157,12 +158,13 @@ drm_clflush_virt_range(void *addr, unsigned long length) if (static_cpu_has(X86_FEATURE_CLFLUSH)) { const int size = boot_cpu_data.x86_clflush_size; void *end = addr + length; + addr = (void *)(((unsigned long)addr) & -size); - mb(); + mb(); /*CLFLUSH is only ordered with a full memory barrier*/ for (; addr < end; addr += size) clflushopt(addr); clflushopt(end - 1); /* force serialisation */ - mb(); + mb(); /*Ensure that evry data cache line entry is flushed*/ return; } diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index b75ae8555baf91d5214d55b516eb5c0de20cadaf..8ebeccdeed2334ebb06f9562668705074c7342b1 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -92,9 +92,12 @@ static DEFINE_MUTEX(kernel_fb_helper_lock); * * Drivers that support a dumb buffer with a virtual address and mmap support, * should try out the generic fbdev emulation using drm_fbdev_generic_setup(). + * It will automatically set up deferred I/O if the driver requires a shadow + * buffer. * - * Setup fbdev emulation by calling drm_fb_helper_fbdev_setup() and tear it - * down by calling drm_fb_helper_fbdev_teardown(). + * For other drivers, setup fbdev emulation by calling + * drm_fb_helper_fbdev_setup() and tear it down by calling + * drm_fb_helper_fbdev_teardown(). * * At runtime drivers should restore the fbdev console by using * drm_fb_helper_lastclose() as their &drm_driver.lastclose callback. @@ -127,8 +130,10 @@ static DEFINE_MUTEX(kernel_fb_helper_lock); * always run in process context since the fb_*() function could be running in * atomic context. If drm_fb_helper_deferred_io() is used as the deferred_io * callback it will also schedule dirty_work with the damage collected from the - * mmap page writes. Drivers can use drm_fb_helper_defio_init() to setup - * deferred I/O (coupled with drm_fb_helper_fbdev_teardown()). + * mmap page writes. + * + * Deferred I/O is not compatible with SHMEM. Such drivers should request an + * fbdev shadow buffer and call drm_fbdev_generic_setup() instead. */ static void drm_fb_helper_restore_lut_atomic(struct drm_crtc *crtc) @@ -679,49 +684,6 @@ void drm_fb_helper_deferred_io(struct fb_info *info, } EXPORT_SYMBOL(drm_fb_helper_deferred_io); -/** - * drm_fb_helper_defio_init - fbdev deferred I/O initialization - * @fb_helper: driver-allocated fbdev helper - * - * This function allocates &fb_deferred_io, sets callback to - * drm_fb_helper_deferred_io(), delay to 50ms and calls fb_deferred_io_init(). - * It should be called from the &drm_fb_helper_funcs->fb_probe callback. - * drm_fb_helper_fbdev_teardown() cleans up deferred I/O. - * - * NOTE: A copy of &fb_ops is made and assigned to &info->fbops. This is done - * because fb_deferred_io_cleanup() clears &fbops->fb_mmap and would thereby - * affect other instances of that &fb_ops. - * - * Returns: - * 0 on success or a negative error code on failure. - */ -int drm_fb_helper_defio_init(struct drm_fb_helper *fb_helper) -{ - struct fb_info *info = fb_helper->fbdev; - struct fb_deferred_io *fbdefio; - struct fb_ops *fbops; - - fbdefio = kzalloc(sizeof(*fbdefio), GFP_KERNEL); - fbops = kzalloc(sizeof(*fbops), GFP_KERNEL); - if (!fbdefio || !fbops) { - kfree(fbdefio); - kfree(fbops); - return -ENOMEM; - } - - info->fbdefio = fbdefio; - fbdefio->delay = msecs_to_jiffies(50); - fbdefio->deferred_io = drm_fb_helper_deferred_io; - - *fbops = *info->fbops; - info->fbops = fbops; - - fb_deferred_io_init(info); - - return 0; -} -EXPORT_SYMBOL(drm_fb_helper_defio_init); - /** * drm_fb_helper_sys_read - wrapper around fb_sys_read * @info: fb_info struct pointer @@ -2356,7 +2318,10 @@ static const struct drm_client_funcs drm_fbdev_client_funcs = { * * Drivers that set the dirty callback on their framebuffer will get a shadow * fbdev buffer that is blitted onto the real buffer. This is done in order to - * make deferred I/O work with all kinds of buffers. + * make deferred I/O work with all kinds of buffers. A shadow buffer can be + * requested explicitly by setting struct drm_mode_config.prefer_shadow or + * struct drm_mode_config.prefer_shadow_fbdev to true beforehand. This is + * required to use generic fbdev emulation with SHMEM helpers. * * This function is safe to call even when there are no connectors present. * Setup will be retried on the next hotplug event. diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 56f42e0f25847e9ef5bdf3b539af24e69ab61630..2f2b889096b07699cdaa8d8179239bb4c9ea9189 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -1106,6 +1106,9 @@ int drm_gem_mmap_obj(struct drm_gem_object *obj, unsigned long obj_size, return -EINVAL; if (obj->funcs && obj->funcs->mmap) { + /* Remove the fake offset */ + vma->vm_pgoff -= drm_vma_node_start(&obj->vma_node); + ret = obj->funcs->mmap(obj, vma); if (ret) return ret; diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_gem_shmem_helper.c index 3bc69b1ffa7dc5554197b99e4b84c99844cb1673..0810d3ef6961ccb59f111bc74f692cab0d7bc87b 100644 --- a/drivers/gpu/drm/drm_gem_shmem_helper.c +++ b/drivers/gpu/drm/drm_gem_shmem_helper.c @@ -541,9 +541,6 @@ int drm_gem_shmem_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma) vma->vm_page_prot = pgprot_decrypted(vma->vm_page_prot); vma->vm_ops = &drm_gem_shmem_vm_ops; - /* Remove the fake offset */ - vma->vm_pgoff -= drm_vma_node_start(&shmem->base.vma_node); - return 0; } EXPORT_SYMBOL_GPL(drm_gem_shmem_mmap); diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c index b86fe0fa9d0501ef08db478851c4746598580a60..666cb4c22bb9b12fcfcf63b6d930d3c614143b33 100644 --- a/drivers/gpu/drm/drm_gem_vram_helper.c +++ b/drivers/gpu/drm/drm_gem_vram_helper.c @@ -3,10 +3,13 @@ #include <drm/drm_debugfs.h> #include <drm/drm_device.h> #include <drm/drm_file.h> +#include <drm/drm_framebuffer.h> #include <drm/drm_gem_ttm_helper.h> #include <drm/drm_gem_vram_helper.h> #include <drm/drm_mode.h> +#include <drm/drm_plane.h> #include <drm/drm_prime.h> +#include <drm/drm_simple_kms_helper.h> #include <drm/ttm/ttm_page_alloc.h> static const struct drm_gem_object_funcs drm_gem_vram_object_funcs; @@ -646,6 +649,129 @@ int drm_gem_vram_driver_dumb_mmap_offset(struct drm_file *file, } EXPORT_SYMBOL(drm_gem_vram_driver_dumb_mmap_offset); +/* + * Helpers for struct drm_plane_helper_funcs + */ + +/** + * drm_gem_vram_plane_helper_prepare_fb() - \ + * Implements &struct drm_plane_helper_funcs.prepare_fb + * @plane: a DRM plane + * @new_state: the plane's new state + * + * During plane updates, this function pins the GEM VRAM + * objects of the plane's new framebuffer to VRAM. Call + * drm_gem_vram_plane_helper_cleanup_fb() to unpin them. + * + * Returns: + * 0 on success, or + * a negative errno code otherwise. + */ +int +drm_gem_vram_plane_helper_prepare_fb(struct drm_plane *plane, + struct drm_plane_state *new_state) +{ + size_t i; + struct drm_gem_vram_object *gbo; + int ret; + + if (!new_state->fb) + return 0; + + for (i = 0; i < ARRAY_SIZE(new_state->fb->obj); ++i) { + if (!new_state->fb->obj[i]) + continue; + gbo = drm_gem_vram_of_gem(new_state->fb->obj[i]); + ret = drm_gem_vram_pin(gbo, DRM_GEM_VRAM_PL_FLAG_VRAM); + if (ret) + goto err_drm_gem_vram_unpin; + } + + return 0; + +err_drm_gem_vram_unpin: + while (i) { + --i; + gbo = drm_gem_vram_of_gem(new_state->fb->obj[i]); + drm_gem_vram_unpin(gbo); + } + return ret; +} +EXPORT_SYMBOL(drm_gem_vram_plane_helper_prepare_fb); + +/** + * drm_gem_vram_plane_helper_cleanup_fb() - \ + * Implements &struct drm_plane_helper_funcs.cleanup_fb + * @plane: a DRM plane + * @old_state: the plane's old state + * + * During plane updates, this function unpins the GEM VRAM + * objects of the plane's old framebuffer from VRAM. Complements + * drm_gem_vram_plane_helper_prepare_fb(). + */ +void +drm_gem_vram_plane_helper_cleanup_fb(struct drm_plane *plane, + struct drm_plane_state *old_state) +{ + size_t i; + struct drm_gem_vram_object *gbo; + + if (!old_state->fb) + return; + + for (i = 0; i < ARRAY_SIZE(old_state->fb->obj); ++i) { + if (!old_state->fb->obj[i]) + continue; + gbo = drm_gem_vram_of_gem(old_state->fb->obj[i]); + drm_gem_vram_unpin(gbo); + } +} +EXPORT_SYMBOL(drm_gem_vram_plane_helper_cleanup_fb); + +/* + * Helpers for struct drm_simple_display_pipe_funcs + */ + +/** + * drm_gem_vram_simple_display_pipe_prepare_fb() - \ + * Implements &struct drm_simple_display_pipe_funcs.prepare_fb + * @pipe: a simple display pipe + * @new_state: the plane's new state + * + * During plane updates, this function pins the GEM VRAM + * objects of the plane's new framebuffer to VRAM. Call + * drm_gem_vram_simple_display_pipe_cleanup_fb() to unpin them. + * + * Returns: + * 0 on success, or + * a negative errno code otherwise. + */ +int drm_gem_vram_simple_display_pipe_prepare_fb( + struct drm_simple_display_pipe *pipe, + struct drm_plane_state *new_state) +{ + return drm_gem_vram_plane_helper_prepare_fb(&pipe->plane, new_state); +} +EXPORT_SYMBOL(drm_gem_vram_simple_display_pipe_prepare_fb); + +/** + * drm_gem_vram_simple_display_pipe_cleanup_fb() - \ + * Implements &struct drm_simple_display_pipe_funcs.cleanup_fb + * @pipe: a simple display pipe + * @old_state: the plane's old state + * + * During plane updates, this function unpins the GEM VRAM + * objects of the plane's old framebuffer from VRAM. Complements + * drm_gem_vram_simple_display_pipe_prepare_fb(). + */ +void drm_gem_vram_simple_display_pipe_cleanup_fb( + struct drm_simple_display_pipe *pipe, + struct drm_plane_state *old_state) +{ + drm_gem_vram_plane_helper_cleanup_fb(&pipe->plane, old_state); +} +EXPORT_SYMBOL(drm_gem_vram_simple_display_pipe_cleanup_fb); + /* * PRIME helpers */ @@ -887,12 +1013,11 @@ static int drm_vram_mm_debugfs(struct seq_file *m, void *data) struct drm_info_node *node = (struct drm_info_node *) m->private; struct drm_vram_mm *vmm = node->minor->dev->vram_mm; struct drm_mm *mm = vmm->bdev.man[TTM_PL_VRAM].priv; - struct ttm_bo_global *glob = vmm->bdev.glob; struct drm_printer p = drm_seq_file_printer(m); - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); drm_mm_print(mm, &p); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); return 0; } diff --git a/drivers/gpu/drm/drm_simple_kms_helper.c b/drivers/gpu/drm/drm_simple_kms_helper.c index 046055719245a2861583f061b8ac0752ecc902ff..15fb516ae2d82338e7122d5fd0afca939cfafc29 100644 --- a/drivers/gpu/drm/drm_simple_kms_helper.c +++ b/drivers/gpu/drm/drm_simple_kms_helper.c @@ -43,7 +43,7 @@ drm_simple_kms_crtc_mode_valid(struct drm_crtc *crtc, /* Anything goes */ return MODE_OK; - return pipe->funcs->mode_valid(crtc, mode); + return pipe->funcs->mode_valid(pipe, mode); } static int drm_simple_kms_crtc_check(struct drm_crtc *crtc, diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.c b/drivers/gpu/drm/gma500/mdfld_dsi_output.c index 03023fa0fb6f9d3c3895e96557e01ae514a319b3..f350ac1ead18213e296e4371dd0f89f3763f7f56 100644 --- a/drivers/gpu/drm/gma500/mdfld_dsi_output.c +++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.c @@ -498,7 +498,7 @@ void mdfld_dsi_output_init(struct drm_device *dev, return; } - /*create a new connetor*/ + /*create a new connector*/ dsi_connector = kzalloc(sizeof(struct mdfld_dsi_connector), GFP_KERNEL); if (!dsi_connector) { DRM_ERROR("No memory"); diff --git a/drivers/gpu/drm/hisilicon/hibmc/hibmc_drm_de.c b/drivers/gpu/drm/hisilicon/hibmc/hibmc_drm_de.c index cc4c41748cfb570649197e2e689f30ae3a4e230e..6527a97f68a3cc003e8b0a9da017cc2c9a8ca847 100644 --- a/drivers/gpu/drm/hisilicon/hibmc/hibmc_drm_de.c +++ b/drivers/gpu/drm/hisilicon/hibmc/hibmc_drm_de.c @@ -96,7 +96,6 @@ static void hibmc_plane_atomic_update(struct drm_plane *plane, { struct drm_plane_state *state = plane->state; u32 reg; - int ret; s64 gpu_addr = 0; unsigned int line_l; struct hibmc_drm_private *priv = plane->dev->dev_private; @@ -109,16 +108,9 @@ static void hibmc_plane_atomic_update(struct drm_plane *plane, hibmc_fb = to_hibmc_framebuffer(state->fb); gbo = drm_gem_vram_of_gem(hibmc_fb->obj); - ret = drm_gem_vram_pin(gbo, DRM_GEM_VRAM_PL_FLAG_VRAM); - if (ret) { - DRM_ERROR("failed to pin bo: %d", ret); - return; - } gpu_addr = drm_gem_vram_offset(gbo); - if (gpu_addr < 0) { - drm_gem_vram_unpin(gbo); - return; - } + if (WARN_ON_ONCE(gpu_addr < 0)) + return; /* Bug: we didn't pin the BO to VRAM in prepare_fb. */ writel(gpu_addr, priv->mmio + HIBMC_CRT_FB_ADDRESS); @@ -157,6 +149,8 @@ static struct drm_plane_funcs hibmc_plane_funcs = { }; static const struct drm_plane_helper_funcs hibmc_plane_helper_funcs = { + .prepare_fb = drm_gem_vram_plane_helper_prepare_fb, + .cleanup_fb = drm_gem_vram_plane_helper_cleanup_fb, .atomic_check = hibmc_plane_atomic_check, .atomic_update = hibmc_plane_atomic_update, }; diff --git a/drivers/gpu/drm/mediatek/mtk_drm_gem.c b/drivers/gpu/drm/mediatek/mtk_drm_gem.c index ca672f1d140de9792c049de6a6b28d7122c93de3..b04a3c2b111e09f7dcebb6bbc628a0f8bf50915a 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_gem.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_gem.c @@ -271,7 +271,7 @@ void *mtk_drm_gem_prime_vmap(struct drm_gem_object *obj) pgprot_writecombine(PAGE_KERNEL)); out: - kfree((void *)sgt); + kfree(sgt); return mtk_gem->kvaddr; } @@ -285,5 +285,5 @@ void mtk_drm_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr) vunmap(vaddr); mtk_gem->kvaddr = 0; - kfree((void *)mtk_gem->pages); + kfree(mtk_gem->pages); } diff --git a/drivers/gpu/drm/panfrost/panfrost_devfreq.c b/drivers/gpu/drm/panfrost/panfrost_devfreq.c index 12ff77dacc9541fb3d7a1af3d9d3f0876f59d8b8..4c4e8a30a1ac3cb7e30662fd4adee69ff0074d5a 100644 --- a/drivers/gpu/drm/panfrost/panfrost_devfreq.c +++ b/drivers/gpu/drm/panfrost/panfrost_devfreq.c @@ -13,97 +13,42 @@ #include "panfrost_gpu.h" #include "panfrost_regs.h" -static void panfrost_devfreq_update_utilization(struct panfrost_device *pfdev, int slot); +static void panfrost_devfreq_update_utilization(struct panfrost_device *pfdev); static int panfrost_devfreq_target(struct device *dev, unsigned long *freq, u32 flags) { - struct panfrost_device *pfdev = platform_get_drvdata(to_platform_device(dev)); - struct dev_pm_opp *opp; - unsigned long old_clk_rate = pfdev->devfreq.cur_freq; - unsigned long target_volt, target_rate; + struct panfrost_device *pfdev = dev_get_drvdata(dev); int err; - opp = devfreq_recommended_opp(dev, freq, flags); - if (IS_ERR(opp)) - return PTR_ERR(opp); - - target_rate = dev_pm_opp_get_freq(opp); - target_volt = dev_pm_opp_get_voltage(opp); - dev_pm_opp_put(opp); - - if (old_clk_rate == target_rate) - return 0; - - /* - * If frequency scaling from low to high, adjust voltage first. - * If frequency scaling from high to low, adjust frequency first. - */ - if (old_clk_rate < target_rate) { - err = regulator_set_voltage(pfdev->regulator, target_volt, - target_volt); - if (err) { - dev_err(dev, "Cannot set voltage %lu uV\n", - target_volt); - return err; - } - } - - err = clk_set_rate(pfdev->clock, target_rate); - if (err) { - dev_err(dev, "Cannot set frequency %lu (%d)\n", target_rate, - err); - regulator_set_voltage(pfdev->regulator, pfdev->devfreq.cur_volt, - pfdev->devfreq.cur_volt); + err = dev_pm_opp_set_rate(dev, *freq); + if (err) return err; - } - if (old_clk_rate > target_rate) { - err = regulator_set_voltage(pfdev->regulator, target_volt, - target_volt); - if (err) - dev_err(dev, "Cannot set voltage %lu uV\n", target_volt); - } - - pfdev->devfreq.cur_freq = target_rate; - pfdev->devfreq.cur_volt = target_volt; + *freq = clk_get_rate(pfdev->clock); return 0; } static void panfrost_devfreq_reset(struct panfrost_device *pfdev) { - ktime_t now = ktime_get(); - int i; - - for (i = 0; i < NUM_JOB_SLOTS; i++) { - pfdev->devfreq.slot[i].busy_time = 0; - pfdev->devfreq.slot[i].idle_time = 0; - pfdev->devfreq.slot[i].time_last_update = now; - } + pfdev->devfreq.busy_time = 0; + pfdev->devfreq.idle_time = 0; + pfdev->devfreq.time_last_update = ktime_get(); } static int panfrost_devfreq_get_dev_status(struct device *dev, struct devfreq_dev_status *status) { - struct panfrost_device *pfdev = platform_get_drvdata(to_platform_device(dev)); - int i; + struct panfrost_device *pfdev = dev_get_drvdata(dev); - for (i = 0; i < NUM_JOB_SLOTS; i++) { - panfrost_devfreq_update_utilization(pfdev, i); - } + panfrost_devfreq_update_utilization(pfdev); status->current_frequency = clk_get_rate(pfdev->clock); - status->total_time = ktime_to_ns(ktime_add(pfdev->devfreq.slot[0].busy_time, - pfdev->devfreq.slot[0].idle_time)); + status->total_time = ktime_to_ns(ktime_add(pfdev->devfreq.busy_time, + pfdev->devfreq.idle_time)); - status->busy_time = 0; - for (i = 0; i < NUM_JOB_SLOTS; i++) { - status->busy_time += ktime_to_ns(pfdev->devfreq.slot[i].busy_time); - } - - /* We're scheduling only to one core atm, so don't divide for now */ - /* status->busy_time /= NUM_JOB_SLOTS; */ + status->busy_time = ktime_to_ns(pfdev->devfreq.busy_time); panfrost_devfreq_reset(pfdev); @@ -119,7 +64,7 @@ static int panfrost_devfreq_get_cur_freq(struct device *dev, unsigned long *freq { struct panfrost_device *pfdev = platform_get_drvdata(to_platform_device(dev)); - *freq = pfdev->devfreq.cur_freq; + *freq = clk_get_rate(pfdev->clock); return 0; } @@ -135,6 +80,7 @@ int panfrost_devfreq_init(struct panfrost_device *pfdev) { int ret; struct dev_pm_opp *opp; + unsigned long cur_freq; ret = dev_pm_opp_of_add_table(&pfdev->pdev->dev); if (ret == -ENODEV) /* Optional, continue without devfreq */ @@ -144,13 +90,13 @@ int panfrost_devfreq_init(struct panfrost_device *pfdev) panfrost_devfreq_reset(pfdev); - pfdev->devfreq.cur_freq = clk_get_rate(pfdev->clock); + cur_freq = clk_get_rate(pfdev->clock); - opp = devfreq_recommended_opp(&pfdev->pdev->dev, &pfdev->devfreq.cur_freq, 0); + opp = devfreq_recommended_opp(&pfdev->pdev->dev, &cur_freq, 0); if (IS_ERR(opp)) return PTR_ERR(opp); - panfrost_devfreq_profile.initial_freq = pfdev->devfreq.cur_freq; + panfrost_devfreq_profile.initial_freq = cur_freq; dev_pm_opp_put(opp); pfdev->devfreq.devfreq = devm_devfreq_add_device(&pfdev->pdev->dev, @@ -174,14 +120,10 @@ void panfrost_devfreq_fini(struct panfrost_device *pfdev) void panfrost_devfreq_resume(struct panfrost_device *pfdev) { - int i; - if (!pfdev->devfreq.devfreq) return; panfrost_devfreq_reset(pfdev); - for (i = 0; i < NUM_JOB_SLOTS; i++) - pfdev->devfreq.slot[i].busy = false; devfreq_resume_device(pfdev->devfreq.devfreq); } @@ -194,9 +136,8 @@ void panfrost_devfreq_suspend(struct panfrost_device *pfdev) devfreq_suspend_device(pfdev->devfreq.devfreq); } -static void panfrost_devfreq_update_utilization(struct panfrost_device *pfdev, int slot) +static void panfrost_devfreq_update_utilization(struct panfrost_device *pfdev) { - struct panfrost_devfreq_slot *devfreq_slot = &pfdev->devfreq.slot[slot]; ktime_t now; ktime_t last; @@ -204,22 +145,27 @@ static void panfrost_devfreq_update_utilization(struct panfrost_device *pfdev, i return; now = ktime_get(); - last = pfdev->devfreq.slot[slot].time_last_update; + last = pfdev->devfreq.time_last_update; - /* If we last recorded a transition to busy, we have been idle since */ - if (devfreq_slot->busy) - pfdev->devfreq.slot[slot].busy_time += ktime_sub(now, last); + if (atomic_read(&pfdev->devfreq.busy_count) > 0) + pfdev->devfreq.busy_time += ktime_sub(now, last); else - pfdev->devfreq.slot[slot].idle_time += ktime_sub(now, last); + pfdev->devfreq.idle_time += ktime_sub(now, last); - pfdev->devfreq.slot[slot].time_last_update = now; + pfdev->devfreq.time_last_update = now; +} + +void panfrost_devfreq_record_busy(struct panfrost_device *pfdev) +{ + panfrost_devfreq_update_utilization(pfdev); + atomic_inc(&pfdev->devfreq.busy_count); } -/* The job scheduler is expected to call this at every transition busy <-> idle */ -void panfrost_devfreq_record_transition(struct panfrost_device *pfdev, int slot) +void panfrost_devfreq_record_idle(struct panfrost_device *pfdev) { - struct panfrost_devfreq_slot *devfreq_slot = &pfdev->devfreq.slot[slot]; + int count; - panfrost_devfreq_update_utilization(pfdev, slot); - devfreq_slot->busy = !devfreq_slot->busy; + panfrost_devfreq_update_utilization(pfdev); + count = atomic_dec_if_positive(&pfdev->devfreq.busy_count); + WARN_ON(count < 0); } diff --git a/drivers/gpu/drm/panfrost/panfrost_devfreq.h b/drivers/gpu/drm/panfrost/panfrost_devfreq.h index e3bc63e82843e48183a644e05871dcc3db403400..0611beffc8d0ef3603e51eb609ff332f48b260fb 100644 --- a/drivers/gpu/drm/panfrost/panfrost_devfreq.h +++ b/drivers/gpu/drm/panfrost/panfrost_devfreq.h @@ -10,6 +10,7 @@ void panfrost_devfreq_fini(struct panfrost_device *pfdev); void panfrost_devfreq_resume(struct panfrost_device *pfdev); void panfrost_devfreq_suspend(struct panfrost_device *pfdev); -void panfrost_devfreq_record_transition(struct panfrost_device *pfdev, int slot); +void panfrost_devfreq_record_busy(struct panfrost_device *pfdev); +void panfrost_devfreq_record_idle(struct panfrost_device *pfdev); #endif /* __PANFROST_DEVFREQ_H__ */ diff --git a/drivers/gpu/drm/panfrost/panfrost_device.h b/drivers/gpu/drm/panfrost/panfrost_device.h index 9c39b9794811012005ebd8d72be17e2bbe94fa79..06713811b92cdf787832e9842ac2572c3faf3e4d 100644 --- a/drivers/gpu/drm/panfrost/panfrost_device.h +++ b/drivers/gpu/drm/panfrost/panfrost_device.h @@ -51,13 +51,6 @@ struct panfrost_features { unsigned long hw_issues[64 / BITS_PER_LONG]; }; -struct panfrost_devfreq_slot { - ktime_t busy_time; - ktime_t idle_time; - ktime_t time_last_update; - bool busy; -}; - struct panfrost_device { struct device *dev; struct drm_device *ddev; @@ -93,9 +86,10 @@ struct panfrost_device { struct { struct devfreq *devfreq; struct thermal_cooling_device *cooling; - unsigned long cur_freq; - unsigned long cur_volt; - struct panfrost_devfreq_slot slot[NUM_JOB_SLOTS]; + ktime_t busy_time; + ktime_t idle_time; + ktime_t time_last_update; + atomic_t busy_count; } devfreq; }; diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c index 33bf25ba506eb26372b639c1d3d62c08dfcc8360..d411eb6c8eb953451fa105ebe4b487373e91e01c 100644 --- a/drivers/gpu/drm/panfrost/panfrost_job.c +++ b/drivers/gpu/drm/panfrost/panfrost_job.c @@ -155,8 +155,7 @@ static void panfrost_job_hw_submit(struct panfrost_job *job, int js) } cfg = panfrost_mmu_as_get(pfdev, &job->file_priv->mmu); - - panfrost_devfreq_record_transition(pfdev, js); + panfrost_devfreq_record_busy(pfdev); job_write(pfdev, JS_HEAD_NEXT_LO(js), jc_head & 0xFFFFFFFF); job_write(pfdev, JS_HEAD_NEXT_HI(js), jc_head >> 32); @@ -404,7 +403,7 @@ static void panfrost_job_timedout(struct drm_sched_job *sched_job) } spin_unlock_irqrestore(&pfdev->js->job_lock, flags); - panfrost_devfreq_record_transition(pfdev, js); + panfrost_devfreq_record_idle(pfdev); panfrost_device_reset(pfdev); for (i = 0; i < NUM_JOB_SLOTS; i++) @@ -467,7 +466,7 @@ static irqreturn_t panfrost_job_irq_handler(int irq, void *data) pfdev->jobs[j] = NULL; panfrost_mmu_as_put(pfdev, &job->file_priv->mmu); - panfrost_devfreq_record_transition(pfdev, j); + panfrost_devfreq_record_idle(pfdev); dma_fence_signal_locked(job->done_fence); pm_runtime_put_autosuspend(pfdev->dev); @@ -568,14 +567,14 @@ int panfrost_job_is_idle(struct panfrost_device *pfdev) struct panfrost_job_slot *js = pfdev->js; int i; + /* Check whether the hardware is idle */ + if (atomic_read(&pfdev->devfreq.busy_count)) + return false; + for (i = 0; i < NUM_JOB_SLOTS; i++) { /* If there are any jobs in the HW queue, we're not idle */ if (atomic_read(&js->queue[i].sched.hw_rq_count)) return false; - - /* Check whether the hardware is idle */ - if (pfdev->devfreq.slot[i].busy) - return false; } return true; diff --git a/drivers/gpu/drm/pl111/pl111_display.c b/drivers/gpu/drm/pl111/pl111_display.c index 024771a4083e1caefdfd770a792c967e6c62ac17..703ddc803c5540c90407efddbde53f4facf8f07f 100644 --- a/drivers/gpu/drm/pl111/pl111_display.c +++ b/drivers/gpu/drm/pl111/pl111_display.c @@ -48,10 +48,10 @@ irqreturn_t pl111_irq(int irq, void *data) } static enum drm_mode_status -pl111_mode_valid(struct drm_crtc *crtc, +pl111_mode_valid(struct drm_simple_display_pipe *pipe, const struct drm_display_mode *mode) { - struct drm_device *drm = crtc->dev; + struct drm_device *drm = pipe->crtc.dev; struct pl111_drm_dev_private *priv = drm->dev_private; u32 cpp = priv->variant->fb_bpp / 8; u64 bw; diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h index a5cb3864d686f7c92405312d829a0ebef14b74cf..27e45a2d6b52cdba60cc41dc0c753671e920b823 100644 --- a/drivers/gpu/drm/qxl/qxl_drv.h +++ b/drivers/gpu/drm/qxl/qxl_drv.h @@ -355,6 +355,8 @@ int qxl_mode_dumb_mmap(struct drm_file *filp, /* qxl ttm */ int qxl_ttm_init(struct qxl_device *qdev); void qxl_ttm_fini(struct qxl_device *qdev); +int qxl_ttm_io_mem_reserve(struct ttm_bo_device *bdev, + struct ttm_mem_reg *mem); /* qxl image */ diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c index ad336c98a0cff2130b9d6869874d94e7c78b8bff..ab72dc3476e9cfc6d0b486376a89dcdbfa01ddd6 100644 --- a/drivers/gpu/drm/qxl/qxl_object.c +++ b/drivers/gpu/drm/qxl/qxl_object.c @@ -167,7 +167,6 @@ int qxl_bo_kmap(struct qxl_bo *bo, void **ptr) void *qxl_bo_kmap_atomic_page(struct qxl_device *qdev, struct qxl_bo *bo, int page_offset) { - struct ttm_mem_type_manager *man = &bo->tbo.bdev->man[bo->tbo.mem.mem_type]; void *rptr; int ret; struct io_mapping *map; @@ -179,9 +178,7 @@ void *qxl_bo_kmap_atomic_page(struct qxl_device *qdev, else goto fallback; - (void) ttm_mem_io_lock(man, false); - ret = ttm_mem_io_reserve(bo->tbo.bdev, &bo->tbo.mem); - ttm_mem_io_unlock(man); + ret = qxl_ttm_io_mem_reserve(bo->tbo.bdev, &bo->tbo.mem); return io_mapping_map_atomic_wc(map, bo->tbo.mem.bus.offset + page_offset); fallback: @@ -212,17 +209,11 @@ void qxl_bo_kunmap(struct qxl_bo *bo) void qxl_bo_kunmap_atomic_page(struct qxl_device *qdev, struct qxl_bo *bo, void *pmap) { - struct ttm_mem_type_manager *man = &bo->tbo.bdev->man[bo->tbo.mem.mem_type]; - if ((bo->tbo.mem.mem_type != TTM_PL_VRAM) && (bo->tbo.mem.mem_type != TTM_PL_PRIV)) goto fallback; io_mapping_unmap_atomic(pmap); - - (void) ttm_mem_io_lock(man, false); - ttm_mem_io_free(bo->tbo.bdev, &bo->tbo.mem); - ttm_mem_io_unlock(man); return; fallback: qxl_bo_kunmap(bo); diff --git a/drivers/gpu/drm/qxl/qxl_release.c b/drivers/gpu/drm/qxl/qxl_release.c index 312216caeea26be0dcc3e1e2cfe8a6324f4597bc..2feca734c7b195fe385dea55417ad9c139c6655d 100644 --- a/drivers/gpu/drm/qxl/qxl_release.c +++ b/drivers/gpu/drm/qxl/qxl_release.c @@ -260,7 +260,7 @@ int qxl_release_reserve_list(struct qxl_release *release, bool no_intr) return 0; ret = ttm_eu_reserve_buffers(&release->ticket, &release->bos, - !no_intr, NULL, true); + !no_intr, NULL); if (ret) return ret; @@ -429,7 +429,6 @@ void qxl_release_unmap(struct qxl_device *qdev, void qxl_release_fence_buffer_objects(struct qxl_release *release) { struct ttm_buffer_object *bo; - struct ttm_bo_global *glob; struct ttm_bo_device *bdev; struct ttm_validate_buffer *entry; struct qxl_device *qdev; @@ -451,18 +450,16 @@ void qxl_release_fence_buffer_objects(struct qxl_release *release) release->id | 0xf0000000, release->base.seqno); trace_dma_fence_emit(&release->base); - glob = bdev->glob; - - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); list_for_each_entry(entry, &release->bos, head) { bo = entry->bo; dma_resv_add_shared_fence(bo->base.resv, &release->base); - ttm_bo_add_to_lru(bo); + ttm_bo_move_to_lru_tail(bo, NULL); dma_resv_unlock(bo->base.resv); } - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); ww_acquire_fini(&release->ticket); } diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c index 54cc5a5b607eac09fd42b671ac643c15a3574999..16a5e903533d940214b5d143da2d0198bd996762 100644 --- a/drivers/gpu/drm/qxl/qxl_ttm.c +++ b/drivers/gpu/drm/qxl/qxl_ttm.c @@ -110,8 +110,8 @@ static void qxl_evict_flags(struct ttm_buffer_object *bo, *placement = qbo->placement; } -static int qxl_ttm_io_mem_reserve(struct ttm_bo_device *bdev, - struct ttm_mem_reg *mem) +int qxl_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 qxl_device *qdev = qxl_get_qdev(bdev); @@ -319,14 +319,11 @@ static int qxl_mm_dump_table(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *)m->private; struct drm_mm *mm = (struct drm_mm *)node->info_ent->data; - struct drm_device *dev = node->minor->dev; - struct qxl_device *rdev = dev->dev_private; - struct ttm_bo_global *glob = rdev->mman.bdev.glob; struct drm_printer p = drm_seq_file_printer(m); - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); drm_mm_print(mm, &p); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); return 0; } #endif diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index b2b076606f54bbc9240944b7ce4546f056780955..67298a0739cb2d23193da4edc6177d702e0aaa00 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -566,7 +566,7 @@ static void radeon_gem_va_update_vm(struct radeon_device *rdev, if (!vm_bos) return; - r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL, true); + r = ttm_eu_reserve_buffers(&ticket, &list, true, NULL); if (r) goto error_free; diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 2abe1eab471f14c0fdf7ceba41ec9b35e74b96bf..140d94cc080d6252fde05129aa4371684a459715 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -542,7 +542,7 @@ int radeon_bo_list_validate(struct radeon_device *rdev, u64 bytes_moved_threshold = radeon_bo_get_threshold_for_moves(rdev); INIT_LIST_HEAD(&duplicates); - r = ttm_eu_reserve_buffers(ticket, head, true, &duplicates, true); + r = ttm_eu_reserve_buffers(ticket, head, true, &duplicates); if (unlikely(r != 0)) { return r; } diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 9a0ee74d82dc2b6c39859781c44e4e4be615cd1e..d4cc7289147e65f742fe0950f7c21ac341a042cb 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -622,43 +622,41 @@ static void drm_sched_process_job(struct dma_fence *f, struct dma_fence_cb *cb) } /** - * drm_sched_cleanup_jobs - destroy finished jobs + * drm_sched_get_cleanup_job - fetch the next finished job to be destroyed * * @sched: scheduler instance * - * Remove all finished jobs from the mirror list and destroy them. + * Returns the next finished job from the mirror list (if there is one) + * ready for it to be destroyed. */ -static void drm_sched_cleanup_jobs(struct drm_gpu_scheduler *sched) +static struct drm_sched_job * +drm_sched_get_cleanup_job(struct drm_gpu_scheduler *sched) { + struct drm_sched_job *job; unsigned long flags; /* Don't destroy jobs while the timeout worker is running */ if (sched->timeout != MAX_SCHEDULE_TIMEOUT && !cancel_delayed_work(&sched->work_tdr)) - return; - + return NULL; - while (!list_empty(&sched->ring_mirror_list)) { - struct drm_sched_job *job; + spin_lock_irqsave(&sched->job_list_lock, flags); - job = list_first_entry(&sched->ring_mirror_list, + job = list_first_entry_or_null(&sched->ring_mirror_list, struct drm_sched_job, node); - if (!dma_fence_is_signaled(&job->s_fence->finished)) - break; - spin_lock_irqsave(&sched->job_list_lock, flags); + if (job && dma_fence_is_signaled(&job->s_fence->finished)) { /* remove job from ring_mirror_list */ list_del_init(&job->node); - spin_unlock_irqrestore(&sched->job_list_lock, flags); - - sched->ops->free_job(job); + } else { + job = NULL; + /* queue timeout for next job */ + drm_sched_start_timeout(sched); } - /* queue timeout for next job */ - spin_lock_irqsave(&sched->job_list_lock, flags); - drm_sched_start_timeout(sched); spin_unlock_irqrestore(&sched->job_list_lock, flags); + return job; } /** @@ -698,12 +696,19 @@ static int drm_sched_main(void *param) struct drm_sched_fence *s_fence; struct drm_sched_job *sched_job; struct dma_fence *fence; + struct drm_sched_job *cleanup_job = NULL; wait_event_interruptible(sched->wake_up_worker, - (drm_sched_cleanup_jobs(sched), + (cleanup_job = drm_sched_get_cleanup_job(sched)) || (!drm_sched_blocked(sched) && (entity = drm_sched_select_entity(sched))) || - kthread_should_stop())); + kthread_should_stop()); + + if (cleanup_job) { + sched->ops->free_job(cleanup_job); + /* queue timeout for next job */ + drm_sched_start_timeout(sched); + } if (!entity) continue; diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c index ea4d59eb896694442685449a96837c46b0ff97a2..6050dc846894257ce4af23d2776256b827a00183 100644 --- a/drivers/gpu/drm/ttm/ttm_agp_backend.c +++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c @@ -51,7 +51,7 @@ struct ttm_agp_backend { static int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem) { struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm); - struct page *dummy_read_page = ttm->bdev->glob->dummy_read_page; + struct page *dummy_read_page = ttm_bo_glob.dummy_read_page; struct drm_mm_node *node = bo_mem->mm_node; struct agp_memory *mem; int ret, cached = (bo_mem->placement & TTM_PL_FLAG_CACHED); diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index dc3f7d69969737433b43fdefedb8be6a72508a5c..8d91b0428af1b58336c88780cc940e25788c2e93 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -51,6 +51,7 @@ static void ttm_bo_global_kobj_release(struct kobject *kobj); DEFINE_MUTEX(ttm_global_mutex); unsigned ttm_bo_glob_use_count; struct ttm_bo_global ttm_bo_glob; +EXPORT_SYMBOL(ttm_bo_glob); static struct attribute ttm_bo_count = { .name = "bo_count", @@ -148,23 +149,21 @@ static void ttm_bo_release_list(struct kref *list_kref) { struct ttm_buffer_object *bo = container_of(list_kref, struct ttm_buffer_object, list_kref); - struct ttm_bo_device *bdev = bo->bdev; size_t acc_size = bo->acc_size; BUG_ON(kref_read(&bo->list_kref)); BUG_ON(kref_read(&bo->kref)); - BUG_ON(atomic_read(&bo->cpu_writers)); BUG_ON(bo->mem.mm_node != NULL); BUG_ON(!list_empty(&bo->lru)); BUG_ON(!list_empty(&bo->ddestroy)); ttm_tt_destroy(bo->ttm); - atomic_dec(&bo->bdev->glob->bo_count); + atomic_dec(&ttm_bo_glob.bo_count); dma_fence_put(bo->moving); if (!ttm_bo_uses_embedded_gem_object(bo)) dma_resv_fini(&bo->base._resv); mutex_destroy(&bo->wu_mutex); bo->destroy(bo); - ttm_mem_global_free(bdev->glob->mem_glob, acc_size); + ttm_mem_global_free(&ttm_mem_glob, acc_size); } static void ttm_bo_add_mem_to_lru(struct ttm_buffer_object *bo, @@ -188,23 +187,17 @@ static void ttm_bo_add_mem_to_lru(struct ttm_buffer_object *bo, if (!(man->flags & TTM_MEMTYPE_FLAG_FIXED) && bo->ttm && !(bo->ttm->page_flags & (TTM_PAGE_FLAG_SG | TTM_PAGE_FLAG_SWAPPED))) { - list_add_tail(&bo->swap, &bdev->glob->swap_lru[bo->priority]); + list_add_tail(&bo->swap, &ttm_bo_glob.swap_lru[bo->priority]); kref_get(&bo->list_kref); } } -void ttm_bo_add_to_lru(struct ttm_buffer_object *bo) -{ - ttm_bo_add_mem_to_lru(bo, &bo->mem); -} -EXPORT_SYMBOL(ttm_bo_add_to_lru); - static void ttm_bo_ref_bug(struct kref *list_kref) { BUG(); } -void ttm_bo_del_from_lru(struct ttm_buffer_object *bo) +static void ttm_bo_del_from_lru(struct ttm_buffer_object *bo) { struct ttm_bo_device *bdev = bo->bdev; bool notify = false; @@ -224,16 +217,6 @@ void ttm_bo_del_from_lru(struct ttm_buffer_object *bo) bdev->driver->del_from_lru_notify(bo); } -void ttm_bo_del_sub_from_lru(struct ttm_buffer_object *bo) -{ - struct ttm_bo_global *glob = bo->bdev->glob; - - spin_lock(&glob->lru_lock); - ttm_bo_del_from_lru(bo); - spin_unlock(&glob->lru_lock); -} -EXPORT_SYMBOL(ttm_bo_del_sub_from_lru); - static void ttm_bo_bulk_move_set_pos(struct ttm_lru_bulk_move_pos *pos, struct ttm_buffer_object *bo) { @@ -248,7 +231,7 @@ void ttm_bo_move_to_lru_tail(struct ttm_buffer_object *bo, dma_resv_assert_held(bo->base.resv); ttm_bo_del_from_lru(bo); - ttm_bo_add_to_lru(bo); + ttm_bo_add_mem_to_lru(bo, &bo->mem); if (bulk && !(bo->mem.placement & TTM_PL_FLAG_NO_EVICT)) { switch (bo->mem.mem_type) { @@ -311,7 +294,7 @@ void ttm_bo_bulk_move_lru_tail(struct ttm_lru_bulk_move *bulk) dma_resv_assert_held(pos->first->base.resv); dma_resv_assert_held(pos->last->base.resv); - lru = &pos->first->bdev->glob->swap_lru[i]; + lru = &ttm_bo_glob.swap_lru[i]; list_bulk_move_tail(lru, &pos->first->swap, &pos->last->swap); } } @@ -475,7 +458,6 @@ static void ttm_bo_flush_all_fences(struct ttm_buffer_object *bo) static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo) { struct ttm_bo_device *bdev = bo->bdev; - struct ttm_bo_global *glob = bdev->glob; int ret; ret = ttm_bo_individualize_resv(bo); @@ -485,16 +467,16 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo) */ dma_resv_wait_timeout_rcu(bo->base.resv, true, false, 30 * HZ); - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); goto error; } - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); ret = dma_resv_trylock(bo->base.resv) ? 0 : -EBUSY; if (!ret) { if (dma_resv_test_signaled_rcu(&bo->base._resv, true)) { ttm_bo_del_from_lru(bo); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); if (bo->base.resv != &bo->base._resv) dma_resv_unlock(&bo->base._resv); @@ -512,7 +494,7 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo) */ if (bo->mem.placement & TTM_PL_FLAG_NO_EVICT) { bo->mem.placement &= ~TTM_PL_FLAG_NO_EVICT; - ttm_bo_add_to_lru(bo); + ttm_bo_move_to_lru_tail(bo, NULL); } dma_resv_unlock(bo->base.resv); @@ -523,7 +505,7 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo) error: kref_get(&bo->list_kref); list_add_tail(&bo->ddestroy, &bdev->ddestroy); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); schedule_delayed_work(&bdev->wq, ((HZ / 100) < 1) ? 1 : HZ / 100); @@ -546,7 +528,6 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, bool interruptible, bool no_wait_gpu, bool unlock_resv) { - struct ttm_bo_global *glob = bo->bdev->glob; struct dma_resv *resv; int ret; @@ -565,7 +546,7 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, if (unlock_resv) dma_resv_unlock(bo->base.resv); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); lret = dma_resv_wait_timeout_rcu(resv, true, interruptible, @@ -576,7 +557,7 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, else if (lret == 0) return -EBUSY; - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); if (unlock_resv && !dma_resv_trylock(bo->base.resv)) { /* * We raced, and lost, someone else holds the reservation now, @@ -586,7 +567,7 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, * delayed destruction would succeed, so just return success * here. */ - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); return 0; } ret = 0; @@ -595,7 +576,7 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, if (ret || unlikely(list_empty(&bo->ddestroy))) { if (unlock_resv) dma_resv_unlock(bo->base.resv); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); return ret; } @@ -603,7 +584,7 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, list_del_init(&bo->ddestroy); kref_put(&bo->list_kref, ttm_bo_ref_bug); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); ttm_bo_cleanup_memtype_use(bo); if (unlock_resv) @@ -618,7 +599,7 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, */ static bool ttm_bo_delayed_delete(struct ttm_bo_device *bdev, bool remove_all) { - struct ttm_bo_global *glob = bdev->glob; + struct ttm_bo_global *glob = &ttm_bo_glob; struct list_head removed; bool empty; @@ -842,13 +823,12 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev, struct ww_acquire_ctx *ticket) { struct ttm_buffer_object *bo = NULL, *busy_bo = NULL; - struct ttm_bo_global *glob = bdev->glob; struct ttm_mem_type_manager *man = &bdev->man[mem_type]; bool locked = false; unsigned i; int ret; - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) { list_for_each_entry(bo, &man->lru[i], lru) { bool busy; @@ -880,7 +860,7 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev, if (!bo) { if (busy_bo) kref_get(&busy_bo->list_kref); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); ret = ttm_mem_evict_wait_busy(busy_bo, ctx, ticket); if (busy_bo) kref_put(&busy_bo->list_kref, ttm_bo_release_list); @@ -896,17 +876,11 @@ static int ttm_mem_evict_first(struct ttm_bo_device *bdev, return ret; } - ttm_bo_del_from_lru(bo); - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); ret = ttm_bo_evict(bo, ctx); - if (locked) { + if (locked) ttm_bo_unreserve(bo); - } else { - spin_lock(&glob->lru_lock); - ttm_bo_add_to_lru(bo); - spin_unlock(&glob->lru_lock); - } kref_put(&bo->list_kref, ttm_bo_release_list); return ret; @@ -1072,12 +1046,10 @@ static int ttm_bo_mem_placement(struct ttm_buffer_object *bo, mem->mem_type = mem_type; mem->placement = cur_flags; - if (bo->mem.mem_type < mem_type && !list_empty(&bo->lru)) { - spin_lock(&bo->bdev->glob->lru_lock); - ttm_bo_del_from_lru(bo); - ttm_bo_add_mem_to_lru(bo, mem); - spin_unlock(&bo->bdev->glob->lru_lock); - } + spin_lock(&ttm_bo_glob.lru_lock); + ttm_bo_del_from_lru(bo); + ttm_bo_add_mem_to_lru(bo, mem); + spin_unlock(&ttm_bo_glob.lru_lock); return 0; } @@ -1168,9 +1140,9 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, error: if (bo->mem.mem_type == TTM_PL_SYSTEM && !list_empty(&bo->lru)) { - spin_lock(&bo->bdev->glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); ttm_bo_move_to_lru_tail(bo, NULL); - spin_unlock(&bo->bdev->glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); } return ret; @@ -1294,9 +1266,9 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev, struct dma_resv *resv, void (*destroy) (struct ttm_buffer_object *)) { + struct ttm_mem_global *mem_glob = &ttm_mem_glob; int ret = 0; unsigned long num_pages; - struct ttm_mem_global *mem_glob = bdev->glob->mem_glob; bool locked; ret = ttm_mem_global_alloc(mem_glob, acc_size, ctx); @@ -1323,7 +1295,6 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev, kref_init(&bo->kref); kref_init(&bo->list_kref); - atomic_set(&bo->cpu_writers, 0); INIT_LIST_HEAD(&bo->lru); INIT_LIST_HEAD(&bo->ddestroy); INIT_LIST_HEAD(&bo->swap); @@ -1357,7 +1328,7 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev, dma_resv_init(&bo->base._resv); drm_vma_node_reset(&bo->base.vma_node); } - atomic_inc(&bo->bdev->glob->bo_count); + atomic_inc(&ttm_bo_glob.bo_count); /* * For ttm_bo_type_device buffers, allocate @@ -1387,11 +1358,9 @@ int ttm_bo_init_reserved(struct ttm_bo_device *bdev, return ret; } - if (resv && !(bo->mem.placement & TTM_PL_FLAG_NO_EVICT)) { - spin_lock(&bdev->glob->lru_lock); - ttm_bo_add_to_lru(bo); - spin_unlock(&bdev->glob->lru_lock); - } + spin_lock(&ttm_bo_glob.lru_lock); + ttm_bo_move_to_lru_tail(bo, NULL); + spin_unlock(&ttm_bo_glob.lru_lock); return ret; } @@ -1489,7 +1458,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev, .flags = TTM_OPT_FLAG_FORCE_ALLOC }; struct ttm_mem_type_manager *man = &bdev->man[mem_type]; - struct ttm_bo_global *glob = bdev->glob; + struct ttm_bo_global *glob = &ttm_bo_glob; struct dma_fence *fence; int ret; unsigned i; @@ -1658,8 +1627,6 @@ static int ttm_bo_global_init(void) goto out; spin_lock_init(&glob->lru_lock); - glob->mem_glob = &ttm_mem_glob; - glob->mem_glob->bo_glob = glob; glob->dummy_read_page = alloc_page(__GFP_ZERO | GFP_DMA32); if (unlikely(glob->dummy_read_page == NULL)) { @@ -1683,10 +1650,10 @@ static int ttm_bo_global_init(void) int ttm_bo_device_release(struct ttm_bo_device *bdev) { + struct ttm_bo_global *glob = &ttm_bo_glob; int ret = 0; unsigned i = TTM_NUM_MEM_TYPES; struct ttm_mem_type_manager *man; - struct ttm_bo_global *glob = bdev->glob; while (i--) { man = &bdev->man[i]; @@ -1755,7 +1722,6 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev, INIT_DELAYED_WORK(&bdev->wq, ttm_bo_delayed_workqueue); INIT_LIST_HEAD(&bdev->ddestroy); bdev->dev_mapping = mapping; - bdev->glob = glob; bdev->need_dma32 = need_dma32; mutex_lock(&ttm_global_mutex); list_add_tail(&bdev->device_list, &glob->device_list); @@ -1835,31 +1801,6 @@ int ttm_bo_wait(struct ttm_buffer_object *bo, } EXPORT_SYMBOL(ttm_bo_wait); -int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait) -{ - int ret = 0; - - /* - * Using ttm_bo_reserve makes sure the lru lists are updated. - */ - - ret = ttm_bo_reserve(bo, true, no_wait, NULL); - if (unlikely(ret != 0)) - return ret; - ret = ttm_bo_wait(bo, true, no_wait); - if (likely(ret == 0)) - atomic_inc(&bo->cpu_writers); - ttm_bo_unreserve(bo); - return ret; -} -EXPORT_SYMBOL(ttm_bo_synccpu_write_grab); - -void ttm_bo_synccpu_write_release(struct ttm_buffer_object *bo) -{ - atomic_dec(&bo->cpu_writers); -} -EXPORT_SYMBOL(ttm_bo_synccpu_write_release); - /** * A buffer object shrink method that tries to swap out the first * buffer object on the bo_global::swap_lru list. @@ -1959,8 +1900,7 @@ void ttm_bo_swapout_all(struct ttm_bo_device *bdev) .no_wait_gpu = false }; - while (ttm_bo_swapout(bdev->glob, &ctx) == 0) - ; + while (ttm_bo_swapout(&ttm_bo_glob, &ctx) == 0); } EXPORT_SYMBOL(ttm_bo_swapout_all); diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c index fe81c565e7ef88e49d7e1c3cceb1e41d19897b24..6b0883a1776e1c8645024379242b3eabfd2b1f33 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_util.c +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c @@ -102,7 +102,6 @@ int ttm_mem_io_lock(struct ttm_mem_type_manager *man, bool interruptible) mutex_lock(&man->io_reserve_mutex); return 0; } -EXPORT_SYMBOL(ttm_mem_io_lock); void ttm_mem_io_unlock(struct ttm_mem_type_manager *man) { @@ -111,7 +110,6 @@ void ttm_mem_io_unlock(struct ttm_mem_type_manager *man) mutex_unlock(&man->io_reserve_mutex); } -EXPORT_SYMBOL(ttm_mem_io_unlock); static int ttm_mem_io_evict(struct ttm_mem_type_manager *man) { @@ -153,7 +151,6 @@ int ttm_mem_io_reserve(struct ttm_bo_device *bdev, } return ret; } -EXPORT_SYMBOL(ttm_mem_io_reserve); void ttm_mem_io_free(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem) @@ -169,7 +166,6 @@ void ttm_mem_io_free(struct ttm_bo_device *bdev, bdev->driver->io_mem_free(bdev, mem); } -EXPORT_SYMBOL(ttm_mem_io_free); int ttm_mem_io_reserve_vm(struct ttm_buffer_object *bo) { @@ -503,7 +499,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo, * TODO: Explicit member copy would probably be better here. */ - atomic_inc(&bo->bdev->glob->bo_count); + atomic_inc(&ttm_bo_glob.bo_count); INIT_LIST_HEAD(&fbo->base.ddestroy); INIT_LIST_HEAD(&fbo->base.lru); INIT_LIST_HEAD(&fbo->base.swap); @@ -511,15 +507,16 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo, mutex_init(&fbo->base.wu_mutex); fbo->base.moving = NULL; drm_vma_node_reset(&fbo->base.base.vma_node); - atomic_set(&fbo->base.cpu_writers, 0); kref_init(&fbo->base.list_kref); kref_init(&fbo->base.kref); fbo->base.destroy = &ttm_transfered_destroy; fbo->base.acc_size = 0; - fbo->base.base.resv = &fbo->base.base._resv; - dma_resv_init(fbo->base.base.resv); - ret = dma_resv_trylock(fbo->base.base.resv); + if (bo->base.resv == &bo->base._resv) + fbo->base.base.resv = &fbo->base.base._resv; + + dma_resv_init(&fbo->base.base._resv); + ret = dma_resv_trylock(&fbo->base.base._resv); WARN_ON(!ret); *new_obj = &fbo->base; @@ -716,7 +713,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, if (ret) return ret; - dma_resv_add_excl_fence(ghost_obj->base.resv, fence); + dma_resv_add_excl_fence(&ghost_obj->base._resv, fence); /** * If we're not moving to fixed memory, the TTM object @@ -729,7 +726,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, else bo->ttm = NULL; - ttm_bo_unreserve(ghost_obj); + dma_resv_unlock(&ghost_obj->base._resv); ttm_bo_put(ghost_obj); } @@ -772,7 +769,7 @@ int ttm_bo_pipeline_move(struct ttm_buffer_object *bo, if (ret) return ret; - dma_resv_add_excl_fence(ghost_obj->base.resv, fence); + dma_resv_add_excl_fence(&ghost_obj->base._resv, fence); /** * If we're not moving to fixed memory, the TTM object @@ -785,7 +782,7 @@ int ttm_bo_pipeline_move(struct ttm_buffer_object *bo, else bo->ttm = NULL; - ttm_bo_unreserve(ghost_obj); + dma_resv_unlock(&ghost_obj->base._resv); ttm_bo_put(ghost_obj); } else if (from->flags & TTM_MEMTYPE_FLAG_FIXED) { @@ -841,7 +838,7 @@ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo) if (ret) return ret; - ret = dma_resv_copy_fences(ghost->base.resv, bo->base.resv); + ret = dma_resv_copy_fences(&ghost->base._resv, bo->base.resv); /* Last resort, wait for the BO to be idle when we are OOM */ if (ret) ttm_bo_wait(bo, false, false); @@ -850,7 +847,7 @@ int ttm_bo_pipeline_gutting(struct ttm_buffer_object *bo) bo->mem.mem_type = TTM_PL_SYSTEM; bo->ttm = NULL; - ttm_bo_unreserve(ghost); + dma_resv_unlock(&ghost->base._resv); ttm_bo_put(ghost); return 0; diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index 79f01c5ff65ef0380ce5bbbff81ada8562a50f89..4b34a278d65b8477fd0287d29d03a88ed8834bf6 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -177,9 +177,9 @@ static vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf) } if (bo->moving != moving) { - spin_lock(&bdev->glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); ttm_bo_move_to_lru_tail(bo, NULL); - spin_unlock(&bdev->glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); } dma_fence_put(moving); } @@ -480,6 +480,13 @@ EXPORT_SYMBOL(ttm_bo_mmap); int ttm_bo_mmap_obj(struct vm_area_struct *vma, struct ttm_buffer_object *bo) { ttm_bo_get(bo); + + /* + * FIXME: &drm_gem_object_funcs.mmap is called with the fake offset + * removed. Add it back here until the rest of TTM works without it. + */ + vma->vm_pgoff += drm_vma_node_start(&bo->base.vma_node); + ttm_bo_mmap_vma_setup(bo, vma); return 0; } diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c index 131dae8f4170aafd3a10c07874ac1e6cf2177eaf..1797f04c05345ba226b5ea566bf4a60ba8f4cd6d 100644 --- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c +++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c @@ -43,37 +43,22 @@ static void ttm_eu_backoff_reservation_reverse(struct list_head *list, } } -static void ttm_eu_del_from_lru_locked(struct list_head *list) -{ - struct ttm_validate_buffer *entry; - - list_for_each_entry(entry, list, head) { - struct ttm_buffer_object *bo = entry->bo; - ttm_bo_del_from_lru(bo); - } -} - void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket, struct list_head *list) { struct ttm_validate_buffer *entry; - struct ttm_bo_global *glob; if (list_empty(list)) return; - entry = list_first_entry(list, struct ttm_validate_buffer, head); - glob = entry->bo->bdev->glob; - - spin_lock(&glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); list_for_each_entry(entry, list, head) { struct ttm_buffer_object *bo = entry->bo; - if (list_empty(&bo->lru)) - ttm_bo_add_to_lru(bo); + ttm_bo_move_to_lru_tail(bo, NULL); dma_resv_unlock(bo->base.resv); } - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); if (ticket) ww_acquire_fini(ticket); @@ -94,18 +79,14 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation); int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, struct list_head *list, bool intr, - struct list_head *dups, bool del_lru) + struct list_head *dups) { - struct ttm_bo_global *glob; struct ttm_validate_buffer *entry; int ret; if (list_empty(list)) return 0; - entry = list_first_entry(list, struct ttm_validate_buffer, head); - glob = entry->bo->bdev->glob; - if (ticket) ww_acquire_init(ticket, &reservation_ww_class); @@ -113,12 +94,7 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, struct ttm_buffer_object *bo = entry->bo; ret = __ttm_bo_reserve(bo, intr, (ticket == NULL), ticket); - if (!ret && unlikely(atomic_read(&bo->cpu_writers) > 0)) { - dma_resv_unlock(bo->base.resv); - - ret = -EBUSY; - - } else if (ret == -EALREADY && dups) { + if (ret == -EALREADY && dups) { struct ttm_validate_buffer *safe = entry; entry = list_prev_entry(entry, head); list_del(&safe->head); @@ -173,11 +149,6 @@ int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, list_add(&entry->head, list); } - if (del_lru) { - spin_lock(&glob->lru_lock); - ttm_eu_del_from_lru_locked(list); - spin_unlock(&glob->lru_lock); - } return 0; } EXPORT_SYMBOL(ttm_eu_reserve_buffers); @@ -187,30 +158,22 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket, struct dma_fence *fence) { struct ttm_validate_buffer *entry; - struct ttm_buffer_object *bo; - struct ttm_bo_global *glob; if (list_empty(list)) return; - bo = list_first_entry(list, struct ttm_validate_buffer, head)->bo; - glob = bo->bdev->glob; - - spin_lock(&glob->lru_lock); - + spin_lock(&ttm_bo_glob.lru_lock); list_for_each_entry(entry, list, head) { - bo = entry->bo; + struct ttm_buffer_object *bo = entry->bo; + if (entry->num_shared) dma_resv_add_shared_fence(bo->base.resv, fence); else dma_resv_add_excl_fence(bo->base.resv, fence); - if (list_empty(&bo->lru)) - ttm_bo_add_to_lru(bo); - else - ttm_bo_move_to_lru_tail(bo, NULL); + ttm_bo_move_to_lru_tail(bo, NULL); dma_resv_unlock(bo->base.resv); } - spin_unlock(&glob->lru_lock); + spin_unlock(&ttm_bo_glob.lru_lock); if (ticket) ww_acquire_fini(ticket); } diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c index 8617958b7ae6b7759e27c6e1ff3235e53060f0a1..acd63b70d8147a0e52b3d005b387bd0f1022c84d 100644 --- a/drivers/gpu/drm/ttm/ttm_memory.c +++ b/drivers/gpu/drm/ttm/ttm_memory.c @@ -275,7 +275,7 @@ static void ttm_shrink(struct ttm_mem_global *glob, bool from_wq, while (ttm_zones_above_swap_target(glob, from_wq, extra)) { spin_unlock(&glob->lock); - ret = ttm_bo_swapout(glob->bo_glob, ctx); + ret = ttm_bo_swapout(&ttm_bo_glob, ctx); spin_lock(&glob->lock); if (unlikely(ret != 0)) break; diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index 627f8dc91d0ed23e0958dfc39d106c967dcd376a..b40a4678c29639c3090be39e64033a46c61ed138 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -1028,7 +1028,7 @@ void ttm_page_alloc_fini(void) static void ttm_pool_unpopulate_helper(struct ttm_tt *ttm, unsigned mem_count_update) { - struct ttm_mem_global *mem_glob = ttm->bdev->glob->mem_glob; + struct ttm_mem_global *mem_glob = &ttm_mem_glob; unsigned i; if (mem_count_update == 0) @@ -1049,7 +1049,7 @@ ttm_pool_unpopulate_helper(struct ttm_tt *ttm, unsigned mem_count_update) int ttm_pool_populate(struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) { - struct ttm_mem_global *mem_glob = ttm->bdev->glob->mem_glob; + struct ttm_mem_global *mem_glob = &ttm_mem_glob; unsigned i; int ret; diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c index 7d78e6deac89737276f6151ee934e9a6614deefd..ff54e7609e8ff994dc38c3500e7c4047926dbefc 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c @@ -886,8 +886,8 @@ static gfp_t ttm_dma_pool_gfp_flags(struct ttm_dma_tt *ttm_dma, bool huge) int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev, struct ttm_operation_ctx *ctx) { + struct ttm_mem_global *mem_glob = &ttm_mem_glob; struct ttm_tt *ttm = &ttm_dma->ttm; - struct ttm_mem_global *mem_glob = ttm->bdev->glob->mem_glob; unsigned long num_pages = ttm->num_pages; struct dma_pool *pool; struct dma_page *d_page; @@ -991,8 +991,8 @@ EXPORT_SYMBOL_GPL(ttm_dma_populate); /* Put all pages in pages list to correct pool to wait for reuse */ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev) { + struct ttm_mem_global *mem_glob = &ttm_mem_glob; struct ttm_tt *ttm = &ttm_dma->ttm; - struct ttm_mem_global *mem_glob = ttm->bdev->glob->mem_glob; struct dma_pool *pool; struct dma_page *d_page, *next; enum pool_type type; diff --git a/drivers/gpu/drm/vboxvideo/vbox_mode.c b/drivers/gpu/drm/vboxvideo/vbox_mode.c index b5604d32122e47f6e67f2825628cd7ea1fc4ffb0..19612132c8a39deb2b718c4c9b1a102c75d65343 100644 --- a/drivers/gpu/drm/vboxvideo/vbox_mode.c +++ b/drivers/gpu/drm/vboxvideo/vbox_mode.c @@ -334,35 +334,6 @@ static void vbox_primary_atomic_disable(struct drm_plane *plane, old_state->src_y >> 16); } -static int vbox_primary_prepare_fb(struct drm_plane *plane, - struct drm_plane_state *new_state) -{ - struct drm_gem_vram_object *gbo; - int ret; - - if (!new_state->fb) - return 0; - - gbo = drm_gem_vram_of_gem(new_state->fb->obj[0]); - ret = drm_gem_vram_pin(gbo, DRM_GEM_VRAM_PL_FLAG_VRAM); - if (ret) - DRM_WARN("Error %d pinning new fb, out of video mem?\n", ret); - - return ret; -} - -static void vbox_primary_cleanup_fb(struct drm_plane *plane, - struct drm_plane_state *old_state) -{ - struct drm_gem_vram_object *gbo; - - if (!old_state->fb) - return; - - gbo = drm_gem_vram_of_gem(old_state->fb->obj[0]); - drm_gem_vram_unpin(gbo); -} - static int vbox_cursor_atomic_check(struct drm_plane *plane, struct drm_plane_state *new_state) { @@ -492,30 +463,6 @@ static void vbox_cursor_atomic_disable(struct drm_plane *plane, mutex_unlock(&vbox->hw_mutex); } -static int vbox_cursor_prepare_fb(struct drm_plane *plane, - struct drm_plane_state *new_state) -{ - struct drm_gem_vram_object *gbo; - - if (!new_state->fb) - return 0; - - gbo = drm_gem_vram_of_gem(new_state->fb->obj[0]); - return drm_gem_vram_pin(gbo, DRM_GEM_VRAM_PL_FLAG_SYSTEM); -} - -static void vbox_cursor_cleanup_fb(struct drm_plane *plane, - struct drm_plane_state *old_state) -{ - struct drm_gem_vram_object *gbo; - - if (!plane->state->fb) - return; - - gbo = drm_gem_vram_of_gem(plane->state->fb->obj[0]); - drm_gem_vram_unpin(gbo); -} - static const u32 vbox_cursor_plane_formats[] = { DRM_FORMAT_ARGB8888, }; @@ -524,8 +471,8 @@ static const struct drm_plane_helper_funcs vbox_cursor_helper_funcs = { .atomic_check = vbox_cursor_atomic_check, .atomic_update = vbox_cursor_atomic_update, .atomic_disable = vbox_cursor_atomic_disable, - .prepare_fb = vbox_cursor_prepare_fb, - .cleanup_fb = vbox_cursor_cleanup_fb, + .prepare_fb = drm_gem_vram_plane_helper_prepare_fb, + .cleanup_fb = drm_gem_vram_plane_helper_cleanup_fb, }; static const struct drm_plane_funcs vbox_cursor_plane_funcs = { @@ -546,8 +493,8 @@ static const struct drm_plane_helper_funcs vbox_primary_helper_funcs = { .atomic_check = vbox_primary_atomic_check, .atomic_update = vbox_primary_atomic_update, .atomic_disable = vbox_primary_atomic_disable, - .prepare_fb = vbox_primary_prepare_fb, - .cleanup_fb = vbox_primary_cleanup_fb, + .prepare_fb = drm_gem_vram_plane_helper_prepare_fb, + .cleanup_fb = drm_gem_vram_plane_helper_cleanup_fb, }; static const struct drm_plane_funcs vbox_primary_plane_funcs = { @@ -890,7 +837,7 @@ static int vbox_connector_init(struct drm_device *dev, } static const struct drm_mode_config_funcs vbox_mode_funcs = { - .fb_create = drm_gem_fb_create, + .fb_create = drm_gem_fb_create_with_dirty, .atomic_check = drm_atomic_helper_check, .atomic_commit = drm_atomic_helper_commit, }; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index aad8d81402595c74f4519194daada8eb9ff958cb..74016a08d1186f6f06d81ab54459555e6b0a0449 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -566,7 +566,7 @@ static void vmw_user_bo_ref_obj_release(struct ttm_base_object *base, switch (ref_type) { case TTM_REF_SYNCCPU_WRITE: - ttm_bo_synccpu_write_release(&user_bo->vbo.base); + atomic_dec(&user_bo->vbo.cpu_writers); break; default: WARN_ONCE(true, "Undefined buffer object reference release.\n"); @@ -682,12 +682,12 @@ static int vmw_user_bo_synccpu_grab(struct vmw_user_buffer_object *user_bo, struct ttm_object_file *tfile, uint32_t flags) { + bool nonblock = !!(flags & drm_vmw_synccpu_dontblock); struct ttm_buffer_object *bo = &user_bo->vbo.base; bool existed; int ret; if (flags & drm_vmw_synccpu_allow_cs) { - bool nonblock = !!(flags & drm_vmw_synccpu_dontblock); long lret; lret = dma_resv_wait_timeout_rcu @@ -700,15 +700,22 @@ static int vmw_user_bo_synccpu_grab(struct vmw_user_buffer_object *user_bo, return 0; } - ret = ttm_bo_synccpu_write_grab - (bo, !!(flags & drm_vmw_synccpu_dontblock)); + ret = ttm_bo_reserve(bo, true, nonblock, NULL); + if (unlikely(ret != 0)) + return ret; + + ret = ttm_bo_wait(bo, true, nonblock); + if (likely(ret == 0)) + atomic_inc(&user_bo->vbo.cpu_writers); + + ttm_bo_unreserve(bo); if (unlikely(ret != 0)) return ret; ret = ttm_ref_object_add(tfile, &user_bo->prime.base, TTM_REF_SYNCCPU_WRITE, &existed, false); if (ret != 0 || existed) - ttm_bo_synccpu_write_release(&user_bo->vbo.base); + atomic_dec(&user_bo->vbo.cpu_writers); return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 29f60e027a38290e6a348bd0ed91698993b672dc..b18842f7308107d3280a4654b936f0bfe602b883 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -102,6 +102,8 @@ struct vmw_fpriv { * @base: The TTM buffer object * @res_list: List of resources using this buffer object as a backing MOB * @pin_count: pin depth + * @cpu_writers: Number of synccpu write grabs. Protected by reservation when + * increased. May be decreased without reservation. * @dx_query_ctx: DX context if this buffer object is used as a DX query MOB * @map: Kmap object for semi-persistent mappings * @res_prios: Eviction priority counts for attached resources @@ -110,6 +112,7 @@ struct vmw_buffer_object { struct ttm_buffer_object base; struct list_head res_list; s32 pin_count; + atomic_t cpu_writers; /* Not ref-counted. Protected by binding_mutex */ struct vmw_resource *dx_query_ctx; /* Protected by reservation */ diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index 5581a7826b4c06fc814d9c028ff6ec68f51c5ecf..6dfe36fb817cc7fa5e8253d8eb03de0f6fe64b89 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -492,8 +492,7 @@ vmw_resource_check_buffer(struct ww_acquire_ctx *ticket, val_buf->bo = &res->backup->base; val_buf->num_shared = 0; list_add_tail(&val_buf->head, &val_list); - ret = ttm_eu_reserve_buffers(ticket, &val_list, interruptible, NULL, - true); + ret = ttm_eu_reserve_buffers(ticket, &val_list, interruptible, NULL); if (unlikely(ret != 0)) goto out_no_reserve; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c index f611b2290a1b9325ac0405936c6cbbfcf414776f..7bff3628fc54d0f0fcfb3fc2feb38eaca4523911 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c @@ -521,6 +521,9 @@ int vmw_validation_bo_validate_single(struct ttm_buffer_object *bo, }; int ret; + if (atomic_read(&vbo->cpu_writers)) + return -EBUSY; + if (vbo->pin_count > 0) return 0; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h index 0e063743dd86de38c84f03e4ced198c5ed0198d9..71ce4b31885025be05b13b93bbcb96b756e84118 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.h @@ -170,7 +170,7 @@ vmw_validation_bo_reserve(struct vmw_validation_context *ctx, bool intr) { return ttm_eu_reserve_buffers(&ctx->ticket, &ctx->bo_list, intr, - NULL, true); + NULL); } /** diff --git a/drivers/gpu/drm/xen/xen_drm_front_kms.c b/drivers/gpu/drm/xen/xen_drm_front_kms.c index 21ad1c359b613f79c07bc798761f5aeaf40b8403..ff506bc99414948fac3b947b9c4d48586df3a92d 100644 --- a/drivers/gpu/drm/xen/xen_drm_front_kms.c +++ b/drivers/gpu/drm/xen/xen_drm_front_kms.c @@ -270,11 +270,12 @@ static void display_update(struct drm_simple_display_pipe *pipe, } static enum drm_mode_status -display_mode_valid(struct drm_crtc *crtc, const struct drm_display_mode *mode) +display_mode_valid(struct drm_simple_display_pipe *pipe, + const struct drm_display_mode *mode) { struct xen_drm_front_drm_pipeline *pipeline = - container_of(crtc, struct xen_drm_front_drm_pipeline, - pipe.crtc); + container_of(pipe, struct xen_drm_front_drm_pipeline, + pipe); if (mode->hdisplay != pipeline->width) return MODE_ERROR; diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h index 8dcc012ccbc853922fac6f08030f65bbe1541e1e..2338e9f94a03e236fb045dabdada0908fe738a18 100644 --- a/include/drm/drm_fb_helper.h +++ b/include/drm/drm_fb_helper.h @@ -235,7 +235,6 @@ void drm_fb_helper_unlink_fbi(struct drm_fb_helper *fb_helper); void drm_fb_helper_deferred_io(struct fb_info *info, struct list_head *pagelist); -int drm_fb_helper_defio_init(struct drm_fb_helper *fb_helper); ssize_t drm_fb_helper_sys_read(struct fb_info *info, char __user *buf, size_t count, loff_t *ppos); diff --git a/include/drm/drm_gem.h b/include/drm/drm_gem.h index e71f75a2ab57feb6dd382ad0792a2bd8200cdb38..97a48165642c1e5191902a32e84ba5e93d90f8f5 100644 --- a/include/drm/drm_gem.h +++ b/include/drm/drm_gem.h @@ -159,8 +159,9 @@ struct drm_gem_object_funcs { * * The callback is used by by both drm_gem_mmap_obj() and * drm_gem_prime_mmap(). When @mmap is present @vm_ops is not - * used, the @mmap callback must set vma->vm_ops instead. - * + * used, the @mmap callback must set vma->vm_ops instead. The @mmap + * callback is always called with a 0 offset. The caller will remove + * the fake offset as necessary. */ int (*mmap)(struct drm_gem_object *obj, struct vm_area_struct *vma); diff --git a/include/drm/drm_gem_vram_helper.h b/include/drm/drm_gem_vram_helper.h index b8ad4531ebb41f3176cfad3d8e1ec371570caa1a..e040541a105f7c702dbcbfe9b0da078524fb9ad4 100644 --- a/include/drm/drm_gem_vram_helper.h +++ b/include/drm/drm_gem_vram_helper.h @@ -13,6 +13,9 @@ #include <linux/kernel.h> /* for container_of() */ struct drm_mode_create_dumb; +struct drm_plane; +struct drm_plane_state; +struct drm_simple_display_pipe; struct drm_vram_mm_funcs; struct filp; struct vm_area_struct; @@ -124,6 +127,28 @@ int drm_gem_vram_driver_dumb_mmap_offset(struct drm_file *file, struct drm_device *dev, uint32_t handle, uint64_t *offset); +/* + * Helpers for struct drm_plane_helper_funcs + */ +int +drm_gem_vram_plane_helper_prepare_fb(struct drm_plane *plane, + struct drm_plane_state *new_state); +void +drm_gem_vram_plane_helper_cleanup_fb(struct drm_plane *plane, + struct drm_plane_state *old_state); + +/* + * Helpers for struct drm_simple_display_pipe_funcs + */ + +int drm_gem_vram_simple_display_pipe_prepare_fb( + struct drm_simple_display_pipe *pipe, + struct drm_plane_state *new_state); + +void drm_gem_vram_simple_display_pipe_cleanup_fb( + struct drm_simple_display_pipe *pipe, + struct drm_plane_state *old_state); + /** * define DRM_GEM_VRAM_DRIVER - default callback functions for \ &struct drm_driver diff --git a/include/drm/drm_simple_kms_helper.h b/include/drm/drm_simple_kms_helper.h index 4d89cd0a60db85b947685cb8944397eba806fa43..15afee9cf049841e42b2213c900779c886038689 100644 --- a/include/drm/drm_simple_kms_helper.h +++ b/include/drm/drm_simple_kms_helper.h @@ -49,7 +49,7 @@ struct drm_simple_display_pipe_funcs { * * drm_mode_status Enum */ - enum drm_mode_status (*mode_valid)(struct drm_crtc *crtc, + enum drm_mode_status (*mode_valid)(struct drm_simple_display_pipe *pipe, const struct drm_display_mode *mode); /** diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h index d2277e06316d85f6957d9f1819ae261901251303..54fa457b26abb9a97ef898ccc1de3631952151f2 100644 --- a/include/drm/ttm/ttm_bo_api.h +++ b/include/drm/ttm/ttm_bo_api.h @@ -147,7 +147,6 @@ struct ttm_tt; * holds a pointer to a persistent shmem object. * @ttm: TTM structure holding system pages. * @evicted: Whether the object was evicted without user-space knowing. - * @cpu_writes: For synchronization. Number of cpu writers. * @lru: List head for the lru list. * @ddestroy: List head for the delayed destroy list. * @swap: List head for swap LRU list. @@ -198,12 +197,6 @@ struct ttm_buffer_object { struct ttm_tt *ttm; bool evicted; - /** - * Members protected by the bo::reserved lock only when written to. - */ - - atomic_t cpu_writers; - /** * Members protected by the bdev::lru_lock. */ @@ -367,30 +360,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo, */ void ttm_bo_put(struct ttm_buffer_object *bo); -/** - * ttm_bo_add_to_lru - * - * @bo: The buffer object. - * - * Add this bo to the relevant mem type lru and, if it's backed by - * system pages (ttms) to the swap list. - * This function must be called with struct ttm_bo_global::lru_lock held, and - * is typically called immediately prior to unreserving a bo. - */ -void ttm_bo_add_to_lru(struct ttm_buffer_object *bo); - -/** - * ttm_bo_del_from_lru - * - * @bo: The buffer object. - * - * Remove this bo from all lru lists used to lookup and reserve an object. - * This function must be called with struct ttm_bo_global::lru_lock held, - * and is usually called just immediately after the bo has been reserved to - * avoid recursive reservation from lru lists. - */ -void ttm_bo_del_from_lru(struct ttm_buffer_object *bo); - /** * ttm_bo_move_to_lru_tail * @@ -441,31 +410,6 @@ void ttm_bo_unlock_delayed_workqueue(struct ttm_bo_device *bdev, int resched); bool ttm_bo_eviction_valuable(struct ttm_buffer_object *bo, const struct ttm_place *place); -/** - * ttm_bo_synccpu_write_grab - * - * @bo: The buffer object: - * @no_wait: Return immediately if buffer is busy. - * - * Synchronizes a buffer object for CPU RW access. This means - * command submission that affects the buffer will return -EBUSY - * until ttm_bo_synccpu_write_release is called. - * - * Returns - * -EBUSY if the buffer is busy and no_wait is true. - * -ERESTARTSYS if interrupted by a signal. - */ -int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait); - -/** - * ttm_bo_synccpu_write_release: - * - * @bo : The buffer object. - * - * Releases a synccpu lock. - */ -void ttm_bo_synccpu_write_release(struct ttm_buffer_object *bo); - /** * ttm_bo_acc_size * diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h index 27b9d5be197f22cd88de4a62630985287ad9a468..cac7a8a0825a75bd9032ce4320ddc212c0b44713 100644 --- a/include/drm/ttm/ttm_bo_driver.h +++ b/include/drm/ttm/ttm_bo_driver.h @@ -423,7 +423,6 @@ extern struct ttm_bo_global { */ struct kobject kobj; - struct ttm_mem_global *mem_glob; struct page *dummy_read_page; spinlock_t lru_lock; @@ -467,7 +466,6 @@ struct ttm_bo_device { * Constant after bo device init / atomic. */ struct list_head device_list; - struct ttm_bo_global *glob; struct ttm_bo_driver *driver; struct ttm_mem_type_manager man[TTM_NUM_MEM_TYPES]; @@ -631,9 +629,6 @@ void ttm_mem_io_free_vm(struct ttm_buffer_object *bo); int ttm_mem_io_lock(struct ttm_mem_type_manager *man, bool interruptible); void ttm_mem_io_unlock(struct ttm_mem_type_manager *man); -void ttm_bo_del_sub_from_lru(struct ttm_buffer_object *bo); -void ttm_bo_add_to_lru(struct ttm_buffer_object *bo); - /** * __ttm_bo_reserve: * @@ -727,15 +722,9 @@ static inline int ttm_bo_reserve(struct ttm_buffer_object *bo, bool interruptible, bool no_wait, struct ww_acquire_ctx *ticket) { - int ret; - WARN_ON(!kref_read(&bo->kref)); - ret = __ttm_bo_reserve(bo, interruptible, no_wait, ticket); - if (likely(ret == 0)) - ttm_bo_del_sub_from_lru(bo); - - return ret; + return __ttm_bo_reserve(bo, interruptible, no_wait, ticket); } /** @@ -762,9 +751,7 @@ static inline int ttm_bo_reserve_slowpath(struct ttm_buffer_object *bo, else dma_resv_lock_slow(bo->base.resv, ticket); - if (likely(ret == 0)) - ttm_bo_del_sub_from_lru(bo); - else if (ret == -EINTR) + if (ret == -EINTR) ret = -ERESTARTSYS; return ret; @@ -779,12 +766,9 @@ static inline int ttm_bo_reserve_slowpath(struct ttm_buffer_object *bo, */ static inline void ttm_bo_unreserve(struct ttm_buffer_object *bo) { - spin_lock(&bo->bdev->glob->lru_lock); - if (list_empty(&bo->lru)) - ttm_bo_add_to_lru(bo); - else - ttm_bo_move_to_lru_tail(bo, NULL); - spin_unlock(&bo->bdev->glob->lru_lock); + spin_lock(&ttm_bo_glob.lru_lock); + ttm_bo_move_to_lru_tail(bo, NULL); + spin_unlock(&ttm_bo_glob.lru_lock); dma_resv_unlock(bo->base.resv); } diff --git a/include/drm/ttm/ttm_execbuf_util.h b/include/drm/ttm/ttm_execbuf_util.h index 7e46cc678e7e7ab48b8dd9538a3864dfae6bbf3a..5a19843bb80d6fcab8075951764189b6f74eb9e6 100644 --- a/include/drm/ttm/ttm_execbuf_util.h +++ b/include/drm/ttm/ttm_execbuf_util.h @@ -99,7 +99,7 @@ extern void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket, extern int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket, struct list_head *list, bool intr, - struct list_head *dups, bool del_lru); + struct list_head *dups); /** * function ttm_eu_fence_buffer_objects. diff --git a/include/drm/ttm/ttm_memory.h b/include/drm/ttm/ttm_memory.h index 3ff48a0a2d7b7728469034430434a311e4bbcb40..c78ea99c42cf3b04ed0fc5f71038917735f84f04 100644 --- a/include/drm/ttm/ttm_memory.h +++ b/include/drm/ttm/ttm_memory.h @@ -65,7 +65,6 @@ struct ttm_mem_zone; extern struct ttm_mem_global { struct kobject kobj; - struct ttm_bo_global *bo_glob; struct workqueue_struct *swap_queue; struct work_struct work; spinlock_t lock; diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h index ec212cb27fdc6c49e3a0ad1859eea6317c65c35e..af73f835c51c4db8f888a04346f360cc72c516c1 100644 --- a/include/linux/dma-buf.h +++ b/include/linux/dma-buf.h @@ -42,6 +42,18 @@ struct dma_buf_ops { */ bool cache_sgt_mapping; + /** + * @dynamic_mapping: + * + * If true the framework makes sure that the map/unmap_dma_buf + * callbacks are always called with the dma_resv object locked. + * + * If false the framework makes sure that the map/unmap_dma_buf + * callbacks are always called without the dma_resv object locked. + * Mutual exclusive with @cache_sgt_mapping. + */ + bool dynamic_mapping; + /** * @attach: * @@ -109,6 +121,9 @@ struct dma_buf_ops { * any other kind of sharing that the exporter might wish to make * available to buffer-users. * + * This is always called with the dmabuf->resv object locked when + * the dynamic_mapping flag is true. + * * Returns: * * A &sg_table scatter list of or the backing storage of the DMA buffer, @@ -267,14 +282,16 @@ struct dma_buf_ops { * struct dma_buf - shared buffer object * @size: size of the buffer * @file: file pointer used for sharing buffers across, and for refcounting. - * @attachments: list of dma_buf_attachment that denotes all devices attached. + * @attachments: list of dma_buf_attachment that denotes all devices attached, + * protected by dma_resv lock. * @ops: dma_buf_ops associated with this buffer object. * @lock: used internally to serialize list manipulation, attach/detach and - * vmap/unmap, and accesses to name + * vmap/unmap * @vmapping_counter: used internally to refcnt the vmaps * @vmap_ptr: the current vmap ptr if vmapping_counter > 0 * @exp_name: name of the exporter; useful for debugging. - * @name: userspace-provided name; useful for accounting and debugging. + * @name: userspace-provided name; useful for accounting and debugging, + * protected by @resv. * @owner: pointer to exporter module; used for refcounting when exporter is a * kernel module. * @list_node: node for dma_buf accounting and debugging. @@ -323,10 +340,12 @@ struct dma_buf { * struct dma_buf_attachment - holds device-buffer attachment data * @dmabuf: buffer for this attachment. * @dev: device attached to the buffer. - * @node: list of dma_buf_attachment. + * @node: list of dma_buf_attachment, protected by dma_resv lock of the dmabuf. * @sgt: cached mapping. * @dir: direction of cached mapping. * @priv: exporter specific attachment data. + * @dynamic_mapping: true if dma_buf_map/unmap_attachment() is called with the + * dma_resv lock held. * * This structure holds the attachment information between the dma_buf buffer * and its user device(s). The list contains one attachment struct per device @@ -343,6 +362,7 @@ struct dma_buf_attachment { struct list_head node; struct sg_table *sgt; enum dma_data_direction dir; + bool dynamic_mapping; void *priv; }; @@ -394,10 +414,40 @@ static inline void get_dma_buf(struct dma_buf *dmabuf) get_file(dmabuf->file); } +/** + * dma_buf_is_dynamic - check if a DMA-buf uses dynamic mappings. + * @dmabuf: the DMA-buf to check + * + * Returns true if a DMA-buf exporter wants to be called with the dma_resv + * locked for the map/unmap callbacks, false if it doesn't wants to be called + * with the lock held. + */ +static inline bool dma_buf_is_dynamic(struct dma_buf *dmabuf) +{ + return dmabuf->ops->dynamic_mapping; +} + +/** + * dma_buf_attachment_is_dynamic - check if a DMA-buf attachment uses dynamic + * mappinsg + * @attach: the DMA-buf attachment to check + * + * Returns true if a DMA-buf importer wants to call the map/unmap functions with + * the dma_resv lock held. + */ +static inline bool +dma_buf_attachment_is_dynamic(struct dma_buf_attachment *attach) +{ + return attach->dynamic_mapping; +} + struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf, - struct device *dev); + struct device *dev); +struct dma_buf_attachment * +dma_buf_dynamic_attach(struct dma_buf *dmabuf, struct device *dev, + bool dynamic_mapping); void dma_buf_detach(struct dma_buf *dmabuf, - struct dma_buf_attachment *dmabuf_attach); + struct dma_buf_attachment *attach); struct dma_buf *dma_buf_export(const struct dma_buf_export_info *exp_info); @@ -409,6 +459,7 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *, enum dma_data_direction); void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *, enum dma_data_direction); +void dma_buf_move_notify(struct dma_buf *dma_buf); int dma_buf_begin_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction dir); int dma_buf_end_cpu_access(struct dma_buf *dma_buf, diff --git a/include/uapi/drm/exynos_drm.h b/include/uapi/drm/exynos_drm.h index 3e59b8382dd8cead4544ab7e0d302170712b70e1..45c6582b3df31dbf1424b3f772252a8c7434c9da 100644 --- a/include/uapi/drm/exynos_drm.h +++ b/include/uapi/drm/exynos_drm.h @@ -68,7 +68,7 @@ struct drm_exynos_gem_info { /** * A structure for user connection request of virtual display. * - * @connection: indicate whether doing connetion or not by user. + * @connection: indicate whether doing connection or not by user. * @extensions: if this value is 1 then the vidi driver would need additional * 128bytes edid data. * @edid: the edid data pointer from user side.