From patchwork Sun Mar 7 20:28:31 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul Cercueil X-Patchwork-Id: 12121063 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 5FFB1C433DB for ; Sun, 7 Mar 2021 20:29:05 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id E0E81650A2 for ; Sun, 7 Mar 2021 20:29:04 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org E0E81650A2 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=crapouillou.net Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 3654C6E1E8; Sun, 7 Mar 2021 20:29:04 +0000 (UTC) Received: from aposti.net (aposti.net [89.234.176.197]) by gabe.freedesktop.org (Postfix) with ESMTPS id 5E9CB6E1E8 for ; Sun, 7 Mar 2021 20:29:03 +0000 (UTC) From: Paul Cercueil To: Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Daniel Vetter Subject: [PATCH v2 1/5] drm: Add and export function drm_gem_cma_create_noncoherent Date: Sun, 7 Mar 2021 20:28:31 +0000 Message-Id: <20210307202835.253907-2-paul@crapouillou.net> In-Reply-To: <20210307202835.253907-1-paul@crapouillou.net> References: <20210307202835.253907-1-paul@crapouillou.net> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-mips@vger.kernel.org, Paul Cercueil , od@zcrc.me, Sam Ravnborg Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This function can be used by drivers that need to create a GEM object with non-coherent backing memory. Creating non-coherent CMA objects is useful on architectures where writing to a buffer with the non-coherent cache attribute set then invalidating the cache is faster than writing to the same buffer with the write-combine cache attribute set. This is the case for instance on some Ingenic SoCs. v2: Add inline doc about why we need this, and improve commit message Signed-off-by: Paul Cercueil --- drivers/gpu/drm/drm_gem_cma_helper.c | 76 +++++++++++++++++++++------- include/drm/drm_gem_cma_helper.h | 2 + 2 files changed, 61 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c index 7942cf05cd93..917b092b23c2 100644 --- a/drivers/gpu/drm/drm_gem_cma_helper.c +++ b/drivers/gpu/drm/drm_gem_cma_helper.c @@ -90,21 +90,10 @@ __drm_gem_cma_create(struct drm_device *drm, size_t size) return ERR_PTR(ret); } -/** - * drm_gem_cma_create - allocate an object with the given size - * @drm: DRM device - * @size: size of the object to allocate - * - * This function creates a CMA GEM object and allocates a contiguous chunk of - * memory as backing store. The backing memory has the writecombine attribute - * set. - * - * Returns: - * A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative - * error code on failure. - */ -struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, - size_t size) +static struct drm_gem_cma_object * +drm_gem_cma_create_with_cache_param(struct drm_device *drm, + size_t size, + bool noncoherent) { struct drm_gem_cma_object *cma_obj; int ret; @@ -115,8 +104,16 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, if (IS_ERR(cma_obj)) return cma_obj; - cma_obj->vaddr = dma_alloc_wc(drm->dev, size, &cma_obj->paddr, - GFP_KERNEL | __GFP_NOWARN); + if (noncoherent) { + cma_obj->vaddr = dma_alloc_noncoherent(drm->dev, size, + &cma_obj->paddr, + DMA_TO_DEVICE, + GFP_KERNEL | __GFP_NOWARN); + + } else { + cma_obj->vaddr = dma_alloc_wc(drm->dev, size, &cma_obj->paddr, + GFP_KERNEL | __GFP_NOWARN); + } if (!cma_obj->vaddr) { drm_dbg(drm, "failed to allocate buffer with size %zu\n", size); @@ -130,6 +127,51 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, drm_gem_object_put(&cma_obj->base); return ERR_PTR(ret); } + +/** + * drm_gem_cma_create_noncoherent - allocate an object with the given size + * and non-coherent cache attribute + * @drm: DRM device + * @size: size of the object to allocate + * + * This function creates a CMA GEM object and allocates a contiguous chunk of + * memory as backing store. The backing memory has the noncoherent attribute + * set. + * + * Creating non-coherent CMA objects is useful on architectures where writing + * to a buffer with the non-coherent cache attribute set then invalidating the + * cache is faster than writing to the same buffer with the write-combine cache + * attribute set. This is the case for instance on some Ingenic SoCs. + * + * Returns: + * A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative + * error code on failure. + */ +struct drm_gem_cma_object * +drm_gem_cma_create_noncoherent(struct drm_device *drm, size_t size) +{ + return drm_gem_cma_create_with_cache_param(drm, size, true); +} +EXPORT_SYMBOL_GPL(drm_gem_cma_create_noncoherent); + +/** + * drm_gem_cma_create - allocate an object with the given size + * @drm: DRM device + * @size: size of the object to allocate + * + * This function creates a CMA GEM object and allocates a contiguous chunk of + * memory as backing store. The backing memory has the writecombine attribute + * set. + * + * Returns: + * A struct drm_gem_cma_object * on success or an ERR_PTR()-encoded negative + * error code on failure. + */ +struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, + size_t size) +{ + return drm_gem_cma_create_with_cache_param(drm, size, false); +} EXPORT_SYMBOL_GPL(drm_gem_cma_create); /** diff --git a/include/drm/drm_gem_cma_helper.h b/include/drm/drm_gem_cma_helper.h index 0a9711caa3e8..360771f5f485 100644 --- a/include/drm/drm_gem_cma_helper.h +++ b/include/drm/drm_gem_cma_helper.h @@ -79,6 +79,8 @@ int drm_gem_cma_dumb_create(struct drm_file *file_priv, /* allocate physical memory */ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, size_t size); +struct drm_gem_cma_object * +drm_gem_cma_create_noncoherent(struct drm_device *drm, size_t size); extern const struct vm_operations_struct drm_gem_cma_vm_ops; From patchwork Sun Mar 7 20:28:32 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul Cercueil X-Patchwork-Id: 12121065 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8A3CDC433E0 for ; Sun, 7 Mar 2021 20:29:11 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 2657365168 for ; Sun, 7 Mar 2021 20:29:11 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 2657365168 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=crapouillou.net Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 83DEF6E7E6; Sun, 7 Mar 2021 20:29:10 +0000 (UTC) Received: from aposti.net (aposti.net [89.234.176.197]) by gabe.freedesktop.org (Postfix) with ESMTPS id 2B0C56E7E6 for ; Sun, 7 Mar 2021 20:29:10 +0000 (UTC) From: Paul Cercueil To: Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Daniel Vetter Subject: [PATCH v2 2/5] drm: Add and export function drm_gem_cma_dumb_create_noncoherent Date: Sun, 7 Mar 2021 20:28:32 +0000 Message-Id: <20210307202835.253907-3-paul@crapouillou.net> In-Reply-To: <20210307202835.253907-1-paul@crapouillou.net> References: <20210307202835.253907-1-paul@crapouillou.net> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-mips@vger.kernel.org, Paul Cercueil , od@zcrc.me, Sam Ravnborg Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This function can be used by drivers to create dumb buffers with non-coherent backing memory. Signed-off-by: Paul Cercueil --- drivers/gpu/drm/drm_gem_cma_helper.c | 37 +++++++++++++++++++++++++--- include/drm/drm_gem_cma_helper.h | 5 ++++ 2 files changed, 38 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c index 917b092b23c2..d100c5f9c140 100644 --- a/drivers/gpu/drm/drm_gem_cma_helper.c +++ b/drivers/gpu/drm/drm_gem_cma_helper.c @@ -181,6 +181,7 @@ EXPORT_SYMBOL_GPL(drm_gem_cma_create); * @drm: DRM device * @size: size of the object to allocate * @handle: return location for the GEM handle + * @noncoherent: allocate object with non-coherent cache attribute * * This function creates a CMA GEM object, allocating a physically contiguous * chunk of memory as backing store. The GEM object is then added to the list @@ -193,13 +194,13 @@ EXPORT_SYMBOL_GPL(drm_gem_cma_create); static struct drm_gem_cma_object * drm_gem_cma_create_with_handle(struct drm_file *file_priv, struct drm_device *drm, size_t size, - uint32_t *handle) + uint32_t *handle, bool noncoherent) { struct drm_gem_cma_object *cma_obj; struct drm_gem_object *gem_obj; int ret; - cma_obj = drm_gem_cma_create(drm, size); + cma_obj = drm_gem_cma_create_with_cache_param(drm, size, noncoherent); if (IS_ERR(cma_obj)) return cma_obj; @@ -276,7 +277,7 @@ int drm_gem_cma_dumb_create_internal(struct drm_file *file_priv, args->size = args->pitch * args->height; cma_obj = drm_gem_cma_create_with_handle(file_priv, drm, args->size, - &args->handle); + &args->handle, false); return PTR_ERR_OR_ZERO(cma_obj); } EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create_internal); @@ -309,11 +310,39 @@ int drm_gem_cma_dumb_create(struct drm_file *file_priv, args->size = args->pitch * args->height; cma_obj = drm_gem_cma_create_with_handle(file_priv, drm, args->size, - &args->handle); + &args->handle, false); return PTR_ERR_OR_ZERO(cma_obj); } EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create); +/** + * drm_gem_cma_dumb_create_noncoherent - create a dumb buffer object with + * non-coherent cache attribute + * @file_priv: DRM file-private structure to create the dumb buffer for + * @drm: DRM device + * @args: IOCTL data + * + * Same as drm_gem_cma_dumb_create, but the dumb buffer object created has + * the non-coherent cache attribute set. + * + * Returns: + * 0 on success or a negative error code on failure. + */ +int drm_gem_cma_dumb_create_noncoherent(struct drm_file *file_priv, + struct drm_device *drm, + struct drm_mode_create_dumb *args) +{ + struct drm_gem_cma_object *cma_obj; + + args->pitch = DIV_ROUND_UP(args->width * args->bpp, 8); + args->size = args->pitch * args->height; + + cma_obj = drm_gem_cma_create_with_handle(file_priv, drm, args->size, + &args->handle, true); + return PTR_ERR_OR_ZERO(cma_obj); +} +EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create_noncoherent); + const struct vm_operations_struct drm_gem_cma_vm_ops = { .open = drm_gem_vm_open, .close = drm_gem_vm_close, diff --git a/include/drm/drm_gem_cma_helper.h b/include/drm/drm_gem_cma_helper.h index 360771f5f485..6b44e7492a63 100644 --- a/include/drm/drm_gem_cma_helper.h +++ b/include/drm/drm_gem_cma_helper.h @@ -76,6 +76,11 @@ int drm_gem_cma_dumb_create(struct drm_file *file_priv, struct drm_device *drm, struct drm_mode_create_dumb *args); +/* create non-coherent memory region for DRM framebuffer */ +int drm_gem_cma_dumb_create_noncoherent(struct drm_file *file_priv, + struct drm_device *drm, + struct drm_mode_create_dumb *args); + /* allocate physical memory */ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm, size_t size); From patchwork Sun Mar 7 20:28:33 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul Cercueil X-Patchwork-Id: 12121067 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C02AFC433DB for ; Sun, 7 Mar 2021 20:29:18 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 6ABE1650A2 for ; Sun, 7 Mar 2021 20:29:18 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 6ABE1650A2 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=crapouillou.net Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id C6BF06E7EA; Sun, 7 Mar 2021 20:29:17 +0000 (UTC) Received: from aposti.net (aposti.net [89.234.176.197]) by gabe.freedesktop.org (Postfix) with ESMTPS id 79D196E7EA for ; Sun, 7 Mar 2021 20:29:16 +0000 (UTC) From: Paul Cercueil To: Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Daniel Vetter Subject: [PATCH v2 3/5] drm: Add and export function drm_gem_cma_mmap_noncoherent Date: Sun, 7 Mar 2021 20:28:33 +0000 Message-Id: <20210307202835.253907-4-paul@crapouillou.net> In-Reply-To: <20210307202835.253907-1-paul@crapouillou.net> References: <20210307202835.253907-1-paul@crapouillou.net> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-mips@vger.kernel.org, Paul Cercueil , od@zcrc.me, Sam Ravnborg Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This function can be used by drivers that need to mmap dumb buffers created with non-coherent backing memory. v2: Use dma_to_phys() since cma_obj->paddr isn't a phys_addr_t but a dma_addr_t. Signed-off-by: Paul Cercueil --- drivers/gpu/drm/drm_gem_cma_helper.c | 67 +++++++++++++++++++++++++--- include/drm/drm_gem_cma_helper.h | 1 + 2 files changed, 63 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c index d100c5f9c140..e39b0464e19d 100644 --- a/drivers/gpu/drm/drm_gem_cma_helper.c +++ b/drivers/gpu/drm/drm_gem_cma_helper.c @@ -10,6 +10,7 @@ */ #include +#include #include #include #include @@ -42,10 +43,20 @@ static const struct drm_gem_object_funcs drm_gem_cma_default_funcs = { .vm_ops = &drm_gem_cma_vm_ops, }; +static const struct drm_gem_object_funcs drm_gem_cma_noncoherent_funcs = { + .free = drm_gem_cma_free_object, + .print_info = drm_gem_cma_print_info, + .get_sg_table = drm_gem_cma_get_sg_table, + .vmap = drm_gem_cma_vmap, + .mmap = drm_gem_cma_mmap_noncoherent, + .vm_ops = &drm_gem_cma_vm_ops, +}; + /** * __drm_gem_cma_create - Create a GEM CMA object without allocating memory * @drm: DRM device * @size: size of the object to allocate + * @noncoherent: if true, will use non-coherent backed memory * * This function creates and initializes a GEM CMA object of the given size, * but doesn't allocate any memory to back the object. @@ -55,7 +66,7 @@ static const struct drm_gem_object_funcs drm_gem_cma_default_funcs = { * error code on failure. */ static struct drm_gem_cma_object * -__drm_gem_cma_create(struct drm_device *drm, size_t size) +__drm_gem_cma_create(struct drm_device *drm, size_t size, bool noncoherent) { struct drm_gem_cma_object *cma_obj; struct drm_gem_object *gem_obj; @@ -68,8 +79,12 @@ __drm_gem_cma_create(struct drm_device *drm, size_t size) if (!gem_obj) return ERR_PTR(-ENOMEM); - if (!gem_obj->funcs) - gem_obj->funcs = &drm_gem_cma_default_funcs; + if (!gem_obj->funcs) { + if (noncoherent) + gem_obj->funcs = &drm_gem_cma_noncoherent_funcs; + else + gem_obj->funcs = &drm_gem_cma_default_funcs; + } cma_obj = container_of(gem_obj, struct drm_gem_cma_object, base); @@ -100,7 +115,7 @@ drm_gem_cma_create_with_cache_param(struct drm_device *drm, size = round_up(size, PAGE_SIZE); - cma_obj = __drm_gem_cma_create(drm, size); + cma_obj = __drm_gem_cma_create(drm, size, noncoherent); if (IS_ERR(cma_obj)) return cma_obj; @@ -503,7 +518,7 @@ drm_gem_cma_prime_import_sg_table(struct drm_device *dev, return ERR_PTR(-EINVAL); /* Create a CMA GEM buffer. */ - cma_obj = __drm_gem_cma_create(dev, attach->dmabuf->size); + cma_obj = __drm_gem_cma_create(dev, attach->dmabuf->size, false); if (IS_ERR(cma_obj)) return ERR_CAST(cma_obj); @@ -579,6 +594,48 @@ int drm_gem_cma_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma) } EXPORT_SYMBOL_GPL(drm_gem_cma_mmap); +/** + * drm_gem_cma_mmap_noncoherent - memory-map a CMA GEM object with + * non-coherent cache attribute + * @filp: file object + * @vma: VMA for the area to be mapped + * + * Just like drm_gem_cma_mmap, but for a GEM object backed by non-coherent + * memory. + * + * Returns: + * 0 on success or a negative error code on failure. + */ +int drm_gem_cma_mmap_noncoherent(struct drm_gem_object *obj, + struct vm_area_struct *vma) +{ + struct drm_gem_cma_object *cma_obj; + unsigned long pfn; + int ret; + + /* + * Clear the VM_PFNMAP flag that was set by drm_gem_mmap(), and set the + * vm_pgoff (used as a fake buffer offset by DRM) to 0 as we want to map + * the whole buffer. + */ + vma->vm_pgoff -= drm_vma_node_start(&obj->vma_node); + vma->vm_flags &= ~VM_PFNMAP; + vma->vm_page_prot = vm_get_page_prot(vma->vm_flags); + + cma_obj = to_drm_gem_cma_obj(obj); + + pfn = PHYS_PFN(dma_to_phys(cma_obj->base.dev->dev, cma_obj->paddr)); + + ret = remap_pfn_range(vma, vma->vm_start, pfn, + vma->vm_end - vma->vm_start, + vma->vm_page_prot); + if (ret) + drm_gem_vm_close(vma); + + return ret; +} +EXPORT_SYMBOL_GPL(drm_gem_cma_mmap_noncoherent); + /** * drm_gem_cma_prime_import_sg_table_vmap - PRIME import another driver's * scatter/gather table and get the virtual address of the buffer diff --git a/include/drm/drm_gem_cma_helper.h b/include/drm/drm_gem_cma_helper.h index 6b44e7492a63..6a3f7e1312cc 100644 --- a/include/drm/drm_gem_cma_helper.h +++ b/include/drm/drm_gem_cma_helper.h @@ -107,6 +107,7 @@ drm_gem_cma_prime_import_sg_table(struct drm_device *dev, struct sg_table *sgt); int drm_gem_cma_vmap(struct drm_gem_object *obj, struct dma_buf_map *map); int drm_gem_cma_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma); +int drm_gem_cma_mmap_noncoherent(struct drm_gem_object *obj, struct vm_area_struct *vma); /** * DRM_GEM_CMA_DRIVER_OPS_WITH_DUMB_CREATE - CMA GEM driver operations From patchwork Sun Mar 7 20:28:34 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul Cercueil X-Patchwork-Id: 12121069 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 378A0C433E0 for ; Sun, 7 Mar 2021 20:29:25 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id DFF4965166 for ; Sun, 7 Mar 2021 20:29:24 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org DFF4965166 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=crapouillou.net Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 5E3CC6E7EC; Sun, 7 Mar 2021 20:29:24 +0000 (UTC) Received: from aposti.net (aposti.net [89.234.176.197]) by gabe.freedesktop.org (Postfix) with ESMTPS id 2F52A6E7EC for ; Sun, 7 Mar 2021 20:29:23 +0000 (UTC) From: Paul Cercueil To: Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Daniel Vetter Subject: [PATCH v2 4/5] drm: Add and export function drm_gem_cma_sync_data Date: Sun, 7 Mar 2021 20:28:34 +0000 Message-Id: <20210307202835.253907-5-paul@crapouillou.net> In-Reply-To: <20210307202835.253907-1-paul@crapouillou.net> References: <20210307202835.253907-1-paul@crapouillou.net> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-mips@vger.kernel.org, Paul Cercueil , od@zcrc.me, Sam Ravnborg Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This function can be used by drivers that use damage clips and have CMA GEM objects backed by non-coherent memory. Calling this function in a plane's .atomic_update ensures that all the data in the backing memory have been written to RAM. Signed-off-by: Paul Cercueil --- drivers/gpu/drm/drm_gem_cma_helper.c | 43 ++++++++++++++++++++++++++++ include/drm/drm_gem_cma_helper.h | 5 ++++ 2 files changed, 48 insertions(+) diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c index e39b0464e19d..fdae54a18670 100644 --- a/drivers/gpu/drm/drm_gem_cma_helper.c +++ b/drivers/gpu/drm/drm_gem_cma_helper.c @@ -18,9 +18,14 @@ #include #include +#include #include #include +#include +#include +#include #include +#include #include /** @@ -684,3 +689,41 @@ drm_gem_cma_prime_import_sg_table_vmap(struct drm_device *dev, return obj; } EXPORT_SYMBOL(drm_gem_cma_prime_import_sg_table_vmap); + +/** + * drm_gem_cma_sync_data - Sync GEM object to non-coherent backing memory + * @dev: DRM device + * @old_state: Old plane state + * @state: New plane state + * + * This function can be used by drivers that use damage clips and have + * CMA GEM objects backed by non-coherent memory. Calling this function + * in a plane's .atomic_update ensures that all the data in the backing + * memory have been written to RAM. + */ +void drm_gem_cma_sync_data(struct device *dev, + struct drm_plane_state *old_state, + struct drm_plane_state *state) +{ + const struct drm_format_info *finfo = state->fb->format; + struct drm_atomic_helper_damage_iter iter; + unsigned int offset, i; + struct drm_rect clip; + dma_addr_t daddr; + + drm_atomic_helper_damage_iter_init(&iter, old_state, state); + + drm_atomic_for_each_plane_damage(&iter, &clip) { + for (i = 0; i < finfo->num_planes; i++) { + daddr = drm_fb_cma_get_gem_addr(state->fb, state, i); + + /* Ignore x1/x2 values, invalidate complete lines */ + offset = clip.y1 * state->fb->pitches[i]; + + dma_sync_single_for_device(dev, daddr + offset, + (clip.y2 - clip.y1) * state->fb->pitches[i], + DMA_TO_DEVICE); + } + } +} +EXPORT_SYMBOL_GPL(drm_gem_cma_sync_data); diff --git a/include/drm/drm_gem_cma_helper.h b/include/drm/drm_gem_cma_helper.h index 6a3f7e1312cc..cdd3fb456916 100644 --- a/include/drm/drm_gem_cma_helper.h +++ b/include/drm/drm_gem_cma_helper.h @@ -7,6 +7,7 @@ #include struct drm_mode_create_dumb; +struct drm_plane_state; /** * struct drm_gem_cma_object - GEM object backed by CMA memory allocations @@ -190,4 +191,8 @@ drm_gem_cma_prime_import_sg_table_vmap(struct drm_device *drm, struct dma_buf_attachment *attach, struct sg_table *sgt); +void drm_gem_cma_sync_data(struct device *dev, + struct drm_plane_state *old_state, + struct drm_plane_state *state); + #endif /* __DRM_GEM_CMA_HELPER_H__ */ From patchwork Sun Mar 7 20:28:35 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul Cercueil X-Patchwork-Id: 12121071 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C6362C433DB for ; Sun, 7 Mar 2021 20:29:31 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 7EA4065165 for ; Sun, 7 Mar 2021 20:29:31 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 7EA4065165 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=crapouillou.net Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=dri-devel-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id E44BD6E7EF; Sun, 7 Mar 2021 20:29:30 +0000 (UTC) Received: from aposti.net (aposti.net [89.234.176.197]) by gabe.freedesktop.org (Postfix) with ESMTPS id 8B3AA6E7EF for ; Sun, 7 Mar 2021 20:29:29 +0000 (UTC) From: Paul Cercueil To: Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Daniel Vetter Subject: [PATCH v2 5/5] drm/ingenic: Add option to alloc cached GEM buffers Date: Sun, 7 Mar 2021 20:28:35 +0000 Message-Id: <20210307202835.253907-6-paul@crapouillou.net> In-Reply-To: <20210307202835.253907-1-paul@crapouillou.net> References: <20210307202835.253907-1-paul@crapouillou.net> MIME-Version: 1.0 X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, linux-mips@vger.kernel.org, Paul Cercueil , od@zcrc.me, Sam Ravnborg Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" With the module parameter ingenic-drm.cached_gem_buffers, it is possible to specify that we want GEM buffers backed by non-coherent memory. This dramatically speeds up software rendering on Ingenic SoCs, even for tasks where write-combine memory should in theory be faster (e.g. simple blits). Leave it disabled by default, since it is specific to one use-case (software rendering). v2: Rework code to work with new DRM APIs regarding plane states Signed-off-by: Paul Cercueil --- drivers/gpu/drm/ingenic/ingenic-drm-drv.c | 49 ++++++++++++++++++++++- drivers/gpu/drm/ingenic/ingenic-drm.h | 4 ++ drivers/gpu/drm/ingenic/ingenic-ipu.c | 14 ++++++- 3 files changed, 63 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/ingenic/ingenic-drm-drv.c b/drivers/gpu/drm/ingenic/ingenic-drm-drv.c index d60e1eefc9d1..ba1ac0fcda74 100644 --- a/drivers/gpu/drm/ingenic/ingenic-drm-drv.c +++ b/drivers/gpu/drm/ingenic/ingenic-drm-drv.c @@ -9,6 +9,7 @@ #include #include #include +#include #include #include #include @@ -23,6 +24,7 @@ #include #include #include +#include #include #include #include @@ -99,6 +101,11 @@ struct ingenic_drm { struct notifier_block clock_nb; }; +static bool ingenic_drm_cached_gem_buf; +module_param_named(cached_gem_buffers, ingenic_drm_cached_gem_buf, bool, 0400); +MODULE_PARM_DESC(cached_gem_buffers, + "Enable fully cached GEM buffers [default=false]"); + static bool ingenic_drm_writeable_reg(struct device *dev, unsigned int reg) { switch (reg) { @@ -410,6 +417,8 @@ static int ingenic_drm_plane_atomic_check(struct drm_plane *plane, old_plane_state->fb->format->format != new_plane_state->fb->format->format)) crtc_state->mode_changed = true; + drm_atomic_helper_check_plane_damage(state, new_plane_state); + return 0; } @@ -541,10 +550,20 @@ static void ingenic_drm_update_palette(struct ingenic_drm *priv, } } +void ingenic_drm_sync_data(struct device *dev, + struct drm_plane_state *old_state, + struct drm_plane_state *state) +{ + if (ingenic_drm_cached_gem_buf) + drm_gem_cma_sync_data(dev, old_state, state); +} + static void ingenic_drm_plane_atomic_update(struct drm_plane *plane, struct drm_atomic_state *state) { struct ingenic_drm *priv = drm_device_get_priv(plane->dev); + struct drm_plane_state *oldstate = drm_atomic_get_old_plane_state(state, + plane); struct drm_plane_state *newstate = drm_atomic_get_new_plane_state(state, plane); struct drm_crtc_state *crtc_state; @@ -554,6 +573,8 @@ static void ingenic_drm_plane_atomic_update(struct drm_plane *plane, u32 fourcc; if (newstate && newstate->fb) { + ingenic_drm_sync_data(priv->dev, oldstate, newstate); + crtc_state = newstate->crtc->state; addr = drm_fb_cma_get_gem_addr(newstate->fb, newstate, 0); @@ -743,6 +764,26 @@ static void ingenic_drm_disable_vblank(struct drm_crtc *crtc) regmap_update_bits(priv->map, JZ_REG_LCD_CTRL, JZ_LCD_CTRL_EOF_IRQ, 0); } +static struct drm_framebuffer * +ingenic_drm_gem_fb_create(struct drm_device *dev, struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + if (ingenic_drm_cached_gem_buf) + return drm_gem_fb_create_with_dirty(dev, file, mode_cmd); + + return drm_gem_fb_create(dev, file, mode_cmd); +} + +static int ingenic_drm_gem_cma_dumb_create(struct drm_file *file_priv, + struct drm_device *drm, + struct drm_mode_create_dumb *args) +{ + if (ingenic_drm_cached_gem_buf) + return drm_gem_cma_dumb_create_noncoherent(file_priv, drm, args); + + return drm_gem_cma_dumb_create(file_priv, drm, args); +} + DEFINE_DRM_GEM_CMA_FOPS(ingenic_drm_fops); static const struct drm_driver ingenic_drm_driver_data = { @@ -755,7 +796,7 @@ static const struct drm_driver ingenic_drm_driver_data = { .patchlevel = 0, .fops = &ingenic_drm_fops, - DRM_GEM_CMA_DRIVER_OPS, + DRM_GEM_CMA_DRIVER_OPS_WITH_DUMB_CREATE(ingenic_drm_gem_cma_dumb_create), .irq_handler = ingenic_drm_irq_handler, }; @@ -805,7 +846,7 @@ static const struct drm_encoder_helper_funcs ingenic_drm_encoder_helper_funcs = }; static const struct drm_mode_config_funcs ingenic_drm_mode_config_funcs = { - .fb_create = drm_gem_fb_create, + .fb_create = ingenic_drm_gem_fb_create, .output_poll_changed = drm_fb_helper_output_poll_changed, .atomic_check = drm_atomic_helper_check, .atomic_commit = drm_atomic_helper_commit, @@ -962,6 +1003,8 @@ static int ingenic_drm_bind(struct device *dev, bool has_components) return ret; } + drm_plane_enable_fb_damage_clips(&priv->f1); + drm_crtc_helper_add(&priv->crtc, &ingenic_drm_crtc_helper_funcs); ret = drm_crtc_init_with_planes(drm, &priv->crtc, primary, @@ -990,6 +1033,8 @@ static int ingenic_drm_bind(struct device *dev, bool has_components) return ret; } + drm_plane_enable_fb_damage_clips(&priv->f0); + if (IS_ENABLED(CONFIG_DRM_INGENIC_IPU) && has_components) { ret = component_bind_all(dev, drm); if (ret) { diff --git a/drivers/gpu/drm/ingenic/ingenic-drm.h b/drivers/gpu/drm/ingenic/ingenic-drm.h index 1b4347f7f084..b6bca356e024 100644 --- a/drivers/gpu/drm/ingenic/ingenic-drm.h +++ b/drivers/gpu/drm/ingenic/ingenic-drm.h @@ -185,6 +185,10 @@ void ingenic_drm_plane_config(struct device *dev, struct drm_plane *plane, u32 fourcc); void ingenic_drm_plane_disable(struct device *dev, struct drm_plane *plane); +void ingenic_drm_sync_data(struct device *dev, + struct drm_plane_state *old_state, + struct drm_plane_state *state); + extern struct platform_driver *ingenic_ipu_driver_ptr; #endif /* DRIVERS_GPU_DRM_INGENIC_INGENIC_DRM_H */ diff --git a/drivers/gpu/drm/ingenic/ingenic-ipu.c b/drivers/gpu/drm/ingenic/ingenic-ipu.c index 5ae6adab8306..7826eab044ba 100644 --- a/drivers/gpu/drm/ingenic/ingenic-ipu.c +++ b/drivers/gpu/drm/ingenic/ingenic-ipu.c @@ -20,6 +20,7 @@ #include #include +#include #include #include #include @@ -285,6 +286,8 @@ static void ingenic_ipu_plane_atomic_update(struct drm_plane *plane, struct drm_atomic_state *state) { struct ingenic_ipu *ipu = plane_to_ingenic_ipu(plane); + struct drm_plane_state *oldstate = drm_atomic_get_old_plane_state(state, + plane); struct drm_plane_state *newstate = drm_atomic_get_new_plane_state(state, plane); const struct drm_format_info *finfo; @@ -317,6 +320,8 @@ static void ingenic_ipu_plane_atomic_update(struct drm_plane *plane, JZ_IPU_CTRL_CHIP_EN | JZ_IPU_CTRL_LCDC_SEL); } + ingenic_drm_sync_data(ipu->master, oldstate, newstate); + /* New addresses will be committed in vblank handler... */ ipu->addr_y = drm_fb_cma_get_gem_addr(newstate->fb, newstate, 0); if (finfo->num_planes > 1) @@ -541,7 +546,7 @@ static int ingenic_ipu_plane_atomic_check(struct drm_plane *plane, if (!new_plane_state->crtc || !crtc_state->mode.hdisplay || !crtc_state->mode.vdisplay) - return 0; + goto out_check_damage; /* Plane must be fully visible */ if (new_plane_state->crtc_x < 0 || new_plane_state->crtc_y < 0 || @@ -558,7 +563,7 @@ static int ingenic_ipu_plane_atomic_check(struct drm_plane *plane, return -EINVAL; if (!osd_changed(new_plane_state, old_plane_state)) - return 0; + goto out_check_damage; crtc_state->mode_changed = true; @@ -592,6 +597,9 @@ static int ingenic_ipu_plane_atomic_check(struct drm_plane *plane, ipu->denom_w = denom_w; ipu->denom_h = denom_h; +out_check_damage: + drm_atomic_helper_check_plane_damage(state, new_plane_state); + return 0; } @@ -773,6 +781,8 @@ static int ingenic_ipu_bind(struct device *dev, struct device *master, void *d) return err; } + drm_plane_enable_fb_damage_clips(plane); + /* * Sharpness settings range is [0,32] * 0 : nearest-neighbor