From patchwork Sat Dec 14 21:57:34 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chris Wilson X-Patchwork-Id: 11292519 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 62767109A for ; Sat, 14 Dec 2019 21:57:45 +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 4771C2073D for ; Sat, 14 Dec 2019 21:57:45 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 4771C2073D Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=chris-wilson.co.uk Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=intel-gfx-bounces@lists.freedesktop.org Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 10C796E3F3; Sat, 14 Dec 2019 21:57:43 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from fireflyinternet.com (mail.fireflyinternet.com [109.228.58.192]) by gabe.freedesktop.org (Postfix) with ESMTPS id 8A1E86E3F3 for ; Sat, 14 Dec 2019 21:57:41 +0000 (UTC) X-Default-Received-SPF: pass (skip=forwardok (res=PASS)) x-ip-name=78.156.65.138; Received: from haswell.alporthouse.com (unverified [78.156.65.138]) by fireflyinternet.com (Firefly Internet (M1)) with ESMTP id 19580963-1500050 for multiple; Sat, 14 Dec 2019 21:57:35 +0000 From: Chris Wilson To: intel-gfx@lists.freedesktop.org Date: Sat, 14 Dec 2019 21:57:34 +0000 Message-Id: <20191214215734.2277302-1-chris@chris-wilson.co.uk> X-Mailer: git-send-email 2.24.0 In-Reply-To: <20191214211241.2243420-1-chris@chris-wilson.co.uk> References: <20191214211241.2243420-1-chris@chris-wilson.co.uk> MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH] drm/i915: Unpin vma->obj on early error X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" If we inherit an error along the fence chain, we skip the main work callback and go straight to the error. In the case of the vma bind worker, we only dropped the pinned pages from the worker. In the process, make sure we call the release earlier rather than wait until the final reference to the fence is dropped (as a reference is kept while being listened upon). Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/gem/i915_gem_clflush.c | 11 ++++------- drivers/gpu/drm/i915/i915_sw_fence_work.c | 5 ++--- drivers/gpu/drm/i915/i915_vma.c | 17 +++++++++++++---- 3 files changed, 19 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_clflush.c b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c index b9f504ba3b32..5448efa77710 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_clflush.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c @@ -26,27 +26,24 @@ static void __do_clflush(struct drm_i915_gem_object *obj) static int clflush_work(struct dma_fence_work *base) { struct clflush *clflush = container_of(base, typeof(*clflush), base); - struct drm_i915_gem_object *obj = fetch_and_zero(&clflush->obj); + struct drm_i915_gem_object *obj = clflush->obj; int err; err = i915_gem_object_pin_pages(obj); if (err) - goto put; + return err; __do_clflush(obj); i915_gem_object_unpin_pages(obj); -put: - i915_gem_object_put(obj); - return err; + return 0; } static void clflush_release(struct dma_fence_work *base) { struct clflush *clflush = container_of(base, typeof(*clflush), base); - if (clflush->obj) - i915_gem_object_put(clflush->obj); + i915_gem_object_put(clflush->obj); } static const struct dma_fence_work_ops clflush_ops = { diff --git a/drivers/gpu/drm/i915/i915_sw_fence_work.c b/drivers/gpu/drm/i915/i915_sw_fence_work.c index 8538ee7a521d..0c27f316ba83 100644 --- a/drivers/gpu/drm/i915/i915_sw_fence_work.c +++ b/drivers/gpu/drm/i915/i915_sw_fence_work.c @@ -37,6 +37,8 @@ fence_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state) break; case FENCE_FREE: + if (f->ops->release) + f->ops->release(f); dma_fence_put(&f->dma); break; } @@ -60,9 +62,6 @@ static void fence_release(struct dma_fence *fence) { struct dma_fence_work *f = container_of(fence, typeof(*f), dma); - if (f->ops->release) - f->ops->release(f); - i915_sw_fence_fini(&f->chain); BUILD_BUG_ON(offsetof(typeof(*f), dma)); diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 6794c742fbbf..878975b37a45 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -292,6 +292,7 @@ i915_vma_instance(struct drm_i915_gem_object *obj, struct i915_vma_work { struct dma_fence_work base; struct i915_vma *vma; + struct drm_i915_gem_object *pin; enum i915_cache_level cache_level; unsigned int flags; }; @@ -306,15 +307,21 @@ static int __vma_bind(struct dma_fence_work *work) if (err) atomic_or(I915_VMA_ERROR, &vma->flags); - if (vma->obj) - __i915_gem_object_unpin_pages(vma->obj); - return err; } +static void __vma_release(struct dma_fence_work *work) +{ + struct i915_vma_work *vw = container_of(work, typeof(*vw), base); + + if (vw->pin) + __i915_gem_object_unpin_pages(vw->pin); +} + static const struct dma_fence_work_ops bind_ops = { .name = "bind", .work = __vma_bind, + .release = __vma_release, }; struct i915_vma_work *i915_vma_work(void) @@ -395,8 +402,10 @@ int i915_vma_bind(struct i915_vma *vma, i915_active_set_exclusive(&vma->active, &work->base.dma); work->base.dma.error = 0; /* enable the queue_work() */ - if (vma->obj) + if (vma->obj) { __i915_gem_object_pin_pages(vma->obj); + work->pin = vma->obj; + } } else { GEM_BUG_ON((bind_flags & ~vma_flags) & vma->vm->bind_async_flags); ret = vma->ops->bind_vma(vma, cache_level, bind_flags);