From patchwork Thu Sep 10 09:51:54 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chris Wilson X-Patchwork-Id: 7153111 Return-Path: X-Original-To: patchwork-intel-gfx@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id F213C9F314 for ; Thu, 10 Sep 2015 09:52:15 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id A7AA920921 for ; Thu, 10 Sep 2015 09:52:14 +0000 (UTC) Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) by mail.kernel.org (Postfix) with ESMTP id 273CB20927 for ; Thu, 10 Sep 2015 09:52:12 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 95EBC6EB03; Thu, 10 Sep 2015 02:52:11 -0700 (PDT) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from fireflyinternet.com (mail.fireflyinternet.com [87.106.93.118]) by gabe.freedesktop.org (Postfix) with ESMTP id 4B3756EACF for ; Thu, 10 Sep 2015 02:52:09 -0700 (PDT) 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 44995223-1500048 for multiple; Thu, 10 Sep 2015 10:52:16 +0100 Received: by haswell.alporthouse.com (sSMTP sendmail emulation); Thu, 10 Sep 2015 10:51:58 +0100 From: Chris Wilson To: intel-gfx@lists.freedesktop.org Date: Thu, 10 Sep 2015 10:51:54 +0100 Message-Id: <1441878715-24948-2-git-send-email-chris@chris-wilson.co.uk> X-Mailer: git-send-email 2.5.1 In-Reply-To: <1441878715-24948-1-git-send-email-chris@chris-wilson.co.uk> References: <1441878715-24948-1-git-send-email-chris@chris-wilson.co.uk> MIME-Version: 1.0 X-Originating-IP: 78.156.65.138 X-Country: code=GB country="United Kingdom" ip=78.156.65.138 Cc: stable@vger.kernel.org Subject: [Intel-gfx] [PATCH v4 2/3] drm/i915: Fix userptr deadlock with aliased GTT mmappings X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.18 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" X-Spam-Status: No, score=-4.2 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_MED, T_RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Micha? Winiarski found a really evil way to trigger a struct_mutex deadlock with userptr. He found that if he allocated a userptr bo and then GTT mmaped another bo, or even itself, at the same address as the userptr using MAP_FIXED, he could then cause a deadlock any time we then had to invalidate the GTT mmappings (so at will). Tvrtko then found by repeatedly allocating GTT mmappings he could alias with an old userptr mmap and also trigger the deadlock. To counter act the deadlock, we make the observation that we only need to take the struct_mutex if the object has any pages to revoke, and that before userspace can alias with the userptr address space, it must have invalidated the userptr->pages. Thus if we can check for those pages outside of the struct_mutex, we can avoid the deadlock. To do so we introduce a separate flag for userptr objects that we can inspect from the mmu-notifier underneath its spinlock. The patch makes one eye-catching change. That is the removal serial=0 after detecting a to-be-freed object inside the invalidate walker. I felt setting serial=0 was a questionable pessimisation: it denies us the chance to reuse the current iterator for the next loop (before it is freed) and being explicit makes the reader question the validity of the locking (since the object-free race could occur elsewhere). The serialisation of the iterator is through the spinlock, if the object is freed before the next loop then the notifier.serial will be incremented and we start the walk from the beginning as we detect the invalid cache. To try and tame the error paths and interactions with the userptr->active flag, we have to do a fair amount of rearranging of get_pages_userptr(). v2: Grammar fixes v3: Reorder set-active so that it is only set when obj->pages is set (and so needs cancellation). Only the order of setting obj->pages and the active-flag is crucial. Calling gup after invalidate-range begin means the userptr sees the new set of backing storage (and so will not need to invalidate its new pages), but we have to be careful not to set the active-flag prior to successfully establishing obj->pages. v4: Take the active->flag early so we know in the mmu-notifier when we have to cancel a pending gup-worker. v5: Rearrange the error path so that is not so convoluted Reported-by: Micha? Winiarski Testcase: igt/gem_userptr_blits/map-fixed* Signed-off-by: Chris Wilson Cc: Micha? Winiarski Cc: Tvrtko Ursulin Cc: stable@vger.kernel.org --- drivers/gpu/drm/i915/i915_gem_userptr.c | 175 ++++++++++++++++++++------------ 1 file changed, 109 insertions(+), 66 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c index 800a5394aa1e..505c608ade7e 100644 --- a/drivers/gpu/drm/i915/i915_gem_userptr.c +++ b/drivers/gpu/drm/i915/i915_gem_userptr.c @@ -59,6 +59,7 @@ struct i915_mmu_object { struct interval_tree_node it; struct list_head link; struct drm_i915_gem_object *obj; + bool active; bool is_linear; }; @@ -114,7 +115,8 @@ restart: obj = mo->obj; - if (!kref_get_unless_zero(&obj->base.refcount)) + if (!mo->active || + !kref_get_unless_zero(&obj->base.refcount)) continue; spin_unlock(&mn->lock); @@ -151,7 +153,8 @@ static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, else it = interval_tree_iter_first(&mn->objects, start, end); if (it != NULL) { - obj = container_of(it, struct i915_mmu_object, it)->obj; + struct i915_mmu_object *mo = + container_of(it, struct i915_mmu_object, it); /* The mmu_object is released late when destroying the * GEM object so it is entirely possible to gain a @@ -160,11 +163,9 @@ static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, * the struct_mutex - and consequently use it after it * is freed and then double free it. */ - if (!kref_get_unless_zero(&obj->base.refcount)) { - spin_unlock(&mn->lock); - serial = 0; - continue; - } + if (mo->active && + kref_get_unless_zero(&mo->obj->base.refcount)) + obj = mo->obj; serial = mn->serial; } @@ -566,6 +567,30 @@ __i915_gem_userptr_set_pages(struct drm_i915_gem_object *obj, } static void +__i915_gem_userptr_set_active(struct drm_i915_gem_object *obj, + bool value) +{ + /* During mm_invalidate_range we need to cancel any userptr that + * overlaps the range being invalidated. Doing so requires the + * struct_mutex, and that risks recursion. In order to cause + * recursion, the user must alias the userptr address space with + * a GTT mmapping (possible with a MAP_FIXED) - then when we have + * to invalidate that mmaping, mm_invalidate_range is called with + * the userptr address *and* the struct_mutex held. To prevent that + * we set a flag under the i915_mmu_notifier spinlock to indicate + * whether this object is valid. + */ +#if defined(CONFIG_MMU_NOTIFIER) + if (obj->userptr.mmu_object == NULL) + return; + + spin_lock(&obj->userptr.mmu_object->mn->lock); + obj->userptr.mmu_object->active = value; + spin_unlock(&obj->userptr.mmu_object->mn->lock); +#endif +} + +static void __i915_gem_userptr_get_pages_worker(struct work_struct *_work) { struct get_pages_work *work = container_of(_work, typeof(*work), work); @@ -613,6 +638,8 @@ __i915_gem_userptr_get_pages_worker(struct work_struct *_work) } } obj->userptr.work = ERR_PTR(ret); + if (ret) + __i915_gem_userptr_set_active(obj, false); } obj->userptr.workers--; @@ -627,11 +654,60 @@ __i915_gem_userptr_get_pages_worker(struct work_struct *_work) } static int +__i915_gem_userptr_get_pages_schedule(struct drm_i915_gem_object *obj, + bool *active) +{ + struct get_pages_work *work; + + /* Spawn a worker so that we can acquire the + * user pages without holding our mutex. Access + * to the user pages requires mmap_sem, and we have + * a strict lock ordering of mmap_sem, struct_mutex - + * we already hold struct_mutex here and so cannot + * call gup without encountering a lock inversion. + * + * Userspace will keep on repeating the operation + * (thanks to EAGAIN) until either we hit the fast + * path or the worker completes. If the worker is + * cancelled or superseded, the task is still run + * but the results ignored. (This leads to + * complications that we may have a stray object + * refcount that we need to be wary of when + * checking for existing objects during creation.) + * If the worker encounters an error, it reports + * that error back to this function through + * obj->userptr.work = ERR_PTR. + */ + if (obj->userptr.workers >= I915_GEM_USERPTR_MAX_WORKERS) + return -EAGAIN; + + work = kmalloc(sizeof(*work), GFP_KERNEL); + if (work == NULL) + return -ENOMEM; + + obj->userptr.work = &work->work; + obj->userptr.workers++; + + work->obj = obj; + drm_gem_object_reference(&obj->base); + + work->task = current; + get_task_struct(work->task); + + INIT_WORK(&work->work, __i915_gem_userptr_get_pages_worker); + schedule_work(&work->work); + + *active = true; + return -EAGAIN; +} + +static int i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj) { const int num_pages = obj->base.size >> PAGE_SHIFT; struct page **pvec; int pinned, ret; + bool active; /* If userspace should engineer that these pages are replaced in * the vma between us binding this page into the GTT and completion @@ -649,6 +725,18 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj) * to the vma (discard or cloning) which should prevent the more * egregious cases from causing harm. */ + if (IS_ERR(obj->userptr.work)) { + /* active flag will have been dropped already by the worker */ + ret = PTR_ERR(obj->userptr.work); + obj->userptr.work = NULL; + return ret; + } + if (obj->userptr.work) + /* active flag should still be held for the pending work */ + return -EAGAIN; + + /* Let the mmu-notifier know that we have begun and need cancellation */ + __i915_gem_userptr_set_active(obj, true); pvec = NULL; pinned = 0; @@ -657,73 +745,27 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj) GFP_TEMPORARY | __GFP_NOWARN | __GFP_NORETRY); if (pvec == NULL) { pvec = drm_malloc_ab(num_pages, sizeof(struct page *)); - if (pvec == NULL) + if (pvec == NULL) { + __i915_gem_userptr_set_active(obj, false); return -ENOMEM; + } } pinned = __get_user_pages_fast(obj->userptr.ptr, num_pages, !obj->userptr.read_only, pvec); } - if (pinned < num_pages) { - if (pinned < 0) { - ret = pinned; - pinned = 0; - } else { - /* Spawn a worker so that we can acquire the - * user pages without holding our mutex. Access - * to the user pages requires mmap_sem, and we have - * a strict lock ordering of mmap_sem, struct_mutex - - * we already hold struct_mutex here and so cannot - * call gup without encountering a lock inversion. - * - * Userspace will keep on repeating the operation - * (thanks to EAGAIN) until either we hit the fast - * path or the worker completes. If the worker is - * cancelled or superseded, the task is still run - * but the results ignored. (This leads to - * complications that we may have a stray object - * refcount that we need to be wary of when - * checking for existing objects during creation.) - * If the worker encounters an error, it reports - * that error back to this function through - * obj->userptr.work = ERR_PTR. - */ - ret = -EAGAIN; - if (obj->userptr.work == NULL && - obj->userptr.workers < I915_GEM_USERPTR_MAX_WORKERS) { - struct get_pages_work *work; - - work = kmalloc(sizeof(*work), GFP_KERNEL); - if (work != NULL) { - obj->userptr.work = &work->work; - obj->userptr.workers++; - - work->obj = obj; - drm_gem_object_reference(&obj->base); - - work->task = current; - get_task_struct(work->task); - - INIT_WORK(&work->work, __i915_gem_userptr_get_pages_worker); - schedule_work(&work->work); - } else - ret = -ENOMEM; - } else { - if (IS_ERR(obj->userptr.work)) { - ret = PTR_ERR(obj->userptr.work); - obj->userptr.work = NULL; - } - } - } - } else { + + active = false; + if (pinned < 0) + ret = pinned; + else if (pinned < num_pages) + ret = __i915_gem_userptr_get_pages_schedule(obj, &active); + else ret = __i915_gem_userptr_set_pages(obj, pvec, num_pages); - if (ret == 0) { - obj->userptr.work = NULL; - pinned = 0; - } + if (ret) { + __i915_gem_userptr_set_active(obj, active); + release_pages(pvec, pinned, 0); } - - release_pages(pvec, pinned, 0); drm_free_large(pvec); return ret; } @@ -734,6 +776,7 @@ i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj) struct sg_page_iter sg_iter; BUG_ON(obj->userptr.work != NULL); + __i915_gem_userptr_set_active(obj, false); if (obj->madv != I915_MADV_WILLNEED) obj->dirty = 0;