From patchwork Thu Aug 15 07:59:48 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chris Wilson X-Patchwork-Id: 11095169 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id CBEC01399 for ; Thu, 15 Aug 2019 08:00:07 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id C18E4271FD for ; Thu, 15 Aug 2019 08:00:07 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id B635E285A0; Thu, 15 Aug 2019 08:00:07 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.2 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_MED autolearn=ham version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id DB5FF2823E for ; Thu, 15 Aug 2019 08:00:04 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id C35586E0AC; Thu, 15 Aug 2019 08:00:02 +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 2AC416E0AC for ; Thu, 15 Aug 2019 08:00:01 +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 18147413-1500050 for multiple; Thu, 15 Aug 2019 08:59:55 +0100 From: Chris Wilson To: intel-gfx@lists.freedesktop.org Date: Thu, 15 Aug 2019 08:59:48 +0100 Message-Id: <20190815075948.9256-1-chris@chris-wilson.co.uk> X-Mailer: git-send-email 2.23.0.rc1 MIME-Version: 1.0 Subject: [Intel-gfx] [PATCH v2] drm/i915: Replace PIN_NONFAULT with calls to PIN_NOEVICT X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.23 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-Virus-Scanned: ClamAV using ClamSMTP When under severe stress for GTT mappable space, the LRU eviction model falls off a cliff. We spend all our time scanning the much large non-mappable vma searching for something within the mappable zone we can evict. Turn this on its head by only using the full vma if it is already pinned in the mappable zone or there is sufficient *free* space to accommodate it(prioritizing speedy reuse). If there is not, immediately fall back to using small chunks (tilerow for GTT mmap, single pages for pwrite/relocation) and using random eviction before doing a full search. Testcase: igt/gem_concurrent_blt References: https://bugs.freedesktop.org/show_bug.cgi?id=110848 Signed-off-by: Chris Wilson --- Keep nonblock around, it serves as a NOWARN flag for the full pin. --- drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 4 ++-- drivers/gpu/drm/i915/gem/i915_gem_mman.c | 8 ++++---- drivers/gpu/drm/i915/i915_gem.c | 8 ++++---- drivers/gpu/drm/i915/i915_gem_evict.c | 8 -------- drivers/gpu/drm/i915/i915_gem_gtt.c | 6 +++++- drivers/gpu/drm/i915/i915_gem_gtt.h | 6 +++--- 6 files changed, 18 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 91512cc6d7a6..669f8e2fffeb 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -1092,8 +1092,8 @@ static void *reloc_iomap(struct drm_i915_gem_object *obj, vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE | - PIN_NONBLOCK | - PIN_NONFAULT); + PIN_NONBLOCK /* NOWARN */ | + PIN_NOEVICT); if (IS_ERR(vma)) { memset(&cache->node, 0, sizeof(cache->node)); err = drm_mm_insert_node_in_range diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c index 1e7311493530..495ff681647b 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c @@ -267,15 +267,15 @@ vm_fault_t i915_gem_fault(struct vm_fault *vmf) /* Now pin it into the GTT as needed */ vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE | - PIN_NONBLOCK | - PIN_NONFAULT); + PIN_NONBLOCK /* NOWARN */ | + PIN_NOSEARCH); if (IS_ERR(vma)) { /* Use a partial view if it is bigger than available space */ struct i915_ggtt_view view = compute_partial_view(obj, page_offset, MIN_CHUNK_PAGES); unsigned int flags; - flags = PIN_MAPPABLE; + flags = PIN_MAPPABLE | PIN_NOSEARCH; if (view.type == I915_GGTT_VIEW_NORMAL) flags |= PIN_NONBLOCK; /* avoid warnings for pinned */ @@ -286,7 +286,7 @@ vm_fault_t i915_gem_fault(struct vm_fault *vmf) obj->frontbuffer_ggtt_origin = ORIGIN_CPU; vma = i915_gem_object_ggtt_pin(obj, &view, 0, 0, flags); - if (IS_ERR(vma) && !view.type) { + if (IS_ERR(vma)) { flags = PIN_MAPPABLE; view.type = I915_GGTT_VIEW_PARTIAL; vma = i915_gem_object_ggtt_pin(obj, &view, 0, 0, flags); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index b4c39a06fee5..c57094471232 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -343,8 +343,8 @@ i915_gem_gtt_pread(struct drm_i915_gem_object *obj, wakeref = intel_runtime_pm_get(&i915->runtime_pm); vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE | - PIN_NONFAULT | - PIN_NONBLOCK); + PIN_NONBLOCK /* NOWARN */ | + PIN_NOEVICT); if (!IS_ERR(vma)) { node.start = i915_ggtt_offset(vma); node.allocated = false; @@ -557,8 +557,8 @@ i915_gem_gtt_pwrite_fast(struct drm_i915_gem_object *obj, vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE | - PIN_NONFAULT | - PIN_NONBLOCK); + PIN_NONBLOCK /* NOWARN */ | + PIN_NOEVICT); if (!IS_ERR(vma)) { node.start = i915_ggtt_offset(vma); node.allocated = false; diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index 1796391bc598..52c86c6e0673 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -61,9 +61,6 @@ mark_free(struct drm_mm_scan *scan, if (i915_vma_is_pinned(vma)) return false; - if (flags & PIN_NONFAULT && i915_vma_has_userfault(vma)) - return false; - list_add(&vma->evict_link, unwind); return drm_mm_scan_add_block(scan, &vma->node); } @@ -330,11 +327,6 @@ int i915_gem_evict_for_node(struct i915_address_space *vm, break; } - if (flags & PIN_NONFAULT && i915_vma_has_userfault(vma)) { - ret = -ENOSPC; - break; - } - /* Overlap of objects in the same batch? */ if (i915_vma_is_pinned(vma)) { ret = -ENOSPC; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index e07c1ae971d7..a3f4a296d65f 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -3747,7 +3747,8 @@ int i915_gem_gtt_insert(struct i915_address_space *vm, if (flags & PIN_NOEVICT) return -ENOSPC; - /* No free space, pick a slot at random. + /* + * No free space, pick a slot at random. * * There is a pathological case here using a GTT shared between * mmap and GPU (i.e. ggtt/aliasing_ppgtt but not full-ppgtt): @@ -3775,6 +3776,9 @@ int i915_gem_gtt_insert(struct i915_address_space *vm, if (err != -ENOSPC) return err; + if (flags & PIN_NOSEARCH) + return -ENOSPC; + /* Randomly selected placement is pinned, do a search */ err = i915_gem_evict_something(vm, size, alignment, color, start, end, flags); diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index 51274483502e..b97a47fc7a68 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -600,9 +600,9 @@ int i915_gem_gtt_insert(struct i915_address_space *vm, u64 start, u64 end, unsigned int flags); /* Flags used by pin/bind&friends. */ -#define PIN_NONBLOCK BIT_ULL(0) -#define PIN_NONFAULT BIT_ULL(1) -#define PIN_NOEVICT BIT_ULL(2) +#define PIN_NOEVICT BIT_ULL(0) +#define PIN_NOSEARCH BIT_ULL(1) +#define PIN_NONBLOCK BIT_ULL(2) #define PIN_MAPPABLE BIT_ULL(3) #define PIN_ZONE_4G BIT_ULL(4) #define PIN_HIGH BIT_ULL(5)