From patchwork Sun Jun 10 19:43:19 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Chris Wilson X-Patchwork-Id: 10456729 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 8463F601F7 for ; Sun, 10 Jun 2018 19:44:07 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 74D6B26E79 for ; Sun, 10 Jun 2018 19:44:07 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 6989E274D0; Sun, 10 Jun 2018 19:44:07 +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 E935126E79 for ; Sun, 10 Jun 2018 19:44:06 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 5F74A6E1FF; Sun, 10 Jun 2018 19:44:06 +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 90BE16E1CD for ; Sun, 10 Jun 2018 19:44:04 +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 12004298-1500050 for multiple; Sun, 10 Jun 2018 20:43:57 +0100 Received: by haswell.alporthouse.com (sSMTP sendmail emulation); Sun, 10 Jun 2018 20:43:57 +0100 From: Chris Wilson To: intel-gfx@lists.freedesktop.org Date: Sun, 10 Jun 2018 20:43:19 +0100 Message-Id: <20180610194325.13467-12-chris@chris-wilson.co.uk> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20180610194325.13467-1-chris@chris-wilson.co.uk> References: <20180610194325.13467-1-chris@chris-wilson.co.uk> X-Originating-IP: 78.156.65.138 X-Country: code=GB country="United Kingdom" ip=78.156.65.138 Subject: [Intel-gfx] [PATCH 11/17] drm/i915/gtt: Lazily allocate page directories for gen7 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: , MIME-Version: 1.0 Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Virus-Scanned: ClamAV using ClamSMTP As we were only supporting aliasing_ppgtt on gen7 for some time, we saved a few checks by preallocating the page directories on creation. However, since we need 2MiB of page directories for each ppgtt, to support arbitrary numbers of user contexts, we need to be more prudent in our allocations, and defer the page allocation until it is used. We don't recover unused pages yet as we found that doing so on the fly (i.e. altering TLB entries) would confuse the GPU. Signed-off-by: Chris Wilson Cc: Joonas Lahtinen Cc: Mika Kuoppala Cc: Matthew Auld Reviewed-by: Matthew Auld Reviewed-by: Joonas Lahtinen --- drivers/gpu/drm/i915/i915_gem_gtt.c | 67 +++++++++++------------------ 1 file changed, 26 insertions(+), 41 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 317f27a9d78e..3dc17793afac 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -190,11 +190,19 @@ int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, return 1; } -static int gen6_ppgtt_bind_vma(struct i915_vma *vma, - enum i915_cache_level cache_level, - u32 unused) +static int ppgtt_bind_vma(struct i915_vma *vma, + enum i915_cache_level cache_level, + u32 unused) { u32 pte_flags; + int err; + + if (!(vma->flags & I915_VMA_LOCAL_BIND)) { + err = vma->vm->allocate_va_range(vma->vm, + vma->node.start, vma->size); + if (err) + return err; + } /* Currently applicable only to VLV */ pte_flags = 0; @@ -206,22 +214,6 @@ static int gen6_ppgtt_bind_vma(struct i915_vma *vma, return 0; } -static int gen8_ppgtt_bind_vma(struct i915_vma *vma, - enum i915_cache_level cache_level, - u32 unused) -{ - int ret; - - if (!(vma->flags & I915_VMA_LOCAL_BIND)) { - ret = vma->vm->allocate_va_range(vma->vm, - vma->node.start, vma->size); - if (ret) - return ret; - } - - return gen6_ppgtt_bind_vma(vma, cache_level, unused); -} - static void ppgtt_unbind_vma(struct i915_vma *vma) { vma->vm->clear_range(vma->vm, vma->node.start, vma->size); @@ -1622,7 +1614,7 @@ static struct i915_hw_ppgtt *gen8_ppgtt_create(struct drm_i915_private *i915) ppgtt->vm.cleanup = gen8_ppgtt_cleanup; ppgtt->debug_dump = gen8_dump_ppgtt; - ppgtt->vm.vma_ops.bind_vma = gen8_ppgtt_bind_vma; + ppgtt->vm.vma_ops.bind_vma = ppgtt_bind_vma; ppgtt->vm.vma_ops.unbind_vma = ppgtt_unbind_vma; ppgtt->vm.vma_ops.set_pages = ppgtt_set_pages; ppgtt->vm.vma_ops.clear_pages = clear_pages; @@ -1776,7 +1768,8 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm, num_entries -= end - pte; - /* Note that the hw doesn't support removing PDE on the fly + /* + * Note that the hw doesn't support removing PDE on the fly * (they are cached inside the context with no means to * invalidate the cache), so we can only reset the PTE * entries back to scratch. @@ -2046,12 +2039,13 @@ static struct i915_hw_ppgtt *gen6_ppgtt_create(struct drm_i915_private *i915) ppgtt->base.vm.total = I915_PDES * GEN6_PTES * PAGE_SIZE; + ppgtt->base.vm.allocate_va_range = gen6_alloc_va_range; ppgtt->base.vm.clear_range = gen6_ppgtt_clear_range; ppgtt->base.vm.insert_entries = gen6_ppgtt_insert_entries; ppgtt->base.vm.cleanup = gen6_ppgtt_cleanup; ppgtt->base.debug_dump = gen6_dump_ppgtt; - ppgtt->base.vm.vma_ops.bind_vma = gen6_ppgtt_bind_vma; + ppgtt->base.vm.vma_ops.bind_vma = ppgtt_bind_vma; ppgtt->base.vm.vma_ops.unbind_vma = ppgtt_unbind_vma; ppgtt->base.vm.vma_ops.set_pages = ppgtt_set_pages; ppgtt->base.vm.vma_ops.clear_pages = clear_pages; @@ -2068,14 +2062,8 @@ static struct i915_hw_ppgtt *gen6_ppgtt_create(struct drm_i915_private *i915) goto err_scratch; } - err = gen6_alloc_va_range(&ppgtt->base.vm, 0, ppgtt->base.vm.total); - if (err) - goto err_vma; - return &ppgtt->base; -err_vma: - i915_vma_destroy(ppgtt->vma); err_scratch: gen6_ppgtt_free_scratch(&ppgtt->base.vm); err_free: @@ -2671,8 +2659,7 @@ static int aliasing_gtt_bind_vma(struct i915_vma *vma, if (flags & I915_VMA_LOCAL_BIND) { struct i915_hw_ppgtt *appgtt = i915->mm.aliasing_ppgtt; - if (!(vma->flags & I915_VMA_LOCAL_BIND) && - appgtt->vm.allocate_va_range) { + if (!(vma->flags & I915_VMA_LOCAL_BIND)) { ret = appgtt->vm.allocate_va_range(&appgtt->vm, vma->node.start, vma->size); @@ -2776,17 +2763,15 @@ int i915_gem_init_aliasing_ppgtt(struct drm_i915_private *i915) goto err_ppgtt; } - if (ppgtt->vm.allocate_va_range) { - /* Note we only pre-allocate as far as the end of the global - * GTT. On 48b / 4-level page-tables, the difference is very, - * very significant! We have to preallocate as GVT/vgpu does - * not like the page directory disappearing. - */ - err = ppgtt->vm.allocate_va_range(&ppgtt->vm, - 0, ggtt->vm.total); - if (err) - goto err_ppgtt; - } + /* + * Note we only pre-allocate as far as the end of the global + * GTT. On 48b / 4-level page-tables, the difference is very, + * very significant! We have to preallocate as GVT/vgpu does + * not like the page directory disappearing. + */ + err = ppgtt->vm.allocate_va_range(&ppgtt->vm, 0, ggtt->vm.total); + if (err) + goto err_ppgtt; i915->mm.aliasing_ppgtt = ppgtt;