From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756101Ab1GAJSZ (ORCPT ); Fri, 1 Jul 2011 05:18:25 -0400 Received: from adelie.canonical.com ([91.189.90.139]:33137 "EHLO adelie.canonical.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756010Ab1GAJSX (ORCPT ); Fri, 1 Jul 2011 05:18:23 -0400 From: Stefan Bader To: Chris Wilson Cc: Eric Anholt , Seth Forshee , Stefan Bader , linux-kernel@vger.kernel.org, stable@kernel.org Subject: [2.6.32+drm33-longterm] Patch "Subject: [PATCH 06/10] drm/i915: Maintain LRU order of inactive objects upon access by CPU (v2)" has been added to staging queue Date: Fri, 1 Jul 2011 10:18:19 +0100 Message-Id: <1309511899-3045-1-git-send-email-stefan.bader@canonical.com> X-Mailer: git-send-email 1.7.4.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This is a note to let you know that I have just added a patch titled Subject: [PATCH 06/10] drm/i915: Maintain LRU order of inactive objects upon access by CPU (v2) to the drm-next branch of the 2.6.32+drm33-longterm tree which can be found at http://git.kernel.org/?p=linux/kernel/git/smb/linux-2.6.32.y-drm33.z.git;a=shortlog;h=refs/heads/drm-next If you, or anyone else, feels it should not be added to the drm33-longterm tree, please reply to this email not later than 8 days after this email was sent. Thanks. -Stefan ------ >>From f8fd3ab5b8bf8f99ea13ebbecabd5c8e42c82948 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 17 Jun 2011 10:04:22 -0500 Subject: [PATCH 06/10] drm/i915: Maintain LRU order of inactive objects upon access by CPU (v2) BugLink: http://bugs.launchpad.net/bugs/599017 In order to reduce the penalty of fallbacks under memory pressure and to avoid a potential immediate ping-pong of evicting a mmaped buffer, we move the object to the tail of the inactive list when a page is freshly faulted or the object is moved into the CPU domain. We choose not to protect the CPU objects from casual eviction, preferring to keep the GPU active for as long as possible. v2: Daniel Vetter found a bug where I forgot that pinned objects are kept off the inactive list. Signed-off-by: Chris Wilson Signed-off-by: Eric Anholt (backported from commit 7d1c4804ae98cdee572d7d10d8a5deaa2e686285 upstream) Signed-off-by: Seth Forshee Signed-off-by: Stefan Bader --- drivers/gpu/drm/i915/i915_gem.c | 16 ++++++++++++++++ 1 files changed, 16 insertions(+), 0 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 2e4ff69..b3c7bd1 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -55,6 +55,14 @@ static int i915_gem_phys_pwrite(struct drm_device *dev, struct drm_gem_object *o static LIST_HEAD(shrink_list); static DEFINE_SPINLOCK(shrink_list_lock); +static inline bool +i915_gem_object_is_inactive(struct drm_i915_gem_object *obj_priv) +{ + return obj_priv->gtt_space && + !obj_priv->active && + obj_priv->pin_count == 0; +} + int i915_gem_do_init(struct drm_device *dev, unsigned long start, unsigned long end) { @@ -1068,6 +1076,11 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data, ret = i915_gem_object_set_to_cpu_domain(obj, write_domain != 0); } + + /* Maintain LRU order of "inactive" objects */ + if (ret == 0 && i915_gem_object_is_inactive(obj_priv)) + list_move_tail(&obj_priv->list, &dev_priv->mm.inactive_list); + drm_gem_object_unreference(obj); mutex_unlock(&dev->struct_mutex); return ret; @@ -1203,6 +1216,9 @@ int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) goto unlock; } + if (i915_gem_object_is_inactive(obj_priv)) + list_move_tail(&obj_priv->list, &dev_priv->mm.inactive_list); + pfn = ((dev->agp->base + obj_priv->gtt_offset) >> PAGE_SHIFT) + page_offset; -- 1.7.4.1