blob: ffe08162ce8f8b8a4a134291a8262c92abf4b120 [file] [log] [blame]
From 6bb51f12925d46843995b4b871bea0baac158920 Mon Sep 17 00:00:00 2001
From: Chris Wilson <>
Date: Fri, 20 Sep 2019 13:18:21 +0100
Subject: [PATCH] drm/i915: Mark contents as dirty on a write fault
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
commit b925708f28c2b7a3a362d709bd7f77bc75c1daac upstream.
Since dropping the set-to-gtt-domain in commit a679f58d0510 ("drm/i915:
Flush pages on acquisition"), we no longer mark the contents as dirty on
a write fault. This has the issue of us then not marking the pages as
dirty on releasing the buffer, which means the contents are not written
out to the swap device (should we ever pick that buffer as a victim).
Notably, this is visible in the dumb buffer interface used for cursors.
Having updated the cursor contents via mmap, and swapped away, if the
shrinker should evict the old cursor, upon next reuse, the cursor would
be invisible.
E.g. echo 80 > /proc/sys/kernel/sysrq ; echo f > /proc/sysrq-trigger
Fixes: a679f58d0510 ("drm/i915: Flush pages on acquisition")
Signed-off-by: Chris Wilson <>
Cc: Matthew Auld <>
Cc: Ville Syrjälä <>
Cc: <> # v5.2+
Reviewed-by: Matthew Auld <>
(cherry picked from commit 5028851cdfdf78dc22eacbc44a0ab0b3f599ee4a)
Signed-off-by: Rodrigo Vivi <>
Signed-off-by: Paul Gortmaker <>
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ad01c92aaf74..3c4e41a837b5 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1908,7 +1908,11 @@ vm_fault_t i915_gem_fault(struct vm_fault *vmf)
list_add(&obj->userfault_link, &dev_priv->mm.userfault_list);
- i915_vma_set_ggtt_write(vma);
+ if (write) {
+ GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+ i915_vma_set_ggtt_write(vma);
+ obj->mm.dirty = true;
+ }