Skip to contentSkip to conversation

drm/i915: add sanity check for partial view creation

PATCH In review Matthew Auld 6 participants 7 messages Feb 29, 2016 – Mar 2, 2016

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: add sanity check for partial view creation

When binding pages for a partial view we should check that the offset + size is valid relative to the size of the gem object.

Cc Joonas Lahtinen <[email protected]>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/i915/i915_gem_gtt.c +4 -0
index 49e4f26..a477bb2 100644
@@ -3500,6 +3500,10 @@ intel_partial_pages(const struct i915_ggtt_view *view,
 	struct sg_page_iter obj_sg_iter;
 	int ret = -ENOMEM;
 
+	if (view->params.partial.offset + view->params.partial.size >
+	    obj->pages->nents)
+		return ERR_PTR(-EINVAL);
+
 	st = kmalloc(sizeof(*st), GFP_KERNEL);
 	if (!st)
 		goto err_st_alloc;
Signature · 6 lines
2.4.3

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

✗ Fi.CI.BAT: warning for drm/i915: add sanity check for partial view creation

== Series Details ==

== Series Details ==

Series: drm/i915: add sanity check for partial view creation
URL : https://patchwork.freedesktop.org/series/3926/
State : warning

== Summary ==

Series 3926v1 drm/i915: add sanity check for partial view creation
http://patchwork.freedesktop.org/api/1.0/series/3926/revisions/1/mbox/

Test kms_flip:

        Subgroup basic-flip-vs-dpms:
                pass       -> DMESG-WARN (hsw-brixbox)
        Subgroup basic-flip-vs-wf_vblank:
                dmesg-warn -> PASS       (hsw-brixbox)
Test kms_pipe_crc_basic:
        Subgroup nonblocking-crc-pipe-b:
                dmesg-warn -> PASS       (snb-x220t)
        Subgroup read-crc-pipe-b:
                pass       -> DMESG-WARN (hsw-gt2)
        Subgroup suspend-read-crc-pipe-c:
                pass       -> DMESG-WARN (bsw-nuc-2)
Test pm_rpm:
        Subgroup basic-pci-d3-state:
                fail       -> DMESG-FAIL (snb-x220t)
                pass       -> DMESG-WARN (hsw-gt2)
                dmesg-warn -> PASS       (byt-nuc)
bdw-nuci7        total:166  pass:155  dwarn:0   dfail:0   fail:0   skip:11 
bdw-ultra        total:169  pass:155  dwarn:0   dfail:0   fail:0   skip:14 
bsw-nuc-2        total:169  pass:137  dwarn:1   dfail:0   fail:1   skip:30 
byt-nuc          total:169  pass:144  dwarn:0   dfail:0   fail:0   skip:25 
hsw-brixbox      total:169  pass:153  dwarn:1   dfail:0   fail:0   skip:15 
hsw-gt2          total:169  pass:156  dwarn:2   dfail:1   fail:0   skip:10 
ilk-hp8440p      total:169  pass:118  dwarn:0   dfail:0   fail:1   skip:50 
ivb-t430s        total:169  pass:153  dwarn:0   dfail:0   fail:1   skip:15 
snb-dellxps      total:169  pass:143  dwarn:2   dfail:0   fail:1   skip:23 
snb-x220t        total:169  pass:144  dwarn:1   dfail:1   fail:1   skip:22 

Results at /archive/results/CI_IGT_test/Patchwork_1495/

a1c053436ca45c5860f842e294ad29d6055bd43b drm-intel-nightly: 2016y-02m-29d-16h-58m-26s UTC integration manifest 2cdbd535042074b7b79b7db1ae18cb632ffd823c drm/i915: add sanity check for partial view creation

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#2 lore raw
In reply to Matthew Auld
Quoted 14 lines · Matthew Auld

When binding pages for a partial view we should check that the offset + size is valid relative to the size of the gem object.

Cc Joonas Lahtinen <[email protected]>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/i915/i915_gem_gtt.c +0 -0
index 49e4f26..a477bb2 100644

drivers/gpu/drm/i915/i915_gem_gtt.c intel_partial_pages(const struct i915_ggtt_view *view, · line 3505

Quoted 6 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +3 -0
@@ -3500,6 +3500,10 @@ intel_partial_pages(const struct i915_ggtt_view *view,
 	struct sg_page_iter obj_sg_iter;
 	int ret = -ENOMEM;
 
+	if (view->params.partial.offset + view->params.partial.size >
+	    obj->pages->nents)
+		return ERR_PTR(-EINVAL);

It seems to me that if we hit this, there must a bug somewhere higher up.

Quoted 11 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +1 -0
+
 	st = kmalloc(sizeof(*st), GFP_KERNEL);
 	if (!st)
 		goto err_st_alloc;
Signature · 6 lines
2.4.3

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer Reviewed-by on v1 1/1
#4 lore raw
In reply to Matthew Auld
Quoted 4 lines · Matthew Auld

When binding pages for a partial view we should check that the offset + size is valid relative to the size of the gem object.

Cc Joonas Lahtinen <[email protected]>
Quoted 20 lines

---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++++

 1 file changed, 4 insertions(+)
drivers/gpu/drm/i915/i915_gem_gtt.c +0 -0
index 49e4f26..a477bb2 100644
@@ -3500,6 +3500,10 @@ intel_partial_pages(const struct i915_ggtt_view *view,
  	struct sg_page_iter obj_sg_iter;
  	int ret = -ENOMEM;

drivers/gpu/drm/i915/i915_gem_gtt.c +4 -0
+	if (view->params.partial.offset + view->params.partial.size >
+	    obj->pages->nents)
+		return ERR_PTR(-EINVAL);
+

  st = kmalloc(sizeof(*st), GFP_KERNEL);
  if (!st)
  goto err_st_alloc;

Signature · 7 lines
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#6 lore raw
In reply to Matthew Auld
Quoted 14 lines · Matthew Auld

When binding pages for a partial view we should check that the offset + size is valid relative to the size of the gem object.

Cc Joonas Lahtinen <[email protected]>
---
  drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++++
  1 file changed, 4 insertions(+)
drivers/gpu/drm/i915/i915_gem_gtt.c +0 -0
index 49e4f26..a477bb2 100644

drivers/gpu/drm/i915/i915_gem_gtt.c intel_partial_pages(const struct i915_ggtt_view *view, · line 3506

Quoted 7 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +4 -0
@@ -3500,6 +3500,10 @@ intel_partial_pages(const struct i915_ggtt_view *view,
  	struct sg_page_iter obj_sg_iter;
  	int ret = -ENOMEM;
 
+	if (view->params.partial.offset + view->params.partial.size >
+	    obj->pages->nents)
+		return ERR_PTR(-EINVAL);
+

obj->pages->nents is not guaranteed to be equal to number of pages but can be less than due sg entry coalescing.

I suggest replacing with a check against "obj->base.size >> PAGE_SHIFT".

drivers/gpu/drm/i915/i915_gem_gtt.c intel_partial_pages(const struct i915_ggtt_view *view, · line 3509

Quoted 4 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +0 -0
  	st = kmalloc(sizeof(*st), GFP_KERNEL);
  	if (!st)
  		goto err_st_alloc;

Regards,

Tvrtko
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#5 lore raw
In reply to Joonas Lahtinen
Quoted 13 lines · Joonas Lahtinen
Quoted 4 lines · Matthew Auld

When binding pages for a partial view we should check that the offset + size is valid relative to the size of the gem object.

Cc Joonas Lahtinen <[email protected]>
Quoted 5 lines

---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++++

 1 file changed, 4 insertions(+)
Quoted 11 lines

drivers/gpu/drm/i915/i915_gem_gtt.c intel_partial_pages(const struct i915_ggtt_view *view, · line 3505

Quoted 11 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +0 -0
index 49e4f26..a477bb2 100644
@@ -3500,6 +3500,10 @@ intel_partial_pages(const struct i915_ggtt_view *view,
  	struct sg_page_iter obj_sg_iter;
  	int ret = -ENOMEM;

drivers/gpu/drm/i915/i915_gem_gtt.c +3 -0
+	if (view->params.partial.offset + view->params.partial.size >
+	    obj->pages->nents)
+		return ERR_PTR(-EINVAL);

Wrong. Tell me again what nents has to do with the object size? -Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#3 lore raw
In reply to Ville Syrjälä
Quoted 11 lines · Ville Syrjälä
Quoted 10 lines · Matthew Auld

When binding pages for a partial view we should check that the offset + size is valid relative to the size of the gem object.

Cc Joonas Lahtinen <[email protected]>

---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 4 ++++

 1 file changed, 4 insertions(+)
Quoted 14 lines

drivers/gpu/drm/i915/i915_gem_gtt.c intel_partial_pages(const struct i915_ggtt_view *view, · line 3505

Quoted 11 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +0 -0
index 49e4f26..a477bb2 100644
@@ -3500,6 +3500,10 @@ intel_partial_pages(const struct i915_ggtt_view *view,
  	struct sg_page_iter obj_sg_iter;
  	int ret = -ENOMEM;

drivers/gpu/drm/i915/i915_gem_gtt.c +3 -0
+	if (view->params.partial.offset + view->params.partial.size >
+	    obj->pages->nents)
+		return ERR_PTR(-EINVAL);

It seems to me that if we hit this, there must a bug somewhere higher up.

Currently yes. This is in preparation of the more widespread support for partial views and was chosen as a good get-to-know-GEM-code candidate.

Regards, Joonas

Quoted 12 lines
Quoted 12 lines

+
  st = kmalloc(sizeof(*st), GFP_KERNEL);
  if (!st)
  goto err_st_alloc;
--
2.4.3

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 7 lines
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx