Skip to contentSkip to conversation

drm/i915: fix i915 running as dom0 under Xen

PATCH In review Juergen Gross 5 participants 10 messages Feb 2, 2017

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

PATCH 1/1 drm/i915: fix i915 running as dom0 under Xen

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.

---
 drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +10 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;
Signature · 6 lines
2.10.2

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw
Quoted 5 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Quoted 8 lines

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

Fixes 920cf4194954 ("drm/i915: Introduce an internal allocator for disposable private objects")
Cc Chris Wilson <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jani Nikula <[email protected]>
Cc <[email protected]> # v4.10-rc1+

We have a nice script for these :-)
-Daniel

Quoted 37 lines

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.

---
 drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +10 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;
Signature · 6 lines
2.10.2

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
#2 lore raw
Quoted 22 lines · Daniel Vetter
Quoted 5 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Quoted 8 lines

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

Fixes 920cf4194954 ("drm/i915: Introduce an internal allocator for disposable private objects")
Cc Chris Wilson <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jani Nikula <[email protected]>
Cc <[email protected]> # v4.10-rc1+
Quoted 1 line

We have a nice script for these :-)

Pffifle. 7453c549f5f648 allowed Xen to change it and silently conflicted with those that already used the previous limits, which didn't land in our tree until v4.10-rc3.
-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
#3 lore raw
In reply to Daniel Vetter
Quoted 22 lines · Daniel Vetter
Quoted 5 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Quoted 8 lines

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

Fixes 920cf4194954 ("drm/i915: Introduce an internal allocator for disposable private objects")
Cc Chris Wilson <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jani Nikula <[email protected]>
Cc <[email protected]> # v4.10-rc1+
Quoted 1 line

We have a nice script for these :-)

Pffifle. 7453c549f5f648 allowed Xen to change it and silently conflicted with those that already used the previous limits, which didn't land in our tree until v4.10-rc3.
-Chris

Signature · 6 lines
Chris Wilson, Intel Open Source Technology Centre

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xen.org/xen-devel
#4 lore raw
In reply to Juergen Gross
Quoted 5 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Quoted 8 lines

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

Fixes 920cf4194954 ("drm/i915: Introduce an internal allocator for disposable private objects")
Cc Chris Wilson <[email protected]>
Cc Tvrtko Ursulin <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Jani Nikula <[email protected]>
Cc <[email protected]> # v4.10-rc1+

We have a nice script for these :-)
-Daniel

Quoted 37 lines

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.

---
 drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +10 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;
Signature · 6 lines
2.10.2

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 8 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xen.org/xen-devel
Reviewed-by on v1 1/1
#5 lore raw
In reply to Juergen Gross
Quoted 34 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.

---
 drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +3 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+

drivers/gpu/drm/i915/i915_gem_internal.c i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj) · line 81

Quoted 10 lines
drivers/gpu/drm/i915/i915_gem_internal.c +7 -0
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;

Looks OK to me. We could bikeshed it to only use swiotlb_max_segment() which I think was the intention when that API was added but can leave that for the future.

Regards,

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

#6 lore raw
Quoted 44 lines · Tvrtko Ursulin
Quoted 42 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.
---
drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--

1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +10 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
Quoted 8 lines

drivers/gpu/drm/i915/i915_gem_internal.c i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj) · line 81

Quoted 2 lines
drivers/gpu/drm/i915/i915_gem_internal.c +0 -0
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;

Looks OK to me. We could bikeshed it to only use swiotlb_max_segment() which I think was the intention when that API was added but can leave that for the future.

Pushed, I imagine this has been added to the list of sg cleanups you have :)
-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
#7 lore raw
In reply to Tvrtko Ursulin
Quoted 44 lines · Tvrtko Ursulin
Quoted 42 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.
---
drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--

1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +10 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
Quoted 8 lines

drivers/gpu/drm/i915/i915_gem_internal.c i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj) · line 81

Quoted 2 lines
drivers/gpu/drm/i915/i915_gem_internal.c +0 -0
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;

Looks OK to me. We could bikeshed it to only use swiotlb_max_segment() which I think was the intention when that API was added but can leave that for the future.

Pushed, I imagine this has been added to the list of sg cleanups you have :)
-Chris

Signature · 6 lines
Chris Wilson, Intel Open Source Technology Centre

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xen.org/xen-devel
Reviewed-by on v1 1/1
#8 lore raw
In reply to Juergen Gross
Quoted 34 lines · Juergen Gross

Commit 920cf4194954ec ("drm/i915: Introduce an internal allocator for disposable private objects") introduced a regression for the kernel running as Xen dom0: when switching to graphics mode a GPU HANG occurred.

Reason seems to be a missing adaption similar to that done in commit 7453c549f5f648 ("swiotlb: Export swiotlb_max_segment to users") to i915_gem_object_get_pages_internal().

So limit the maximum page order to be used according to the maximum swiotlb segment size instead to the complete swiotlb size.

---
Please consider for 4.10 as otherwise 4.10 will be unusable as Xen dom0 with i915 graphics.

---
 drivers/gpu/drm/i915/i915_gem_internal.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/i915_gem_internal.c +3 -2
index 4b3ff3e..d09c749 100644
@@ -66,8 +66,16 @@ i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
 
 	max_order = MAX_ORDER;
 #ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) /* minimum max swiotlb size is IO_TLB_SEGSIZE */
-		max_order = min(max_order, ilog2(IO_TLB_SEGPAGES));
+	if (swiotlb_nr_tbl()) {
+		unsigned int max_segment;
+

drivers/gpu/drm/i915/i915_gem_internal.c i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj) · line 81

Quoted 10 lines
drivers/gpu/drm/i915/i915_gem_internal.c +7 -0
+		max_segment = swiotlb_max_segment();
+		if (max_segment) {
+			max_segment = max_t(unsigned int, max_segment,
+					    PAGE_SIZE) >> PAGE_SHIFT;
+			max_order = min(max_order, ilog2(max_segment));
+		}
+	}
 #endif
 
 	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;

Looks OK to me. We could bikeshed it to only use swiotlb_max_segment() which I think was the intention when that API was added but can leave that for the future.

Regards,

Tvrtko

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xen.org/xen-devel

#9 lore raw

✗ Fi.CI.BAT: failure for drm/i915: fix i915 running as dom0 under Xen

In reply to Juergen Gross
== Series Details ==

== Series Details ==

Series: drm/i915: fix i915 running as dom0 under Xen
URL : https://patchwork.freedesktop.org/series/18979/
State : failure

== Summary ==

Series 18979v1 drm/i915: fix i915 running as dom0 under Xen
https://patchwork.freedesktop.org/api/1.0/series/18979/revisions/1/mbox/

Test gem_exec_flush:

        Subgroup basic-wb-ro-default:
                pass       -> INCOMPLETE (fi-byt-j1900)
Test kms_pipe_crc_basic:
        Subgroup nonblocking-crc-pipe-c-frame-sequence:
                fail       -> PASS       (fi-skl-6770hq)
fi-bdw-5557u     total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-bsw-n3050     total:247  pass:208  dwarn:0   dfail:0   fail:0   skip:39 
fi-bxt-j4205     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-bxt-t5700     total:78   pass:65   dwarn:0   dfail:0   fail:0   skip:12 
fi-byt-j1900     total:56   pass:46   dwarn:0   dfail:0   fail:0   skip:9  
fi-byt-n2820     total:247  pass:216  dwarn:0   dfail:0   fail:0   skip:31 
fi-hsw-4770      total:247  pass:228  dwarn:0   dfail:0   fail:0   skip:19 
fi-hsw-4770r     total:247  pass:228  dwarn:0   dfail:0   fail:0   skip:19 
fi-ivb-3520m     total:247  pass:226  dwarn:0   dfail:0   fail:0   skip:21 
fi-ivb-3770      total:247  pass:226  dwarn:0   dfail:0   fail:0   skip:21 
fi-kbl-7500u     total:247  pass:224  dwarn:0   dfail:0   fail:2   skip:21 
fi-skl-6260u     total:247  pass:234  dwarn:0   dfail:0   fail:0   skip:13 
fi-skl-6700hq    total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-skl-6700k     total:247  pass:222  dwarn:4   dfail:0   fail:0   skip:21 
fi-skl-6770hq    total:247  pass:234  dwarn:0   dfail:0   fail:0   skip:13 
fi-snb-2520m     total:247  pass:216  dwarn:0   dfail:0   fail:0   skip:31 
fi-snb-2600      total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 

ce4c6eb9b0d270df56813af310cd918298ac7ba2 drm-tip: 2017y-02m-02d-10h-12m-30s UTC integration manifest a372572 drm/i915: fix i915 running as dom0 under Xen

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3673/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx