[CI 1/6] drm/i915: Mark all non-vma being inserted into the address spaces

[CI 1/6] drm/i915: Mark all non-vma being inserted into the address spaces

[Chris Wilson]

We need to distinguish between full i915_vma structs and simple
drm_mm_nodes when considering eviction (i.e. we must be careful not to
treat a mere drm_mm_node as a much larger i915_vma causing memory
corruption, if we are lucky). To do this, color these not-a-vma with -1
(I915_COLOR_UNEVICTABLE).

v2...v200: New name for -1.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/gvt/aperture_gm.c     |  7 +++++--
 drivers/gpu/drm/i915/i915_drv.h            |  2 ++
 drivers/gpu/drm/i915/i915_gem.c            |  3 ++-
 drivers/gpu/drm/i915/i915_gem_execbuffer.c |  2 +-
 drivers/gpu/drm/i915/i915_gem_gtt.c        | 11 ++++++-----
 5 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c
index 0d41ebc4aea6..7d33b607bc89 100644
--- a/drivers/gpu/drm/i915/gvt/aperture_gm.c
+++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c
@@ -73,12 +73,15 @@ static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm)
 	mutex_lock(&dev_priv->drm.struct_mutex);
 search_again:
 	ret = drm_mm_insert_node_in_range_generic(&dev_priv->ggtt.base.mm,
-						  node, size, 4096, 0,
+						  node, size, 4096,
+						  I915_COLOR_UNEVICTABLE,
 						  start, end, search_flag,
 						  alloc_flag);
 	if (ret) {
 		ret = i915_gem_evict_something(&dev_priv->ggtt.base,
-					       size, 4096, 0, start, end, 0);
+					       size, 4096,
+					       I915_COLOR_UNEVICTABLE,
+					       start, end, 0);
 		if (ret == 0 && ++retried < 3)
 			goto search_again;
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 74658acb307b..a0c67916ec20 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -934,6 +934,8 @@ enum i915_cache_level {
 	I915_CACHE_WT, /* hsw:gt3e WriteThrough for scanouts */
 };
 
+#define I915_COLOR_UNEVICTABLE (-1) /* a non-vma sharing the address space */
+
 #define DEFAULT_CONTEXT_HANDLE 0
 
 /**
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 59065ae0b153..ca0bb837a57f 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -70,7 +70,8 @@ insert_mappable_node(struct i915_ggtt *ggtt,
 {
 	memset(node, 0, sizeof(*node));
 	return drm_mm_insert_node_in_range_generic(&ggtt->base.mm, node,
-						   size, 0, -1,
+						   size, 0,
+						   I915_COLOR_UNEVICTABLE,
 						   0, ggtt->mappable_end,
 						   DRM_MM_SEARCH_DEFAULT,
 						   DRM_MM_CREATE_DEFAULT);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 985142cc6cc9..5f164adc837c 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -437,7 +437,7 @@ static void *reloc_iomap(struct drm_i915_gem_object *obj,
 			memset(&cache->node, 0, sizeof(cache->node));
 			ret = drm_mm_insert_node_in_range_generic
 				(&ggtt->base.mm, &cache->node,
-				 4096, 0, 0,
+				 4096, 0, I915_COLOR_UNEVICTABLE,
 				 0, ggtt->mappable_end,
 				 DRM_MM_SEARCH_DEFAULT,
 				 DRM_MM_CREATE_DEFAULT);
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index c0042471ad87..88ddca24afdb 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -2071,15 +2071,15 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt)
 		return ret;
 
 alloc:
-	ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm,
-						  &ppgtt->node, GEN6_PD_SIZE,
-						  GEN6_PD_ALIGN, 0,
+	ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm, &ppgtt->node,
+						  GEN6_PD_SIZE, GEN6_PD_ALIGN,
+						  I915_COLOR_UNEVICTABLE,
 						  0, ggtt->base.total,
 						  DRM_MM_TOPDOWN);
 	if (ret == -ENOSPC && !retried) {
 		ret = i915_gem_evict_something(&ggtt->base,
 					       GEN6_PD_SIZE, GEN6_PD_ALIGN,
-					       I915_CACHE_NONE,
+					       I915_COLOR_UNEVICTABLE,
 					       0, ggtt->base.total,
 					       0);
 		if (ret)
@@ -2755,7 +2755,8 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 	/* Reserve a mappable slot for our lockless error capture */
 	ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm,
 						  &ggtt->error_capture,
-						  4096, 0, -1,
+						  4096, 0,
+						  I915_COLOR_UNEVICTABLE,
 						  0, ggtt->mappable_end,
 						  0, 0);
 	if (ret)
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[CI 2/6] drm/i915: Fix i915_gem_evict_for_vma (soft-pinning)

[Chris Wilson]

Soft-pinning depends upon being able to check for availabilty of an
interval and evict overlapping object from a drm_mm range manager very
quickly. Currently it uses a linear list, and so performance is dire and
not suitable as a general replacement. Worse, the current code will oops
if it tries to evict an active buffer.

It also helps if the routine reports the correct error codes as expected
by its callers and emits a tracepoint upon use.

For posterity since the wrong patch was pushed (i.e. that missed these
key points and had known bugs), this is the changelog that should have
been on commit 506a8e87d8d2 ("drm/i915: Add soft-pinning API for
execbuffer"):

Userspace can pass in an offset that it presumes the object is located
at. The kernel will then do its utmost to fit the object into that
location. The assumption is that userspace is handling its own object
locations (for example along with full-ppgtt) and that the kernel will
rarely have to make space for the user's requests.

This extends the DRM_IOCTL_I915_GEM_EXECBUFFER2 to do the following:
* if the user supplies a virtual address via the execobject->offset
  *and* sets the EXEC_OBJECT_PINNED flag in execobject->flags, then
  that object is placed at that offset in the address space selected
  by the context specifier in execbuffer.
* the location must be aligned to the GTT page size, 4096 bytes
* as the object is placed exactly as specified, it may be used by this
  execbuffer call without relocations pointing to it

It may fail to do so if:
* EINVAL is returned if the object does not have a 4096 byte aligned
  address
* the object conflicts with another pinned object (either pinned by
  hardware in that address space, e.g. scanouts in the aliasing ppgtt)
  or within the same batch.
  EBUSY is returned if the location is pinned by hardware
  EINVAL is returned if the location is already in use by the batch
* EINVAL is returned if the object conflicts with its own alignment (as meets
  the hardware requirements) or if the placement of the object does not fit
  within the address space

All other execbuffer errors apply.

Presence of this execbuf extension may be queried by passing
I915_PARAM_HAS_EXEC_SOFTPIN to DRM_IOCTL_I915_GETPARAM and checking for
a reported value of 1 (or greater).

v2: Combine the hole/adjusted-hole ENOSPC checks
v3: More color, more splitting, more blurb.

Fixes: 506a8e87d8d2 ("drm/i915: Add soft-pinning API for execbuffer")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_drv.h            |   3 +-
 drivers/gpu/drm/i915/i915_gem_evict.c      | 104 ++++++++++++++++++++++-------
 drivers/gpu/drm/i915/i915_gem_execbuffer.c |   1 +
 drivers/gpu/drm/i915/i915_trace.h          |  28 ++++++++
 drivers/gpu/drm/i915/i915_vma.c            |   2 +-
 5 files changed, 111 insertions(+), 27 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index a0c67916ec20..605247baa7d1 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -3341,7 +3341,8 @@ int __must_check i915_gem_evict_something(struct i915_address_space *vm,
 					  unsigned cache_level,
 					  u64 start, u64 end,
 					  unsigned flags);
-int __must_check i915_gem_evict_for_vma(struct i915_vma *target);
+int __must_check i915_gem_evict_for_vma(struct i915_vma *vma,
+					unsigned int flags);
 int i915_gem_evict_vm(struct i915_address_space *vm, bool do_idle);
 
 /* belongs in i915_gem_gtt.h */
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index 739ede7c89ed..fa40100146ea 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -212,45 +212,99 @@ i915_gem_evict_something(struct i915_address_space *vm,
 	return ret;
 }
 
-int
-i915_gem_evict_for_vma(struct i915_vma *target)
+/**
+ * i915_gem_evict_for_vma - Evict vmas to make room for binding a new one
+ * @target: address space and range to evict for
+ * @flags: additional flags to control the eviction algorithm
+ *
+ * This function will try to evict vmas that overlap the target node.
+ *
+ * To clarify: This is for freeing up virtual address space, not for freeing
+ * memory in e.g. the shrinker.
+ */
+int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags)
 {
-	struct drm_mm_node *node, *next;
+	LIST_HEAD(eviction_list);
+	struct drm_mm_node *node;
+	u64 start = target->node.start;
+	u64 end = start + target->node.size;
+	struct i915_vma *vma, *next;
+	bool check_color;
+	int ret = 0;
 
 	lockdep_assert_held(&target->vm->i915->drm.struct_mutex);
+	trace_i915_gem_evict_vma(target, flags);
+
+	check_color = target->vm->mm.color_adjust;
+	if (check_color) {
+		/* Expand search to cover neighbouring guard pages (or lack!) */
+		if (start > target->vm->start)
+			start -= 4096;
+		if (end < target->vm->start + target->vm->total)
+			end += 4096;
+	}
 
-	list_for_each_entry_safe(node, next,
-			&target->vm->mm.head_node.node_list,
-			node_list) {
-		struct i915_vma *vma;
-		int ret;
-
-		if (node->start + node->size <= target->node.start)
-			continue;
-		if (node->start >= target->node.start + target->node.size)
+	drm_mm_for_each_node_in_range(node, &target->vm->mm, start, end) {
+		/* If we find any non-objects (!vma), we cannot evict them */
+		if (node->color == I915_COLOR_UNEVICTABLE) {
+			ret = -ENOSPC;
 			break;
+		}
 
 		vma = container_of(node, typeof(*vma), node);
 
-		if (i915_vma_is_pinned(vma)) {
-			if (!vma->exec_entry || i915_vma_pin_count(vma) > 1)
-				/* Object is pinned for some other use */
-				return -EBUSY;
+		/* If we are using coloring to insert guard pages between
+		 * different cache domains within the address space, we have
+		 * to check whether the objects on either side of our range
+		 * abutt and conflict. If they are in conflict, then we evict
+		 * those as well to make room for our guard pages.
+		 */
+		if (check_color) {
+			if (vma->node.start + vma->node.size == target->node.start) {
+				if (vma->node.color == target->node.color)
+					continue;
+			}
+			if (vma->node.start == target->node.start + target->node.size) {
+				if (vma->node.color == target->node.color)
+					continue;
+			}
+		}
 
-			/* We need to evict a buffer in the same batch */
-			if (vma->exec_entry->flags & EXEC_OBJECT_PINNED)
-				/* Overlapping fixed objects in the same batch */
-				return -EINVAL;
+		if (flags & PIN_NONBLOCK &&
+		    (i915_vma_is_pinned(vma) || i915_vma_is_active(vma))) {
+			ret = -ENOSPC;
+			break;
+		}
 
-			return -ENOSPC;
+		/* Overlap of objects in the same batch? */
+		if (i915_vma_is_pinned(vma)) {
+			ret = -ENOSPC;
+			if (vma->exec_entry &&
+			    vma->exec_entry->flags & EXEC_OBJECT_PINNED)
+				ret = -EINVAL;
+			break;
 		}
 
-		ret = i915_vma_unbind(vma);
-		if (ret)
-			return ret;
+		/* Never show fear in the face of dragons!
+		 *
+		 * We cannot directly remove this node from within this
+		 * iterator and as with i915_gem_evict_something() we employ
+		 * the vma pin_count in order to prevent the action of
+		 * unbinding one vma from freeing (by dropping its active
+		 * reference) another in our eviction list.
+		 */
+		__i915_vma_pin(vma);
+		list_add(&vma->exec_list, &eviction_list);
 	}
 
-	return 0;
+	list_for_each_entry_safe(vma, next, &eviction_list, exec_list) {
+		list_del_init(&vma->exec_list);
+		__i915_vma_unpin(vma);
+		if (ret == 0)
+			ret = i915_vma_unbind(vma);
+	}
+
+	return ret;
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 5f164adc837c..d665a33229bd 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -274,6 +274,7 @@ static void eb_destroy(struct eb_vmas *eb)
 				       exec_list);
 		list_del_init(&vma->exec_list);
 		i915_gem_execbuffer_unreserve_vma(vma);
+		vma->exec_entry = NULL;
 		i915_vma_put(vma);
 	}
 	kfree(eb);
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
index 240067705e1c..18ae37c411fd 100644
--- a/drivers/gpu/drm/i915/i915_trace.h
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -450,6 +450,34 @@ TRACE_EVENT(i915_gem_evict_vm,
 	    TP_printk("dev=%d, vm=%p", __entry->dev, __entry->vm)
 );
 
+TRACE_EVENT(i915_gem_evict_vma,
+	    TP_PROTO(struct i915_vma *vma, unsigned int flags),
+	    TP_ARGS(vma, flags),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(struct i915_address_space *, vm)
+			     __field(u64, start)
+			     __field(u64, size)
+			     __field(unsigned long, color)
+			     __field(unsigned int, flags)
+			    ),
+
+	    TP_fast_assign(
+			   __entry->dev = vma->vm->i915->drm.primary->index;
+			   __entry->vm = vma->vm;
+			   __entry->start = vma->node.start;
+			   __entry->size = vma->node.size;
+			   __entry->color = vma->node.color;
+			   __entry->flags = flags;
+			  ),
+
+	    TP_printk("dev=%d, vm=%p, start=%llx size=%llx, color=%lx, flags=%x",
+		      __entry->dev, __entry->vm,
+		      __entry->start, __entry->size,
+		      __entry->color, __entry->flags)
+);
+
 TRACE_EVENT(i915_gem_ring_sync_to,
 	    TP_PROTO(struct drm_i915_gem_request *to,
 		     struct drm_i915_gem_request *from),
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index 4c91a68ecb6d..bc077e51b3e7 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -401,7 +401,7 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
 		vma->node.color = obj->cache_level;
 		ret = drm_mm_reserve_node(&vma->vm->mm, &vma->node);
 		if (ret) {
-			ret = i915_gem_evict_for_vma(vma);
+			ret = i915_gem_evict_for_vma(vma, flags);
 			if (ret == 0)
 				ret = drm_mm_reserve_node(&vma->vm->mm, &vma->node);
 			if (ret)
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[CI 3/6] drm/i915: Tidy i915_gem_valid_gtt_space()

[Chris Wilson]

We can replace a couple of tests with an assertion that the passed in
node is already allocated (as matches the existing call convention) and
by a small bit of refactoring we can bring the line lengths to under
80cols.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_vma.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index bc077e51b3e7..37c3eebe8316 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -297,10 +297,14 @@ void __i915_vma_set_map_and_fenceable(struct i915_vma *vma)
 		vma->flags &= ~I915_VMA_CAN_FENCE;
 }
 
-bool i915_gem_valid_gtt_space(struct i915_vma *vma,
-			      unsigned long cache_level)
+static bool color_differs(struct drm_mm_node *node, unsigned long color)
 {
-	struct drm_mm_node *gtt_space = &vma->node;
+	return node->allocated && node->color != color;
+}
+
+bool i915_gem_valid_gtt_space(struct i915_vma *vma, unsigned long cache_level)
+{
+	struct drm_mm_node *node = &vma->node;
 	struct drm_mm_node *other;
 
 	/*
@@ -313,18 +317,16 @@ bool i915_gem_valid_gtt_space(struct i915_vma *vma,
 	if (vma->vm->mm.color_adjust == NULL)
 		return true;
 
-	if (!drm_mm_node_allocated(gtt_space))
-		return true;
-
-	if (list_empty(&gtt_space->node_list))
-		return true;
+	/* Only valid to be called on an already inserted vma */
+	GEM_BUG_ON(!drm_mm_node_allocated(node));
+	GEM_BUG_ON(list_empty(&node->node_list));
 
-	other = list_entry(gtt_space->node_list.prev, struct drm_mm_node, node_list);
-	if (other->allocated && !other->hole_follows && other->color != cache_level)
+	other = list_prev_entry(node, node_list);
+	if (color_differs(other, cache_level) && !other->hole_follows)
 		return false;
 
-	other = list_entry(gtt_space->node_list.next, struct drm_mm_node, node_list);
-	if (other->allocated && !gtt_space->hole_follows && other->color != cache_level)
+	other = list_next_entry(node, node_list);
+	if (color_differs(other, cache_level) && !node->hole_follows)
 		return false;
 
 	return true;
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[CI 4/6] drm/i915: Implement local atomic_state_free callback

[Chris Wilson]

As we use debugobjects to track the lifetime of fences within our atomic
state, we ideally want to mark those objects as freed along with their
containers. This merits us hookin into config->funcs->atomic_state_free
for this purpose.

This allows us to enable debugobjects for sw-fences without triggering
known issues.

Fixes: fc1584059d6c ("drm/i915: Integrate i915_sw_fence with debugobjects")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 231f099bbd34..a88c810dbf6b 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -15994,6 +15994,17 @@ intel_user_framebuffer_create(struct drm_device *dev,
 	return fb;
 }
 
+static void intel_atomic_state_free(struct drm_atomic_state *state)
+{
+	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
+
+	drm_atomic_state_default_release(state);
+
+	i915_sw_fence_fini(&intel_state->commit_ready);
+
+	kfree(state);
+}
+
 static const struct drm_mode_config_funcs intel_mode_funcs = {
 	.fb_create = intel_user_framebuffer_create,
 	.output_poll_changed = intel_fbdev_output_poll_changed,
@@ -16001,6 +16012,7 @@ static const struct drm_mode_config_funcs intel_mode_funcs = {
 	.atomic_commit = intel_atomic_commit,
 	.atomic_state_alloc = intel_atomic_state_alloc,
 	.atomic_state_clear = intel_atomic_state_clear,
+	.atomic_state_free = intel_atomic_state_free,
 };
 
 /**
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[CI 5/6] drm/i915: Enable swfence debugobject support for i915.ko

[Chris Wilson]

Only once the debugobject symbols are exported can we enable support for
debugging swfences when i915 is built as a module. Requires commit
2617fdca3f68 ("lib/debugobjects: export for use in modules")

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/Kconfig.debug | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug
index cc4ad576e6e9..597648c7a645 100644
--- a/drivers/gpu/drm/i915/Kconfig.debug
+++ b/drivers/gpu/drm/i915/Kconfig.debug
@@ -24,7 +24,7 @@ config DRM_I915_DEBUG
         select X86_MSR # used by igt/pm_rpm
         select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks)
         select DRM_DEBUG_MM if DRM=y
-        select DRM_I915_SW_FENCE_DEBUG_OBJECTS if DRM_I915=y
+	select DRM_I915_SW_FENCE_DEBUG_OBJECTS
         default n
         help
           Choose this option to turn on extra driver debugging that may affect
@@ -48,7 +48,7 @@ config DRM_I915_DEBUG_GEM
 
 config DRM_I915_SW_FENCE_DEBUG_OBJECTS
         bool "Enable additional driver debugging for fence objects"
-        depends on DRM_I915=y
+        depends on DRM_I915
         select DEBUG_OBJECTS
         default n
         help
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[CI 6/6] drm/i915/execlists: Use list_safe_reset_next() instead of opencoding

[Chris Wilson]

list.h provides a macro for updating the next element in a safe
list-iter, so let's use it so that it is hopefully clearer to the reader
about the unusual behaviour, and also easier to grep.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
---
 drivers/gpu/drm/i915/intel_lrc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 67aec8f33c1d..8b412880e88c 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -733,7 +733,7 @@ static void execlists_schedule(struct drm_i915_gem_request *request, int prio)
 			if (prio > READ_ONCE(p->signaler->priority))
 				list_move_tail(&p->dfs_link, &dfs);
 
-		p = list_next_entry(dep, dfs_link);
+		list_safe_reset_next(dep, p, dfs_link);
 		if (!RB_EMPTY_NODE(&pt->node))
 			continue;
 
-- 
2.11.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

✓ Fi.CI.BAT: success for series starting with [CI,1/6] drm/i915: Mark all non-vma being inserted into the address spaces

[Patchwork]

== Series Details ==

Series: series starting with [CI,1/6] drm/i915: Mark all non-vma being inserted into the address spaces
URL   : https://patchwork.freedesktop.org/series/16367/
State : success

== Summary ==

Series 16367v1 Series without cover letter
https://patchwork.freedesktop.org/api/1.0/series/16367/revisions/1/mbox/

Test gem_busy:
        Subgroup basic-hang-default:
                fail       -> PASS       (fi-hsw-4770r)

fi-bdw-5557u     total:247  pass:232  dwarn:0   dfail:0   fail:0   skip:15 
fi-bsw-n3050     total:247  pass:207  dwarn:0   dfail:0   fail:0   skip:40 
fi-bxt-t5700     total:247  pass:219  dwarn:0   dfail:0   fail:0   skip:28 
fi-byt-j1900     total:247  pass:219  dwarn:0   dfail:0   fail:0   skip:28 
fi-byt-n2820     total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 
fi-hsw-4770      total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-hsw-4770r     total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-ilk-650       total:247  pass:194  dwarn:0   dfail:0   fail:0   skip:53 
fi-ivb-3520m     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-ivb-3770      total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-kbl-7500u     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-skl-6260u     total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-skl-6700hq    total:247  pass:226  dwarn:0   dfail:0   fail:0   skip:21 
fi-skl-6770hq    total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-snb-2520m     total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 
fi-snb-2600      total:247  pass:214  dwarn:0   dfail:0   fail:0   skip:33 

a947529bb652fdcf50e8c0e8ee5102b737bae23e drm-tip: 2016y-12m-05d-14h-27m-06s UTC integration manifest
d4e7c8b drm/i915/execlists: Use list_safe_reset_next() instead of opencoding
ef20a3e drm/i915: Enable swfence debugobject support for i915.ko
41105d8 drm/i915: Implement local atomic_state_free callback
c975d12 drm/i915: Tidy i915_gem_valid_gtt_space()
eb67726 drm/i915: Fix i915_gem_evict_for_vma (soft-pinning)
226ff7b drm/i915: Mark all non-vma being inserted into the address spaces

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3195/
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[CI 1/6] drm/i915: Mark all non-vma being inserted into the address spaces

[Chris Wilson]

We need to distinguish between full i915_vma structs and simple
drm_mm_nodes when considering eviction (i.e. we must be careful not to
treat a mere drm_mm_node as a much larger i915_vma causing memory
corruption, if we are lucky). To do this, color these not-a-vma with -1
(I915_COLOR_UNEVICTABLE).

v2...v200: New name for -1.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/gvt/aperture_gm.c     |  7 +++++--
 drivers/gpu/drm/i915/i915_drv.h            |  2 ++
 drivers/gpu/drm/i915/i915_gem.c            |  3 ++-
 drivers/gpu/drm/i915/i915_gem_execbuffer.c |  2 +-
 drivers/gpu/drm/i915/i915_gem_gtt.c        | 11 ++++++-----
 5 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c
index 0d41ebc4aea6..7d33b607bc89 100644
--- a/drivers/gpu/drm/i915/gvt/aperture_gm.c
+++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c
@@ -73,12 +73,15 @@ static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm)
 	mutex_lock(&dev_priv->drm.struct_mutex);
 search_again:
 	ret = drm_mm_insert_node_in_range_generic(&dev_priv->ggtt.base.mm,
-						  node, size, 4096, 0,
+						  node, size, 4096,
+						  I915_COLOR_UNEVICTABLE,
 						  start, end, search_flag,
 						  alloc_flag);
 	if (ret) {
 		ret = i915_gem_evict_something(&dev_priv->ggtt.base,
-					       size, 4096, 0, start, end, 0);
+					       size, 4096,
+					       I915_COLOR_UNEVICTABLE,
+					       start, end, 0);
 		if (ret == 0 && ++retried < 3)
 			goto search_again;
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index ca9786c05bb2..aad336789b00 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -934,6 +934,8 @@ enum i915_cache_level {
 	I915_CACHE_WT, /* hsw:gt3e WriteThrough for scanouts */
 };
 
+#define I915_COLOR_UNEVICTABLE (-1) /* a non-vma sharing the address space */
+
 #define DEFAULT_CONTEXT_HANDLE 0
 
 /**
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 59065ae0b153..ca0bb837a57f 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -70,7 +70,8 @@ insert_mappable_node(struct i915_ggtt *ggtt,
 {
 	memset(node, 0, sizeof(*node));
 	return drm_mm_insert_node_in_range_generic(&ggtt->base.mm, node,
-						   size, 0, -1,
+						   size, 0,
+						   I915_COLOR_UNEVICTABLE,
 						   0, ggtt->mappable_end,
 						   DRM_MM_SEARCH_DEFAULT,
 						   DRM_MM_CREATE_DEFAULT);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 985142cc6cc9..5f164adc837c 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -437,7 +437,7 @@ static void *reloc_iomap(struct drm_i915_gem_object *obj,
 			memset(&cache->node, 0, sizeof(cache->node));
 			ret = drm_mm_insert_node_in_range_generic
 				(&ggtt->base.mm, &cache->node,
-				 4096, 0, 0,
+				 4096, 0, I915_COLOR_UNEVICTABLE,
 				 0, ggtt->mappable_end,
 				 DRM_MM_SEARCH_DEFAULT,
 				 DRM_MM_CREATE_DEFAULT);
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index c0042471ad87..88ddca24afdb 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -2071,15 +2071,15 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt)
 		return ret;
 
 alloc:
-	ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm,
-						  &ppgtt->node, GEN6_PD_SIZE,
-						  GEN6_PD_ALIGN, 0,
+	ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm, &ppgtt->node,
+						  GEN6_PD_SIZE, GEN6_PD_ALIGN,
+						  I915_COLOR_UNEVICTABLE,
 						  0, ggtt->base.total,
 						  DRM_MM_TOPDOWN);
 	if (ret == -ENOSPC && !retried) {
 		ret = i915_gem_evict_something(&ggtt->base,
 					       GEN6_PD_SIZE, GEN6_PD_ALIGN,
-					       I915_CACHE_NONE,
+					       I915_COLOR_UNEVICTABLE,
 					       0, ggtt->base.total,
 					       0);
 		if (ret)
@@ -2755,7 +2755,8 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 	/* Reserve a mappable slot for our lockless error capture */
 	ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm,
 						  &ggtt->error_capture,
-						  4096, 0, -1,
+						  4096, 0,
+						  I915_COLOR_UNEVICTABLE,
 						  0, ggtt->mappable_end,
 						  0, 0);
 	if (ret)
-- 
2.10.2

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

✓ Fi.CI.BAT: success for series starting with [CI,1/6] drm/i915: Mark all non-vma being inserted into the address spaces

[Patchwork]

== Series Details ==

Series: series starting with [CI,1/6] drm/i915: Mark all non-vma being inserted into the address spaces
URL   : https://patchwork.freedesktop.org/series/16353/
State : success

== Summary ==

Series 16353v1 Series without cover letter
https://patchwork.freedesktop.org/api/1.0/series/16353/revisions/1/mbox/


fi-bdw-5557u     total:247  pass:232  dwarn:0   dfail:0   fail:0   skip:15 
fi-bsw-n3050     total:247  pass:207  dwarn:0   dfail:0   fail:0   skip:40 
fi-byt-j1900     total:247  pass:219  dwarn:0   dfail:0   fail:0   skip:28 
fi-byt-n2820     total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 
fi-hsw-4770      total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-hsw-4770r     total:247  pass:227  dwarn:0   dfail:0   fail:0   skip:20 
fi-ilk-650       total:247  pass:194  dwarn:0   dfail:0   fail:0   skip:53 
fi-ivb-3520m     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-ivb-3770      total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-kbl-7500u     total:247  pass:225  dwarn:0   dfail:0   fail:0   skip:22 
fi-skl-6260u     total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-skl-6700hq    total:247  pass:226  dwarn:0   dfail:0   fail:0   skip:21 
fi-skl-6700k     total:247  pass:223  dwarn:3   dfail:0   fail:0   skip:21 
fi-skl-6770hq    total:247  pass:233  dwarn:0   dfail:0   fail:0   skip:14 
fi-snb-2520m     total:247  pass:215  dwarn:0   dfail:0   fail:0   skip:32 
fi-snb-2600      total:247  pass:214  dwarn:0   dfail:0   fail:0   skip:33 

a17bd268c6f9e1076a9aea44573ff58b0caeb44d drm-tip: 2016y-12m-05d-11h-19m-41s UTC integration manifest
b8835aa drm/i915/execlists: Use list_safe_reset_next() instead of opencoding
a19bacf drm/i915: Enable swfence debugobject support for i915.ko
9ede387 drm/i915: Mark the atomic commit_ready fence as freed
d2f003b drm/i915: Tidy i915_gem_valid_gtt_space()
8c51b92 drm/i915: Fix i915_gem_evict_for_vma (soft-pinning)
91db816 drm/i915: Mark all non-vma being inserted into the address spaces

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3188/
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx