drm: Add colouring to the range allocator
In order to support snoopable memory on non-LLC architectures (so that
we can bind vgem objects into the i915 GATT for example), we have to
avoid the prefetcher on the GPU from crossing memory domains and so
prevent allocation of a snoopable PTE immediately following an uncached
PTE. To do that, we need to extend the range allocator with support for
tracking and segregating different node colours.
This will be used by i915 to segregate memory domains within the GTT.
v2: Now with more drm_mm helpers and less driver interference.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Airlie <airlied@redhat.com
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@gmail.com>
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index d58e69d..fbe0842 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -354,7 +354,7 @@
/* Get a DRM GEM mmap offset allocated... */
list->file_offset_node = drm_mm_search_free(&mm->offset_manager,
- obj->size / PAGE_SIZE, 0, 0);
+ obj->size / PAGE_SIZE, 0, false);
if (!list->file_offset_node) {
DRM_ERROR("failed to allocate offset for bo %d\n", obj->name);
diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index 961fb54..9bb82f7 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -118,45 +118,53 @@
static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
struct drm_mm_node *node,
- unsigned long size, unsigned alignment)
+ unsigned long size, unsigned alignment,
+ unsigned long color)
{
struct drm_mm *mm = hole_node->mm;
- unsigned long tmp = 0, wasted = 0;
unsigned long hole_start = drm_mm_hole_node_start(hole_node);
unsigned long hole_end = drm_mm_hole_node_end(hole_node);
+ unsigned long adj_start = hole_start;
+ unsigned long adj_end = hole_end;
BUG_ON(!hole_node->hole_follows || node->allocated);
- if (alignment)
- tmp = hole_start % alignment;
+ if (mm->color_adjust)
+ mm->color_adjust(hole_node, color, &adj_start, &adj_end);
- if (!tmp) {
+ if (alignment) {
+ unsigned tmp = adj_start % alignment;
+ if (tmp)
+ adj_start += alignment - tmp;
+ }
+
+ if (adj_start == hole_start) {
hole_node->hole_follows = 0;
- list_del_init(&hole_node->hole_stack);
- } else
- wasted = alignment - tmp;
+ list_del(&hole_node->hole_stack);
+ }
- node->start = hole_start + wasted;
+ node->start = adj_start;
node->size = size;
node->mm = mm;
+ node->color = color;
node->allocated = 1;
INIT_LIST_HEAD(&node->hole_stack);
list_add(&node->node_list, &hole_node->node_list);
- BUG_ON(node->start + node->size > hole_end);
+ BUG_ON(node->start + node->size > adj_end);
+ node->hole_follows = 0;
if (node->start + node->size < hole_end) {
list_add(&node->hole_stack, &mm->hole_stack);
node->hole_follows = 1;
- } else {
- node->hole_follows = 0;
}
}
struct drm_mm_node *drm_mm_get_block_generic(struct drm_mm_node *hole_node,
unsigned long size,
unsigned alignment,
+ unsigned long color,
int atomic)
{
struct drm_mm_node *node;
@@ -165,7 +173,7 @@
if (unlikely(node == NULL))
return NULL;
- drm_mm_insert_helper(hole_node, node, size, alignment);
+ drm_mm_insert_helper(hole_node, node, size, alignment, color);
return node;
}
@@ -181,11 +189,11 @@
{
struct drm_mm_node *hole_node;
- hole_node = drm_mm_search_free(mm, size, alignment, 0);
+ hole_node = drm_mm_search_free(mm, size, alignment, false);
if (!hole_node)
return -ENOSPC;
- drm_mm_insert_helper(hole_node, node, size, alignment);
+ drm_mm_insert_helper(hole_node, node, size, alignment, 0);
return 0;
}
@@ -194,50 +202,57 @@
static void drm_mm_insert_helper_range(struct drm_mm_node *hole_node,
struct drm_mm_node *node,
unsigned long size, unsigned alignment,
+ unsigned long color,
unsigned long start, unsigned long end)
{
struct drm_mm *mm = hole_node->mm;
- unsigned long tmp = 0, wasted = 0;
unsigned long hole_start = drm_mm_hole_node_start(hole_node);
unsigned long hole_end = drm_mm_hole_node_end(hole_node);
+ unsigned long adj_start = hole_start;
+ unsigned long adj_end = hole_end;
BUG_ON(!hole_node->hole_follows || node->allocated);
- if (hole_start < start)
- wasted += start - hole_start;
- if (alignment)
- tmp = (hole_start + wasted) % alignment;
+ if (mm->color_adjust)
+ mm->color_adjust(hole_node, color, &adj_start, &adj_end);
- if (tmp)
- wasted += alignment - tmp;
+ if (adj_start < start)
+ adj_start = start;
- if (!wasted) {
- hole_node->hole_follows = 0;
- list_del_init(&hole_node->hole_stack);
+ if (alignment) {
+ unsigned tmp = adj_start % alignment;
+ if (tmp)
+ adj_start += alignment - tmp;
}
- node->start = hole_start + wasted;
+ if (adj_start == hole_start) {
+ hole_node->hole_follows = 0;
+ list_del(&hole_node->hole_stack);
+ }
+
+ node->start = adj_start;
node->size = size;
node->mm = mm;
+ node->color = color;
node->allocated = 1;
INIT_LIST_HEAD(&node->hole_stack);
list_add(&node->node_list, &hole_node->node_list);
- BUG_ON(node->start + node->size > hole_end);
+ BUG_ON(node->start + node->size > adj_end);
BUG_ON(node->start + node->size > end);
+ node->hole_follows = 0;
if (node->start + node->size < hole_end) {
list_add(&node->hole_stack, &mm->hole_stack);
node->hole_follows = 1;
- } else {
- node->hole_follows = 0;
}
}
struct drm_mm_node *drm_mm_get_block_range_generic(struct drm_mm_node *hole_node,
unsigned long size,
unsigned alignment,
+ unsigned long color,
unsigned long start,
unsigned long end,
int atomic)
@@ -248,7 +263,7 @@
if (unlikely(node == NULL))
return NULL;
- drm_mm_insert_helper_range(hole_node, node, size, alignment,
+ drm_mm_insert_helper_range(hole_node, node, size, alignment, color,
start, end);
return node;
@@ -267,11 +282,11 @@
struct drm_mm_node *hole_node;
hole_node = drm_mm_search_free_in_range(mm, size, alignment,
- start, end, 0);
+ start, end, false);
if (!hole_node)
return -ENOSPC;
- drm_mm_insert_helper_range(hole_node, node, size, alignment,
+ drm_mm_insert_helper_range(hole_node, node, size, alignment, 0,
start, end);
return 0;
@@ -336,27 +351,23 @@
static int check_free_hole(unsigned long start, unsigned long end,
unsigned long size, unsigned alignment)
{
- unsigned wasted = 0;
-
if (end - start < size)
return 0;
if (alignment) {
unsigned tmp = start % alignment;
if (tmp)
- wasted = alignment - tmp;
+ start += alignment - tmp;
}
- if (end >= start + size + wasted) {
- return 1;
- }
-
- return 0;
+ return end >= start + size;
}
-struct drm_mm_node *drm_mm_search_free(const struct drm_mm *mm,
- unsigned long size,
- unsigned alignment, int best_match)
+struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm,
+ unsigned long size,
+ unsigned alignment,
+ unsigned long color,
+ bool best_match)
{
struct drm_mm_node *entry;
struct drm_mm_node *best;
@@ -368,10 +379,17 @@
best_size = ~0UL;
list_for_each_entry(entry, &mm->hole_stack, hole_stack) {
+ unsigned long adj_start = drm_mm_hole_node_start(entry);
+ unsigned long adj_end = drm_mm_hole_node_end(entry);
+
+ if (mm->color_adjust) {
+ mm->color_adjust(entry, color, &adj_start, &adj_end);
+ if (adj_end <= adj_start)
+ continue;
+ }
+
BUG_ON(!entry->hole_follows);
- if (!check_free_hole(drm_mm_hole_node_start(entry),
- drm_mm_hole_node_end(entry),
- size, alignment))
+ if (!check_free_hole(adj_start, adj_end, size, alignment))
continue;
if (!best_match)
@@ -385,14 +403,15 @@
return best;
}
-EXPORT_SYMBOL(drm_mm_search_free);
+EXPORT_SYMBOL(drm_mm_search_free_generic);
-struct drm_mm_node *drm_mm_search_free_in_range(const struct drm_mm *mm,
- unsigned long size,
- unsigned alignment,
- unsigned long start,
- unsigned long end,
- int best_match)
+struct drm_mm_node *drm_mm_search_free_in_range_generic(const struct drm_mm *mm,
+ unsigned long size,
+ unsigned alignment,
+ unsigned long color,
+ unsigned long start,
+ unsigned long end,
+ bool best_match)
{
struct drm_mm_node *entry;
struct drm_mm_node *best;
@@ -410,6 +429,13 @@
end : drm_mm_hole_node_end(entry);
BUG_ON(!entry->hole_follows);
+
+ if (mm->color_adjust) {
+ mm->color_adjust(entry, color, &adj_start, &adj_end);
+ if (adj_end <= adj_start)
+ continue;
+ }
+
if (!check_free_hole(adj_start, adj_end, size, alignment))
continue;
@@ -424,7 +450,7 @@
return best;
}
-EXPORT_SYMBOL(drm_mm_search_free_in_range);
+EXPORT_SYMBOL(drm_mm_search_free_in_range_generic);
/**
* Moves an allocation. To be used with embedded struct drm_mm_node.
@@ -437,6 +463,7 @@
new->mm = old->mm;
new->start = old->start;
new->size = old->size;
+ new->color = old->color;
old->allocated = 0;
new->allocated = 1;
@@ -452,9 +479,12 @@
* Warning: As long as the scan list is non-empty, no other operations than
* adding/removing nodes to/from the scan list are allowed.
*/
-void drm_mm_init_scan(struct drm_mm *mm, unsigned long size,
- unsigned alignment)
+void drm_mm_init_scan(struct drm_mm *mm,
+ unsigned long size,
+ unsigned alignment,
+ unsigned long color)
{
+ mm->scan_color = color;
mm->scan_alignment = alignment;
mm->scan_size = size;
mm->scanned_blocks = 0;
@@ -474,11 +504,14 @@
* Warning: As long as the scan list is non-empty, no other operations than
* adding/removing nodes to/from the scan list are allowed.
*/
-void drm_mm_init_scan_with_range(struct drm_mm *mm, unsigned long size,
+void drm_mm_init_scan_with_range(struct drm_mm *mm,
+ unsigned long size,
unsigned alignment,
+ unsigned long color,
unsigned long start,
unsigned long end)
{
+ mm->scan_color = color;
mm->scan_alignment = alignment;
mm->scan_size = size;
mm->scanned_blocks = 0;
@@ -522,17 +555,21 @@
hole_start = drm_mm_hole_node_start(prev_node);
hole_end = drm_mm_hole_node_end(prev_node);
+
+ adj_start = hole_start;
+ adj_end = hole_end;
+
+ if (mm->color_adjust)
+ mm->color_adjust(prev_node, mm->scan_color, &adj_start, &adj_end);
+
if (mm->scan_check_range) {
- adj_start = hole_start < mm->scan_start ?
- mm->scan_start : hole_start;
- adj_end = hole_end > mm->scan_end ?
- mm->scan_end : hole_end;
- } else {
- adj_start = hole_start;
- adj_end = hole_end;
+ if (adj_start < mm->scan_start)
+ adj_start = mm->scan_start;
+ if (adj_end > mm->scan_end)
+ adj_end = mm->scan_end;
}
- if (check_free_hole(adj_start , adj_end,
+ if (check_free_hole(adj_start, adj_end,
mm->scan_size, mm->scan_alignment)) {
mm->scan_hit_start = hole_start;
mm->scan_hit_size = hole_end;
@@ -616,6 +653,8 @@
mm->head_node.size = start - mm->head_node.start;
list_add_tail(&mm->head_node.hole_stack, &mm->hole_stack);
+ mm->color_adjust = NULL;
+
return 0;
}
EXPORT_SYMBOL(drm_mm_init);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 2b54142..0fdb3d2 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2748,8 +2748,8 @@
if (map_and_fenceable)
free_space =
drm_mm_search_free_in_range(&dev_priv->mm.gtt_space,
- size, alignment, 0,
- dev_priv->mm.gtt_mappable_end,
+ size, alignment,
+ 0, dev_priv->mm.gtt_mappable_end,
0);
else
free_space = drm_mm_search_free(&dev_priv->mm.gtt_space,
@@ -2760,7 +2760,7 @@
obj->gtt_space =
drm_mm_get_block_range_generic(free_space,
size, alignment, 0,
- dev_priv->mm.gtt_mappable_end,
+ 0, dev_priv->mm.gtt_mappable_end,
0);
else
obj->gtt_space =
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index ae7c24e..eba0308 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -78,11 +78,12 @@
INIT_LIST_HEAD(&unwind_list);
if (mappable)
- drm_mm_init_scan_with_range(&dev_priv->mm.gtt_space, min_size,
- alignment, 0,
- dev_priv->mm.gtt_mappable_end);
+ drm_mm_init_scan_with_range(&dev_priv->mm.gtt_space,
+ min_size, alignment, 0,
+ 0, dev_priv->mm.gtt_mappable_end);
else
- drm_mm_init_scan(&dev_priv->mm.gtt_space, min_size, alignment);
+ drm_mm_init_scan(&dev_priv->mm.gtt_space,
+ min_size, alignment, 0);
/* First see if there is a large enough contiguous idle region... */
list_for_each_entry(obj, &dev_priv->mm.inactive_list, mm_list) {