aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/i915/gem/i915_gem_pages.c
diff options
context:
space:
mode:
authorMaarten Lankhorst <maarten.lankhorst@linux.intel.com>2021-03-23 16:49:52 +0100
committerDaniel Vetter <daniel.vetter@ffwll.ch>2021-03-24 11:39:59 +0100
commit0edbb9ba1bfe70b38850d4a712e83487525f0640 (patch)
treed959315ba9e18ce610c55cd31ad567bbbac90d32 /drivers/gpu/drm/i915/gem/i915_gem_pages.c
parentdrm/i915: Pin timeline map after first timeline pin, v4. (diff)
downloadlinux-dev-0edbb9ba1bfe70b38850d4a712e83487525f0640.tar.xz
linux-dev-0edbb9ba1bfe70b38850d4a712e83487525f0640.zip
drm/i915: Move cmd parser pinning to execbuffer
We need to get rid of allocations in the cmd parser, because it needs to be called from a signaling context, first move all pinning to execbuf, where we already hold all locks. Allocate jump_whitelist in the execbuffer, and add annotations around intel_engine_cmd_parser(), to ensure we only call the command parser without allocating any memory, or taking any locks we're not supposed to. Because i915_gem_object_get_page() may also allocate memory, add a path to i915_gem_object_get_sg() that prevents memory allocations, and walk the sg list manually. It should be similarly fast. This has the added benefit of being able to catch all memory allocation errors before the point of no return, and return -ENOMEM safely to the execbuf submitter. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20210323155059.628690-4-maarten.lankhorst@linux.intel.com
Diffstat (limited to 'drivers/gpu/drm/i915/gem/i915_gem_pages.c')
-rw-r--r--drivers/gpu/drm/i915/gem/i915_gem_pages.c21
1 files changed, 17 insertions, 4 deletions
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
index 43028f3539a6..d44b72dd13fe 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_pages.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
@@ -448,7 +448,8 @@ struct scatterlist *
__i915_gem_object_get_sg(struct drm_i915_gem_object *obj,
struct i915_gem_object_page_iter *iter,
unsigned int n,
- unsigned int *offset)
+ unsigned int *offset,
+ bool allow_alloc)
{
const bool dma = iter == &obj->mm.get_dma_page;
struct scatterlist *sg;
@@ -470,6 +471,9 @@ __i915_gem_object_get_sg(struct drm_i915_gem_object *obj,
if (n < READ_ONCE(iter->sg_idx))
goto lookup;
+ if (!allow_alloc)
+ goto manual_lookup;
+
mutex_lock(&iter->lock);
/* We prefer to reuse the last sg so that repeated lookup of this
@@ -519,7 +523,16 @@ scan:
if (unlikely(n < idx)) /* insertion completed by another thread */
goto lookup;
- /* In case we failed to insert the entry into the radixtree, we need
+ goto manual_walk;
+
+manual_lookup:
+ idx = 0;
+ sg = obj->mm.pages->sgl;
+ count = __sg_page_count(sg);
+
+manual_walk:
+ /*
+ * In case we failed to insert the entry into the radixtree, we need
* to look beyond the current sg.
*/
while (idx + count <= n) {
@@ -566,7 +579,7 @@ i915_gem_object_get_page(struct drm_i915_gem_object *obj, unsigned int n)
GEM_BUG_ON(!i915_gem_object_has_struct_page(obj));
- sg = i915_gem_object_get_sg(obj, n, &offset);
+ sg = i915_gem_object_get_sg(obj, n, &offset, true);
return nth_page(sg_page(sg), offset);
}
@@ -592,7 +605,7 @@ i915_gem_object_get_dma_address_len(struct drm_i915_gem_object *obj,
struct scatterlist *sg;
unsigned int offset;
- sg = i915_gem_object_get_sg_dma(obj, n, &offset);
+ sg = i915_gem_object_get_sg_dma(obj, n, &offset, true);
if (len)
*len = sg_dma_len(sg) - (offset << PAGE_SHIFT);