drm/i915: Choose not to evict faultable objects from the GGTT
Often times we do not want to evict mapped objects from the GGTT as
these are quite expensive to teardown and frequently reused (causing an
equally, if not more so, expensive setup). In particular, when faulting
in a new object we want to avoid evicting an active object, or else we
may trigger a page-fault-of-doom as we ping-pong between evicting two
objects.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-26-chris@chris-wilson.co.uk
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 9811980..a13290a 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1706,6 +1706,7 @@
struct i915_vma *vma;
pgoff_t page_offset;
unsigned long pfn;
+ unsigned int flags;
int ret;
/* We don't use vmf->pgoff since that has the fake offset */
@@ -1735,9 +1736,16 @@
goto err_unlock;
}
+ /* If the object is smaller than a couple of partial vma, it is
+ * not worth only creating a single partial vma - we may as well
+ * clear enough space for the full object.
+ */
+ flags = PIN_MAPPABLE;
+ if (obj->base.size > 2 * MIN_CHUNK_PAGES << PAGE_SHIFT)
+ flags |= PIN_NONBLOCK | PIN_NONFAULT;
+
/* Now pin it into the GTT as needed */
- vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0,
- PIN_MAPPABLE | PIN_NONBLOCK);
+ vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, flags);
if (IS_ERR(vma)) {
struct i915_ggtt_view view;
unsigned int chunk_size;