Skip to contentSkip to conversation

drm/i915: disable cpu relocs on ilk and earlier

PATCH In review Daniel Vetter 2 participants 2 messages Oct 10, 2012

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

PATCH 1/1 drm/i915: disable cpu relocs on ilk and earlier

They seem to be implicated in render corruptions. And up to now no one really seems to understand the issue, so let's just disable them for now. Most of the machines exhibiting this issue have only a 128 gtt mmio window, so increased pressure on the mappable part (and so higher chance for cpu relocs) seems to be the key.

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=852210
Cc: Dave Airlie <[email protected]>
Cc: [email protected]
[Note for stable, there's a minor conflict in the context since we've tuned cpu relocs on gen6+ a bit in 3.7.]

---
 drivers/gpu/drm/i915/i915_gem_execbuffer.c |    6 ++++++
 1 file changed, 6 insertions(+)
drivers/gpu/drm/i915/i915_gem_execbuffer.c +6 -0
index 6a2f3e5..4052ac5 100644
@@ -95,6 +95,12 @@ eb_destroy(struct eb_objects *eb)
 
 static inline int use_cpu_reloc(struct drm_i915_gem_object *obj)
 {
+	/* cpu relocs are implicated in some not-yet-understood render
+	 * corruptions on at least ilk, but probably also gm45. Until we know
+	 * what's going on, just disable them. */
+	if (INTEL_INFO(obj->base.dev)->gen < 6)
+		return false;
+
 	return (obj->base.write_domain == I915_GEM_DOMAIN_CPU ||
 		!obj->map_and_fenceable ||
 		obj->cache_level != I915_CACHE_NONE);
Signature · 1 line
1.7.10.4
Acked-by on v1 1/1
#1 lore raw
Quoted 7 lines · Daniel Vetter

They seem to be implicated in render corruptions. And up to now no one really seems to understand the issue, so let's just disable them for now. Most of the machines exhibiting this issue have only a 128 gtt mmio window, so increased pressure on the mappable part (and so higher chance for cpu relocs) seems to be the key.

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=852210

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55112

Quoted 5 lines

Cc: Dave Airlie <[email protected]>
Cc: [email protected]
[Note for stable, there's a minor conflict in the context since we've tuned cpu relocs on gen6+ a bit in 3.7.]

My only query is that the fdo reporter said that "kernel 3.6 + drm-next" didn't trigger the error. So I'd like some confirmation that this patch is indeed needed for 3.7, but meh. (If only because that might give some insight as to what the root cause is.)

-Chris

Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre