* re: drm/i915: Decouple execbuf uAPI from internal implementation
@ 2016-01-28 22:30 Dan Carpenter
2016-01-29 11:15 ` [Intel-gfx] " Tvrtko Ursulin
0 siblings, 1 reply; 2+ messages in thread
From: Dan Carpenter @ 2016-01-28 22:30 UTC (permalink / raw)
To: tvrtko.ursulin; +Cc: intel-gfx, dri-devel
Hello Tvrtko Ursulin,
The patch de1add360522: "drm/i915: Decouple execbuf uAPI from
internal implementation" from Jan 15, 2016, leads to the following
static checker warning:
drivers/gpu/drm/i915/i915_gem_execbuffer.c:1411 eb_select_ring()
warn: buffer overflow 'dev_priv->ring' 5 <= 16385
drivers/gpu/drm/i915/i915_gem_execbuffer.c
1397 if (user_ring_id == I915_EXEC_BSD && HAS_BSD2(dev_priv)) {
1398 unsigned int bsd_idx = args->flags & I915_EXEC_BSD_MASK;
1399
1400 if (bsd_idx == I915_EXEC_BSD_DEFAULT) {
1401 bsd_idx = gen8_dispatch_bsd_ring(dev_priv, file);
1402 } else if (bsd_idx >= I915_EXEC_BSD_RING1 &&
1403 bsd_idx <= I915_EXEC_BSD_RING2) {
1404 bsd_idx--;
^^^^^^^^^
This should probablye be "bsd_idx = (bsd_idx >> 13) - 1;" or something.
1405 } else {
1406 DRM_DEBUG("execbuf with unknown bsd ring: %u\n",
1407 bsd_idx);
1408 return -EINVAL;
1409 }
1410
1411 *ring = &dev_priv->ring[_VCS(bsd_idx)];
Otherwise we're way past the end of this array.
1412 } else {
1413 *ring = &dev_priv->ring[user_ring_map[user_ring_id]];
1414 }
1415
1416 if (!intel_ring_initialized(*ring)) {
1417 DRM_DEBUG("execbuf with invalid ring: %u\n", user_ring_id);
1418 return -EINVAL;
1419 }
regards,
dan carpenter
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [Intel-gfx] drm/i915: Decouple execbuf uAPI from internal implementation
2016-01-28 22:30 drm/i915: Decouple execbuf uAPI from internal implementation Dan Carpenter
@ 2016-01-29 11:15 ` Tvrtko Ursulin
0 siblings, 0 replies; 2+ messages in thread
From: Tvrtko Ursulin @ 2016-01-29 11:15 UTC (permalink / raw)
To: Dan Carpenter, tvrtko.ursulin; +Cc: intel-gfx, dri-devel
Hi Dan,
On 28/01/16 22:30, Dan Carpenter wrote:
> Hello Tvrtko Ursulin,
>
> The patch de1add360522: "drm/i915: Decouple execbuf uAPI from
> internal implementation" from Jan 15, 2016, leads to the following
> static checker warning:
>
> drivers/gpu/drm/i915/i915_gem_execbuffer.c:1411 eb_select_ring()
> warn: buffer overflow 'dev_priv->ring' 5 <= 16385
>
> drivers/gpu/drm/i915/i915_gem_execbuffer.c
> 1397 if (user_ring_id == I915_EXEC_BSD && HAS_BSD2(dev_priv)) {
> 1398 unsigned int bsd_idx = args->flags & I915_EXEC_BSD_MASK;
> 1399
> 1400 if (bsd_idx == I915_EXEC_BSD_DEFAULT) {
> 1401 bsd_idx = gen8_dispatch_bsd_ring(dev_priv, file);
> 1402 } else if (bsd_idx >= I915_EXEC_BSD_RING1 &&
> 1403 bsd_idx <= I915_EXEC_BSD_RING2) {
> 1404 bsd_idx--;
> ^^^^^^^^^
> This should probablye be "bsd_idx = (bsd_idx >> 13) - 1;" or something.
>
> 1405 } else {
> 1406 DRM_DEBUG("execbuf with unknown bsd ring: %u\n",
> 1407 bsd_idx);
> 1408 return -EINVAL;
> 1409 }
> 1410
> 1411 *ring = &dev_priv->ring[_VCS(bsd_idx)];
>
> Otherwise we're way past the end of this array.
Yep, we have already found this and fixed it in "drm/i915: Fix VCS ring
selection after uapi decoupling".
Thanks for the report, it is very useful!
Regards,
Tvrtko
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2016-01-29 11:15 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-01-28 22:30 drm/i915: Decouple execbuf uAPI from internal implementation Dan Carpenter
2016-01-29 11:15 ` [Intel-gfx] " Tvrtko Ursulin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox