[Intel-gfx] [PATCH] drm/i915: Wrap around the tail offset before setting ring->tail

[Tvrtko Ursulin]

On 11/06/2018 09:32, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2018-06-11 09:28:22)
>>
>> On 08/06/2018 18:25, Chris Wilson wrote:
>>> The HW only accepts offsets within ring->size, and fails peculiarly if
>>> the RING_HEAD or RING_TAIL is set to ring->size. Therefore whenever we
>>> set ring->head/ring->tail we want to make sure it is within value (using
>>> intel_ring_wrap()).
>>>
>>> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
>>> Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
>>> Cc: Mika Kuoppala <mika.kuoppala at linux.intel.com>
>>> Cc: Matthew Auld <matthew.william.auld at gmail.com>
>>> Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>> ---
>>>    drivers/gpu/drm/i915/intel_ringbuffer.c |  5 +++++
>>>    drivers/gpu/drm/i915/intel_ringbuffer.h | 12 ++++++++++++
>>>    2 files changed, 17 insertions(+)
>>>
>>> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
>>> index 6ac3b65373fe..9fac0e0f078e 100644
>>> --- a/drivers/gpu/drm/i915/intel_ringbuffer.c
>>> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
>>> @@ -496,6 +496,10 @@ static int init_ring_common(struct intel_engine_cs *engine)
>>>                DRM_DEBUG_DRIVER("%s initialization failed [head=%08x], fudging\n",
>>>                                 engine->name, I915_READ_HEAD(engine));
>>>    
>>> +     /* Check that the ring offsets point within the ring! */
>>> +     GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->head));
>>> +     GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->tail));
>>> +
>>>        intel_ring_update_space(ring);
>>>        I915_WRITE_HEAD(engine, ring->head);
>>>        I915_WRITE_TAIL(engine, ring->tail);
>>> @@ -1064,6 +1068,7 @@ int intel_ring_pin(struct intel_ring *ring,
>>>    
>>>    void intel_ring_reset(struct intel_ring *ring, u32 tail)
>>>    {
>>> +     tail = intel_ring_wrap(ring, tail);
>>>        ring->tail = tail;
>>>        ring->head = tail;
>>>        ring->emit = tail;
>>> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
>>> index b44c67849749..1d8140ac2016 100644
>>> --- a/drivers/gpu/drm/i915/intel_ringbuffer.h
>>> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
>>> @@ -805,6 +805,18 @@ static inline u32 intel_ring_wrap(const struct intel_ring *ring, u32 pos)
>>>        return pos & (ring->size - 1);
>>>    }
>>>    
>>> +static inline bool
>>> +intel_ring_offset_valid(const struct intel_ring *ring, u32 pos)
>>> +{
>>> +     if (pos & -ring->size) /* must be strictly within the ring */
>>> +             return false;
>>> +
>>> +     if (!IS_ALIGNED(pos, 8)) /* must be qword aligned */
>>> +             return false;
>>> +
>>> +     return true;
>>> +}
>>
>> Looks like we have assert_ring_tail_valid and intel_ring_set_tail
>> already in the tree. So could just use the latter in intel_ring_reset if
>> needed.
> 
> No, because that is only setting the tail. The problem also lies in that
> we set RING_HEAD and so need to be careful about the value we assign to
> ring->head.
> 
>> But also since intel_ring_reset is only setting the tail to
>> either zero or existing ring->tail it sounds like the check would be
>> better placed where the tail advances?
> 
> Which we do. The problem is not the tail...

Silly me. Can you just make use of intel_ring_offset_valid from 
assert_ring_tail_valid so we don't duplicate the checks?

Regards,

Tvrtko