[Intel-gfx] [PATCH v2 1/4] drm/i915: Initialize hardware semaphore state on ring init

[Mika Kuoppala]

Chris Wilson <chris at chris-wilson.co.uk> writes:

> On Tue, 18 Dec 2012 19:26:18 +0200, Mika Kuoppala <mika.kuoppala at linux.intel.com> wrote:
>> Hardware status page needs to have proper seqno set
>> as our initial seqno can be arbitrary. If initial seqno is close
>> to wrap boundary on init and i915_seqno_passed() (31bit space)
>> refers to hw status page which contains zero, errorneous result
>> will be returned.
>> 
>> v2: clear mboxes and set hws page directly instead of going
>> through rings. Suggested by Chris Wilson.
>> 
>> v3: hws needs to be updated for all gens. Noticed by Chris
>> Wilson.
>> 
>> References: https://bugs.freedesktop.org/show_bug.cgi?id=58230
>> Signed-off-by: Mika Kuoppala <mika.kuoppala at intel.com>
>
>> -int intel_ring_handle_seqno_wrap(struct intel_ring_buffer *ring)
>> +void intel_ring_init_seqno(struct intel_ring_buffer *ring, u32 seqno)
>>  {
>> -	int ret;
>> +	struct drm_i915_private *dev_priv = ring->dev->dev_private;
>>  
>>  	BUG_ON(ring->outstanding_lazy_request);
>>  
>> -	if (INTEL_INFO(ring->dev)->gen < 6)
>> -		return 0;
>> -
>> -	ret = __intel_ring_begin(ring, 6 * sizeof(uint32_t));
>> -	if (ret)
>> -		return ret;
>> -
>> -	/* Leaving a stale, pre-wrap seqno behind in the mboxes will result in
>> -	 * post-wrap semaphore waits completing immediately. Clear them. */
>> -	update_mboxes(ring, ring->signal_mbox[0]);
>> -	update_mboxes(ring, ring->signal_mbox[1]);
>> -	intel_ring_advance(ring);
>> +	if (INTEL_INFO(ring->dev)->gen >= 6) {
>> +		I915_WRITE(RING_SYNC_0(ring->mmio_base), 0);
>> +		I915_WRITE(RING_SYNC_1(ring->mmio_base), 0);
>> +	}
>>  
>> -	return 0;
>> +	intel_write_status_page(ring, I915_GEM_HWS_INDEX, seqno);
>
> There's yet another complication here: not all gen use the HWS. Ironlake
> uses a pipecontrol write to another page -- though in principle we could
> write to the HWS instead.
>

I posted a new series which adds ring::set_seqno to take care per
architecture seqno writing...
 
> Any one feel like that challenge?

... and i will rip the ring:set_seqno out when ilk uses hws.

-Mika

> -Chris
>
> -- 
> Chris Wilson, Intel Open Source Technology Centre