[Intel-gfx] [PATCH 09/21] drm/i915: Make i915_check_and_clear_faults take uncore

[Tvrtko Ursulin]

On 06/06/2019 10:57, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2019-06-06 10:36:27)
>> From: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>
>> Continuing the conversion and elimination of implicit dev_priv.
>>
>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>> Suggested-by: Rodrigo Vivi <rodrigo.vivi at intel.com>
>> ---
>>   drivers/gpu/drm/i915/gt/intel_engine_cs.c |  2 +-
>>   drivers/gpu/drm/i915/i915_drv.c           |  2 +-
>>   drivers/gpu/drm/i915/i915_gem_gtt.c       | 34 ++++++++++++-----------
>>   drivers/gpu/drm/i915/i915_gem_gtt.h       |  2 +-
>>   4 files changed, 21 insertions(+), 19 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>> index 0e9b74f52503..3554d0dd7b1a 100644
>> --- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>> +++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>> @@ -453,7 +453,7 @@ int intel_engines_init_mmio(struct drm_i915_private *i915)
>>   
>>          RUNTIME_INFO(i915)->num_engines = hweight32(mask);
>>   
>> -       i915_check_and_clear_faults(i915);
>> +       i915_check_and_clear_faults(&i915->uncore);
> 
> I am not sold on that. Especially as it is then unwrapped back to i915.

It isn't really, not on the logical level. This is the body:

void i915_check_and_clear_faults(struct intel_uncore *uncore)
{
	struct drm_i915_private *i915 = uncore_to_i915(uncore);

	/* From GEN8 onwards we only have one 'All Engine Fault Register' */
	if (INTEL_GEN(i915) >= 8)
		gen8_check_faults(uncore);
	else if (INTEL_GEN(i915) >= 6)
		gen6_check_faults(uncore);
	else
		return;

	uncore_clear_error_registers(uncore, ALL_ENGINES);
}

So the idea being i915 is used only for "what gen am I checkes",
while the actual functionality operates on uncore.

Regards,

Tvrtko