[Intel-gfx] [PATCH 1/4] drm/i915: Do a synchronous switch-to-kernel-context on idling

[Tvrtko Ursulin]

On 08/03/2019 12:46, Chris Wilson wrote:
> Quoting Chris Wilson (2019-03-08 12:44:16)
>> Quoting Chris Wilson (2019-03-08 09:36:54)
>>> +static bool switch_to_kernel_context_sync(struct drm_i915_private *i915)
>>> +{
>>> +       bool result = true;
>>> +
>>> +       /*
>>> +        * Even if we fail to switch, give whatever is running a small chance
>>> +        * to save itself before we report the failure. Yes, this may be a
>>> +        * false positive due to e.g. ENOMEM, caveat emptor!
>>> +        */
>>> +       if (i915_gem_switch_to_kernel_context(i915))
>>> +               result = false;
>>> +
>>> +       if (i915_gem_wait_for_idle(i915,
>>> +                                  I915_WAIT_LOCKED |
>>> +                                  I915_WAIT_FOR_IDLE_BOOST,
>>> +                                  I915_GEM_IDLE_TIMEOUT))
>>> +               result = false;
>>> +
>>> +       if (result) {
>>> +               assert_kernel_context_is_current(i915);
>>> +       } else {
>>
>>          if (intel_has_gpu_reset()) {
>>
>>> +               /* Forcibly cancel outstanding work and leave the gpu quiet. */
>>> +               dev_err(i915->drm.dev,
>>> +                       "Failed to idle engines, declaring wedged!\n");
>>> +               GEM_TRACE_DUMP();
>>
>>          }
>>
>> To prevent it spewing just for gem_eio???
>>
>> I think the error message is important, but gem_eio is a bit of an arm
>> twister.
> 
> On the other hand, gem_eio does show that this can be user invoked so is
> anything above a dev_notice() required?

Varying log level depending on intel_has_gpu_reset sounds like the answer.

Regards,

Tvrtko