[Intel-gfx] [PATCH 3/3] drm/i915: Only grab correct forcewake for the engine with execlists

[Chris Wilson]

On Thu, Apr 07, 2016 at 03:52:46PM +0100, Tvrtko Ursulin wrote:
> 
> 
> On 07/04/16 15:35, Chris Wilson wrote:
> >On Thu, Apr 07, 2016 at 03:05:40PM +0100, Tvrtko Ursulin wrote:
> >>diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
> >>index a1db6a02cf23..cac387f38cf6 100644
> >>--- a/drivers/gpu/drm/i915/intel_lrc.c
> >>+++ b/drivers/gpu/drm/i915/intel_lrc.c
> >>@@ -418,6 +418,7 @@ static void execlists_submit_requests(struct drm_i915_gem_request *rq0,
> >>  				      struct drm_i915_gem_request *rq1)
> >>  {
> >>  	struct drm_i915_private *dev_priv = rq0->i915;
> >>+	unsigned int fw_domains = rq0->engine->fw_domains_elsp;
> >
> >So with a slightly different layout of fw that nest the elsp fw within
> >the tasklet handler's fw I would have a preamble like:
> >
> >fw_domains = 0;
> >for_each_reg({ELSP, WRITE},
> >	     {CONTEXT_STATUS_BUF, READ},
> >	     {CONTEXT_STATUS_PTR, READ | WRITE})
> >	fw_domains |= intel_reg_fw_domains(dev_priv, reg, direction);
> >engine->execlist_fw_domains = fw_domains;
> 
> I actually considered this (or-ing together for all registers)..
> might as well do it now.
> 
> >Hmm, we have a name clash with i915_reg_t i915_mmio_reg and
> >intel_uncore_forcewake_get()
> >
> >intel_uncore_forcewake_for_mmio()
> >i915_mmio_reg_fw_domains()
> 
> intel_uncore_forcewake_for_reg ?

for_read / for_write if you wish to stick with two functions,
for_reg if go with a combined.

We shall leave the i915_mmio_reg_blah() for another day when there is a
clear direction for i915_reg_t.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre