| Message ID | 20180919195544.1511-4-chris@chris-wilson.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [01/40] drm: Use default dma_fence hooks where possible for null syncobj | expand |
On 19/09/2018 20:55, Chris Wilson wrote: > Once we have flushed the first request through the system to both load a > context and record the default state; tell the GPU to park and idle > itself, putting itself immediately (hopefully at least) into a > powersaving state, and allowing ourselves to start from known state > after setting up all our bookkeeping. > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> > --- > drivers/gpu/drm/i915/i915_gem.c | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > index a94d5a308c4d..6b347ffb996b 100644 > --- a/drivers/gpu/drm/i915/i915_gem.c > +++ b/drivers/gpu/drm/i915/i915_gem.c > @@ -5417,6 +5417,14 @@ static int __intel_engines_record_defaults(struct drm_i915_private *i915) > > assert_kernel_context_is_current(i915); > > + /* > + * Immediately park the GPU so that we enable powersaving and > + * treat it as idle. The next time we issue a request, we will > + * unpark and start using the engine->pinned_default_state, otherwise > + * it is in limbo and an early reset may fail. > + */ > + __i915_gem_park(i915); > + > for_each_engine(engine, i915, id) { > struct i915_vma *state; > void *vaddr; > Presumably the previous patch will make this pass CI? I would prefer the parking is done from the caller level but we couldn't agree on this minor detail so anyway: Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Regards, Tvrtko
Quoting Tvrtko Ursulin (2018-09-20 15:02:36) > > On 19/09/2018 20:55, Chris Wilson wrote: > > Once we have flushed the first request through the system to both load a > > context and record the default state; tell the GPU to park and idle > > itself, putting itself immediately (hopefully at least) into a > > powersaving state, and allowing ourselves to start from known state > > after setting up all our bookkeeping. > > > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > > Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> > > --- > > drivers/gpu/drm/i915/i915_gem.c | 8 ++++++++ > > 1 file changed, 8 insertions(+) > > > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > > index a94d5a308c4d..6b347ffb996b 100644 > > --- a/drivers/gpu/drm/i915/i915_gem.c > > +++ b/drivers/gpu/drm/i915/i915_gem.c > > @@ -5417,6 +5417,14 @@ static int __intel_engines_record_defaults(struct drm_i915_private *i915) > > > > assert_kernel_context_is_current(i915); > > > > + /* > > + * Immediately park the GPU so that we enable powersaving and > > + * treat it as idle. The next time we issue a request, we will > > + * unpark and start using the engine->pinned_default_state, otherwise > > + * it is in limbo and an early reset may fail. > > + */ > > + __i915_gem_park(i915); > > + > > for_each_engine(engine, i915, id) { > > struct i915_vma *state; > > void *vaddr; > > > > Presumably the previous patch will make this pass CI? Aye. There was an earlier series of just this pair to check that I had caught all the missing wakerefs uncovered by idling on init. > I would prefer the parking is done from the caller level but we couldn't > agree on this minor detail so anyway: I wanted to postpone that idea until we have the rc6 setup refactored, as I think that may open some more interesting avenues of how to handle this. -Chris
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index a94d5a308c4d..6b347ffb996b 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -5417,6 +5417,14 @@ static int __intel_engines_record_defaults(struct drm_i915_private *i915) assert_kernel_context_is_current(i915); + /* + * Immediately park the GPU so that we enable powersaving and + * treat it as idle. The next time we issue a request, we will + * unpark and start using the engine->pinned_default_state, otherwise + * it is in limbo and an early reset may fail. + */ + __i915_gem_park(i915); + for_each_engine(engine, i915, id) { struct i915_vma *state; void *vaddr;
Once we have flushed the first request through the system to both load a context and record the default state; tell the GPU to park and idle itself, putting itself immediately (hopefully at least) into a powersaving state, and allowing ourselves to start from known state after setting up all our bookkeeping. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> --- drivers/gpu/drm/i915/i915_gem.c | 8 ++++++++ 1 file changed, 8 insertions(+)