drm/i915: fixup runtime PM handling v2
authorJesse Barnes <jbarnes@virtuousgeek.org>
Wed, 23 Sep 2015 21:37:17 +0000 (14:37 -0700)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Wed, 30 Sep 2015 08:20:01 +0000 (10:20 +0200)
According to the PCI docs and Rafael, we don't need to be doing explicit
enables and disables in our init and teardown routines, as they're taken
care of by the PCI core.  So drop the pm_runtime_disable() at teardown
and pm_runtime_set_active() at init.

This fixes one failure of the basic-pci-d3-state test on my BYT.

v2: drop extra get_noresume() and put_noidle() (Rafael)

Signed-off-by: Jesse Barnes <jbarnes at virtuousgeek.org>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Acked-by: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drivers/gpu/drm/i915/intel_runtime_pm.c

index 4a815bb6cfcaf42b3252e2d286f6ce7f0c57dd45..e1fdbabaf2bfa80b7398ed45aa8db7c24e71c0e8 100644 (file)
@@ -1828,7 +1828,6 @@ static void intel_runtime_pm_disable(struct drm_i915_private *dev_priv)
 
        /* Make sure we're not suspended first. */
        pm_runtime_get_sync(device);
-       pm_runtime_disable(device);
 }
 
 /**
@@ -2120,8 +2119,6 @@ void intel_runtime_pm_enable(struct drm_i915_private *dev_priv)
        if (!HAS_RUNTIME_PM(dev))
                return;
 
-       pm_runtime_set_active(device);
-
        /*
         * RPM depends on RC6 to save restore the GT HW context, so make RC6 a
         * requirement.
This page took 0.038551 seconds and 5 git commands to generate.