On Mon, May 26, 2014 at 02:46:32PM +0300, ville.syrjala@linux.intel.com wrote:
From: Ville Syrjälä ville.syrjala@linux.intel.com
The new watermaek update mechanism requires interrupts to work correctly. Because of this we need interrupts while disabling crtcs during suspend. So move the irq disable to happen a bit later.
This also avoid clobbering the vblank.last count in case the vblank interrupt was already disabled earlier by the timer. In that case drm_vblank_off() will need .last to be correct so that it can update the user visible vblank counter value approapriately.
v2: Note vblank counter in commit message
Signed-off-by: Ville Syrjälä ville.syrjala@linux.intel.com
First round of this patch had a question from me whether we have sufficient amounts of WARN_ON(! or !!dev->irqs_enabled) sprinkled all over the place ... Do we have that?
I'm specifically thinking of places where we cancel some work/time lauchned by interrupts and must have the guarantee in place that it can't rearm.
Otoh we have places where we absolutely want interrupts to still work, e.g. in gpu_idle.
Sprinkling some modeset amounts of WARNS in a follow-up would make me really happy. -Daniel
drivers/gpu/drm/i915/i915_drv.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 4619c9e..21554a0 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -524,7 +524,6 @@ static int i915_drm_freeze(struct drm_device *dev) return error; }
drm_irq_uninstall(dev);dev_priv->enable_hotplug_processing = false;
intel_disable_gt_powersave(dev);
@@ -541,6 +540,8 @@ static int i915_drm_freeze(struct drm_device *dev) } mutex_unlock(&dev->mode_config.mutex);
drm_irq_uninstall(dev);- intel_modeset_suspend_hw(dev); }
-- 1.8.5.5
Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx