From 855d30b402b91f09c90f65c34ec91debaae8cf3a Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Wed, 10 Sep 2014 17:36:09 +0200 Subject: [PATCH] drm: Only update final vblank count when precise ts is available MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Drivers without a hardware vblank counter simply can't account for the vblanks that happened while the vblank interrupt was off. To check this grab a vblank timestamp and if the result is dubious follow the normal save-and-disable logic. Drivers should prevent this by setting vblank_disable_allowed = false, but since running vblank interrupts constantly is not good for power consumption most drivers lie. Testing for precise vblank timestamps is the next best thing we can check for. Suggested-by: Mario Kleiner Cc: Mario Kleiner Cc: Matt Roper Cc: Ville Syrjälä Reviewed-by: Mario Kleiner Signed-off-by: Daniel Vetter --- drivers/gpu/drm/drm_irq.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c index 87d148c..ad699b7 100644 --- a/drivers/gpu/drm/drm_irq.c +++ b/drivers/gpu/drm/drm_irq.c @@ -148,8 +148,15 @@ static void vblank_disable_and_save(struct drm_device *dev, int crtc) * has been ticking all along until this time. This makes the * count account for the entire time between drm_vblank_on() and * drm_vblank_off(). + * + * But only do this if precise vblank timestamps are available. + * Otherwise we might read a totally bogus timestamp since drivers + * lacking precise timestamp support rely upon sampling the system clock + * at vblank interrupt time. Which obviously won't work out well if the + * vblank interrupt is disabled. */ - if (!vblank->enabled) { + if (!vblank->enabled && + drm_get_last_vbltimestamp(dev, crtc, &tvblank, 0) > 0) { drm_update_vblank_count(dev, crtc); spin_unlock_irqrestore(&dev->vblank_time_lock, irqflags); return; -- 2.7.4