| Message ID | 1436311737-18270-6-git-send-email-rodrigo.vivi@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, Jul 07, 2015 at 04:28:56PM -0700, Rodrigo Vivi wrote: > fbdev_set_par is called when fbcon is taking over control, but > frontbuffer was being invalidated only on the first time when > moving obj to GTT domain. > However on the following calls write domain was already GTT > so invalidate was never called again. > > The issue was mainly on boot with plymouth doing a splash screen > when returning to the console frontbuffer wans't being invalidated > causing missed screen updates with PSR enabled. > > Cc: Daniel Vetter <daniel.vetter@ffwll.ch> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> > --- > drivers/gpu/drm/i915/intel_fbdev.c | 10 ++++++++-- > 1 file changed, 8 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c > index 2a1724e..a76cebc 100644 > --- a/drivers/gpu/drm/i915/intel_fbdev.c > +++ b/drivers/gpu/drm/i915/intel_fbdev.c > @@ -50,6 +50,7 @@ static int intel_fbdev_set_par(struct fb_info *info) > struct drm_fb_helper *fb_helper = info->par; > struct intel_fbdev *ifbdev = > container_of(fb_helper, struct intel_fbdev, helper); > + struct drm_i915_gem_object *obj = ifbdev->fb->obj; > int ret; > > ret = drm_fb_helper_set_par(info); > @@ -63,8 +64,13 @@ static int intel_fbdev_set_par(struct fb_info *info) > * now until we solve this for real. > */ > mutex_lock(&fb_helper->dev->struct_mutex); > - ret = i915_gem_object_set_to_gtt_domain(ifbdev->fb->obj, > - true); > + if (obj) { > + if (obj->base.write_domain != I915_GEM_DOMAIN_GTT) > + ret = i915_gem_object_set_to_gtt_domain(obj, > + true); > + else > + intel_fb_obj_invalidate(obj, ORIGIN_GTT); Unconditional intel_fb_obj_invalidate is what we need here really. And for reference maybe mention the commit that fixed this for the set_domain ioctl: commit 031b698a77a70a6c394568034437b5486a44e868 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Jun 26 19:35:16 2015 +0200 drm/i915: Unconditionally do fb tracking invalidate in set_domain > + } > mutex_unlock(&fb_helper->dev->struct_mutex); > } > > -- > 2.1.0 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index 2a1724e..a76cebc 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -50,6 +50,7 @@ static int intel_fbdev_set_par(struct fb_info *info) struct drm_fb_helper *fb_helper = info->par; struct intel_fbdev *ifbdev = container_of(fb_helper, struct intel_fbdev, helper); + struct drm_i915_gem_object *obj = ifbdev->fb->obj; int ret; ret = drm_fb_helper_set_par(info); @@ -63,8 +64,13 @@ static int intel_fbdev_set_par(struct fb_info *info) * now until we solve this for real. */ mutex_lock(&fb_helper->dev->struct_mutex); - ret = i915_gem_object_set_to_gtt_domain(ifbdev->fb->obj, - true); + if (obj) { + if (obj->base.write_domain != I915_GEM_DOMAIN_GTT) + ret = i915_gem_object_set_to_gtt_domain(obj, + true); + else + intel_fb_obj_invalidate(obj, ORIGIN_GTT); + } mutex_unlock(&fb_helper->dev->struct_mutex); }
fbdev_set_par is called when fbcon is taking over control, but frontbuffer was being invalidated only on the first time when moving obj to GTT domain. However on the following calls write domain was already GTT so invalidate was never called again. The issue was mainly on boot with plymouth doing a splash screen when returning to the console frontbuffer wans't being invalidated causing missed screen updates with PSR enabled. Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> --- drivers/gpu/drm/i915/intel_fbdev.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-)