| Message ID | 20210930001409.254817-9-jose.souza@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2,1/9] drm/i915/display/psr: Handle plane and pipe restrictions at every page flip | expand |
On Wed, Sep 29, 2021 at 05:14:09PM -0700, José Roberto de Souza wrote: > Not waiting for vblank counter to increment could potentialy cause > issues as commits after the one that needs a modeset could change > state of entities that are not committed into hardware yet. > > Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> > Signed-off-by: José Roberto de Souza <jose.souza@intel.com> > --- > drivers/gpu/drm/i915/display/intel_display.c | 17 +++++++++++++---- > 1 file changed, 13 insertions(+), 4 deletions(-) > > diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c > index 9e9db1b0a907b..9c49b6031cb5e 100644 > --- a/drivers/gpu/drm/i915/display/intel_display.c > +++ b/drivers/gpu/drm/i915/display/intel_display.c > @@ -10494,15 +10494,24 @@ static int intel_atomic_commit(struct drm_device *dev, > * FIXME doing watermarks and fb cleanup from a vblank worker > * (assuming we had any) would solve these problems. > */ > - if (DISPLAY_VER(dev_priv) < 9 && state->base.legacy_cursor_update) { > + if (state->base.legacy_cursor_update) { > struct intel_crtc_state *new_crtc_state; > struct intel_crtc *crtc; > int i; > > - for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) > - if (new_crtc_state->wm.need_postvbl_update || > - new_crtc_state->update_wm_post) > + for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) { > + if (intel_crtc_needs_modeset(new_crtc_state)) { > + state->base.legacy_cursor_update = false; > + break; > + } This combination should be impossible to get. Are you seeing this happening in some case? > + > + if (DISPLAY_VER(dev_priv) < 9 && > + (new_crtc_state->wm.need_postvbl_update || > + new_crtc_state->update_wm_post)) { > state->base.legacy_cursor_update = false; > + break; > + } > + } > } > > ret = intel_atomic_prepare_commit(state); > -- > 2.33.0
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 9e9db1b0a907b..9c49b6031cb5e 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -10494,15 +10494,24 @@ static int intel_atomic_commit(struct drm_device *dev, * FIXME doing watermarks and fb cleanup from a vblank worker * (assuming we had any) would solve these problems. */ - if (DISPLAY_VER(dev_priv) < 9 && state->base.legacy_cursor_update) { + if (state->base.legacy_cursor_update) { struct intel_crtc_state *new_crtc_state; struct intel_crtc *crtc; int i; - for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) - if (new_crtc_state->wm.need_postvbl_update || - new_crtc_state->update_wm_post) + for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) { + if (intel_crtc_needs_modeset(new_crtc_state)) { + state->base.legacy_cursor_update = false; + break; + } + + if (DISPLAY_VER(dev_priv) < 9 && + (new_crtc_state->wm.need_postvbl_update || + new_crtc_state->update_wm_post)) { state->base.legacy_cursor_update = false; + break; + } + } } ret = intel_atomic_prepare_commit(state);
Not waiting for vblank counter to increment could potentialy cause issues as commits after the one that needs a modeset could change state of entities that are not committed into hardware yet. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: José Roberto de Souza <jose.souza@intel.com> --- drivers/gpu/drm/i915/display/intel_display.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-)