On Wed, Nov 01, 2017 at 04:21:02PM +0200, Jani Nikula wrote: > diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c > index e4751f92b342..e0a190a0f029 100644 > --- a/drivers/gpu/drm/nouveau/nv50_display.c > +++ b/drivers/gpu/drm/nouveau/nv50_display.c > @@ -2688,7 +2688,6 @@ nv50_audio_enable(struct drm_encoder *encoder, struct drm_display_mode *mode) > if (!drm_detect_monitor_audio(nv_connector->edid)) > return; > > - drm_edid_to_eld(&nv_connector->base, nv_connector->edid); > memcpy(args.data, nv_connector->base.eld, sizeof(args.data)); > > nvif_mthd(disp->disp, 0, &args, This being the only call outside a .get_modes() hook means this is the only one that might change some behaviour. Looks like nv_connector->edid gets updated in .detect() as one might expect, so in theory we might get a mismatch between the drm_detect_monitor_audio(edid) check above and the actual eld contents here if .detect() gets called without .get_modes(). Not a problem for the probe helper's .fill_modes(), but eg. the poll helper does do that. I guess we could always consider doing the edid_to_eld() already from .detect()/.force() for all drivers if we think there would be a good reason to populate the ELD even if .get_modes()/.fill_modes() hasn't been called. -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
