On Fri, 15 Apr 2011 15:13:02 -0400 Adam Jackson <ajax at redhat.com> wrote:
> On 4/15/11 2:40 PM, Jesse Barnes wrote: > > > @@ -1461,6 +1462,15 @@ static void drm_add_display_info(struct edid *edid, > > info->bpp = 0; > > break; > > } > > + > > + if (edid->features & DRM_EDID_FEATURE_RGB) > > + info->color_formats = DRM_COLOR_FORMAT_RGB444; > > + else if (edid->features & DRM_EDID_FEATURE_RGB_YCRCB444) > > + info->color_formats = DRM_COLOR_FORMAT_RGB444 | > > + DRM_COLOR_FORMAT_YCRCB444; > > + else if (edid->features & DRM_EDID_FEATURE_RGB_YCRCB) > > + info->color_formats = DRM_COLOR_FORMAT_RGB444 | > > + DRM_COLOR_FORMAT_YCRCB444 | DRM_COLOR_FORMAT_YCRCB422; > > Overcomplicated (RGB is numerically 0, YCRCB is just two other values > or'd together) and wrong (doesn't handle YCbCr 4:2:2 alone). You want: Arg, of course I have to mask out the value first, I'll fix that (my current test display conveniently sets RGB_YCRCB so I missed this in testing). > > info->color_formats = DRM_COLOR_FORMAT_RGB444; > if (edid->features & DRM_EDID_FEATURE_YCRCB444) > info->color_formats |= DRM_COLOR_FORMAT_YCBCR444; > if (edid->features & DRM_EDID_FEATURE_YCRCB422) > info->color_formats |= DRM_COLOR_FORMAT_YCBCR422; > > ... which is corrected to not include RGB uselessly in the > DRM_EDID_FEATURE_* tokens. I should have noticed that in your first > patch, whoops. I don't think EDID supports that? The docs I have here imply that either RGB, RGB + YCrCb444 or RGB + YCrCb444 + YCrCb422 are the only things we can report. Or is there a CEA block extension that allows for more granularity? -- Jesse Barnes, Intel Open Source Technology Center
