On 10/16/20 7:44 PM, Sam Ravnborg wrote: > Hi Marek. Hi, > On Sat, Oct 03, 2020 at 01:07:26AM +0200, Marek Vasut wrote: >> The drm_display_mode_to_videomode() does not populate DISPLAY_FLAGS_DE_LOW >> or DISPLAY_FLAGS_PIXDATA_NEGEDGE flags in struct videomode. > > So after reading this paragrahp I assumed this patch would fix this, alas > there is no changes to drm_modes.c. > Rather than introducing hacks could we try to fix the root cause here? Nope, this is a driver fix, hence the drm/stm tags. > Root cause - as you point out is that there is missing some flags in > the conversion. > Another point is that the stm driver could drop the use of > drm_display_mode_to_videomode() - all the information is available in drm_display_mode > and using drm_display_mode_to_videomode() does not help much. I kinda already have such a patch in the tree to remove this, but I want to have a patch which fixes the problem described here and can be easily backported to stable too. > If the point is to have easier access to hsync and friendns maybe we > should have a few helpers operating direct on display_mode and not this > conversion to video_mode which is something that belongs to fbdev and > should IMO not be used inside a drm driver. Yep, that is for another patch though, because if that one breaks something, I want to be able to bisect it. _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
