On 08/08/14 03:25 PM, Alex Deucher wrote: > On Fri, Aug 8, 2014 at 2:52 PM, Connor Behan <[email protected]> > wrote: >> On 08/08/14 02:14 PM, Alex Deucher wrote: >> >> >> >> Since the RandR commit, we no longer parse the detailed timing table. >> However, this code probably should have been moved instead of deleted. >> The detect hook is a good place to gather this information for DFP >> monitors and the mode_fixup hook is a good place to apply it. >> >> You can drop this patch. With randr we are able to get the display >> timing directly from the EDID and program the hw with that directly. >> This crap was just a vestige of the old days when X didn't rely on >> EDIDs. >> >> Alex >> >> I started writing this patch because r128_output->PanelXRes and >> r128_output->PanelYRes (needed for programming registers) were not being set >> for DVI outputs. If I've understood you correctly, the proper thing to do in >> a randr driver is to set them to mode->HDisplay and mode-VDisplay >> respectively. I will make a patch that does this. > You can ignore the panel stuff for DVI. It was really just an > intermediate place to store the LVDS panel timing from the vbios. > Just use the mode timing (mode->Hdisplay, Vdisplay, etc.) directly. > > Alex > I have tried to store the panel dimensions in mode->HDisplay and mode->VDisplay but the X server keeps undoing this and making them equal to mode->CrtcHDisplay and mode->CrtcVDisplay. Is it really a good idea to store native timings in a non-native mode?
I think I need to keep PanelXRes and PanelYRes around and set them according to three methods: 1. Read the VBIOS. 2. Read the panel size registers. 3. Loop through the DDC gathered modes, looking for M_T_DRIVER | M_T_PREFERRED and then read off HDisplay, VDisplay. 1 and 2 will work for LVDS while 2 and 3 will work for DVI.
signature.asc
Description: OpenPGP digital signature
_______________________________________________ xorg-driver-ati mailing list [email protected] http://lists.x.org/mailman/listinfo/xorg-driver-ati
