On Thursday 18 December 2008 16:34:11 Olivier Galibert wrote: > On Thu, Dec 18, 2008 at 03:47:23PM -0500, Marc Ferland wrote: > > Seems like the modeline is adjusted by the driver _even_ if I specified > > the LVDSFixedMode "false" option. I really don't know what to do next... > > How can I force my modeline to the driver? > > You can't. Looking at the code, there is no way to tell the system > you know what you're doing. ->mode_fixup is called unconditionally, > and i830_lvds_mode_fixup has no toggle to switch itself off. That > mode fixup function takes lvds_fixed_mode as gospel, and seems to do > part of the chipset programming, so you can't just disable it anyway. > Strange. lvds_fixed_mode itself can only come from bios or ddc, not > config from what I can see. > > It's weird there isn't a generic high level way to tell X not to trust > the roms and trust the config instead, when it comes to modelines, > dpi, things like that. >
Thanks for the info Oliver! I looked up the i830_lvds_mode_fixup function and added a dummy "return TRUE;" just after the : > /* If we don't have a panel mode there's not much we can do */ > if (pI830->lvds_fixed_mode == NULL) > return TRUE; check. Now the modeline seems to be taken into account (from what I see from the register dump log). At least it works for me now... I don't know if anyone from Intel could tell me if this is the default behavior? Like Oliver, I also think this is pretty strange. What's the purpose of having modelines in xorg.conf if there overwritten by the BIOS.... In my case, I use an LCD with no EDID info (this is for an embedded platform). And the BIOS that comes with the board has wrong modelines hardcoded into it. Regards, Marc _______________________________________________ xorg mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/xorg
