Mailinglist Archive: radeonhd (308 mails)

["Adam J. Richter" <adam@xxxxxxxxxxxxx>]

On Wed, Mar 19, 2008 at 11:51:50PM -0700, Adam J. Richter wrote:
> On Mon, Mar 17, 2008 at 02:42:40PM +0100, Matthias Hopf wrote:
> > Current git of radeonhd should support dual link DVI as well, though
> > this is untested ATM.
>
>       I just tried the radeonhd git head (pulled minutes ago) on a
> 30 inch 2560x1600 HP lp3065 monitor, and the screen is black with the
> green LED flashing.  This same screen displays BIOS messages, lilo
> text and the character mode linux console fine, as I think there is
> some built in video stretching done in character mode (for example, in
> character mode, my 1920x1200 monitor always claims to be receiving a
> 1920x1200 signal rather than 640x480 or something similar that one
> would normally associate with character mode).  xrandr, correctly
> describes the video modes for my 1920x1200 monitor on the VGA cable
> and the 2560x1600 monitor on VGA, and indicates that 2560x1600 is
> selected, and the virtual screen size is expanded to 2560x1600.  So, I
> think this is a mode setting issue and not some misconfiguration on my
> part.
>
>       My new lp3065 can select between three dual-link DVI input
> cables, so I should be able to plug it into an nvidia card for now to
> use it immediately while still being available to test new changes
> through a cable connect to my AMD motherboard.  So, if anyone wants me
> to try a patch and report results, I am happy to do so when I have time.
>
>       Also, if anyone would like to point me to the relevant page
> numbers of the recently released X docs for the rs690 in
> http://www.x.org/docs/AMD/ for enabling dual link, I may be able to
> try to work on this myself during the weekend.  I had previously
> skimmed AMD's recently released documents but didn't immediately find
> a bit to set to activate dual link.

        Oops, sorry for sending the above message to the wrong mailing
list (xf86-driver-ati@xxxxxxxxxxx instead of radeonhd@xxxxxxxxxxxx).

        I also have one additional data point to add to the above
message.  Using xrandr to select the 1280x800 mode for the monitor
does not result in viewable video, even though this mode should not
require dual link.  This is different from the behavior of the radeon
driver in xf86-driver-ati (also the git head version), which displays
bad video at 2560x1600 but displays correctly if I use xrandr to
change the video mode to 1280x800.

        Anyhow, I hope this information is helpful.

Adam Richter
-- 
To unsubscribe, e-mail: radeonhd+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: radeonhd+help@xxxxxxxxxxxx