New DRM model

[Dave Dodge]

On Tue, Feb 10, 2004 at 04:08:21PM +1100, Jaymz Julian wrote:
> On Mon, Feb 09, 2004 at 06:18:31PM -0800, Jon Smirl wrote:
[...]
> Be aware, btw, that some monitors return bogus EDID, so you need to
> have this be be able to be overridden.

I'll second this, as I have one of these EDID liars. Otherwise it's a
great monitor. Its circuitry can readily handle 1920x1440 resolution,
but its EDID only gets you to 1600x1200. Any X server I use needs
to be able to override or ignore EDID on demand so that I can get the
better resolution.

> > It's a lot of work supporting non-EDID monitors. Are they common
> > enough anymore to bother? I retired all of mine about four years
> > ago.

Don't forget about KVM switches. They don't all pass EDID even if the
monitor supports it. In fact that's how I found out about my monitor's
problem in the first place -- I took my KVM out of the loop one day
for some unrelated testing and when I logged in my desktop came up
too small.

Also it should be possible to boot, even to graphical login, without
a powered-on monitor attached at the moment. Especially in remote-reboot
or power loss situations.

Example: some configurations of Sparc 20 hardware would disable video
output (but not keyboard input) at boot time if a monitor wasn't
attached. So one of these sits somewhere in a rack with monitor, and
is remotely rebooted. What the admin doesn't know is that someone else
working on the rack has temporarily attached the monitor to a
different nearby system. So our Sparc quietly comes up with no video
output. Then a few months later it starts having some unrelated
trouble, and the admin rushes to the rack to use a console login to
investigate...

Been there, done that, not fun.
                                                  -Dave Dodge