Evergreen success story!

Alex Deucher alexdeucher at gmail.com
Mon Feb 22 08:03:44 PST 2010


On Sun, Feb 21, 2010 at 9:46 AM, Dave Witbrodt <dawitbro at sbcglobal.net> wrote:
> Alex Deucher wrote:
>>
>> On Fri, Feb 19, 2010 at 10:57 PM, Dave Witbrodt <dawitbro at sbcglobal.net>
>> wrote:
>
> [...]
>>>
>>> I tried a week ago, and when I got black screens I gave up too quickly
>>> and pulled the HD 5750, replacing it with my slightly older HD 4850.
>>> This time I also got a black screen, but (as before) Windows Vista works
>>> perfectly, so I know the hardware is not defective.
>>>
>>> As it turns out, I use a KVM device that allows me to switch between 4
>>> machines using Dsub VGA connectors.  There are no VGA connectors on this
>>> HD 5750, so I was using a DVI-to-Dsub converter that came with the card
>>> so I could continue to switch between machines.  After logging in via
>>> SSH, and making sure I saved 'dmesg' and 'Xorg.0.log', I thought about
>>> that converter possibly messing up the situation.
>>>
>>> I turned the monitor off, plugged in a DVI cable, and rebooted.  Voila!
>>>  Working KMS!  No acceleration, of course, but I'm just glad to see the
>>> ball toward open source Evergreen support starting to roll!  Switching
>>> between X and VT works fine, as well.  And (unlike last week) I am
>>> getting the proper 1920x1200 resolution for my monitor.
>>>
>> KVMs are always trouble.  If you use one, your best bet is to force an
>> X configuration via config file as they often prevent the driver from
>> detecting any outputs.  The UMS driver lights up one of the ports if
>> nothing is detected, you just happened to get lucky that it lit up the
>> right port for you, with KMS, we don't.
>
> My KVM worked fine with KMS + HD 4850.  There must be features missing in
> the Evergreen KMS compared to the support for earlier devices.
>
> At any rate, using DVI instead of Dsub cured my black screen issues.  I
> still use the KVM, but have to manually switch the monitor between Dsub and
> DVI.

Ah, ok.  It could be an issue with the load detection code then.  I'll
take a look.

Alex


More information about the xorg-driver-ati mailing list