[PATCH savage] Refactor BIOS modes retrieval to only call VBEGetVBEInfo once

Tormod Volden lists.tormod at gmail.com
Fri Feb 17 12:24:11 PST 2012


On Thu, Feb 16, 2012 at 11:35 PM, Julien Cristau <jcristau at debian.org> wrote:
> On Wed, Feb 15, 2012 at 23:07:12 +0100, Tormod Volden wrote:
>
>> From: Tormod Volden <debian.tormod at gmail.com>
>>
>> Signed-off-by: Tormod Volden <debian.tormod at gmail.com>
>> ---
>>
>> The change should be fine by itself, but the reason I changed this is
>> that with the latest stack (upgraded to 1.12RC etc) the second call to
>> VBEGetVBEInfo would return corrupted values. There, I fixed it!(TM) :)
>>
>> So was the issue due to bugs in the savage driver, or is there something
>> more general VBE stuff broken? I also noticed that the vesa driver messes
>> up now, the computer hardlocks when I try to change VT. I haven't had time
>> to investigate that much further so I don't know if it is related.
>>
>> Anyone else seeing strange stuff with VBE-enabled drivers?
>>
> ajax was having fun in vesa-land recently AIUI (by which I mean was
> trying to make it work again with 1.12), so it seems like it's not just
> you.

Julien, thanks! /me googles "ajax vesa"

Ajax, was that this issue:
"libpciaccess-lol-dev-port.patch: Don't use /dev/port since the kernel insists"
http://pkgs.fedoraproject.org/gitweb/?p=libpciaccess.git;a=commitdiff;h=c2d89ecd4628ee6953986ee564a0052b34771178

or is there more to it?

That patch does not fix my vesa lock-up. BTW, the vesa lock-up happens
on xserver 1.11.3rc and 1.12rc, kernel 3.0.16 and 3.2.6.

Cheers,
Tormod


>
> Cheers,
> Julien


More information about the xorg-devel mailing list