2003-05-02 22:55:28

by Balram Adlakha

[permalink] [raw]
Subject: wrong screen size with fbcon [2.5.68]

I posted about this when 2.5.68 was released but very few people
responded.
While using the framebuffer console driver, I seems that the screen is
set to something like 1024x775 instead of 1024x768. I cannot see the
bottom of my screen that is...
I just checked the latest bk taken from kernel.org and still hasn't been
fixed.
The fb console was working perfectly till 2.5.67.

If it is of any relevance, I'm using an nvidia tnt2 card. Has nobody
noticed this problem?
--


2003-05-03 00:39:06

by Greg KH

[permalink] [raw]
Subject: Re: wrong screen size with fbcon [2.5.68]

On Sat, May 03, 2003 at 04:36:37AM +0530, Balram Adlakha wrote:
> I posted about this when 2.5.68 was released but very few people
> responded.
> While using the framebuffer console driver, I seems that the screen is
> set to something like 1024x775 instead of 1024x768. I cannot see the
> bottom of my screen that is...
> I just checked the latest bk taken from kernel.org and still hasn't been
> fixed.
> The fb console was working perfectly till 2.5.67.
>
> If it is of any relevance, I'm using an nvidia tnt2 card. Has nobody
> noticed this problem?

I have, and posted pretty much the same report a few weeks ago. Time to
go write up a bug in bugzilla.kernel.org...

greg k-h

2003-05-03 15:58:57

by Greg KH

[permalink] [raw]
Subject: Re: wrong screen size with fbcon [2.5.68]

On Sat, May 03, 2003 at 03:05:04PM +0100, Riley Williams wrote:
> Hi Greg.
>
> >> I posted about this when 2.5.68 was released but very few people
> >> responded. While using the framebuffer console driver, I seems
> >> that the screen is set to something like 1024x775 instead of
> >> 1024x768. I cannot see the bottom of my screen that is...
> >>
> >> I just checked the latest bk taken from kernel.org and still
> >> hasn't been fixed. The fb console was working perfectly till
> >> 2.5.67.
> >>
> >> If it is of any relevance, I'm using an NVIDIA tnt2 card. Has
> >> nobody noticed this problem?
>
> > I have, and posted pretty much the same report a few weeks ago.
> > Time to go write up a bug in bugzilla.kernel.org...
>
> The probable response will be WONTFIX on the grounds that the NVIDIA
> driver is a binary only driver and, as such, can't be fixed...

No, I'm using the VESA framebuffer driver that is in the kernel. This
same thing happens on two different machines with two different video
cards.

No binary only driver for me, don't you think I would know better? :)

thanks,

greg k-h

2003-05-03 16:13:12

by James Simmons

[permalink] [raw]
Subject: Re: wrong screen size with fbcon [2.5.68]


> > >> I posted about this when 2.5.68 was released but very few people
> > >> responded. While using the framebuffer console driver, I seems
> > >> that the screen is set to something like 1024x775 instead of
> > >> 1024x768. I cannot see the bottom of my screen that is...
> > >>
> > >> I just checked the latest bk taken from kernel.org and still
> > >> hasn't been fixed. The fb console was working perfectly till
> > >> 2.5.67.
> > >>
> > >> If it is of any relevance, I'm using an NVIDIA tnt2 card. Has
> > >> nobody noticed this problem?
> >
> > > I have, and posted pretty much the same report a few weeks ago.
> > > Time to go write up a bug in bugzilla.kernel.org...
> >
> > The probable response will be WONTFIX on the grounds that the NVIDIA
> > driver is a binary only driver and, as such, can't be fixed...
>
> No, I'm using the VESA framebuffer driver that is in the kernel. This
> same thing happens on two different machines with two different video
> cards.
>
> No binary only driver for me, don't you think I would know better? :)

I think it has to do with the EDID code. The code works but for the intel
platform I'm grabbing data form the BIOS. I have seen 3 responses. The
first is the EDID header is corrupt. This is what happens to me. The
second is it gives wrong data. The screen that the EDID reports is
different from the one we request via the VESA mode setting BIOS call.
THe final is it actually works. The way to fix this is to remove the EDID
code in vesafb.c. I will push a patch for that soon.