Re: Why so few widescreen framebuffer drivers?
- From: phil-news-nospam@xxxxxxxx
- Date: 9 Sep 2007 20:58:24 GMT
On Thu, 06 Sep 2007 19:10:47 +0100 M?ns Rullg?rd <mans@xxxxxxxxx> wrote:
| phil-news-nospam@xxxxxxxx writes:
|
|> On Sun, 02 Sep 2007 20:21:25 +0100 M?ns Rullg?rd <mans@xxxxxxxxx> wrote:
|>
|> | The intel810/915/945 driver doesn't allow arbitrary mode settings on
|> | the LVDS (LCD) output. It only supports setting the mode once during
|> | bootup, and then only to one of the modes reported by the video BIOS.
|> | As I explained in another post, the video BIOS frequently has values
|> | incorrect for the connected display.
|>
|> Sounds like another driver that needs to be tweaked.
|>
|> IMHO, _every_ aspect of information reported by hardware to the software
|> must have a means to manually override it. In so many cases the information
|> is just wrong.
|
| IIRC, the driver uses VESA BIOS calls or something similar to set the
| mode on the LVDS output (but not on the analogue output). This call
| takes a mode number as argument, and the correct mode is simply not in
| the table. It would of course be possible to add code to the driver
| for rewriting the mode table, just like the userspace tool is able to
| do this. It just takes some work, and nobody has yet felt it worth
| the effort.
If I were to do some coding on this, I'd rather access video card registers
directly, rather than depend on BIOS code. Of course that means knowing the
register structure via proper documentation.
|> Every piece of configuration information the kernel obtains from hardware
|> anywhere needs to have some kind of designator to it. And this means in
|> every driver. And there needs to be a way to override any of it if so
|> desired. Means to do this can include an extra source file to compile them
|> in (in case there are a lot of them), and the parameter string at kernel
|> start time. And they all need to be stored somewhere, possible available
|> in /proc or somewhere, to cross check with.
|
| A lot is available through sysfs.
It needs to all be there.
--
|---------------------------------------/----------------------------------|
| Phil Howard KA9WGN (ka9wgn.ham.org) / Do not send to the address below |
| first name lower case at ipal.net / spamtrap-2007-09-09-1555@xxxxxxxx |
|------------------------------------/-------------------------------------|
.
- References:
- Why so few widescreen framebuffer drivers?
- From: useful_infos
- Re: Why so few widescreen framebuffer drivers?
- From: Jan Kandziora
- Re: Why so few widescreen framebuffer drivers?
- From: Måns Rullgård
- Re: Why so few widescreen framebuffer drivers?
- From: phil-news-nospam
- Re: Why so few widescreen framebuffer drivers?
- From: Måns Rullgård
- Why so few widescreen framebuffer drivers?
- Prev by Date: Re: writing drivers using C++
- Next by Date: Re: writing drivers using C++
- Previous by thread: Re: Why so few widescreen framebuffer drivers?
- Next by thread: PLEASE, PLEASE HELP - Confused by kernel paging error
- Index(es):
Relevant Pages
|
