Re: [Q] is framebuffer console code in 2.5.4x functional ?

[Andrey Panin <pazke@orbita1.ru>]

[-- Attachment #1: Type: text/plain, Size: 3598 bytes --]

On Thu, Nov 21, 2002 at 05:24:44PM +0000, James Simmons wrote:
> 
> 
> > > > console doesn't show a single pixel.
> > > 
> > > :-( Can you post your .config file. 
> > 
> > Attached.
> 
> Hm. Strange. It should work. Can you get serial console working? 

Forgot to mention, I've seen message:
	fbcon_setup: No support for fontwidth 8
in /var/log/dmesg. 

I found this printk() in fbcon_setup(), but i can't even imagine 
why it happens.

Sorry for not providing full info, but when i wrote first letter 
my head was full of gory thoughts about visws legacy irq handling :(
I can post complete log tomorrow.
   
> > I did some changes to sgivwfb.c to make it compilable, patch attached.
> > Can you take a look at it ?
> 
> Applied your patch to the BK tree. 
Good.
  
> > > I will be posting a new fbdev patch today against 2.5.48 today. Giev it a 
> > > try.
> > 
> > Didn't see proposed fbdev patch yet :(
> 
>     Sorry about that. You are not the only one that has asked me. Also I 
> keep getting lots of error reports about drivers being broken. The problem 
> is having enough time. For example I haven't found the time to create this 
> patch. This brings up a serious point which I have been wrestling with. The 
> framebuffer layer has been broken for a long time durning the 2.5.X cycle 
> The problem is both maintainers of this subsystem, Geert and I, both have 
> very little time to work on it. For both of us we don't work on the 
> framebuffer code for a living. I work with wireless networking cards. I 
> work 8 hours a day on networking code and travel 3 hours total every day 
> to work. Including eating a sleeping and I have at most 1 to 2 hours a day 
> to work on the framebuffer stuff. Weekends I have to do other survial 
> things like buy food. So the framebuffer developement has gone at a 
> snail pace and will continue to do so unless things change. I estimate 
> about 20+ more versions before the framebuffer layer properly works. 
>     It pains me that this is happening. I really enjoy working on the 
> framebuffer and console layer. So I have been thinking about what to 
> do ? One which is the most likely is to step down from maintaintership 
> and hope someone else who can devote there full time and energy to it 
> can take over. Will someone else take over? I seriously doubt it. We all 
> have to make a living and that means working on things the linux industry 
> cares about which is only server stuff. So I except the framebuffer layer 
> will go into serious code decay. So the best situtation which I except to 
> happen is that I finish as much as I can for the fbdev layer and then 
> step down. 
>     I have tried to look for work locallly (can't really affored to move 
> cross country very few years) relating to the framebuffer layer. In my 
> search I only found one company that seemed interested in this developement, 
> strangeberry (http://www.strangeberry.com). I sent them my resume but 
> never heard from them. As for funding I serious doubt that would happen 
> since it isn't server related. The reality is for proper maintiance of any 
> subsystem you need people hired to solely work to keep it going. 
> Unfortunely the framebuffer layer is one of those few ones that doesn't 
> have that.

I understand this situation perfectly, looks like it's almost common for
developers working in "not so importatnt for servers" subsystems :(

-- 
Andrey Panin            | Embedded systems software developer
pazke@orbita1.ru        | PGP key: wwwkeys.eu.pgp.net

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]