pciaccess clean up

[Tiago Vignatti]

Ian Romanick escreveu:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Adam Jackson wrote:
> | On Mon, 2008-03-03 at 09:52 -0800, Ian Romanick wrote:
> |
> |> This is by design.  In the pre-PCI-rework discussions we decided that it
> |> was pure evil for the X server to muck about with PCI device / bus
> |> enables.  I thought that the int10 boot-strap was still enabled, though.
> |
> | Well, it has to go somewhere.  In the ideal world I have kernel drivers
> | for everything and that handles enabling.  Instead...
> |
> | So until I get an ideal world, I'd like that the server enable my PCI
> | devices please.  Because what we've got right now, where trying to post
> | a secondary card wedges the machine, is clearly crap.
> 
> Fair enough.  The question then become one of where to put it.  Should
> we put in the X server (like it used to be) or in libpciaccess.  I can
> see arguments for either place.

We talked today via IRC about this. Given that we don't have kernel 
drivers for all devices, Ajax suggested that it must be interesting to 
have a helper application external to X server. Maybe extend the vbetool 
[0] to support multiple cards. I'll try to dig deeper until the end of 
this week on this.


Thank you,

[0] I have to see how exactly vbetool works but seems that link the 
server with a "multicard aware version" of libx86 is enough. BTW, libx86 
could be ported to fd.o with the autoconf "delights" and all the 
applications (vbetool, coreboot, xorg, etc) that rely on this library 
would be more happy...

-- 
Tiago Vignatti
C3SL - Centro de Computação Científica e Software Livre
www.c3sl.ufpr.br