I noticed on the kernel status doc that a revamp of kernel pcmcia is in
the works. Can anyone elaborate?
For example, does this mean that 16-bit PCMCIA cards will use hotplug?
Does this mean that 32-bit cards will stop requiring cardmgr to simply
bind devices to drivers (is it too much to ask that the driver know what
it drives :)? The only card of mine (and I have a few 3coms, xircoms,
lucents) that I think works sanely is my xircom 32-bit cardbus card...
it would be nice if they all worked that way. Even though I know that
this will not be possible for the 16-bit pcmcia cards, IMHO atleast all
the hotplug devices should use ONE daemon.
I'm curious and excited by the prospect... if there is anything I can do
to help (even as a newbie I can certainly help change
unregister/register functions or add any device tables, etc), please let
me know.
[email protected] said:
> I noticed on the kernel status doc that a revamp of kernel pcmcia is
> in the works. Can anyone elaborate?
The PCMCIA code makes me want to vomit. It wants printing out and ritually
burning, while taking care not to inhale the fumes which might affect your
brain.
> For example, does this mean that 16-bit PCMCIA cards will use hotplug?
That is the plan.
> Does this mean that 32-bit cards will stop requiring cardmgr to simply
> bind devices to drivers (is it too much to ask that the driver know
> what it drives :)?
That is also the plan.
Unfortunately the plan is all we have at the moment, other than a few
hundred lines of untested core device/driver registration code and untested
CIS-parsing code. I threw that together hoping it would work like stone
soup -- but it hasn't worked yet, so it's going to have to wait till I have
more time to play.
--
dwmw2
On Fri, 2002-07-26 at 10:03, David Woodhouse wrote:
> Unfortunately the plan is all we have at the moment, other than a few
> hundred lines of untested core device/driver registration code and untested
> CIS-parsing code. I threw that together hoping it would work like stone
> soup -- but it hasn't worked yet, so it's going to have to wait till I have
> more time to play.
Maybe if you tell people where your soup pot is 8)
[email protected] said:
> On Fri, 2002-07-26 at 10:03, David Woodhouse wrote:
> > Unfortunately the plan is all we have at the moment, other than a few
> > hundred lines of untested core device/driver registration code and
> > untested CIS-parsing code. I threw that together hoping it would work like
> > stone soup -- but it hasn't worked yet, so it's going to have to wait till
> > I have more time to play.
> Maybe if you tell people where your soup pot is 8)
I was sort of hoping to get it in slightly better shape before mentioning
it to the great unwashed, but who cares...
cvs -d :ext:[email protected] login
(password anoncvs)
cvs -d :ext:[email protected] co pcmcia
http://lists.infradead.org/mailman/listinfo/linux-pcmcia
http://lists.infradead.org/mailman/listinfo/linux-pcmcia-cvs
--
dwmw2