Date: Tue, 05 May 1998 23:26:27 -0700 From: Mike Smith <mike@smith.net.au> To: Archie Cobbs <archie@whistle.com> Cc: stefan@promo.de (Stefan Bethke), luigi@labinfo.iet.unipi.it, freebsd-hackers@FreeBSD.ORG Subject: Re: ISA-PnP w\o BIOS support? Message-ID: <199805060626.XAA01524@antipodes.cdrom.com> In-Reply-To: Your message of "Tue, 05 May 1998 17:28:35 PDT." <199805060028.RAA18956@bubba.whistle.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> I agree with Stefan's sentiment. It seems like this should be
> the paradigm:
We actually thrashed this out last year; you might be able to dig the
thread up again. Mostly it was Luigi, Stefan Esser and myself throwing
ideas around.
> CASE #1: Kernel config file looks like this:
> --------------------------------------------
>
> device foo0
>
> In this case, the kernel automatically configures stuff for driver
> foo0. If the BIOS has already configured the card, and that's
> acceptable, the kernel can go along. Otherwise, the kernel should
> pick resources that are free and configure the card itself.
>
>
> CASE #2: Kernel config file looks like this:
> --------------------------------------------
>
> device foo0 port 0x220 irq 7 vector foointr
>
> Kernel should override whatever is already configured in the
> card (if anything) with the given values.
>
>
> Does this make sense to anyone else besides me? What are the
> technical things that need to be done to make it happen?
This actually falls a little short. It's easier to look at it from a
different angle, too.
For each PnP device/function in the system
If the device/function is not supported (unknown, no driver present)
deconfigure device (if required)
continue
else
while resources required
allocate new resources
test resources (using PnP test facility if available)
endwhile
if resources available
allocate a new driver instance, assign resources
else
continue
endif
endfor
For each explicitly-configured ISA devices
if resource config does not conflict with PnP-configured device
allocate resources
if probe succeeds
allocate new driver instance
else
free resources
endif
endif
endfor
So
device foo0
will support any number of 'foo' devices.
device foo0 at isa? port "IO_FOO0"
will support any number of 'foo' devices, as well as probing for one at
IO_FOO0 after the PnP probe.
The following things have to happen to make this a reality:
- The PnP BIOS call interface needs to be finished. I have done some
work based on Jonathan Lemon's code, but he needs to finish working
out the APM interaction problems. This is absolutely critical as it
allows us to use PnP to locate system devices and thus determine
which resources are free. It also lets us find onboard peripherals.
- Devices need to start ignoring NFOO and using malloc() to obtain
storage for softc structures. They also need to start passing
pointers to these structures around instead of unit numbers. This
has been started with the changed arguments to ISA interrupt
handlers (taking void * instead of int).
- A PnP enumerator needs to be written to traverse the PnP data. It
has to handle all three basic states (No BIOS, BIOS configuring
boot-only devices, BIOS configuring everything). ISA PnP code
already exists to handle the "No BIOS" and "boot-only" cases.
- A mechanism for loading PnP ID:driver mappings would be good.
Putting them in the driver works OK but is suboptimal. Note that
there is a table of "generic" PnP device identifiers as well that
should be used whenever possible. The "extras" stuff that I did for
the splash screen code can be used for this.
--
\\ Sometimes you're ahead, \\ Mike Smith
\\ sometimes you're behind. \\ mike@smith.net.au
\\ The race is long, and in the \\ msmith@freebsd.org
\\ end it's only with yourself. \\ msmith@cdrom.com
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805060626.XAA01524>