Re: My $0.02 on devd and devfs

david parsons (orc@pell.portland.or.us)
12 Oct 1999 09:35:46 -0700

In article <linux.kernel.199910120618.DAA04453@sleipnir.valparaiso.cl>,
Horst von Brand <vonbrand@sleipnir.valparaiso.cl> wrote:
>orc@pell.portland.or.us (david parsons)
>> In article <linux.kernel.380254A4.F0C32162@pobox.com>,
>> Jeff Garzik <jgarzik@pobox.com> wrote:
>> >Khimenko Victor wrote:
>
>> >> See above. Try to turn on printer, do `cat something > /dev/printer/0`
>> >> with devfs and explain how it'll be handled with devd.
>
>> >Are you implying here that /dev/printer/0 is created at open() time?
>> >IMHO that is a bad idea.
>
>> I think that if /dev/printer/0 is controlled by a lkm'ed driver,
>> that's exactly what happens, no matter what. I'd expect that in
>> the devd case, you'd have to populate /dev with every possible
>> device so that you could actually do the open, have the kernel tell
>> kmod to load the driver, then the information would go off to devd
>> to, ummm....
>
>Nope. devd is notified when the device shows up.

How?

If /dev/printer/0 will be a PC-style parallel port, and I've
built the driver as a module, the device has shown up when the
machine is built, but the driver is sitting there waiting for kmod
to drag it into core.

If the file /dev/printer/0 doesn't exist, I can't open it and can't
tell kmod to load the driver.

So what I need to do is either discard modules altogether or do a
speculative load of every driver module in the system at bootup
so that I can generate the appropriate detect messages for devd
to populate /dev. If I have to jump through all of these
hoops, what do I gain over just populating /dev with every possible
device?

____
david parsons \bi/ devd: just say no.
\/

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/