On Wed, Jun 21, 2000, Frank van Maarseveen <F.vanMaarseveen@inter.NL.net> wrote:
> On Mon, Jun 19, 2000 at 06:54:28PM -0400, Johannes Erdfelt wrote:
> > On Mon, Jun 19, 2000, Frank van Maarseveen <F.vanMaarseveen@inter.NL.net> wrote:
> > > Also, I think it is best for usbdevfs to make it part of /proc.
> > > The name is a bit misleading because there are no device nodes
> > > such as with devfs and mounting it on /proc/bus/usb is awful.
> >
> > This is incorrect. Just because the files don't have a major/minor
> > doesn't mean they don't have the same functionality.
> > Having a virtual FS means we don't need major/minor numbers.
>
> Correct, but that addresses a very fundamental issue. Major/minor have
> been there because a kernel is good in numbers. In userland /dev attaches
> names to the numbers. What you suggest here is a whole new ballgame.
It really isn't. When you call open("/dev/sda") the kernel has to do
something with the name.
With usbdevfs, there's still the step of dealing with the name, it just
resolves inode numbers internally instead of resolving major/minor
numbers. The kernel does the same amount of work, just using different
numbers.
The win in that, is removing the administration portion completely (no
major/minor numbers to track, etc).
Of course, this is very similar to devfs and is probably not best to be
discussed in this thread.
> It is
> interesting though, but what I saw in /proc/bus/usb was not a replacement
> for any device file such as /dev/usb/lp0. It only provided for some debug
> info.
You are correct, it does not replace /dev/usb/lp0. However, it is much
more than debug info. It's the whole userspace API. libusb uses it to
communicate to devices. gphoto uses libusb to communicate to digital
still cameras through it.
> Also, the tree merely contained numbered files and no descriptive
> names.
Yes, this was something I wanted to change, but we didn't want to change
it just for the sake of changing it. It's really not that important.
If we change the mountpoint to /dev/usb or something, then I think the
change is worth it.
> Then there is the same issue as with devfs: persistent state.
Bingo!
I don't like usbdevfs either, but it has a purpose right now albeit
with many flaws. We are working on replacing it, but it's an uphill
battle against all of the anti-devfs people since we'd be reinventing
devfs.
> So, unless I missed something usbdevfs is not a replacement for /dev.
> It is only an addition for /proc.
It is not a replacement for /dev no. It's supposed to be an addition to
/dev and an addition to /proc. Just happens to be in /proc because of
some legacy reasons.
It's the whole userspace API, kinda like the SCSI generic interface. I'd
like to keep the devices and drivers file in /proc, and move the rest to
/dev.
JE
-
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/
This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:22 EST