[6065] in linux-scsi channel archive
Re: SCSI generic interface addition
daemon@ATHENA.MIT.EDU (Monty)
Thu Mar 11 03:00:15 1999
From: Monty <xiphmont@mit.edu>
To: Eric Youngdale <eric@andante.jic.com>,
Douglas Gilbert <dgilbert@interlog.com>,
Bernard Hatt <bmh@arkady.demon.co.uk>, linux-scsi@vger.rutgers.edu,
Richard Gooch <rgooch@atnf.csiro.au>
In-reply-to: Your message of "Thu, 11 Mar 1999 15:17:08 +1100."
<199903110417.PAA30642@vindaloo.atnf.CSIRO.AU>
Date: Thu, 11 Mar 1999 02:42:44 EST
>> My main objection to this was that it never seemed obvious
>> that this is something that really had to be in the kernel. There
>> has got to be a user-space solution to this issue that wouldn't
>> bloat the kernel.
I'll begin by saying that I'm one of the big proponents of never
putting anything new in the kernel (every few weeks I have to smack
down someone who wants cdparanoia in the kernel. Jesus!). However, I
strongly believe devfs *does* belong in the kernel.
The points in favor of devfs in general are basically that it makes
finding things easier (a big win for Paranoia and application
developers) and avoids mismatches between the state of /dev and actual
device support in the kernel (which *is* a constant support pain for
Paranoia as users are constantly forgetting to add /dev entries after
adding generic scsi support, and then they send bug reports). The
current devfs naming policy is the only policy both Joerg and I like
for generic SCSI addressing as an aside :-) It makes a clean, complete
dev directory an instant reality.
The userspace implementations I've seen discussed were not as robust
as a kernel based implementation; more points of possible failure.
Devfs is no more bloat that procfs (you could do most of /proc in
userspace, right?), and I'd agrue that both are equally
useful/necessary.
I'm not anywhere approaching 'inner circle' either. I'm not sure who
to take this up with, but I can probably start by talking to Ted at
least :-) Note, though, that I'm not necessarily an advantage to have
as an ally on these things....
Monty
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
