[5902] in linux-scsi channel archive
Re: WG: AW: cdrecord problems on recent Linux versions
daemon@ATHENA.MIT.EDU (Joerg Schilling)
Fri Feb 12 08:52:35 1999
Date: Fri, 12 Feb 1999 14:35:59 +0100 (MET)
From: Joerg Schilling <schilling@fokus.gmd.de>
To: rgooch@atnf.csiro.au, schilling@fokus.gmd.de
Cc: Dominik.Stadler@btk.de, bsc@fleggaard.dk,
cdwrite@lists.debian.org, dgilbert@interlog.com,
heiko_eissfeldt@z.detesystem.de, linux-scsi@vger.rutgers.edu,
michaels@stochastik.rwth-aachen.de, xiphmont@mit.edu
>From rgooch@vindaloo.atnf.CSIRO.AU Fri Feb 12 01:50:02 1999
>Joerg Schilling writes:
>> >From rgooch@vindaloo.atnf.CSIRO.AU Thu Feb 11 21:51:13 1999
>> >In the case I noticed recently, I think it was done correctly. Someone
>> >upgraded the FORE ATM drivers on a system (the new drivers use
>> >different /dev entries), booted with -r and the old device entries in
>> >/dev were still around. This caused much confusion.
>>
>> >It may be a bug due to FORE not doing something right (like adding to
>> >one of those scripts), but either way, it points to a bug/design flaw
>> >in Solaris. Problems like this should not happen. Linux devfs doesn't
>> >suffer from this problem.
>>
>> This definitely must be a FORE bug. They then seem not to provide a correct
>> /etc/devlink.tab which is needed to make devlinks work properly.
>> I know that there are other bugs in the FORE installation.
>That may well be. My point is that I consider such a system
>fragile. It also puts the same information in two places: in the
>kernel and in the scripts. This is one of the things I dislike with
>the existing Linux device number scheme: there are two separate
>repositories of the same information, one in the kernel sources and
>one in MAKEDEV (or /dev if you prefer). Things can get out of sync.
>No, I don't consider Documentation/devices.txt to be the
>repository. It may be *intended* that way, but it is not how it
>works. The *real* repositories are in the kernel sources and the
>source to MAKEDEV (or in /dev).
Important is, that the /devices directory is not for direct use.
There may be left over nodes in there but this doesn't matter if
the synlinks from /dev are removed correctly. I see no discrepance
between this. There is only one place where the knowledge about device
nodes and their names is located: the driver
When it calls ddi_create_minor_node(), it created the /device
nodes as needed.
Jörg
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
js@cs.tu-berlin.de (uni) If you don't have iso-8859-1
schilling@fokus.gmd.de (work) chars I am J"org Schilling
URL: http://www.fokus.gmd.de/usr/schilling ftp://ftp.fokus.gmd.de/pub/unix
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
