From owner-cvs-all  Fri Oct 13  2:27: 2 2000
Delivered-To: cvs-all@freebsd.org
Received: from netplex.com.au (adsl-64-163-195-99.dsl.snfc21.pacbell.net [64.163.195.99])
	by hub.freebsd.org (Postfix) with ESMTP
	id 5DF9A37B502; Fri, 13 Oct 2000 02:26:59 -0700 (PDT)
Received: from netplex.com.au (peter@localhost [127.0.0.1])
	by netplex.com.au (8.11.0/8.9.3) with ESMTP id e9D9QrG38035;
	Fri, 13 Oct 2000 02:26:53 -0700 (PDT)
	(envelope-from peter@netplex.com.au)
Message-Id: <200010130926.e9D9QrG38035@netplex.com.au>
X-Mailer: exmh version 2.1.1 10/15/1999
To: Dag-Erling Smorgrav <des@ofug.org>
Cc: dima@rdy.com, ache@FreeBSD.org, cvs-all@FreeBSD.org
Subject: Re: cvs commit: src/etc rc 
In-Reply-To: <xzp66mxnmcf.fsf@flood.ping.uio.no> 
Date: Fri, 13 Oct 2000 02:26:52 -0700
From: Peter Wemm <peter@netplex.com.au>
Sender: owner-cvs-all@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

Dag-Erling Smorgrav wrote:
> Peter Wemm <peter@netplex.com.au> writes:
> > IMHO, the fix was what was needed given the current internals of
> > savecore(8).  Do you see it differently?
> 
> Yes - savecore(8) or devname(3) needs to be fixed so that /etc/rc
> doesn't need to be broken like this. savecore(8) is often run manually
> in single-user mode and should not depend on dev_mkdb(8); I think it's
> a bug in devname(3) that it does (devname(3) should use dev.db as a
> fallback in case the sysctl doesn't work, rather than the reverse).

Actually, I thought it was the other way around..  In single user,
/var/run had not been mounted and rm -rf'ed yet.

devname.c:
char *
devname(dev, type)
...
        /* First check the DB file. */
        r = xdevname(dev, type);
        if (r != NULL)
                return (r);
        /* Then ask the kernel. */
        if ((type & S_IFMT) == S_IFCHR) {
                j = sizeof(buf);
                i = sysctlbyname("kern.devname", buf, &j, &dev, sizeof (dev));
...

I suspect that what happened is that you booted, mounted the filesystems
(thereby bringing the previous stale dev.db back in /var/run), ran savecore.
Or, you used DEVFS.

The sysctl *does not work* unless devfs is active.  devname() depends on
/var/run/dev.db being present in order to work, and the old savecore/dev_mkdb
ordering was wrong for non-devfs systems.  Whether devname() used dev.db
or sysctl first didn't matter in this case since devfs wasn't there.  Yes,
it probably should use sysctl first.  Anyway:

[2:20am]~src/sbin/savecore-176# ./dumpdev
dumpdev = /dev/rda0s1b

[2:20am]~src/sbin/savecore-177# rm /var/run/dev.db 
[2:20am]~src/sbin/savecore-178# ./dumpdev
dumpdev = /dev/#C:13:0x20001

This is a crude cut/paste from the savecore internals.  This machine does
not have DEVFS.  This had a kernel and world built yesterday.

Cheers,
-Peter
--
Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au
"All of this is for nothing if we don't go to the stars" - JMS/B5



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message