Re: system/611: fsck_ffs(8) and ffs_vfsops.c:ffs_mount() are inconsistent (fwd)

[baker hamilton <garath_(_at_)_code_(_dot_)_ridgefield_(_dot_)_org>]

The following reply was made to PR system/611; it has been noted by GNATS.

From: baker hamilton <garath_(_at_)_code_(_dot_)_ridgefield_(_dot_)_org>
To: gnats_(_at_)_openbsd_(_dot_)_org
Cc:  Subject: Re: system/611: fsck_ffs(8) and ffs_vfsops.c:ffs_mount() are inconsistent  (fwd)
Date: Sun, 4 Oct 1998 15:15:24 -0400 (EDT)

 ---------- Forwarded message ----------
 Date: Sun, 04 Oct 1998 10:25:45 -0700
 From: Robert Earl <rearl_(_at_)_teleport_(_dot_)_com>
 To: baker hamilton <garath_(_at_)_code_(_dot_)_ridgefield_(_dot_)_org>
 Subject: Re: fsck_ffs(8) and ffs_vfsops.c:ffs_mount() are inconsistent 
 
 >> >How-To-Repeat:
 >>	  Mount one or more filesystems read-only (/dev/sd0d on /usr
 >>	  is a good candidate). After a system crash, fsck -p to preen
 >> 	  the filesystems; fsck_ffs(8) skips /dev/rsd0d since it was 
 >>	  never dirtied. 
 >>	  Now run "mount -a" and the kernel code will report "/dev/wd0d:
 >
 >/dev/wd0d? Where did that come from? :)
 
 Yeah, actually for some reason, I decided to use sd0 for my example
 when my messages are actually coming from wd0. /var/log/messages shows
 this happens after crashes and also, apparently, after clean reboots:
 
 Sep 16 07:25:00 arkham /bsd: /dev/wd1d: file system not clean; please fsck(8)
 Sep 21 19:34:16 arkham /bsd: /dev/wd1d: file system not clean; please fsck(8)
 Sep 29 19:51:36 arkham /bsd: /dev/wd0j: file system not clean; please fsck(8)
 Oct  1 04:32:10 arkham /bsd: /dev/wd0j: file system not clean; please fsck(8)
 Oct  1 04:32:10 arkham /bsd: /dev/wd0h: file system not clean; please fsck(8)
 Oct  1 04:32:10 arkham /bsd: /dev/wd0f: file system not clean; please fsck(8)
 Oct  1 04:32:13 arkham savecore: reboot after panic: vref used where vget required
 Oct  2 05:48:59 arkham /bsd: /dev/wd1d: file system not clean; please fsck(8)
 Oct  2 05:48:59 arkham /bsd: /dev/wd0h: file system not clean; please fsck(8)
 Oct  2 14:19:20 arkham /bsd: /dev/wd0j: file system not clean; please fsck(8)
 Oct  2 14:19:21 arkham /bsd: /dev/wd0h: file system not clean; please fsck(8)
 Oct  2 14:19:21 arkham /bsd: /dev/wd0f: file system not clean; please fsck(8)
 Oct  2 14:19:24 arkham savecore: reboot after panic: vnode already on free list
 Oct  3 19:56:55 arkham /bsd: /dev/wd0h: file system not clean; please fsck(8)
 Oct  4 08:58:25 arkham /bsd: /dev/wd1d: file system not clean; please fsck(8)
 Oct  4 08:58:25 arkham /bsd: /dev/wd0f: file system not clean; please fsck(8)
 
 
 >Sorry for pointing it out if its only a typo, but I'm  having a difficult
 >time reproducing this.
 
 I should have mentioned something else in my report, that I am using my
 own wdc_isa and wd drivers, merged from NetBSD's new stuff, and with some
 PIO mode 3/4 support added. Hopefully this is low-level enough that it
 is not affecting filesystem flags like this, but I never know...
 
 --robert