[plug] Hardware problem or kernel bug?

[Lineone]

----- Original Message -----
From: "Cameron Patrick" <cameron at patrick.wattle.id.au>
To: "PLUG List" <plug at plug.linux.org.au>
Sent: Thursday, March 13, 2003 1:59 AM
Subject: [plug] Hardware problem or kernel bug?


> Howdy,
>
> Has anyone here run into filesystem corruption and/or kernel panics on
> 2.4.20 with Reiserfs?
>
> The messages below seem rather ominous, and I'm wondering whether it is
> likely to be a hardware problem or a kernel bug.  If it's the latter, it
> looks like it might be reiserfs.
>
> TIA,
>
> CP.
>
> --------------------------------------------------------------------------
---
>
> (root at plato ~) chgrp
> bash: chgrp: command not found
>
> (root at plato ~) dpkg -L fileutils | grep chgrp
> /bin/chgrp
> /usr/share/man/man1/chgrp.1.gz
>
> (root at plato ~) apt-get install fileutils
> Reading Package Lists... Done
> Building Dependency Tree... Done
> Sorry, fileutils is already the newest version.
> 0 packages upgraded, 0 newly installed, 0 to remove and 0  not upgraded.
>
> (root at plato ~) apt-get --reinstall install fileutils
> Reading Package Lists... Done
> Building Dependency Tree... Done
> 0 packages upgraded, 0 newly installed, 1 reinstalled, 0 to remove and 0
> not upgraded.
> Need to get 710kB of archives. After unpacking 0B will be used.
> Do you want to continue? [Y/n]
> Get:1 http://ftp.uwa.edu.au stable/main fileutils 4.1-10 [710kB]
> Fetched 710kB in 4m17s (2759B/s)
> (Reading database ... 11393 files and directories currently installed.)
> Preparing to replace fileutils 4.1-10 (using
> .../fileutils_4.1-10_i386.deb) ...
> Unpacking replacement fileutils ...
> E: Sub-process /usr/bin/dpkg received a segmentation fault.
>
> (root at plato ~) dpkg-deb: subprocess paste killed by signal (Broken pipe)
>
> Message from syslogd at plato at Thu Mar 13 09:33:26 2003 ...
> plato kernel: vs-7042: entry_points_to_object: entry must be
> visiblekernel BUG at prints.c:334!
>
>
>
And here is the code that triggered it:

static int entry_points_to_object (const char * name, int len, struct
reiserfs_dir_entry * de, struct inode * inode)
{
    if (!de_still_valid (name, len, de))
        return 0;

,---8<------8<---
|   if (inode) {
|       if (!de_visible (de->de_deh + de->de_entry_num))
|           reiserfs_panic (0, "vs-7042: entry_points_to_object: entry must
be visible");
|       return (de->de_objectid == inode->i_ino) ? 1 : 0;
|   }
`---8<------8<---

    /* this must be added hidden entry */
    if (de_visible (de->de_deh + de->de_entry_num))
        reiserfs_panic (0, "vs-7043: entry_points_to_object: entry must be
visible");

    return 1;
}

Looks like a bug to me, but the reiserfs code expected it to happen and
caught it.