Re: Jfs bug ?

From: Dave Kleikamp (shaggy@austin.ibm.com)
Date: Tue Aug 28 2001 - 15:32:42 EST


This should be discussed in jfs-discussion@www-124.ibm.com since
jfs is not included in either Linus' or Alan's kernels.

> Hi
> I just formated one of mine linux partitions with jfs
> I had the following problems
>
> When restoring from backup
>
> jfs_strtoUCS: char2uni returned -22.
> charset = iso8859-1, char = 0xffffff99
> jfs_strtoUCS: char2uni returned -22.
> charset = iso8859-1, char = 0xffffff99
> jfs_strtoUCS: char2uni returned -22.
> charset = iso8859-1, char = 0xffffff99
> jfs_strtoUCS: char2uni returned -22.
> charset = iso8859-1, char = 0xffffff99

First, char is simply "99", but it is being sign-extended.

A pathname in the backup has the character 0x99 which is
apparently not valid in the iso8859-1 codepage. Can you tell
me what filesystem the backup was made from, and do you know
what codepage it used? Currently, JFS uses whatever codepage
is compiled into the kernel as CONFIG_NLS_DEFAULT. There
should probably be a mount flag that JFS uses to override this
value, but we don't have that yet.

Anyway, JFS probably substituted a "?" for the 0x99 in the
pathname, and created the file without any further problems.

I'm not sure about the hangs you described. I'm currently
looking at some hangs that usually only occur under heavy
stress. It would be useful to have a stack trace from the
hanging threads, especially the [jfsCommit] thread.

-- 
David Kleikamp
IBM Linux Technology Center
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Aug 31 2001 - 21:00:31 EST