Hello list members,
On Wed, 14 May 2003 16:36, DAve Kleikamp wrote:
> On Wednesday 14 May 2003 09:57, Thomas A. Cort wrote:
> > [EMAIL PROTECTED] u0]# fsck.jfs -nv jfs_fs
> > fsck.jfs version 1.1.2, 25-Mar-2003
> > Bus error
> > [EMAIL PROTECTED] u0]#
> If I'm not mistaken, sparc64 will get a bus error if a long is not
> properly aligned on an 8-byte boundary. Can you run fsck.jfs under gdb
> and determine where it's trapping?
> I don't know what kind of success anybody has had with JFS on sparc64.
Same problem here on an Ultra-1 running 2.4.22:
(gdb) run -nv /dev/loop0
Starting program: /sbin/fsck.jfs -nv /dev/loop0
/sbin/fsck.jfs version 1.1.3, 05-Sep-2003
processing started: 9/10/2003 16.43.6
The current device is: /dev/loop0
Open(...READONLY...) returned rc = 0
Primary superblock is valid.
The type of file system for the device is JFS.
Block size in bytes: 4096
Filesystem size in blocks: 8192
**Phase 1 - Check Blocks, Files/Directories, and Directory Entries
Program received signal SIGBUS, Bus error.
0x0002f3b8 in establish_wsp_block_map_ctl () at fsckwsp.c:1503
1503 fsckwsp.c: No such file or directory.
in fsckwsp.c
(gdb) where
#0 0x0002f3b8 in establish_wsp_block_map_ctl () at fsckwsp.c:1503
#1 0x0002f18c in establish_wsp_block_map () at fsckwsp.c:1378
#2 0x0002ed34 in establish_agg_workspace () at fsckwsp.c:1039
#3 0x000365b8 in phase1_processing () at xchkdsk.c:2036
#4 0x00034014 in main (argc=3, argv=0xeffffde4) at xchkdsk.c:346
(gdb)
Line 1503 in fsckwsp.c reads as follows:
strncpy(agg_recptr->blkmp_ctlptr->hdr.eyecatcher,
fbmh_eyecatcher_string, strlen(fbmh_eyecatcher_string));
And ../libfs/fsckwsp.h defines the following:
#define fbmh_eyecatcher_string "wspblkmp"
Guessing that the defined string may be 9 characters long including a
terminating \0, and thus would not fit into
agg_recptr->blkmp_ctlptr->hdr.eyecatcher, I simply shortened it by one
character and recompiled.
After that the fsck_jfs crashed a few lines futher down, 1519 to be precise,
which reads
Current_Time = time(NULL);
At this point, I gave up because I had absolutely no idea what could be wrong
with this line :-) To me this looks like something goes wrong a little bit
earlier and the actual crash happens on a not directly related line. Would
that be possible with the alignment issue that Dave mentioned?
Thanks,
Knut
_______________________________________________
Jfs-discussion mailing list
[EMAIL PROTECTED]
http://www-124.ibm.com/developerworks/oss/mailman/listinfo/jfs-discussion
