Segmentation fault in reiserfsck --rebuild-tree

Segmentation fault in reiserfsck --rebuild-tree

[Matt Roberts]

File system went a little fooey, and rebuild-tree was the suggested fix.

On running it the first time it went through around 40% before falling over.  
Further runs don't make it past the 0% marker.

This is running SuSE 8.2 from the CD (rescue mode), which I think is Kernel 
2.4.20.  The machine is an AMD Duron 1.3Ghz, 512MB RAM.

I compiled (on another system) the latest version of the utility programs and 
tried those, but the same result occurs.

Below is the details of what I've run/received thus far:

NB:  Yes I can rebuild the fs and have backups /but/ I don't think thats a 
good reason to ignore a crash!  If this utility worked correctly it would have 
saved me (and probably others) a whole lotta time.  So in the attempt at 
making the world a better place, I present this bug report! ;)

----------------------------

debugreiserfs /dev/hda1

Filesystem state: consistency is not checked after last mounting

Reiserfs super block in block 16 on 0x301 of format 3.6 with standard journal
Count of blocks on the device: 4719085
Number of bitmaps: 145
Blocksize: 4096
Free blocks (count of blocks - used [journal, bitmaps, data, reserved] 
blocks): 4719085
Root block: 0
Filesystem is cleanly umounted
Tree height: 65535
Hash function used to sort names: "r5"
Objectid map size 104, max 972
Journal parameters:
	Device [0x0]
	Magic [0x0]
	Size 8193 blocks (including 1 for journal header) (first block 18)
	Max transaction length 1024 blocks
	Max batch size 900 blocks
	Max commit age 30
Blocks reserved by journal: 0
Fs state field: 0x3:
	FATAL corruptions exist.
	 some corruptions exist.
sb_version: 2
inode generation number: 1002179
UUID: e297274f-5544-44ca-a114-1aa6c6d6c922
LABEL: 
Set flags in SB:
	ATTRIBUTES CLEAN


-----

reiserfsck --rebuild-tree /dev/hda1

Will rebuild the filesystem (/dev/hda1) tree

Replaying journal..
0 transactions replayed

#####
reiserfsck --rebuild-tree started at Nov  1 10:45:16
#####

Pass 0:
Loading on-disk bitmap.. ok, 3596247 blocks marked used
Skipping 8355 (super block, journal, bitmaps) 3587892 blocks will be read
0%Segmentation fault                             left 3518484, 13881 /sec

Nov  1 10:45:21 Rescue kernel:  <1>Unable to handle kernel NULL pointer 
dereference at virtual address 00000808
Nov  1 10:45:21 Rescue kernel:  printing eip:
Nov  1 10:45:21 Rescue kernel: c013437b
Nov  1 10:45:21 Rescue kernel: *pde = 00000000
Nov  1 10:45:21 Rescue kernel: Oops: 0000 2.4.20-4GB #1 Mon Mar 17 17:54:44 
UTC 2003
Nov  1 10:45:21 Rescue kernel: CPU:    0
Nov  1 10:45:21 Rescue kernel: EIP:    0010:[<c013437b>]    Not tainted
Nov  1 10:45:21 Rescue kernel: EFLAGS: 00010206
Nov  1 10:45:21 Rescue kernel: eax: 00000800   ebx: 00000004   ecx: 00000011   
edx: 00007b56
Nov  1 10:45:21 Rescue kernel: esi: c288fcc8   edi: c179ed58   ebp: 00014f94   
esp: c05b9f28
Nov  1 10:45:21 Rescue kernel: ds: 0018   es: 0018   ss: 0018
Nov  1 10:45:21 Rescue kernel: Process reiserfsck (pid: 523, 
stackpage=c05b9000)
Nov  1 10:45:21 Rescue kernel: Stack: 00001000 00000000 00000000 00000000 
c288fc20 00000000 c0134910 c05b9f7c 
Nov  1 10:45:21 Rescue kernel:        c40c9ec0 c40c9ea0 00000000 080dc4c8 
00000000 080dc4c8 c0134b05 c40c9ea0 
Nov  1 10:45:21 Rescue kernel:        c40c9ec0 c05b9f7c c0134910 00000000 
c0346f40 00000000 00001000 080dc4c8 
Nov  1 10:45:21 Rescue kernel: Call Trace:    [<c0134910>] [<c0134b05>] 
[<c0134910>] [<c0144a18>] [<c0108c33>]
Nov  1 10:45:21 Rescue kernel: Code: 39 70 08 75 f4 39 68 0c 75 ef 89 c3 85 c0 
0f 84 4e 02 00 00 

Re: Segmentation fault in reiserfsck --rebuild-tree

[Vitaly Fertman]

Hi, 

> I compiled (on another system) the latest version of the utility programs and 
> tried those, but the same result occurs.

you had connected the drive to that another system before trying 
reiserfsck? Or copied the complied reiserfsck to the computer with
the problem harddrive? If the later, did you comple reiserfsprogs 
statically?

> reiserfsck --rebuild-tree /dev/hda1
>
> Will rebuild the filesystem (/dev/hda1) tree
>
> Replaying journal..
> 0 transactions replayed
>
> #####
> reiserfsck --rebuild-tree started at Nov  1 10:45:16
> #####
>
> Pass 0:
> Loading on-disk bitmap.. ok, 3596247 blocks marked used
> Skipping 8355 (super block, journal, bitmaps) 3587892 blocks will be read
> 0%Segmentation fault                             left 3518484, 13881 /sec

would you pack the metadata of the problem partition with:
    debugreiserfs -p /dev/hda1 | bzip2 -c > hda1.bz2
and make it available for downloading, I will debug the problem
locally then.

-- 
Thanks,
Vitaly Fertman