Subject: kern/18112: Crash mounting file system with very large blocks
To: None <gnats-bugs@gnats.netbsd.org>
From: None <thorpej@shagadelic.org>
List: netbsd-bugs
Date: 08/29/2002 17:07:50
>Number:         18112
>Category:       kern
>Synopsis:       Crash mounting file system with very large blocks
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Aug 29 17:08:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     Jason R Thorpe
>Release:        NetBSD 1.6F (+ nathanw_sa)
>Organization:
Wasabi Systems, Inc.
>Environment:
	
	
System: NetBSD npwr.fast-100.shagadelic.org 1.6F NetBSD 1.6F (TEAMASA_NPWR) #14: Thu Aug 29 16:42:55 PDT 2002     thorpej@yeah-baby.shagadelic.org:/u1/hack/nathanw_sa/src/sys/arch/evbarm/compile/TEAMASA_NPWR evbarm
Architecture: arm
Machine: evbarm
>Description:
	When attempting to mount a 17G file sytem using 64K/16K blocks,
	the system crashes:

[u]vm_fault(0xc041f264, c6d2e000, 1, 0) -> d
Unhandled trap (frame = 0xc7ea4ca4)
Data abort: 'Translation fault (page)' status=007 address=c6d2e000 PC=c03d9d0c
Stopped in pid 119.1 (mount_ffs) at     memmove+0x40:   ldmia   r1!, {r3-r4, r12
, r14}
db> trace
ffs_mountfs+0xc
        scp=0xc02b10b8 rlv=0xc02b0808 (ffs_mount+0x1f8)
        rsp=0xc7ea4d74 rfp=0xc7ea4e28
        r10=0xc7df0748 r9=0x00000000
        r8=0xc7e3e924 r7=0xc1487a00 r6=0xc7ea4e4c r5=0x00000000
        r4=0x00000000
ffs_mount+0xc
        scp=0xc02b061c rlv=0xc0316af4 (sys_mount+0x414)
        rsp=0xc7ea4e2c rfp=0xc7ea4f30
        r10=0xc03e2648 r9=0x00000000
        r8=0xc7df0748 r7=0xc7ea4fb8 r6=0xc7e94d68 r5=0x00000000
        r4=0xc1487a00
sys_mount+0xc
        scp=0xc03166ec rlv=0xc03937a8 (syscall+0x194)
        rsp=0xc7ea4f34 rfp=0xc7ea4f98
        r10=0xc7dcd880 r9=0x00000004
        r8=0xc7ea4fb4 r7=0x0000ba9c r6=0x00000004 r5=0x00000015
        r4=0xc7ea4f48
...

	The PC and the stack trace are inconsistent, leading me to
	believe the stack has gotten smashed somewhere along the way.

>How-To-Repeat:
	Here is the disklabel describing the file system:

# /dev/rsd1c:
type: SCSI
disk: ST318452LW
label: Npwr-1-sd1
flags:
bytes/sector: 512
sectors/track: 484
tracks/cylinder: 4
sectors/cylinder: 1936
cylinders: 18496
total sectors: 35843670
rpm: 3600
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0           # microseconds
track-to-track seek: 0  # microseconds
drivedata: 0 

3 partitions:
#        size    offset     fstype  [fsize bsize cpg/sgs]
 a:  35843670         0     4.2BSD  16384 65536  4296   # (Cyl.    0 - 18514*)
 c:  35843670         0     unused      0     0         # (Cyl.    0 - 18514*)

	The "cpg" was computed by newfs.

>Fix:
	Not provided.
>Release-Note:
>Audit-Trail:
>Unformatted: