kern/43727: panic on use of fs with bsize 65536 fsize 8192

[dieter.NetBSD%pandora.be@localhost]

>Number:         43727
>Category:       kern
>Synopsis:       panic on use of fs with bsize 65536 fsize 8192
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Aug 09 09:55:00 +0000 2010
>Originator:     dieter roelants
>Release:        -current
>Organization:
>Environment:
NetBSD  5.99.38 NetBSD 5.99.38 (XEN3_DOMU) #0: Mon Aug  9 11:25:45 CEST 2010  
dieter%netbsd-u2.inflamm.able.be@localhost:/build/obj/sys/arch/amd64/compile/XEN3_DOMU
 amd64


>Description:
I don't know if this should be a port-xen PR, I don't feel like panicing my 
real machine atm... I'll assume it's MI.

I was in the progress of adding a new "disk" to one of my NetBSD PV domains. It 
panics whenever I try to use a partition on that new disk with bsize 65536 and 
fsize 8192. Other parameters (tried 32768/4096 and 16384/2048) seem to work 
fine. With "use" I mean everything that happens after mount, i.e. df, umount, 
etc. The device backing the disk in the host is a LVM, but I assume that 
doesn't matter.

The panic looks like this:
# mount /dev/xbd1a /mnt        
# umount /mnt
panic: buf mem pool index 7
fatal breakpoint trap in supervisor mode
trap type 1 code 0 rip ffffffff8012c205 cs e030 rflags 246 cr2  7f7fffffcf38 
cpl 0 rsp ffffa00045cb2690
Stopped in pid 16.1 (umount) at netbsd:breakpoint+0x5:  leave
breakpoint() at netbsd:breakpoint+0x5
panic() at netbsd:panic+0x28f
biowait() at netbsd:biowait
allocbuf() at netbsd:allocbuf+0x21
getblk() at netbsd:getblk+0x12e
bio_doread() at netbsd:bio_doread+0x1d
bread() at netbsd:bread+0x1b
ffs_vget() at netbsd:ffs_vget+0x208
ufs_root() at netbsd:ufs_root+0x1c
do_lookup() at netbsd:do_lookup+0x4c9
namei() at netbsd:namei+0x274
do_sys_stat() at netbsd:do_sys_stat+0x44
sys___lstat50() at netbsd:sys___lstat50+0x25
syscall() at netbsd:syscall+0xa8
ds          0xd821
es          0x6b59
fs          0x2670
gs          0x2d4d
rdi         0
rsi         0xd
rbp         0xffffa00045cb2690
rbx         0xffffa00045cb26a0
rdx         0
rcx         0x8
rax         0x1
r8          0xffffffff805a0d40  cpu_info_primary
r9          0
r10         0xffffa00045cb25b0
r11         0x1
r12         0x104
r13         0xffffffff8041d821  copyright+0x14101
r14         0xffffa000020696b8
r15         0
rip         0xffffffff8012c205  breakpoint+0x5
cs          0xe030
rflags      0x246
rsp         0xffffa00045cb2690
ss          0xe02b
netbsd:breakpoint+0x5:  leave
db> sync
syncing disks... 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 giving up
Printing vnodes for busy buffers
vnode @ 0xffffa0004653b900, flags (0x30<MPSAFE,LOCKSWORK>)
        tag VT_UFS(1), type VBLK(3), usecount 1, writecount 0, holdcount 3
        freelisthd 0x0, mount 0xffffa000456a2408, data 0xffffa00046539400 lock 
0xffffa0004653ba08
        tag VT_UFS, ino 742385, on dev 142, 0 flags 0x0, nlink 1
        mode 060640, owner 0, group 5, size 0
giving up

xbd1a is:
file system: /dev/rxbd1a
format  FFSv1
endian  little-endian
magic   11954           time    Mon Aug  9 11:32:59 2010
superblock location     8192    id      [ 4c5fcb4b 12e45b39 ]
cylgrp  dynamic inodes  4.4BSD  sblock  FFSv2   fslevel 4
nbfree  2056261 ndir    1       nifree  4116477 nffree  14
ncg     40      size    16515072        blocks  16450103
bsize   65536   shift   16      mask    0xffff0000
fsize   8192    shift   13      mask    0xffffe000
frag    8       shift   3       fsbtodb 4
bpg     51610   fpg     412880  ipg     102912
minfree 5%      optim   time    maxcontig 1     maxbpg  16384
symlinklen 60   contigsumsize 0
maxfilesize 0x04001000400bffff
nindir  16384   inopb   512
avgfilesize 16384       avgfpdir 64
sblkno  8       cblkno  16      iblkno  24      dblkno  1632
sbsize  8192    cgsize  65536
csaddr  1632    cssize  8192
cgrotor 0       fmod    0       ronly   0       clean   0x01
wapbl version 0x0       location 0      flags 0x0
wapbl loc0 0    loc1 0  loc2 0  loc3 0
flags   none
fsmnt   
volname         swuid   0

The wapbl isn't the issue, I tried with and without, and I also disabled it on 
/ (where /dev resides). I also tried FFSv2.

The machine has 1024 MB.
>How-To-Repeat:
As above.
>Fix: