Re: kern/12404 (panic: ffs_alloccg: map corrupted)

[Havard Eidnes <he%uninett.no@localhost>]

It just happened again (twice today), stack backtrace is:

start = 1, len = 11798, fs = /u
offset=3086 3086
cg 324
panic: ffs_alloccg: map corrupted
Stopped in pid 5659.1 (i386--netbsdelf-) at     netbsd:breakpoint+0x1:  ret
db{1}> 
db{1}> tra
breakpoint(c0976e84,144,c0e,c358c0d4,c0a738e0) at netbsd:breakpoint+0x1
ffs_alloccgblk(1d2a8e8,0,5,4000,ffffffff) at netbsd:ffs_alloccgblk
ffs_alloccg(d13a229c,144,1d2a8e8,0,2800) at netbsd:ffs_alloccg+0x258
ffs_hashalloc(1d2a8e8,0,2800,c03991e0,0) at netbsd:ffs_hashalloc+0x3a
ffs_alloc(d13a229c,0,0,1d2a8e8,0) at netbsd:ffs_alloc+0x1ae
ffs_balloc_ufs1(271c,cde09540,0,0,cde09540) at netbsd:ffs_balloc_ufs1+0x487
ffs_balloc(def842a0,0,0,271c,cde09540) at netbsd:ffs_balloc+0x71
ufs_gop_alloc(def842a0,0,0,271c,0) at netbsd:ufs_gop_alloc+0xc4
ufs_balloc_range(def842a0,0,0,271c,0) at netbsd:ufs_balloc_range+0x288
ffs_write(ce118b34,cf021a80,ce118b4c,c0494252,def842a0) at 
netbsd:ffs_write+0x9c7
VOP_WRITE(def842a0,ce118bac,10,cde09540,d63ea574) at netbsd:VOP_WRITE+0x4e
vn_write(d48402d8,d4840304,ce118bac,cde09540,1) at netbsd:vn_write+0xdf
dofilewrite(28,d48402d8,bbacc000,271c,d4840304) at netbsd:dofilewrite+0x75
sys_write(cf021a80,ce118c30,ce118c58,bfbec400,bfbec000) at netbsd:sys_write+0x9c
syscall(ce118c78,b3,ab,1f,1f) at netbsd:syscall+0x134

This is on a 2x HTT Xeon (appears to have 4 CPUs) running 4.99.40.

I'll keep the core-dump for a few weeks if someone is willing to look
closer into what might be the problem.

Regards,

- Håvard