something write-locked afs_bufferLock (pid 14694) and isn't letting go.

adb> 0t14694$<pid2stack
?(0xffe157f5,0xf2e7120,0x716c87a4,0x0,0x5,0x8)
ufs_iinactive(0x716c86c0,0x704e1380,0x20,0x716c87a4,0x707d4c64,0x21) + c
ufs_rmidle(0x716c87a8,0x716c86c0,0x104469fc,0x10446a14,0x0,0x0) + a0
ufs_iget(0x701a3a70,0x778e,0x716c87a4,0x716c8748,0x716c86c0,0x800003) + 150
getinode(0x701bfb08,0x800003,0x778e,0x407a95c4,0x704e1380,0x407a9630) + 8c
igetinode(0x701bfb08,0x800003,0x778e,0x407a9638,0x704e1380,0x0) + 50
osi_UfsOpen(0x778e,0x10435164,0xc0f10,0xc,0x707d4ccc,0x722ec4d0) + 60
osi_UFSOpen(0x778e,0xf8,0x707d4ccc,0x800,0x707d5450,0x64) + a0
DRead(0x407a98cc,0xf8,0x722ec4d0,0x800,0x0,0x707d4ccc) + 59c
afs_dir_GetBlob(0x407a98cc,0x3e0f,0x0,0x16,0x707d4c64,0x21) + 14
FindItem(0x407a98cc,0x407a99f8,0x707ecca0,0x707ecca2,0x707f6800,0x16) + 110
afs_dir_LookupOffset(0x407a98cc,0x407a99f8,0x407a98c0,0x407a98e0,0x407a98d0,0x1) + 1c
afs_lookup(0x708f15a8,0x709a09ec,0x407a99f4,0x407a9c08,0x0,0x0) + 8b0
gafs_lookup(0x708f15a8,0x407a99f8,0x407a99f4,0x407a9c08,0x0,0x705e9e90) + 48
lookuppnvp(0x0,0x10436260,0x704e1380,0x10433d7c,0x407a9c7c,0x0) + 268
lookuppn(0x407a9c08,0x0,0x0,0x0,0x407a9c7c,0x708f15a8) + 10c
lookupname(0x0,0x0,0x0,0x0,0x407a9c7c,0x2227a) + 9c

crash> p #14694
PROC TABLE SIZE = 11962
SLOT ST  PID  PPID  PGID   SID   UID PRI   NAME        FLAGS
 324 s 14694  6589 14694  6589  2430  60 du             load

adb> $<kps
[...]
0x71ebee8c:     PID=14694
0x704e1384:     UID=2430
0x726e3120:     process args=du -k .

crash> ui 716c86c0
UFS INODE MAX TABLE SIZE = 10000
ADDR      MAJ/MIN     INUMB RCNT LINK  UID   GID          SIZE    MODE  FLAGS
716c86c0   32,   3    30606   2    1     0     0       1007616  f---600 md mt rf

other processes seem to be blocked like this:

adb> 0t17150$<pid2stack
?(0x7307e26c,0x701d9508,0x737aec40,0x0,0x1,0x73918ab8)
wait_for_lock(0x7307e1e0,0x7307e26c,0x1000,0x701d9508,0x1000c87c,0x0) + 30
flk_wait_execute_request(0x7307e1e0,0x7307e1e0,0x701d9508,0x0,0x414d,0x737d07f8) + 188
flk_process_request(0x0,0x7307e1e0,0x1,0x7307ff40,0x7345ac18,0x7307e640) + 750
reclock(0x701d9508,0x415a1c40,0x2000,0x7307e1e0,0x2,0x0) + 3e4
ufs_frlock(0x7345ac18,0x7,0x415a1c40,0x3,0x0,0x0) + 70
fcntl(0x0,0x0,0xffbefc48,0x0,0x0,0x3) + 358

crash> p #17150
PROC TABLE SIZE = 11962
SLOT ST  PID  PPID  PGID   SID   UID PRI   NAME        FLAGS
 820 s 17150 17149   228   228     0  30 cleanup        load

(not much useful information to be extracted because the call
arguments are no longer valid)

back to the first process holding afs_bufferLock:

all-night-tool.mit.edu# find /var/usr/vice -inum 30606
<un-interruptably blocked>

adb> 0t21248$<pid2stack
?(0xffe13dfa,0xeb66b4b9,0x716c87a4,0x0,0x7,0xb)
ufs_iget(0x701a3a70,0x778e,0x716c87a4,0x716c8748,0x716c86c0,0x800003) + 138
ufs_dirlook(0x709c5d48,0x709c5da0,0x0,0xdf60,0x778e,0x709c5cc0) + 408
ufs_lookup(0x0,0x4213d9f8,0x4213d9f4,0x748885a0,0x0,0x709c5cc0) + 100
lookuppnvp(0x0,0x10436260,0x748885a0,0x10433d7c,0x4213dc7c,0x0) + 268
lookuppn(0x4213dc08,0x0,0x0,0x0,0x4213dc7c,0x709c5d48) + 10c
lookupname(0x0,0x0,0x0,0x0,0x4213dc7c,0xffbef440) + 9c
lstat64(0xffbef440,0xffbef020,0xffff8872,0x2d0,0x26cc8,0x3c0) + 14

all-night-tool.mit.edu# ls -id /var/usr/vice /var/usr/vice/cache
         2 /var/usr/vice             27008 /var/usr/vice/cache
all-night-tool.mit.edu# ls -i /var/usr/vice/cache | grep 30606
     30606 V3595

adb> 778e=D
                30606

so we're blocking on the same inode.. interesting..