still problems w/ uvm_pagealloc_pgfl

[Kurt Schreiner <ks%ub.uni-mainz.de@localhost>]

Hi,

playing w/ a -current kernel (from Feb 13) I got the panic
I described in 
http://mail-index.netbsd.org/current-users/2008/01/28/0008.html
and
http://mail-index.netbsd.org/current-users/2008/01/25/0016.html
again.
I traced it down to the TAILQ_REMOVE call at line 1030 of src/sys/uvm/uvm_page.c
Defining QUEUEDEBUG gives a little more information (see below)
but I've still no idea what's really going on.
As a datapoint: If the file system where i/o ist going to is mounted
w/o softdeps, I can't trigger the panic, so softdep is at least part of
the game. On the other hand the 2nd panic below happens while savecore is
writing to the dump device... (This is a dual Atlon 2222 machine.)

Kurt


Feb 14 12:43:07 iwilma savecore: reboot after panic: 666 1665 1664 1663 1662 
1661 1660 1659 1658 1657 1656 1655 1654 1653 1652 16
panic: TAILQ_* forw 0xffff80000708e780 /u/NetBSD/src/sys/uvm/uvm_page.c:1030
Stopped in pid 5509.1 (savecore) at     netbsd:breakpoint+0x5:  leave
db{0}> bt
breakpoint() at netbsd:breakpoint+0x5
panic() at netbsd:panic+0x1ec
uvm_pagealloc_pgfl() at netbsd:uvm_pagealloc_pgfl+0x156
uvm_pagealloc_strat() at netbsd:uvm_pagealloc_strat+0x1d9
uvn_findpage() at netbsd:uvn_findpage+0xae
uvn_findpages() at netbsd:uvn_findpages+0x7c
genfs_getpages() at netbsd:genfs_getpages+0x599
VOP_GETPAGES() at netbsd:VOP_GETPAGES+0x8e
ubc_alloc() at netbsd:ubc_alloc+0x248
ubc_uiomove() at netbsd:ubc_uiomove+0x7e
ffs_write() at netbsd:ffs_write+0x805
VOP_WRITE() at netbsd:VOP_WRITE+0x72
vn_write() at netbsd:vn_write+0xe7
dofilewrite() at netbsd:dofilewrite+0x8a
sys_write() at netbsd:sys_write+0x9f
syscall() at netbsd:syscall+0xaf
db{0}> reboot 0x104

dumping to dev 4,17 offset 16031
dump 8190 8189 8188 8187 8186 ..... 7941
panic: TAILQ_* forw 0xffff80000708e780 /u/NetBSD/src/sys/uvm/uvm_page.c:1030
Stopped in pid 5509.1 (savecore) at     netbsd:breakpoint+0x5:  leave
db{0}> reboot 0x104

dumping to dev 4,17 offset 16031
dump kernel: double fault trap, code=0
Stopped in pid 5509.1 (savecore) at     netbsd:kprintf+0x1d:    movl    %esi,0xf
fffffffffffff7c(%rbp)
db{0}> reboot 0x104

And here the machine hangs. Powerbutton is needed to bring
it up again.