On Mon, 2002-03-11 at 14:24, Dave Alden wrote:
> Hi,
>
> On Tue, Mar 12, 2002 at 06:56:42AM +1100, Keith Owens wrote:
> > Under kdb, do
> > set LINES 2000
> > ps
> > bta
> > The fact that you can do control-A and get to kdb means that it is not
> > a complete kernel lockup, rather some code is sitting in interruptible
> > wait. ps and bta will show which tasks are waiting and why.
>
> Ok -- here it is. :-)
Yep, there it is:
Stack traceback for pid 859
EBP EIP Function(args)
0xf60cdbb8 0xc01145c2 schedule+0x322 (0x0, 0xf60cc000, 0x0, 0x0, 0x0)
kernel .text 0xc0100000 0xc01142a0 0xc01145f0
0xc01325c2 kmap_high+0x92 (0xc0136671, 0x0, 0x0, 0xb10, 0x0)
kernel .text 0xc0100000 0xc0132530 0xc0132680
0xc022721b __pb_block_prepare_write_async+0x7b (0xee4eb6c0,
0xc27e3740, 0x0, 0x1000, 0x1)
kernel .text 0xc0100000 0xc02271a0 0xc0227430
0xc022657e pagebuf_iozero+0x10e (0xee4eb6c0, 0xee2998c0, 0x0,
0x1000, 0xb12000)
kernel .text 0xc0100000 0xc0226470 0xc02266b0
0xc022c501 xfs_zero_eof+0x391 (0xee4eb7e4, 0xee7c1d3c, 0xb10000,
0x0, 0xb0a000)
kernel .text 0xc0100000 0xc022c170 0xc022c590
0xc022c790 xfs_write+0x200 (0xee7c1c44, 0xf60cde30, 0x8001, 0x0, 0x0)
kernel .text 0xc0100000 0xc022c590 0xc022ca10
0xc0228ae6 linvfs_write+0x2b6 (0xf60cdeb8, 0xf42640e8, 0x2000,
0xf60cded8, 0x0)
kernel .text 0xc0100000 0xc0228830 0xc0228b30
0xc0191515 nfsd_write+0x145 (0x0, 0x0, 0xf5d33e40, 0xc281f5c0,
0xc0445a80)
kernel .text 0xc0100000 0xc01913d0 0xc0191690
0xf60cded8 0xc02d63bb sock_sendmsg+0x6b (0xf60e0c00, 0xf60c8014, 0x0, 0x27, 0x7)
kernel .text 0xc0100000 0xc02d6350 0xc02d63e0
0xc031c2b9 svc_process+0x329 (0xf6101240, 0xf60e0c00, 0xf60cc570,
0xf6101240, 0xf60d7fd0)
kernel .text 0xc0100000 0xc031bf90 0xc031c470
0xc018d96b nfsd+0x1bb
kernel .text 0xc0100000 0xc018d7b0 0xc018daa0
0xc0105746 kernel_thread+0x26
kernel .text 0xc0100000 0xc0105720 0xc0105750
I was in there recently... let me take a look.
-Eric
--
Eric Sandeen XFS for Linux http://oss.sgi.com/projects/xfs
sandeen@xxxxxxx SGI, Inc.
|