On Wed, Aug 28, 2002 at 08:54:42PM +0200, Manuel Bouyer wrote:
> OK, I can now reproduce this as well. With GENERIC I get a panic in pool_get,
> with GENERIC_DIAGNOSTIC a panic on filesystem corruption ("bad dir").
> I can reproduce this with a i386 NFS server, but not with a sparc64.

I just got a different panic:
pool pnbufpl: putting with none out
panic: pool_put

in pagedaemon. I got a core dump:
(gdb) where
#0  0x1 in ?? ()
#1  0xc036b52f in cpu_reboot ()
#2  0xc0263fa9 in db_reboot_cmd ()
#3  0xc0263c84 in db_command ()
#4  0xc0263883 in db_command_loop ()
#5  0xc0267354 in db_trap ()
#6  0xc03682e8 in kdb_trap ()
#7  0xc0375b2f in trap ()
#8  0xc0100e81 in calltrap ()
#9  0xc028f979 in panic ()
#10 0xc028e36a in pool_do_put ()
#11 0xc028f594 in pool_cache_do_invalidate ()
#12 0xc028f654 in pool_cache_reclaim ()
#13 0xc028ea95 in pool_reclaim ()
#14 0xc028ed69 in pool_drain ()
#15 0xc03628a6 in uvm_pageout ()

This, as well as the previous panics, looks like memory corruption.

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--