Interesting panic

[John Klos <john%ziaspace.com@localhost>]

I have a pretty heavily used machine running netbsd-4 compiled from 
24-September. Yesterday, it had a few hours of VERY heavy usage - over a 
hundred php-running Apache processes vying for CPU. At one point, it 
paniced:

panic: pool_get: kqueuepl: page empty
Begin traceback...
0xd79fee80: at pool_get+0x2b8
0xd79feea0: at sys_kqueue+0xa0
0xd79feee0: at syscall_plain+0x238
0xd79fef40: user SC trap #344 by 0xefb8a640: srr1=0xd032
            r1=0xffffd570 cr=0x24004028 xer=0 ctr=0xefb8a638
End traceback...
syncing disks... trap: kernel ISI by 0x93a100a0 (SRR1 0x40009032), lr: 
0x309fac
panic: trap
Begin traceback...
0xd562ba70: at trap+0x124
0xd562baf0: kernel ISI trap by _prop_array_pool+0x93478360: 
srr1=0x40009032
            r1=0xd562bbb0 cr=0x22042044 xer=0 ctr=0x93a100a0
0xd562bbb0: at knote+0x6c
0xd562bbd0: at sowakeup+0x44
0xd562bbf0: at sonewconn+0x194
0xd562bc20: at syn_cache_get+0x1a0
0xd562bc70: at tcp_input+0xde8
0xd562bde0: at tcp6_input+0x8c
0xd562bdf0: at ip6_input+0xb60
0xd562be50: at ip6intr+0x70
0xd562be80: at softintr__run+0x58
0xd562bea0: at do_pending_int+0x1f4
0xd562bee0: at splx+0x28
0xd562bef0: at wakeup+0x34
0xd562bf10: at uvm_aiodone_daemon+0x124
0xd562bf40: at cpu_switchto+0x44
0xd562bf50: at _prop_array_pool+0x3c683bc
saved LR(0x53a) is invalid.End traceback...
dumpsys: TBD
sd0(esiop1:0:0:0):  Check Condition on CDB: 0x35 00 00 00 00 00 00 00 00 
00
    SENSE KEY:  Media Error
     ASC/ASCQ:  Data Synchronization Mark Error

sd0: cache synchronization failed
rebooting


What, specifically, is kqueuepl? And how does one go about ensuring that 
one won't run out of it?

Thanks,
John Klos