Jumbo allocation failures on msk0

[Stuart Brooks <stuartb%cat.co.za@localhost>]

Hi,

I am running NetBSD 3 and am seeing a lot of these errors on an msk0 card. Is 
this something to worry about? There is a lot of UDP traffic being received 
(approx 50Mbps) as well as a lot of disk activity as the data is being dumped 
to a RAID device via a scsi RAID controller.  I am seeing strange happenings, 
telnet sessions to this unit sometimes just lock up and never seem to return 
amongst other things.

Details:
Mar 25 13:04:17 HASSELT01 /netbsd: msk0 jumbo allocation failed -- packet 
dropped!
Mar 25 13:14:00 HASSELT01 last message repeated 2632 times
Mar 25 13:23:41 HASSELT01 last message repeated 3269 times
Mar 25 13:30:50 HASSELT01 last message repeated 1867 times

Mar 25 15:43:21 HASSELT01 /netbsd: NetBSD 3.1_STABLE (V5_GENERIC) #22: Thu Feb 
26 17:42:43 SAST 2009
Mar 25 15:43:21 HASSELT01 /netbsd: 
root@netbsd30_release:/usr/src/sys/arch/i386/compile/V5_GENERIC
Mar 25 15:43:21 HASSELT01 /netbsd: total memory = 503 MB

Mar 25 15:43:21 HASSELT01 /netbsd: mskc0 at pci2 dev 0 function 0, Marvell 
Yukon-2 EC rev. A3 (0x2): irq 15
Mar 25 15:43:21 HASSELT01 /netbsd: msk0 at mskc0 port A: Ethernet address 
00:01:29:09:23:2c
Mar 25 15:43:21 HASSELT01 /netbsd: msk_jumbo_buf = 0xcbb24000

Is there anything I can do to prevent this from happening - should I configure 
my switch to disable jumbo packets or is there something else I can try? And 
would running NetBSD 4 deal with this?

Any help or ideas would be appreciated,
Thanks
 Stuart