witness warn

I'm running a -CURRENT built earlier today:
FreeBSD 5.2-CURRENT #0: Tue Jul  6 15:34:51 EEST 2004

While NFS-sharing parts of my disk to a Solaris machine, I saw these
messages flashing by on my console:

Jul  6 17:41:04 orion kernel: malloc(M_WAITOK) of "Mbuf", forcing M_NOWAIT with the following non-sleepable locks held:
Jul  6 17:41:04 orion kernel: exclusive sleep mutex inp (tcpinp) r = 0 (0xc174b630) locked _at_ /usr/src/sys/netinet/tcp_usrreq.c:592
Jul  6 17:41:04 orion kernel: exclusive sleep mutex tcp r = 0 (0xc06a6f6c) locked _at_ /usr/src/sys/netinet/tcp_usrreq.c:592
Jul  6 17:41:04 orion kernel: Stack backtrace:
Jul  6 17:41:04 orion kernel: backtrace(2,c16dbd00,c0c44b40,0,d1b7fad8) at backtrace+0x12
Jul  6 17:41:04 orion kernel: witness_warn(5,0,c06360cb,c0622260) at witness_warn+0x18e
Jul  6 17:41:04 orion kernel: uma_zalloc_arg(c0c44b40,d1b7faf0,2) at uma_zalloc_arg+0x38
Jul  6 17:41:04 orion kernel: nfs_realign(c190bca4,28,c1769a30,80000094,c1769a30) at nfs_realign+0x92
Jul  6 17:41:04 orion kernel: nfsrv_getstream(c190bc80,1) at nfsrv_getstream+0x2dd
Jul  6 17:41:04 orion kernel: nfsrv_rcv(c17699e0,c190bc80,1) at nfsrv_rcv+0x147
Jul  6 17:41:04 orion kernel: sowakeup(c17699e0,c1769a30) at sowakeup+0xb1
Jul  6 17:41:04 orion kernel: tcp_input(c16dbd00,14,0,14,8200060a) at tcp_input+0x249e
Jul  6 17:41:04 orion kernel: ip_input(c16dbd00) at ip_input+0x832
Jul  6 17:41:04 orion kernel: netisr_processqueue(c06a6178,c144dbc0,c1451500,d1b7fd1c,c04c4814) at netisr_processqueue+0x6e
Jul  6 17:41:04 orion kernel: swi_net(0) at swi_net+0x85
Jul  6 17:41:04 orion kernel: ithread_loop(c1451500,d1b7fd48,c1451500,c04c46e0,0) at ithread_loop+0x134
Jul  6 17:41:04 orion kernel: fork_exit(c04c46e0,c1451500,d1b7fd48) at fork_exit+0x98
Jul  6 17:41:04 orion kernel: fork_trampoline() at fork_trampoline+0x8
Jul  6 17:41:04 orion kernel: --- trap 0x1, eip = 0, esp = 0xd1b7fd7c, ebp = 0 ---

How do I go about investigating the cause of this?

Giorgos