Xen41 dom0 panic when downloading iso with wget

[Roger Pau Monné <roger.pau%entel.upc.edu@localhost>]

Hello,

Today I was playing with Xen41, on a NetBSD-current machine, and got
the following panic while downloading a iso with wget at 5M/s:

panic: pool 'pvpl' is IPL_NONE, but called from interrupt context

cpu0: Begin traceback...
copyright() at ffffffff809a3162
rb_tree_find_node() at netbsd:rb_tree_find_node+0x4a
?() at ffffffff
pool_cache_get_paddr() at netbsd:pool_cache_get_paddr+0x125
pmap_enter_ma() at netbsd:pmap_enter_ma+0x157
pmap_enter() at netbsd:pmap_enter+0x5a
_bus_dmamem_map() at netbsd:_bus_dmamem_map+0xe7
_bus_dma_alloc_bouncebuf() at netbsd:_bus_dma_alloc_bouncebuf+0xa0
_bus_dmamap_create() at netbsd:_bus_dmamap_create+0x178
bnx_alloc_pkts() at netbsd:bnx_alloc_pkts+0x66
bnx_tx_encap() at netbsd:bnx_tx_encap+0x5c
bnx_start() at netbsd:bnx_start+0x7e
ifq_enqueue() at netbsd:ifq_enqueue+0xbe
ether_output() at netbsd:ether_output+0x657
ip_output() at netbsd:ip_output+0xba0
tcp_output() at netbsd:tcp_output+0x1503
tcp_input() at netbsd:tcp_input+0x328d
ip_input() at netbsd:ip_input+0x452
ipintr() at netbsd:ipintr+0x70
softint_thread() at netbsd:softint_thread+0x92
cpu0: End traceback...

dump to dev 4,1 not possible
rebooting...
(XEN) Domain 0 shutdown: rebooting machine.

I'm running Xen41/amd64 using the XEN3_DOM0 kernel and the following
boot options:

menu=Boot Xen4.1 (serial):load /netbsd.xen console=com0;multiboot
/xen.gz dom0_mem=512M dom0_max_vcpus=1 dom0_vcpus_pin console=com1
com1=115200,8n1

I don't really know if this issue is related to the network card (I
see on the stack calls to bnx functions, which are probably related to
my Broadcom Netextreme II card), but I think it's best to ask on this
list first.

Thanks and Regards, Roger.