[OpenAFS] Weird conflict between openafs kernel module and nvidia driver on Ubuntu

[Marcus Watts]

The failure case includes this:
[   25.904896] pan0: Dropping NETIF_F_UFO since no NETIF_F_HW_CSUM feature.
[   26.129274] Starting AFS cache scan...Memory cache: Allocating 64 dcache entries...foun
d 0 non-empty cache files (0%).
[   34.396041] afs: Lost contact with volume location server 75.146.134.33 in cell davison
linehome.name
[   34.396050] afs: Lost contact with volume location server 75.146.134.33 in cell davison
linehome.name
[   37.952175] allocation failed: out of vmalloc space - use vmalloc=<size> to increase si
ze.
[   37.952744] NVRM: RmInitAdapter failed! (0x23:0xffffffff:688)
[   37.952753] NVRM: rm_init_adapter(0) failed
[   39.550637] ADDRCONF(NETDEV_UP): eth0: link is not ready
[   39.552383] e1000: eth0: e1000_watchdog: NIC Link is Up 100 Mbps Full Duplex, Flow Cont
rol: RX/TX
[   39.552749] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[   39.708497] NET: Registered protocol family 17
[   41.088684] allocation failed: out of vmalloc space - use vmalloc=<size> to increase si
ze.
[   41.089244] NVRM: RmInitAdapter failed! (0x23:0xffffffff:688)
[   41.089252] NVRM: rm_init_adapter(0) failed
[   44.201590] allocation failed: out of vmalloc space - use vmalloc=<size> to increase si
ze.
[   44.202140] NVRM: RmInitAdapter failed! (0x23:0xffffffff:688)
[   44.202149] NVRM: rm_init_adapter(0) failed
[   49.936008] eth0: no IPv6 routers present

The success case has no message about "out of vmalloc" space.

The alternation of messages makes me think this is related to the NVRM
messages, even though they aren't prefixed with NVRM.

My guess is that setting "vmalloc=XXX" 
for some sufficiently large XXX might just work.

Another possibility; looks like this is using a memory cache.
Using an "on-disk" cache might fix the problem, or perhaps
using a smaller memory cache.

				-Marcus Watts