NFS: server asp, readdir reply truncated
NFS: nr=40, slots=4, len=9
or in some cases:
NFS: server asp, readdir reply truncated
NFS: nr=33, slots=-1, len=0
and from that point on we cannot perform any readdir on that file
system (they all return empty lists).
The problem only seems to happen with mounts done with rsize=1024 (and
wsize=1024), rsize=512 or rsize=8192 both seem to work fine (though
the latter may simply be that the readdir data fits into the buffer...
I can reproduce this with 2.2.7ac4 and 2.2.10ac2 (though I've not
tested a non-ac version yet).
Looking at the source of 2.2.10ac2 the printk seems to be trying to
detect that the server wants the buffer size specified in longwords
(as it replied with a larger buffer than we have space for).
I can provide network dumps if anyone understands why this might
happen or why 2.0 kernels never had a problem with these particular
servers...
-- Jon Peatfield, DAMTP, Computer Officer, University of Cambridge Telephone: +44 1223 3 37852 Mail: J.S.Peatfield@damtp.cam.ac.uk- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/