svcrpc: take advantage of tcp autotuning
authorOlga Kornievskaia <aglo@citi.umich.edu>
Tue, 21 Oct 2008 18:13:47 +0000 (14:13 -0400)
committerJ. Bruce Fields <bfields@redhat.com>
Thu, 7 Apr 2011 18:36:40 +0000 (14:36 -0400)
commit9660439861aa8dbd5e2b8087f33e20760c2c9afc
tree612460d547807056e10cebcd97156a555601d6c3
parent31d68ef65c7d49def19c1bae4e01b87d66cf5a56
svcrpc: take advantage of tcp autotuning

Allow the NFSv4 server to make use of TCP autotuning behaviour, which
was previously disabled by setting the sk_userlocks variable.

Set the receive buffers to be big enough to receive the whole RPC
request, and set this for the listening socket, not the accept socket.

Remove the code that readjusts the receive/send buffer sizes for the
accepted socket. Previously this code was used to influence the TCP
window management behaviour, which is no longer needed when autotuning
is enabled.

This can improve IO bandwidth on networks with high bandwidth-delay
products, where a large tcp window is required.  It also simplifies
performance tuning, since getting adequate tcp buffers previously
required increasing the number of nfsd threads.

Signed-off-by: Olga Kornievskaia <aglo@citi.umich.edu>
Cc: Jim Rees <rees@umich.edu>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
net/sunrpc/svcsock.c