On Apr 21, 2014, at 6:01 PM, Chuck Lever <[email protected]> wrote:
> Section 4 of RFC 5667 (NFS/RDMA) says: > >> The server MUST ignore any Read list for other NFS procedures, >> as well as additional Read list entries beyond the first in the >> list. > > Our XDR code adds a zero pad at the end of NFS WRITEs and SYMLINKs > whose content is not a multiple of 4 octets long. xprtrdma treats > the tail buffer containing the zero pad as a separate read chunk, > which the server ignores. > > Enable the pad optimization so our NFS client avoids sending zeroes > the server is just going to ignore. Looks like the Linux NFS/RDMA server is not spec compliant here. When pad optimization is enabled, the Linux NFS/RDMA server returns GARBAGEARGS for WRITEs whose length does not align with an XDR_QUAD. Dropping this one for now. > Signed-off-by: Chuck Lever <[email protected]> > --- > > net/sunrpc/xprtrdma/transport.c | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c > index 1eb9c46..f94a6c4 100644 > --- a/net/sunrpc/xprtrdma/transport.c > +++ b/net/sunrpc/xprtrdma/transport.c > @@ -73,7 +73,7 @@ static unsigned int xprt_rdma_max_inline_read = > RPCRDMA_DEF_INLINE; > static unsigned int xprt_rdma_max_inline_write = RPCRDMA_DEF_INLINE; > static unsigned int xprt_rdma_inline_write_padding; > static unsigned int xprt_rdma_memreg_strategy = RPCRDMA_FRMR; > - int xprt_rdma_pad_optimize = 0; > +int xprt_rdma_pad_optimize = 1; > > #ifdef RPC_DEBUG > > > -- > To unsubscribe from this list: send the line "unsubscribe linux-rdma" in > the body of a message to [email protected] > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Chuck Lever chuck[dot]lever[at]oracle[dot]com -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
