> The client regression was due to additional XDR sanity checking
> (64bd577ea0021f5903505de061b3b7d8a785ee94) that exposed a latent bug in
> the NFSRDMA client. The bug is that if there were inline data, then the
> rpcrdma_fixup_inline function would would incorrectly set the XDR page_len
> to zero. Since the decode_read3resok logic previously computed this
> value from the reported len - header len, the bug was not symptomatic.
Oops: The client bug has already been fixed by Chuck's patch which is already
in Trond's nfs-for-next branch:
----
commit 2b7bbc963da8d076f263574af4138b5df2e1581f
Author: Chuck Lever <chuck.lever@xxxxxxxxxx>
Date: Wed Mar 12 12:51:30 2014 -0400
SUNRPC: Fix large reads on NFS/RDMA
----
So I'll repost Tom's server-side fix, after applying it on nfs-for-next and verifying things
still work.
Steve.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
