From: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org> To: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org Subject: [PATCH v2 1/9] svcrdma: Do not send XDR roundup bytes for a write chunk Date: Mon, 08 Feb 2016 12:24:31 -0500 [thread overview] Message-ID: <20160208172430.13423.18896.stgit@klimt.1015granger.net> (raw) In-Reply-To: <20160208171756.13423.14384.stgit-Hs+gFlyCn65vLzlybtyyYzGyq/o6K9yX@public.gmane.org> When the Linux server writes an odd-length data item into a Write chunk, it finishes with XDR pad bytes. If the data item is smaller than the Write chunk, the pad bytes are written at the end of the data item, but still inside the chunk. That can expose these zero bytes to the RPC consumer on the client. XDR pad bytes are inserted in order to preserve the XDR alignment of the next XDR data item in an XDR stream. But Write chunks do not appear in the payload XDR stream, and only one data item is allowed in each chunk. So XDR padding is unneeded. Thus the server should not write XDR pad bytes in Write chunks. I believe this is not an operational problem. Short NFS READs that are returned in a Write chunk would be affected by this issue. But they happen only when the read request goes past the EOF. Those are zero bytes anyway, and there's no file data in the client's buffer past EOF. Otherwise, current NFS clients provide a separate extra segment for catching XDR padding. If an odd-size data item fills the chunk, then the XDR pad will be written to the extra segment. Signed-off-by: Chuck Lever <chuck.lever-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org> Reviewed-By: Devesh Sharma <devesh.sharma-dY08KVG/lbpWk0Htik3J/w@public.gmane.org> Tested-By: Devesh Sharma <devesh.sharma-dY08KVG/lbpWk0Htik3J/w@public.gmane.org> --- net/sunrpc/xprtrdma/svc_rdma_sendto.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c index df57f3c..8591314 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c +++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c @@ -308,7 +308,7 @@ static int send_write_chunks(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp, struct svc_rdma_req_map *vec) { - u32 xfer_len = rqstp->rq_res.page_len + rqstp->rq_res.tail[0].iov_len; + u32 xfer_len = rqstp->rq_res.page_len; int write_len; u32 xdr_off; int chunk_off; -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html
WARNING: multiple messages have this Message-ID (diff)
From: Chuck Lever <chuck.lever@oracle.com> To: linux-rdma@vger.kernel.org, linux-nfs@vger.kernel.org Subject: [PATCH v2 1/9] svcrdma: Do not send XDR roundup bytes for a write chunk Date: Mon, 08 Feb 2016 12:24:31 -0500 [thread overview] Message-ID: <20160208172430.13423.18896.stgit@klimt.1015granger.net> (raw) In-Reply-To: <20160208171756.13423.14384.stgit@klimt.1015granger.net> When the Linux server writes an odd-length data item into a Write chunk, it finishes with XDR pad bytes. If the data item is smaller than the Write chunk, the pad bytes are written at the end of the data item, but still inside the chunk. That can expose these zero bytes to the RPC consumer on the client. XDR pad bytes are inserted in order to preserve the XDR alignment of the next XDR data item in an XDR stream. But Write chunks do not appear in the payload XDR stream, and only one data item is allowed in each chunk. So XDR padding is unneeded. Thus the server should not write XDR pad bytes in Write chunks. I believe this is not an operational problem. Short NFS READs that are returned in a Write chunk would be affected by this issue. But they happen only when the read request goes past the EOF. Those are zero bytes anyway, and there's no file data in the client's buffer past EOF. Otherwise, current NFS clients provide a separate extra segment for catching XDR padding. If an odd-size data item fills the chunk, then the XDR pad will be written to the extra segment. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Reviewed-By: Devesh Sharma <devesh.sharma@broadcom.com> Tested-By: Devesh Sharma <devesh.sharma@broadcom.com> --- net/sunrpc/xprtrdma/svc_rdma_sendto.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c index df57f3c..8591314 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c +++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c @@ -308,7 +308,7 @@ static int send_write_chunks(struct svcxprt_rdma *xprt, struct svc_rqst *rqstp, struct svc_rdma_req_map *vec) { - u32 xfer_len = rqstp->rq_res.page_len + rqstp->rq_res.tail[0].iov_len; + u32 xfer_len = rqstp->rq_res.page_len; int write_len; u32 xdr_off; int chunk_off;
next prev parent reply other threads:[~2016-02-08 17:24 UTC|newest] Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top 2016-02-08 17:24 [PATCH v2 0/9] NFS/RDMA server patches for v4.6 Chuck Lever 2016-02-08 17:24 ` Chuck Lever [not found] ` <20160208171756.13423.14384.stgit-Hs+gFlyCn65vLzlybtyyYzGyq/o6K9yX@public.gmane.org> 2016-02-08 17:24 ` Chuck Lever [this message] 2016-02-08 17:24 ` [PATCH v2 1/9] svcrdma: Do not send XDR roundup bytes for a write chunk Chuck Lever [not found] ` <20160208172430.13423.18896.stgit-Hs+gFlyCn65vLzlybtyyYzGyq/o6K9yX@public.gmane.org> 2016-02-08 19:21 ` J. Bruce Fields 2016-02-08 19:21 ` J. Bruce Fields [not found] ` <20160208192119.GF17411-uC3wQj2KruNg9hUCZPvPmw@public.gmane.org> 2016-02-08 20:03 ` Chuck Lever 2016-02-08 20:03 ` Chuck Lever 2016-02-08 17:24 ` [PATCH v2 2/9] svcrdma: svc_rdma_post_recv() should close connection on error Chuck Lever 2016-02-08 17:24 ` Chuck Lever 2016-02-08 17:24 ` [PATCH v2 3/9] rpcrdma: Add RPCRDMA_HDRLEN_ERR Chuck Lever 2016-02-08 17:24 ` Chuck Lever 2016-02-08 17:24 ` [PATCH v2 4/9] svcrdma: Make RDMA_ERROR messages work Chuck Lever 2016-02-08 17:24 ` Chuck Lever 2016-02-08 17:25 ` [PATCH v2 5/9] svcrdma: Use correct XID in error replies Chuck Lever 2016-02-08 17:25 ` Chuck Lever 2016-02-08 17:25 ` [PATCH v2 6/9] svcrdma: Hook up the logic to return ERR_CHUNK Chuck Lever 2016-02-08 17:25 ` Chuck Lever 2016-02-08 17:25 ` [PATCH v2 7/9] svcrdma: Remove close_out exit path Chuck Lever 2016-02-08 17:25 ` Chuck Lever 2016-02-08 17:25 ` [PATCH v2 8/9] svcrdma: Use new CQ API for RPC-over-RDMA server receive CQs Chuck Lever 2016-02-08 17:25 ` Chuck Lever 2016-02-08 17:25 ` [PATCH v2 9/9] svcrdma: Use new CQ API for RPC-over-RDMA server send CQs Chuck Lever 2016-02-08 17:25 ` Chuck Lever [not found] ` <20160208172537.13423.97332.stgit-Hs+gFlyCn65vLzlybtyyYzGyq/o6K9yX@public.gmane.org> 2016-02-11 18:46 ` Steve Wise 2016-02-11 18:46 ` Steve Wise 2016-02-11 18:48 ` Chuck Lever 2016-02-11 18:48 ` Chuck Lever
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20160208172430.13423.18896.stgit@klimt.1015granger.net \ --to=chuck.lever-qhclzuegtsvqt0dzr+alfa@public.gmane.org \ --cc=linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \ --cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.