nfsd: Fix NFSv4 READ on RDMA when using readv

svcrdma expects that the payload falls precisely into the xdr_buf
page vector. This does not seem to be the case for
nfsd4_encode_readv().

This code is called only when fops->splice_read is missing or when
RQ_SPLICE_OK is clear, so it's not a noticeable problem in many
common cases.

Add new transport method: ->xpo_read_payload so that when a READ
payload does not fit exactly in rq_res's page vector, the XDR
encoder can inform the RPC transport exactly where that payload is,
without the payload's XDR pad.

That way, when a Write chunk is present, the transport knows what
byte range in the Reply message is supposed to be matched with the
chunk.

Note that the Linux NFS server implementation of NFS/RDMA can
currently handle only one Write chunk per RPC-over-RDMA message.
This simplifies the implementation of this fix.

Fixes: b04209806384 ("nfsd4: allow exotic read compounds")
Buglink: https://bugzilla.kernel.org/show_bug.cgi?id=198053
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>

1 files changed, 39 insertions, 1 deletions

diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c
index f3f108090aa4..a11983c2056f 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c
@@ -858,7 +858,18 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
 
 	if (wr_lst) {
 		/* XXX: Presume the client sent only one Write chunk */
-		ret = svc_rdma_send_write_chunk(rdma, wr_lst, xdr);
+		unsigned long offset;
+		unsigned int length;
+
+		if (rctxt->rc_read_payload_length) {
+			offset = rctxt->rc_read_payload_offset;
+			length = rctxt->rc_read_payload_length;
+		} else {
+			offset = xdr->head[0].iov_len;
+			length = xdr->page_len;
+		}
+		ret = svc_rdma_send_write_chunk(rdma, wr_lst, xdr, offset,
+						length);
 		if (ret < 0)
 			goto err2;
 		svc_rdma_xdr_encode_write_list(rdma_resp, wr_lst, ret);
@@ -900,3 +911,30 @@ out:
 	ret = -ENOTCONN;
 	goto out;
 }
+
+/**
+ * svc_rdma_read_payload - special processing for a READ payload
+ * @rqstp: svc_rqst to operate on
+ * @offset: payload's byte offset in @xdr
+ * @length: size of payload, in bytes
+ *
+ * Returns zero on success.
+ *
+ * For the moment, just record the xdr_buf location of the READ
+ * payload. svc_rdma_sendto will use that location later when
+ * we actually send the payload.
+ */
+int svc_rdma_read_payload(struct svc_rqst *rqstp, unsigned int offset,
+			  unsigned int length)
+{
+	struct svc_rdma_recv_ctxt *rctxt = rqstp->rq_xprt_ctxt;
+
+	/* XXX: Just one READ payload slot for now, since our
+	 * transport implementation currently supports only one
+	 * Write chunk.
+	 */
+	rctxt->rc_read_payload_offset = offset;
+	rctxt->rc_read_payload_length = length;
+
+	return 0;
+}