Merge tag 'nfsd-4.7' of git://linux-nfs.org/~bfields/linux

Pull nfsd updates from Bruce Fields:
 "A very quiet cycle for nfsd, mainly just an RDMA update from Chuck
  Lever"

* tag 'nfsd-4.7' of git://linux-nfs.org/~bfields/linux:
  sunrpc: fix stripping of padded MIC tokens
  svcrpc: autoload rdma module
  svcrdma: Generalize svc_rdma_xdr_decode_req()
  svcrdma: Eliminate code duplication in svc_rdma_recvfrom()
  svcrdma: Drain QP before freeing svcrdma_xprt
  svcrdma: Post Receives only for forward channel requests
  svcrdma: Remove superfluous line from rdma_read_chunks()
  svcrdma: svc_rdma_put_context() is invoked twice in Send error path
  svcrdma: Do not add XDR padding to xdr_buf page vector
  svcrdma: Support IPv6 with NFS/RDMA
  nfsd: handle seqid wraparound in nfsd4_preprocess_layout_stateid
  Remove unnecessary allocation

11 files changed, 90 insertions, 72 deletions

diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c
index 93d5853f8c99..dba2ff8eaa68 100644
--- a/fs/nfsd/nfs3xdr.c
+++ b/fs/nfsd/nfs3xdr.c
@@ -379,7 +379,7 @@ nfs3svc_decode_writeargs(struct svc_rqst *rqstp, __be32 *p,
 	 */
 	hdr = (void*)p - rqstp->rq_arg.head[0].iov_base;
 	dlen = rqstp->rq_arg.head[0].iov_len + rqstp->rq_arg.page_len
-		- hdr;
+		+ rqstp->rq_arg.tail[0].iov_len - hdr;
 	/*
 	 * Round the length of the data which was specified up to
 	 * the next multiple of XDR units and then compare that
diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c
index 825c7bc8d789..953c0755cb37 100644
--- a/fs/nfsd/nfs4layouts.c
+++ b/fs/nfsd/nfs4layouts.c
@@ -289,7 +289,7 @@ nfsd4_preprocess_layout_stateid(struct svc_rqst *rqstp,
 
 		status = nfserr_bad_stateid;
 		mutex_lock(&ls->ls_mutex);
-		if (stateid->si_generation > stid->sc_stateid.si_generation)
+		if (nfsd4_stateid_generation_after(stateid, &stid->sc_stateid))
 			goto out_unlock_stid;
 		if (layout_type != ls->ls_layout_type)
 			goto out_unlock_stid;
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 0462eeddfff9..f5f82e145018 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -4651,12 +4651,6 @@ grace_disallows_io(struct net *net, struct inode *inode)
 	return opens_in_grace(net) && mandatory_lock(inode);
 }
 
-/* Returns true iff a is later than b: */
-static bool stateid_generation_after(stateid_t *a, stateid_t *b)
-{
-	return (s32)(a->si_generation - b->si_generation) > 0;
-}
-
 static __be32 check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_session)
 {
 	/*
@@ -4670,7 +4664,7 @@ static __be32 check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_s
 		return nfs_ok;
 
 	/* If the client sends us a stateid from the future, it's buggy: */
-	if (stateid_generation_after(in, ref))
+	if (nfsd4_stateid_generation_after(in, ref))
 		return nfserr_bad_stateid;
 	/*
 	 * However, we could see a stateid from the past, even from a
diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
index c050c53036a6..986e51e5ceac 100644
--- a/fs/nfsd/state.h
+++ b/fs/nfsd/state.h
@@ -573,6 +573,11 @@ enum nfsd4_cb_op {
 	NFSPROC4_CLNT_CB_SEQUENCE,
 };
 
+/* Returns true iff a is later than b: */
+static inline bool nfsd4_stateid_generation_after(stateid_t *a, stateid_t *b)
+{
+	return (s32)(a->si_generation - b->si_generation) > 0;
+}
 
 struct nfsd4_compound_state;
 struct nfsd_net;
diff --git a/include/linux/sunrpc/svc_rdma.h b/include/linux/sunrpc/svc_rdma.h
index 3081339968c3..d6917b896d3a 100644
--- a/include/linux/sunrpc/svc_rdma.h
+++ b/include/linux/sunrpc/svc_rdma.h
@@ -199,7 +199,7 @@ extern int svc_rdma_handle_bc_reply(struct rpc_xprt *xprt,
 				    struct xdr_buf *rcvbuf);
 
 /* svc_rdma_marshal.c */
-extern int svc_rdma_xdr_decode_req(struct rpcrdma_msg *, struct svc_rqst *);
+extern int svc_rdma_xdr_decode_req(struct xdr_buf *);
 extern int svc_rdma_xdr_encode_error(struct svcxprt_rdma *,
 				     struct rpcrdma_msg *,
 				     enum rpcrdma_errcode, __be32 *);
diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c
index 1095be9c80ab..e085f5ae1548 100644
--- a/net/sunrpc/auth_gss/svcauth_gss.c
+++ b/net/sunrpc/auth_gss/svcauth_gss.c
@@ -569,10 +569,9 @@ gss_svc_searchbyctx(struct cache_detail *cd, struct xdr_netobj *handle)
 	struct rsc *found;
 
 	memset(&rsci, 0, sizeof(rsci));
-	if (dup_to_netobj(&rsci.handle, handle->data, handle->len))
-		return NULL;
+	rsci.handle.data = handle->data;
+	rsci.handle.len = handle->len;
 	found = rsc_lookup(cd, &rsci);
-	rsc_free(&rsci);
 	if (!found)
 		return NULL;
 	if (cache_check(cd, &found->h, NULL))
@@ -857,8 +856,8 @@ unwrap_integ_data(struct svc_rqst *rqstp, struct xdr_buf *buf, u32 seq, struct g
 		goto out;
 	if (svc_getnl(&buf->head[0]) != seq)
 		goto out;
-	/* trim off the mic at the end before returning */
-	xdr_buf_trim(buf, mic.len + 4);
+	/* trim off the mic and padding at the end before returning */
+	xdr_buf_trim(buf, round_up_to_quad(mic.len) + 4);
 	stat = 0;
 out:
 	kfree(mic.data);
diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
index 7422f28818b2..f5572e31d518 100644
--- a/net/sunrpc/svc_xprt.c
+++ b/net/sunrpc/svc_xprt.c
@@ -244,13 +244,12 @@ void svc_add_new_perm_xprt(struct svc_serv *serv, struct svc_xprt *new)
 	svc_xprt_received(new);
 }
 
-int svc_create_xprt(struct svc_serv *serv, const char *xprt_name,
+int _svc_create_xprt(struct svc_serv *serv, const char *xprt_name,
 		    struct net *net, const int family,
 		    const unsigned short port, int flags)
 {
 	struct svc_xprt_class *xcl;
 
-	dprintk("svc: creating transport %s[%d]\n", xprt_name, port);
 	spin_lock(&svc_xprt_class_lock);
 	list_for_each_entry(xcl, &svc_xprt_class_list, xcl_list) {
 		struct svc_xprt *newxprt;
@@ -274,12 +273,28 @@ int svc_create_xprt(struct svc_serv *serv, const char *xprt_name,
 	}
  err:
 	spin_unlock(&svc_xprt_class_lock);
-	dprintk("svc: transport %s not found\n", xprt_name);
-
 	/* This errno is exposed to user space.  Provide a reasonable
 	 * perror msg for a bad transport. */
 	return -EPROTONOSUPPORT;
 }
+
+int svc_create_xprt(struct svc_serv *serv, const char *xprt_name,
+		    struct net *net, const int family,
+		    const unsigned short port, int flags)
+{
+	int err;
+
+	dprintk("svc: creating transport %s[%d]\n", xprt_name, port);
+	err = _svc_create_xprt(serv, xprt_name, net, family, port, flags);
+	if (err == -EPROTONOSUPPORT) {
+		request_module("svc%s", xprt_name);
+		err = _svc_create_xprt(serv, xprt_name, net, family, port, flags);
+	}
+	if (err)
+		dprintk("svc: transport %s not found, err %d\n",
+			xprt_name, err);
+	return err;
+}
 EXPORT_SYMBOL_GPL(svc_create_xprt);
 
 /*
diff --git a/net/sunrpc/xprtrdma/svc_rdma_marshal.c b/net/sunrpc/xprtrdma/svc_rdma_marshal.c
index 765bca47c74d..0ba9887f3e22 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_marshal.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_marshal.c
@@ -145,19 +145,32 @@ static __be32 *decode_reply_array(__be32 *va, __be32 *vaend)
 	return (__be32 *)&ary->wc_array[nchunks];
 }
 
-int svc_rdma_xdr_decode_req(struct rpcrdma_msg *rmsgp, struct svc_rqst *rqstp)
+/**
+ * svc_rdma_xdr_decode_req - Parse incoming RPC-over-RDMA header
+ * @rq_arg: Receive buffer
+ *
+ * On entry, xdr->head[0].iov_base points to first byte in the
+ * RPC-over-RDMA header.
+ *
+ * On successful exit, head[0] points to first byte past the
+ * RPC-over-RDMA header. For RDMA_MSG, this is the RPC message.
+ * The length of the RPC-over-RDMA header is returned.
+ */
+int svc_rdma_xdr_decode_req(struct xdr_buf *rq_arg)
 {
+	struct rpcrdma_msg *rmsgp;
 	__be32 *va, *vaend;
 	unsigned int len;
 	u32 hdr_len;
 
 	/* Verify that there's enough bytes for header + something */
-	if (rqstp->rq_arg.len <= RPCRDMA_HDRLEN_ERR) {
+	if (rq_arg->len <= RPCRDMA_HDRLEN_ERR) {
 		dprintk("svcrdma: header too short = %d\n",
-			rqstp->rq_arg.len);
+			rq_arg->len);
 		return -EINVAL;
 	}
 
+	rmsgp = (struct rpcrdma_msg *)rq_arg->head[0].iov_base;
 	if (rmsgp->rm_vers != rpcrdma_version) {
 		dprintk("%s: bad version %u\n", __func__,
 			be32_to_cpu(rmsgp->rm_vers));
@@ -189,10 +202,10 @@ int svc_rdma_xdr_decode_req(struct rpcrdma_msg *rmsgp, struct svc_rqst *rqstp)
 			be32_to_cpu(rmsgp->rm_body.rm_padded.rm_thresh);
 
 		va = &rmsgp->rm_body.rm_padded.rm_pempty[4];
-		rqstp->rq_arg.head[0].iov_base = va;
+		rq_arg->head[0].iov_base = va;
 		len = (u32)((unsigned long)va - (unsigned long)rmsgp);
-		rqstp->rq_arg.head[0].iov_len -= len;
-		if (len > rqstp->rq_arg.len)
+		rq_arg->head[0].iov_len -= len;
+		if (len > rq_arg->len)
 			return -EINVAL;
 		return len;
 	default:
@@ -205,7 +218,7 @@ int svc_rdma_xdr_decode_req(struct rpcrdma_msg *rmsgp, struct svc_rqst *rqstp)
 	 * chunk list and a reply chunk list.
 	 */
 	va = &rmsgp->rm_body.rm_chunks[0];
-	vaend = (__be32 *)((unsigned long)rmsgp + rqstp->rq_arg.len);
+	vaend = (__be32 *)((unsigned long)rmsgp + rq_arg->len);
 	va = decode_read_list(va, vaend);
 	if (!va) {
 		dprintk("svcrdma: failed to decode read list\n");
@@ -222,10 +235,9 @@ int svc_rdma_xdr_decode_req(struct rpcrdma_msg *rmsgp, struct svc_rqst *rqstp)
 		return -EINVAL;
 	}
 
-	rqstp->rq_arg.head[0].iov_base = va;
+	rq_arg->head[0].iov_base = va;
 	hdr_len = (unsigned long)va - (unsigned long)rmsgp;
-	rqstp->rq_arg.head[0].iov_len -= hdr_len;
-
+	rq_arg->head[0].iov_len -= hdr_len;
 	return hdr_len;
 }
 
diff --git a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
index fbe7444e7de6..2c25606f2561 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_recvfrom.c
@@ -447,10 +447,8 @@ static int rdma_read_chunks(struct svcxprt_rdma *xprt,
 	head->arg.len = rqstp->rq_arg.len;
 	head->arg.buflen = rqstp->rq_arg.buflen;
 
-	ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0];
-	position = be32_to_cpu(ch->rc_position);
-
 	/* RDMA_NOMSG: RDMA READ data should land just after RDMA RECV data */
+	position = be32_to_cpu(ch->rc_position);
 	if (position == 0) {
 		head->arg.pages = &head->pages[0];
 		page_offset = head->byte_len;
@@ -488,7 +486,7 @@ static int rdma_read_chunks(struct svcxprt_rdma *xprt,
 	if (page_offset & 3) {
 		u32 pad = 4 - (page_offset & 3);
 
-		head->arg.page_len += pad;
+		head->arg.tail[0].iov_len += pad;
 		head->arg.len += pad;
 		head->arg.buflen += pad;
 		page_offset += pad;
@@ -510,11 +508,10 @@ static int rdma_read_chunks(struct svcxprt_rdma *xprt,
 	return ret;
 }
 
-static int rdma_read_complete(struct svc_rqst *rqstp,
-			      struct svc_rdma_op_ctxt *head)
+static void rdma_read_complete(struct svc_rqst *rqstp,
+			       struct svc_rdma_op_ctxt *head)
 {
 	int page_no;
-	int ret;
 
 	/* Copy RPC pages */
 	for (page_no = 0; page_no < head->count; page_no++) {
@@ -550,23 +547,6 @@ static int rdma_read_complete(struct svc_rqst *rqstp,
 	rqstp->rq_arg.tail[0] = head->arg.tail[0];
 	rqstp->rq_arg.len = head->arg.len;
 	rqstp->rq_arg.buflen = head->arg.buflen;
-
-	/* Free the context */
-	svc_rdma_put_context(head, 0);
-
-	/* XXX: What should this be? */
-	rqstp->rq_prot = IPPROTO_MAX;
-	svc_xprt_copy_addrs(rqstp, rqstp->rq_xprt);
-
-	ret = rqstp->rq_arg.head[0].iov_len
-		+ rqstp->rq_arg.page_len
-		+ rqstp->rq_arg.tail[0].iov_len;
-	dprintk("svcrdma: deferred read ret=%d, rq_arg.len=%u, "
-		"rq_arg.head[0].iov_base=%p, rq_arg.head[0].iov_len=%zu\n",
-		ret, rqstp->rq_arg.len,	rqstp->rq_arg.head[0].iov_base,
-		rqstp->rq_arg.head[0].iov_len);
-
-	return ret;
 }
 
 /* By convention, backchannel calls arrive via rdma_msg type
@@ -624,7 +604,8 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp)
 				  dto_q);
 		list_del_init(&ctxt->dto_q);
 		spin_unlock_bh(&rdma_xprt->sc_rq_dto_lock);
-		return rdma_read_complete(rqstp, ctxt);
+		rdma_read_complete(rqstp, ctxt);
+		goto complete;
 	} else if (!list_empty(&rdma_xprt->sc_rq_dto_q)) {
 		ctxt = list_entry(rdma_xprt->sc_rq_dto_q.next,
 				  struct svc_rdma_op_ctxt,
@@ -655,7 +636,7 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp)
 
 	/* Decode the RDMA header. */
 	rmsgp = (struct rpcrdma_msg *)rqstp->rq_arg.head[0].iov_base;
-	ret = svc_rdma_xdr_decode_req(rmsgp, rqstp);
+	ret = svc_rdma_xdr_decode_req(&rqstp->rq_arg);
 	if (ret < 0)
 		goto out_err;
 	if (ret == 0)
@@ -682,6 +663,7 @@ int svc_rdma_recvfrom(struct svc_rqst *rqstp)
 		return 0;
 	}
 
+complete:
 	ret = rqstp->rq_arg.head[0].iov_len
 		+ rqstp->rq_arg.page_len
 		+ rqstp->rq_arg.tail[0].iov_len;
diff --git a/net/sunrpc/xprtrdma/svc_rdma_sendto.c b/net/sunrpc/xprtrdma/svc_rdma_sendto.c
index 4f1b1c4f45f9..54d533300620 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_sendto.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_sendto.c
@@ -463,25 +463,21 @@ static int send_reply(struct svcxprt_rdma *rdma,
 		      struct svc_rqst *rqstp,
 		      struct page *page,
 		      struct rpcrdma_msg *rdma_resp,
-		      struct svc_rdma_op_ctxt *ctxt,
 		      struct svc_rdma_req_map *vec,
 		      int byte_count)
 {
+	struct svc_rdma_op_ctxt *ctxt;
 	struct ib_send_wr send_wr;
 	u32 xdr_off;
 	int sge_no;
 	int sge_bytes;
 	int page_no;
 	int pages;
-	int ret;
-
-	ret = svc_rdma_repost_recv(rdma, GFP_KERNEL);
-	if (ret) {
-		svc_rdma_put_context(ctxt, 0);
-		return -ENOTCONN;
-	}
+	int ret = -EIO;
 
 	/* Prepare the context */
+	ctxt = svc_rdma_get_context(rdma);
+	ctxt->direction = DMA_TO_DEVICE;
 	ctxt->pages[0] = page;
 	ctxt->count = 1;
 
@@ -565,8 +561,7 @@ static int send_reply(struct svcxprt_rdma *rdma,
  err:
 	svc_rdma_unmap_dma(ctxt);
 	svc_rdma_put_context(ctxt, 1);
-	pr_err("svcrdma: failed to send reply, rc=%d\n", ret);
-	return -EIO;
+	return ret;
 }
 
 void svc_rdma_prep_reply_hdr(struct svc_rqst *rqstp)
@@ -585,7 +580,6 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
 	int ret;
 	int inline_bytes;
 	struct page *res_page;
-	struct svc_rdma_op_ctxt *ctxt;
 	struct svc_rdma_req_map *vec;
 
 	dprintk("svcrdma: sending response for rqstp=%p\n", rqstp);
@@ -598,8 +592,6 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
 	rp_ary = svc_rdma_get_reply_array(rdma_argp, wr_ary);
 
 	/* Build an req vec for the XDR */
-	ctxt = svc_rdma_get_context(rdma);
-	ctxt->direction = DMA_TO_DEVICE;
 	vec = svc_rdma_get_req_map(rdma);
 	ret = svc_rdma_map_xdr(rdma, &rqstp->rq_res, vec, wr_ary != NULL);
 	if (ret)
@@ -635,7 +627,12 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
 		inline_bytes -= ret;
 	}
 
-	ret = send_reply(rdma, rqstp, res_page, rdma_resp, ctxt, vec,
+	/* Post a fresh Receive buffer _before_ sending the reply */
+	ret = svc_rdma_post_recv(rdma, GFP_KERNEL);
+	if (ret)
+		goto err1;
+
+	ret = send_reply(rdma, rqstp, res_page, rdma_resp, vec,
 			 inline_bytes);
 	if (ret < 0)
 		goto err1;
@@ -648,7 +645,8 @@ int svc_rdma_sendto(struct svc_rqst *rqstp)
 	put_page(res_page);
  err0:
 	svc_rdma_put_req_map(rdma, vec);
-	svc_rdma_put_context(ctxt, 0);
+	pr_err("svcrdma: Could not send reply, err=%d. Closing transport.\n",
+	       ret);
 	set_bit(XPT_CLOSE, &rdma->sc_xprt.xpt_flags);
 	return -ENOTCONN;
 }
diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c
index 90668969d559..dd9440137834 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_transport.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c
@@ -789,7 +789,7 @@ static struct svc_xprt *svc_rdma_create(struct svc_serv *serv,
 	int ret;
 
 	dprintk("svcrdma: Creating RDMA socket\n");
-	if (sa->sa_family != AF_INET) {
+	if ((sa->sa_family != AF_INET) && (sa->sa_family != AF_INET6)) {
 		dprintk("svcrdma: Address family %d is not supported.\n", sa->sa_family);
 		return ERR_PTR(-EAFNOSUPPORT);
 	}
@@ -805,6 +805,16 @@ static struct svc_xprt *svc_rdma_create(struct svc_serv *serv,
 		goto err0;
 	}
 
+	/* Allow both IPv4 and IPv6 sockets to bind a single port
+	 * at the same time.
+	 */
+#if IS_ENABLED(CONFIG_IPV6)
+	ret = rdma_set_afonly(listen_id, 1);
+	if (ret) {
+		dprintk("svcrdma: rdma_set_afonly failed = %d\n", ret);
+		goto err1;
+	}
+#endif
 	ret = rdma_bind_addr(listen_id, sa);
 	if (ret) {
 		dprintk("svcrdma: rdma_bind_addr failed = %d\n", ret);
@@ -1073,7 +1083,7 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt)
 		newxprt->sc_dev_caps |= SVCRDMA_DEVCAP_READ_W_INV;
 
 	/* Post receive buffers */
-	for (i = 0; i < newxprt->sc_rq_depth; i++) {
+	for (i = 0; i < newxprt->sc_max_requests; i++) {
 		ret = svc_rdma_post_recv(newxprt, GFP_KERNEL);
 		if (ret) {
 			dprintk("svcrdma: failure posting receive buffers\n");
@@ -1170,6 +1180,9 @@ static void __svc_rdma_free(struct work_struct *work)
 
 	dprintk("svcrdma: %s(%p)\n", __func__, rdma);
 
+	if (rdma->sc_qp && !IS_ERR(rdma->sc_qp))
+		ib_drain_qp(rdma->sc_qp);
+
 	/* We should only be called from kref_put */
 	if (atomic_read(&xprt->xpt_ref.refcount) != 0)
 		pr_err("svcrdma: sc_xprt still in use? (%d)\n",