Merge tag 'nfs-for-5.8-2' of git://git.linux-nfs.org/projects/anna/linux-nfs

Pull NFS client bugfixes from Anna Schumaker:
 "Stable Fixes:
   - xprtrdma: Fix handling of RDMA_ERROR replies
   - sunrpc: Fix rollback in rpc_gssd_dummy_populate()
   - pNFS/flexfiles: Fix list corruption if the mirror count changes
   - NFSv4: Fix CLOSE not waiting for direct IO completion
   - SUNRPC: Properly set the @subbuf parameter of xdr_buf_subsegment()

  Other Fixes:
   - xprtrdma: Fix a use-after-free with r_xprt->rx_ep
   - Fix other xprtrdma races during disconnect
   - NFS: Fix memory leak of export_path"

* tag 'nfs-for-5.8-2' of git://git.linux-nfs.org/projects/anna/linux-nfs:
  SUNRPC: Properly set the @subbuf parameter of xdr_buf_subsegment()
  NFSv4 fix CLOSE not waiting for direct IO compeletion
  pNFS/flexfiles: Fix list corruption if the mirror count changes
  nfs: Fix memory leak of export_path
  sunrpc: fixed rollback in rpc_gssd_dummy_populate()
  xprtrdma: Fix handling of RDMA_ERROR replies
  xprtrdma: Clean up disconnect
  xprtrdma: Clean up synopsis of rpcrdma_flush_disconnect()
  xprtrdma: Use re_connect_status safely in rpcrdma_xprt_connect()
  xprtrdma: Prevent dereferencing r_xprt->rx_ep after it is freed

11 files changed, 75 insertions, 49 deletions

diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 1b79dd5cf661..3d113cf8908a 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -267,8 +267,6 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq)
 {
 	struct inode *inode = dreq->inode;
 
-	inode_dio_end(inode);
-
 	if (dreq->iocb) {
 		long res = (long) dreq->error;
 		if (dreq->count != 0) {
@@ -280,7 +278,10 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq)
 
 	complete(&dreq->completion);
 
+	igrab(inode);
 	nfs_direct_req_release(dreq);
+	inode_dio_end(inode);
+	iput(inode);
 }
 
 static void nfs_direct_read_completion(struct nfs_pgio_header *hdr)
@@ -410,8 +411,10 @@ static ssize_t nfs_direct_read_schedule_iovec(struct nfs_direct_req *dreq,
 	 * generic layer handle the completion.
 	 */
 	if (requested_bytes == 0) {
-		inode_dio_end(inode);
+		igrab(inode);
 		nfs_direct_req_release(dreq);
+		inode_dio_end(inode);
+		iput(inode);
 		return result < 0 ? result : -EIO;
 	}
 
@@ -864,8 +867,10 @@ static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq,
 	 * generic layer handle the completion.
 	 */
 	if (requested_bytes == 0) {
-		inode_dio_end(inode);
+		igrab(inode);
 		nfs_direct_req_release(dreq);
+		inode_dio_end(inode);
+		iput(inode);
 		return result < 0 ? result : -EIO;
 	}
 
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index f96367a2463e..ccd6c1637b27 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -83,6 +83,7 @@ nfs_file_release(struct inode *inode, struct file *filp)
 	dprintk("NFS: release(%pD2)\n", filp);
 
 	nfs_inc_stats(inode, NFSIOS_VFSRELEASE);
+	inode_dio_wait(inode);
 	nfs_file_clear_open_context(filp);
 	return 0;
 }
diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c
index 7d399f72ebbb..de03e440b7ee 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.c
+++ b/fs/nfs/flexfilelayout/flexfilelayout.c
@@ -907,9 +907,8 @@ retry:
 		goto out_mds;
 
 	/* Use a direct mapping of ds_idx to pgio mirror_idx */
-	if (WARN_ON_ONCE(pgio->pg_mirror_count !=
-	    FF_LAYOUT_MIRROR_COUNT(pgio->pg_lseg)))
-		goto out_mds;
+	if (pgio->pg_mirror_count != FF_LAYOUT_MIRROR_COUNT(pgio->pg_lseg))
+		goto out_eagain;
 
 	for (i = 0; i < pgio->pg_mirror_count; i++) {
 		mirror = FF_LAYOUT_COMP(pgio->pg_lseg, i);
@@ -931,7 +930,10 @@ retry:
 			(NFS_MOUNT_SOFT|NFS_MOUNT_SOFTERR))
 		pgio->pg_maxretrans = io_maxretrans;
 	return;
-
+out_eagain:
+	pnfs_generic_pg_cleanup(pgio);
+	pgio->pg_error = -EAGAIN;
+	return;
 out_mds:
 	trace_pnfs_mds_fallback_pg_init_write(pgio->pg_inode,
 			0, NFS4_MAX_UINT64, IOMODE_RW,
@@ -941,6 +943,7 @@ out_mds:
 	pgio->pg_lseg = NULL;
 	pgio->pg_maxretrans = 0;
 	nfs_pageio_reset_write_mds(pgio);
+	pgio->pg_error = -EAGAIN;
 }
 
 static unsigned int
diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c
index a3ab6e219061..873342308dc0 100644
--- a/fs/nfs/nfs4namespace.c
+++ b/fs/nfs/nfs4namespace.c
@@ -308,6 +308,7 @@ static int try_location(struct fs_context *fc,
 	if (IS_ERR(export_path))
 		return PTR_ERR(export_path);
 
+	kfree(ctx->nfs_server.export_path);
 	ctx->nfs_server.export_path = export_path;
 
 	source = kmalloc(len + 1 + ctx->nfs_server.export_path_len + 1,
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index 39e14d5edaf1..e9d0953522f0 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -1317,6 +1317,7 @@ rpc_gssd_dummy_populate(struct dentry *root, struct rpc_pipe *pipe_data)
 	q.len = strlen(gssd_dummy_clnt_dir[0].name);
 	clnt_dentry = d_hash_and_lookup(gssd_dentry, &q);
 	if (!clnt_dentry) {
+		__rpc_depopulate(gssd_dentry, gssd_dummy_clnt_dir, 0, 1);
 		pipe_dentry = ERR_PTR(-ENOENT);
 		goto out;
 	}
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index 6f7d82fb1eb0..be11d672b5b9 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -1118,6 +1118,7 @@ xdr_buf_subsegment(struct xdr_buf *buf, struct xdr_buf *subbuf,
 		base = 0;
 	} else {
 		base -= buf->head[0].iov_len;
+		subbuf->head[0].iov_base = buf->head[0].iov_base;
 		subbuf->head[0].iov_len = 0;
 	}
 
@@ -1130,6 +1131,8 @@ xdr_buf_subsegment(struct xdr_buf *buf, struct xdr_buf *subbuf,
 		base = 0;
 	} else {
 		base -= buf->page_len;
+		subbuf->pages = buf->pages;
+		subbuf->page_base = 0;
 		subbuf->page_len = 0;
 	}
 
@@ -1141,6 +1144,7 @@ xdr_buf_subsegment(struct xdr_buf *buf, struct xdr_buf *subbuf,
 		base = 0;
 	} else {
 		base -= buf->tail[0].iov_len;
+		subbuf->tail[0].iov_base = buf->tail[0].iov_base;
 		subbuf->tail[0].iov_len = 0;
 	}
 
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index ef997880e17a..b647562a26dd 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -367,7 +367,7 @@ static void frwr_wc_fastreg(struct ib_cq *cq, struct ib_wc *wc)
 	trace_xprtrdma_wc_fastreg(wc, frwr);
 	/* The MR will get recycled when the associated req is retransmitted */
 
-	rpcrdma_flush_disconnect(cq, wc);
+	rpcrdma_flush_disconnect(cq->cq_context, wc);
 }
 
 /**
@@ -452,7 +452,7 @@ static void frwr_wc_localinv(struct ib_cq *cq, struct ib_wc *wc)
 	trace_xprtrdma_wc_li(wc, frwr);
 	__frwr_release_mr(wc, mr);
 
-	rpcrdma_flush_disconnect(cq, wc);
+	rpcrdma_flush_disconnect(cq->cq_context, wc);
 }
 
 /**
@@ -474,7 +474,7 @@ static void frwr_wc_localinv_wake(struct ib_cq *cq, struct ib_wc *wc)
 	__frwr_release_mr(wc, mr);
 	complete(&frwr->fr_linv_done);
 
-	rpcrdma_flush_disconnect(cq, wc);
+	rpcrdma_flush_disconnect(cq->cq_context, wc);
 }
 
 /**
@@ -582,7 +582,7 @@ static void frwr_wc_localinv_done(struct ib_cq *cq, struct ib_wc *wc)
 	smp_rmb();
 	rpcrdma_complete_rqst(rep);
 
-	rpcrdma_flush_disconnect(cq, wc);
+	rpcrdma_flush_disconnect(cq->cq_context, wc);
 }
 
 /**
diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 2081c8fbfa48..935bbef2f7be 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -1349,8 +1349,7 @@ rpcrdma_decode_error(struct rpcrdma_xprt *r_xprt, struct rpcrdma_rep *rep,
 			be32_to_cpup(p), be32_to_cpu(rep->rr_xid));
 	}
 
-	r_xprt->rx_stats.bad_reply_count++;
-	return -EREMOTEIO;
+	return -EIO;
 }
 
 /* Perform XID lookup, reconstruction of the RPC reply, and
@@ -1387,13 +1386,11 @@ out:
 	spin_unlock(&xprt->queue_lock);
 	return;
 
-/* If the incoming reply terminated a pending RPC, the next
- * RPC call will post a replacement receive buffer as it is
- * being marshaled.
- */
 out_badheader:
 	trace_xprtrdma_reply_hdr(rep);
 	r_xprt->rx_stats.bad_reply_count++;
+	rqst->rq_task->tk_status = status;
+	status = 0;
 	goto out;
 }
 
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 0c4af7f5e241..14165b673b20 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -242,7 +242,7 @@ xprt_rdma_connect_worker(struct work_struct *work)
 
 	rc = rpcrdma_xprt_connect(r_xprt);
 	xprt_clear_connecting(xprt);
-	if (r_xprt->rx_ep && r_xprt->rx_ep->re_connect_status > 0) {
+	if (!rc) {
 		xprt->connect_cookie++;
 		xprt->stat.connect_count++;
 		xprt->stat.connect_time += (long)jiffies -
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 2ae348377806..2198c8ec8dff 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -84,7 +84,8 @@ static void rpcrdma_rep_destroy(struct rpcrdma_rep *rep);
 static void rpcrdma_reps_unmap(struct rpcrdma_xprt *r_xprt);
 static void rpcrdma_mrs_create(struct rpcrdma_xprt *r_xprt);
 static void rpcrdma_mrs_destroy(struct rpcrdma_xprt *r_xprt);
-static int rpcrdma_ep_destroy(struct rpcrdma_ep *ep);
+static void rpcrdma_ep_get(struct rpcrdma_ep *ep);
+static int rpcrdma_ep_put(struct rpcrdma_ep *ep);
 static struct rpcrdma_regbuf *
 rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction,
 		     gfp_t flags);
@@ -97,7 +98,8 @@ static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb);
  */
 static void rpcrdma_xprt_drain(struct rpcrdma_xprt *r_xprt)
 {
-	struct rdma_cm_id *id = r_xprt->rx_ep->re_id;
+	struct rpcrdma_ep *ep = r_xprt->rx_ep;
+	struct rdma_cm_id *id = ep->re_id;
 
 	/* Flush Receives, then wait for deferred Reply work
 	 * to complete.
@@ -108,6 +110,8 @@ static void rpcrdma_xprt_drain(struct rpcrdma_xprt *r_xprt)
 	 * local invalidations.
 	 */
 	ib_drain_sq(id->qp);
+
+	rpcrdma_ep_put(ep);
 }
 
 /**
@@ -126,23 +130,27 @@ static void rpcrdma_qp_event_handler(struct ib_event *event, void *context)
 	trace_xprtrdma_qp_event(ep, event);
 }
 
+/* Ensure xprt_force_disconnect() is invoked exactly once when a
+ * connection is closed or lost. (The important thing is it needs
+ * to be invoked "at least" once).
+ */
+static void rpcrdma_force_disconnect(struct rpcrdma_ep *ep)
+{
+	if (atomic_add_unless(&ep->re_force_disconnect, 1, 1))
+		xprt_force_disconnect(ep->re_xprt);
+}
+
 /**
  * rpcrdma_flush_disconnect - Disconnect on flushed completion
- * @cq: completion queue
+ * @r_xprt: transport to disconnect
  * @wc: work completion entry
  *
  * Must be called in process context.
  */
-void rpcrdma_flush_disconnect(struct ib_cq *cq, struct ib_wc *wc)
+void rpcrdma_flush_disconnect(struct rpcrdma_xprt *r_xprt, struct ib_wc *wc)
 {
-	struct rpcrdma_xprt *r_xprt = cq->cq_context;
-	struct rpc_xprt *xprt = &r_xprt->rx_xprt;
-
-	if (wc->status != IB_WC_SUCCESS &&
-	    r_xprt->rx_ep->re_connect_status == 1) {
-		r_xprt->rx_ep->re_connect_status = -ECONNABORTED;
-		xprt_force_disconnect(xprt);
-	}
+	if (wc->status != IB_WC_SUCCESS)
+		rpcrdma_force_disconnect(r_xprt->rx_ep);
 }
 
 /**
@@ -156,11 +164,12 @@ static void rpcrdma_wc_send(struct ib_cq *cq, struct ib_wc *wc)
 	struct ib_cqe *cqe = wc->wr_cqe;
 	struct rpcrdma_sendctx *sc =
 		container_of(cqe, struct rpcrdma_sendctx, sc_cqe);
+	struct rpcrdma_xprt *r_xprt = cq->cq_context;
 
 	/* WARNING: Only wr_cqe and status are reliable at this point */
 	trace_xprtrdma_wc_send(sc, wc);
-	rpcrdma_sendctx_put_locked((struct rpcrdma_xprt *)cq->cq_context, sc);
-	rpcrdma_flush_disconnect(cq, wc);
+	rpcrdma_sendctx_put_locked(r_xprt, sc);
+	rpcrdma_flush_disconnect(r_xprt, wc);
 }
 
 /**
@@ -195,7 +204,7 @@ static void rpcrdma_wc_receive(struct ib_cq *cq, struct ib_wc *wc)
 	return;
 
 out_flushed:
-	rpcrdma_flush_disconnect(cq, wc);
+	rpcrdma_flush_disconnect(r_xprt, wc);
 	rpcrdma_rep_destroy(rep);
 }
 
@@ -239,7 +248,6 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event)
 {
 	struct sockaddr *sap = (struct sockaddr *)&id->route.addr.dst_addr;
 	struct rpcrdma_ep *ep = id->context;
-	struct rpc_xprt *xprt = ep->re_xprt;
 
 	might_sleep();
 
@@ -263,10 +271,9 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event)
 		/* fall through */
 	case RDMA_CM_EVENT_ADDR_CHANGE:
 		ep->re_connect_status = -ENODEV;
-		xprt_force_disconnect(xprt);
 		goto disconnected;
 	case RDMA_CM_EVENT_ESTABLISHED:
-		kref_get(&ep->re_kref);
+		rpcrdma_ep_get(ep);
 		ep->re_connect_status = 1;
 		rpcrdma_update_cm_private(ep, &event->param.conn);
 		trace_xprtrdma_inline_thresh(ep);
@@ -288,8 +295,8 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event)
 	case RDMA_CM_EVENT_DISCONNECTED:
 		ep->re_connect_status = -ECONNABORTED;
 disconnected:
-		xprt_force_disconnect(xprt);
-		return rpcrdma_ep_destroy(ep);
+		rpcrdma_force_disconnect(ep);
+		return rpcrdma_ep_put(ep);
 	default:
 		break;
 	}
@@ -345,7 +352,7 @@ out:
 	return ERR_PTR(rc);
 }
 
-static void rpcrdma_ep_put(struct kref *kref)
+static void rpcrdma_ep_destroy(struct kref *kref)
 {
 	struct rpcrdma_ep *ep = container_of(kref, struct rpcrdma_ep, re_kref);
 
@@ -369,13 +376,18 @@ static void rpcrdma_ep_put(struct kref *kref)
 	module_put(THIS_MODULE);
 }
 
+static noinline void rpcrdma_ep_get(struct rpcrdma_ep *ep)
+{
+	kref_get(&ep->re_kref);
+}
+
 /* Returns:
  *     %0 if @ep still has a positive kref count, or
  *     %1 if @ep was destroyed successfully.
  */
-static int rpcrdma_ep_destroy(struct rpcrdma_ep *ep)
+static noinline int rpcrdma_ep_put(struct rpcrdma_ep *ep)
 {
-	return kref_put(&ep->re_kref, rpcrdma_ep_put);
+	return kref_put(&ep->re_kref, rpcrdma_ep_destroy);
 }
 
 static int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
@@ -492,7 +504,7 @@ static int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
 	return 0;
 
 out_destroy:
-	rpcrdma_ep_destroy(ep);
+	rpcrdma_ep_put(ep);
 	rdma_destroy_id(id);
 out_free:
 	kfree(ep);
@@ -519,10 +531,13 @@ retry:
 		return rc;
 	ep = r_xprt->rx_ep;
 
-	ep->re_connect_status = 0;
 	xprt_clear_connected(xprt);
-
 	rpcrdma_reset_cwnd(r_xprt);
+
+	/* Bump the ep's reference count while there are
+	 * outstanding Receives.
+	 */
+	rpcrdma_ep_get(ep);
 	rpcrdma_post_recvs(r_xprt, true);
 
 	rc = rpcrdma_sendctxs_create(r_xprt);
@@ -552,8 +567,6 @@ retry:
 	rpcrdma_mrs_create(r_xprt);
 
 out:
-	if (rc)
-		ep->re_connect_status = rc;
 	trace_xprtrdma_connect(r_xprt, rc);
 	return rc;
 }
@@ -587,7 +600,7 @@ void rpcrdma_xprt_disconnect(struct rpcrdma_xprt *r_xprt)
 	rpcrdma_mrs_destroy(r_xprt);
 	rpcrdma_sendctxs_destroy(r_xprt);
 
-	if (rpcrdma_ep_destroy(ep))
+	if (rpcrdma_ep_put(ep))
 		rdma_destroy_id(id);
 
 	r_xprt->rx_ep = NULL;
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index 0a16fdb09b2c..43974ef39a50 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -82,6 +82,7 @@ struct rpcrdma_ep {
 	unsigned int		re_max_inline_recv;
 	int			re_async_rc;
 	int			re_connect_status;
+	atomic_t		re_force_disconnect;
 	struct ib_qp_init_attr	re_attr;
 	wait_queue_head_t       re_connect_wait;
 	struct rpc_xprt		*re_xprt;
@@ -446,7 +447,7 @@ extern unsigned int xprt_rdma_memreg_strategy;
 /*
  * Endpoint calls - xprtrdma/verbs.c
  */
-void rpcrdma_flush_disconnect(struct ib_cq *cq, struct ib_wc *wc);
+void rpcrdma_flush_disconnect(struct rpcrdma_xprt *r_xprt, struct ib_wc *wc);
 int rpcrdma_xprt_connect(struct rpcrdma_xprt *r_xprt);
 void rpcrdma_xprt_disconnect(struct rpcrdma_xprt *r_xprt);