nfs: skip commit in releasepage if we're freeing memory for fs-related reasons

commit 5cf02d0 upstream.

We've had some reports of a deadlock where rpciod ends up with a stack
trace like this:

    PID: 2507   TASK: ffff88103691ab40  CPU: 14  COMMAND: "rpciod/14"
     #0 [ffff8810343bf2f0] schedule at ffffffff814dabd9
     #1 [ffff8810343bf3b8] nfs_wait_bit_killable at ffffffffa038fc04 [nfs]
     #2 [ffff8810343bf3c8] __wait_on_bit at ffffffff814dbc2f
     #3 [ffff8810343bf418] out_of_line_wait_on_bit at ffffffff814dbcd8
     #4 [ffff8810343bf488] nfs_commit_inode at ffffffffa039e0c1 [nfs]
     #5 [ffff8810343bf4f8] nfs_release_page at ffffffffa038bef6 [nfs]
     #6 [ffff8810343bf528] try_to_release_page at ffffffff8110c670
     #7 [ffff8810343bf538] shrink_page_list.clone.0 at ffffffff81126271
     #8 [ffff8810343bf668] shrink_inactive_list at ffffffff81126638
     #9 [ffff8810343bf818] shrink_zone at ffffffff8112788f
    torvalds#10 [ffff8810343bf8c8] do_try_to_free_pages at ffffffff81127b1e
    torvalds#11 [ffff8810343bf958] try_to_free_pages at ffffffff8112812f
    torvalds#12 [ffff8810343bfa08] __alloc_pages_nodemask at ffffffff8111fdad
    torvalds#13 [ffff8810343bfb28] kmem_getpages at ffffffff81159942
    torvalds#14 [ffff8810343bfb58] fallback_alloc at ffffffff8115a55a
    torvalds#15 [ffff8810343bfbd8] ____cache_alloc_node at ffffffff8115a2d9
    torvalds#16 [ffff8810343bfc38] kmem_cache_alloc at ffffffff8115b09b
    torvalds#17 [ffff8810343bfc78] sk_prot_alloc at ffffffff81411808
    torvalds#18 [ffff8810343bfcb8] sk_alloc at ffffffff8141197c
    torvalds#19 [ffff8810343bfce8] inet_create at ffffffff81483ba6
    torvalds#20 [ffff8810343bfd38] __sock_create at ffffffff8140b4a7
    torvalds#21 [ffff8810343bfd98] xs_create_sock at ffffffffa01f649b [sunrpc]
    torvalds#22 [ffff8810343bfdd8] xs_tcp_setup_socket at ffffffffa01f6965 [sunrpc]
    torvalds#23 [ffff8810343bfe38] worker_thread at ffffffff810887d0
    torvalds#24 [ffff8810343bfee8] kthread at ffffffff8108dd96
    torvalds#25 [ffff8810343bff48] kernel_thread at ffffffff8100c1ca

rpciod is trying to allocate memory for a new socket to talk to the
server. The VM ends up calling ->releasepage to get more memory, and it
tries to do a blocking commit. That commit can't succeed however without
a connected socket, so we deadlock.

Fix this by setting PF_FSTRANS on the workqueue task prior to doing the
socket allocation, and having nfs_release_page check for that flag when
deciding whether to do a commit call. Also, set PF_FSTRANS
unconditionally in rpc_async_schedule since that function can also do
allocations sometimes.

Signed-off-by: Jeff Layton <[email protected]>
Signed-off-by: Trond Myklebust <[email protected]>
Signed-off-by: Ben Hutchings <[email protected]>

Author: jtlayton

fs/nfs/file.c

@@ -452,8 +452,11 @@ static int nfs_release_page(struct page *page, gfp_t gfp)
 
 	dfprintk(PAGECACHE, "NFS: release_page(%p)\n", page);
 
-	/* Only do I/O if gfp is a superset of GFP_KERNEL */
-	if (mapping && (gfp & GFP_KERNEL) == GFP_KERNEL) {
+	/* Only do I/O if gfp is a superset of GFP_KERNEL, and we're not
+	 * doing this memory reclaim for a fs-related allocation.
+	 */
+	if (mapping && (gfp & GFP_KERNEL) == GFP_KERNEL &&
+	    !(current->flags & PF_FSTRANS)) {
 		int how = FLUSH_SYNC;
 
 		/* Don't let kswapd deadlock waiting for OOM RPC calls */

net/sunrpc/sched.c

@@ -755,7 +755,9 @@ void rpc_execute(struct rpc_task *task)
 
 static void rpc_async_schedule(struct work_struct *work)
 {
+	current->flags |= PF_FSTRANS;
 	__rpc_execute(container_of(work, struct rpc_task, u.tk_work));
+	current->flags &= ~PF_FSTRANS;
 }
 
 /**

net/sunrpc/xprtrdma/transport.c

@@ -200,6 +200,7 @@ xprt_rdma_connect_worker(struct work_struct *work)
 	int rc = 0;
 
 	if (!xprt->shutdown) {
+		current->flags |= PF_FSTRANS;
 		xprt_clear_connected(xprt);
 
 		dprintk("RPC:       %s: %sconnect\n", __func__,
@@ -212,10 +213,10 @@ xprt_rdma_connect_worker(struct work_struct *work)
 
 out:
 	xprt_wake_pending_tasks(xprt, rc);
-
 out_clear:
 	dprintk("RPC:       %s: exit\n", __func__);
 	xprt_clear_connecting(xprt);
+	current->flags &= ~PF_FSTRANS;
 }
 
 /*

net/sunrpc/xprtsock.c

@@ -1895,6 +1895,8 @@ static void xs_local_setup_socket(struct work_struct *work)
 	if (xprt->shutdown)
 		goto out;
 
+	current->flags |= PF_FSTRANS;
+
 	clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
 	status = __sock_create(xprt->xprt_net, AF_LOCAL,
 					SOCK_STREAM, 0, &sock, 1);
@@ -1928,6 +1930,7 @@ static void xs_local_setup_socket(struct work_struct *work)
 out:
 	xprt_clear_connecting(xprt);
 	xprt_wake_pending_tasks(xprt, status);
+	current->flags &= ~PF_FSTRANS;
 }
 
 static void xs_udp_finish_connecting(struct rpc_xprt *xprt, struct socket *sock)
@@ -1970,6 +1973,8 @@ static void xs_udp_setup_socket(struct work_struct *work)
 	if (xprt->shutdown)
 		goto out;
 
+	current->flags |= PF_FSTRANS;
+
 	/* Start by resetting any existing state */
 	xs_reset_transport(transport);
 	sock = xs_create_sock(xprt, transport,
@@ -1988,6 +1993,7 @@ static void xs_udp_setup_socket(struct work_struct *work)
 out:
 	xprt_clear_connecting(xprt);
 	xprt_wake_pending_tasks(xprt, status);
+	current->flags &= ~PF_FSTRANS;
 }
 
 /*
@@ -2113,6 +2119,8 @@ static void xs_tcp_setup_socket(struct work_struct *work)
 	if (xprt->shutdown)
 		goto out;
 
+	current->flags |= PF_FSTRANS;
+
 	if (!sock) {
 		clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
 		sock = xs_create_sock(xprt, transport,
@@ -2162,6 +2170,7 @@ static void xs_tcp_setup_socket(struct work_struct *work)
 	case -EINPROGRESS:
 	case -EALREADY:
 		xprt_clear_connecting(xprt);
+		current->flags &= ~PF_FSTRANS;
 		return;
 	case -EINVAL:
 		/* Happens, for instance, if the user specified a link
@@ -2174,6 +2183,7 @@ static void xs_tcp_setup_socket(struct work_struct *work)
 out:
 	xprt_clear_connecting(xprt);
 	xprt_wake_pending_tasks(xprt, status);
+	current->flags &= ~PF_FSTRANS;
 }
 
 /**