cifs: fix potential deadlock in direct reclaim

The srv_mutex is used during writeback so cifs should ensure that
allocations done when that mutex is held are done with GFP_NOFS, to
avoid having direct reclaim ending up waiting for the same mutex and
causing a deadlock.  This is detected by lockdep with the splat below:

 ======================================================
 WARNING: possible circular locking dependency detected
 5.18.0 #70 Not tainted
 ------------------------------------------------------
 kswapd0/49 is trying to acquire lock:
 ffff8880195782e0 (&tcp_ses->srv_mutex){+.+.}-{3:3}, at: compound_send_recv

 but task is already holding lock:
 ffffffffa98e66c0 (fs_reclaim){+.+.}-{0:0}, at: balance_pgdat

 which lock already depends on the new lock.

 the existing dependency chain (in reverse order) is:

 -> #1 (fs_reclaim){+.+.}-{0:0}:
        fs_reclaim_acquire
        kmem_cache_alloc_trace
        __request_module
        crypto_alg_mod_lookup
        crypto_alloc_tfm_node
        crypto_alloc_shash
        cifs_alloc_hash
        smb311_crypto_shash_allocate
        smb311_update_preauth_hash
        compound_send_recv
        cifs_send_recv
        SMB2_negotiate
        smb2_negotiate
        cifs_negotiate_protocol
        cifs_get_smb_ses
        cifs_mount
        cifs_smb3_do_mount
        smb3_get_tree
        vfs_get_tree
        path_mount
        __x64_sys_mount
        do_syscall_64
        entry_SYSCALL_64_after_hwframe

 -> #0 (&tcp_ses->srv_mutex){+.+.}-{3:3}:
        __lock_acquire
        lock_acquire
        __mutex_lock
        mutex_lock_nested
        compound_send_recv
        cifs_send_recv
        SMB2_write
        smb2_sync_write
        cifs_write
        cifs_writepage_locked
        cifs_writepage
        shrink_page_list
        shrink_lruvec
        shrink_node
        balance_pgdat
        kswapd
        kthread
        ret_from_fork

 other info that might help us debug this:

  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock(fs_reclaim);
                                lock(&tcp_ses->srv_mutex);
                                lock(fs_reclaim);
   lock(&tcp_ses->srv_mutex);

  *** DEADLOCK ***

 1 lock held by kswapd0/49:
  #0: ffffffffa98e66c0 (fs_reclaim){+.+.}-{0:0}, at: balance_pgdat

 stack backtrace:
 CPU: 2 PID: 49 Comm: kswapd0 Not tainted 5.18.0 #70
 Call Trace:
  <TASK>
  dump_stack_lvl
  dump_stack
  print_circular_bug.cold
  check_noncircular
  __lock_acquire
  lock_acquire
  __mutex_lock
  mutex_lock_nested
  compound_send_recv
  cifs_send_recv
  SMB2_write
  smb2_sync_write
  cifs_write
  cifs_writepage_locked
  cifs_writepage
  shrink_page_list
  shrink_lruvec
  shrink_node
  balance_pgdat
  kswapd
  kthread
  ret_from_fork
  </TASK>

Fix this by using the memalloc_nofs_save/restore APIs around the places
where the srv_mutex is held.  Do this in a wrapper function for the
lock/unlock of the srv_mutex, and rename the srv_mutex to avoid missing
call sites in the conversion.

Note that there is another lockdep warning involving internal crypto
locks, which was masked by this problem and is visible after this fix,
see the discussion in this thread:

 https://lore.kernel.org/all/20220523123755.GA13668@axis.com/

Link: https://lore.kernel.org/r/CANT5p=rqcYfYMVHirqvdnnca4Mo+JQSw5Qu12v=kPfpk5yhhmg@mail.gmail.com/
Reported-by: Shyam Prasad N <nspmangalore@gmail.com>
Suggested-by: Lars Persson <larper@axis.com>
Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com>
Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de>
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

1 files changed, 20 insertions, 20 deletions

diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c
index 05eca41e3b1e..bfc9bd55870a 100644
--- a/fs/cifs/transport.c
+++ b/fs/cifs/transport.c
@@ -822,7 +822,7 @@ cifs_call_async(struct TCP_Server_Info *server, struct smb_rqst *rqst,
 	} else
 		instance = exist_credits->instance;
 
-	mutex_lock(&server->srv_mutex);
+	cifs_server_lock(server);
 
 	/*
 	 * We can't use credits obtained from the previous session to send this
@@ -830,14 +830,14 @@ cifs_call_async(struct TCP_Server_Info *server, struct smb_rqst *rqst,
 	 * return -EAGAIN in such cases to let callers handle it.
 	 */
 	if (instance != server->reconnect_instance) {
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		add_credits_and_wake_if(server, &credits, optype);
 		return -EAGAIN;
 	}
 
 	mid = server->ops->setup_async_request(server, rqst);
 	if (IS_ERR(mid)) {
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		add_credits_and_wake_if(server, &credits, optype);
 		return PTR_ERR(mid);
 	}
@@ -868,7 +868,7 @@ cifs_call_async(struct TCP_Server_Info *server, struct smb_rqst *rqst,
 		cifs_delete_mid(mid);
 	}
 
-	mutex_unlock(&server->srv_mutex);
+	cifs_server_unlock(server);
 
 	if (rc == 0)
 		return 0;
@@ -1109,7 +1109,7 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
 	 * of smb data.
 	 */
 
-	mutex_lock(&server->srv_mutex);
+	cifs_server_lock(server);
 
 	/*
 	 * All the parts of the compound chain belong obtained credits from the
@@ -1119,7 +1119,7 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
 	 * handle it.
 	 */
 	if (instance != server->reconnect_instance) {
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		for (j = 0; j < num_rqst; j++)
 			add_credits(server, &credits[j], optype);
 		return -EAGAIN;
@@ -1131,7 +1131,7 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
 			revert_current_mid(server, i);
 			for (j = 0; j < i; j++)
 				cifs_delete_mid(midQ[j]);
-			mutex_unlock(&server->srv_mutex);
+			cifs_server_unlock(server);
 
 			/* Update # of requests on wire to server */
 			for (j = 0; j < num_rqst; j++)
@@ -1163,7 +1163,7 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
 		server->sequence_number -= 2;
 	}
 
-	mutex_unlock(&server->srv_mutex);
+	cifs_server_unlock(server);
 
 	/*
 	 * If sending failed for some reason or it is an oplock break that we
@@ -1190,9 +1190,9 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
 	if ((ses->ses_status == SES_NEW) || (optype & CIFS_NEG_OP) || (optype & CIFS_SESS_OP)) {
 		spin_unlock(&cifs_tcp_ses_lock);
 
-		mutex_lock(&server->srv_mutex);
+		cifs_server_lock(server);
 		smb311_update_preauth_hash(ses, server, rqst[0].rq_iov, rqst[0].rq_nvec);
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 
 		spin_lock(&cifs_tcp_ses_lock);
 	}
@@ -1266,9 +1266,9 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses,
 			.iov_len = resp_iov[0].iov_len
 		};
 		spin_unlock(&cifs_tcp_ses_lock);
-		mutex_lock(&server->srv_mutex);
+		cifs_server_lock(server);
 		smb311_update_preauth_hash(ses, server, &iov, 1);
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		spin_lock(&cifs_tcp_ses_lock);
 	}
 	spin_unlock(&cifs_tcp_ses_lock);
@@ -1385,11 +1385,11 @@ SendReceive(const unsigned int xid, struct cifs_ses *ses,
 	   and avoid races inside tcp sendmsg code that could cause corruption
 	   of smb data */
 
-	mutex_lock(&server->srv_mutex);
+	cifs_server_lock(server);
 
 	rc = allocate_mid(ses, in_buf, &midQ);
 	if (rc) {
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		/* Update # of requests on wire to server */
 		add_credits(server, &credits, 0);
 		return rc;
@@ -1397,7 +1397,7 @@ SendReceive(const unsigned int xid, struct cifs_ses *ses,
 
 	rc = cifs_sign_smb(in_buf, server, &midQ->sequence_number);
 	if (rc) {
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		goto out;
 	}
 
@@ -1411,7 +1411,7 @@ SendReceive(const unsigned int xid, struct cifs_ses *ses,
 	if (rc < 0)
 		server->sequence_number -= 2;
 
-	mutex_unlock(&server->srv_mutex);
+	cifs_server_unlock(server);
 
 	if (rc < 0)
 		goto out;
@@ -1530,18 +1530,18 @@ SendReceiveBlockingLock(const unsigned int xid, struct cifs_tcon *tcon,
 	   and avoid races inside tcp sendmsg code that could cause corruption
 	   of smb data */
 
-	mutex_lock(&server->srv_mutex);
+	cifs_server_lock(server);
 
 	rc = allocate_mid(ses, in_buf, &midQ);
 	if (rc) {
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		return rc;
 	}
 
 	rc = cifs_sign_smb(in_buf, server, &midQ->sequence_number);
 	if (rc) {
 		cifs_delete_mid(midQ);
-		mutex_unlock(&server->srv_mutex);
+		cifs_server_unlock(server);
 		return rc;
 	}
 
@@ -1554,7 +1554,7 @@ SendReceiveBlockingLock(const unsigned int xid, struct cifs_tcon *tcon,
 	if (rc < 0)
 		server->sequence_number -= 2;
 
-	mutex_unlock(&server->srv_mutex);
+	cifs_server_unlock(server);
 
 	if (rc < 0) {
 		cifs_delete_mid(midQ);