Merge tag '6.3-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

Pull more cifs updates from Steve French:

 - xfstest generic/208 fix (memory leak)

 - minor netfs fix (to address smatch warning)

 - a DFS fix for stable

 - a reconnect race fix

 - two multichannel fixes

 - RDMA (smbdirect) fix

 - two additional writeback fixes from David

* tag '6.3-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: Fix memory leak in direct I/O
  cifs: prevent data race in cifs_reconnect_tcon()
  cifs: improve checking of DFS links over STATUS_OBJECT_NAME_INVALID
  iov: Fix netfs_extract_user_to_sg()
  cifs: Fix cifs_write_back_from_locked_folio()
  cifs: reuse cifs_match_ipaddr for comparison of dstaddr too
  cifs: match even the scope id for ipv6 addresses
  cifs: Fix an uninitialised variable
  cifs: Add some missing xas_retry() calls

10 files changed, 190 insertions, 157 deletions

diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index b7a36ebd0f2f..e2eff66eefab 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -667,11 +667,21 @@ static inline int get_dfs_path(const unsigned int xid, struct cifs_ses *ses,
 int match_target_ip(struct TCP_Server_Info *server,
 		    const char *share, size_t share_len,
 		    bool *result);
-
-int cifs_dfs_query_info_nonascii_quirk(const unsigned int xid,
-				       struct cifs_tcon *tcon,
-				       struct cifs_sb_info *cifs_sb,
-				       const char *dfs_link_path);
+int cifs_inval_name_dfs_link_error(const unsigned int xid,
+				   struct cifs_tcon *tcon,
+				   struct cifs_sb_info *cifs_sb,
+				   const char *full_path,
+				   bool *islink);
+#else
+static inline int cifs_inval_name_dfs_link_error(const unsigned int xid,
+				   struct cifs_tcon *tcon,
+				   struct cifs_sb_info *cifs_sb,
+				   const char *full_path,
+				   bool *islink)
+{
+	*islink = false;
+	return 0;
+}
 #endif
 
 static inline int cifs_create_options(struct cifs_sb_info *cifs_sb, int options)
@@ -684,5 +694,6 @@ static inline int cifs_create_options(struct cifs_sb_info *cifs_sb, int options)
 
 struct super_block *cifs_get_tcon_super(struct cifs_tcon *tcon);
 void cifs_put_tcon_super(struct super_block *sb);
+int cifs_wait_for_server_reconnect(struct TCP_Server_Info *server, bool retry);
 
 #endif			/* _CIFSPROTO_H */
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index a24e4ddf8043..a43c78396dd8 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -72,7 +72,6 @@ cifs_reconnect_tcon(struct cifs_tcon *tcon, int smb_command)
 	struct cifs_ses *ses;
 	struct TCP_Server_Info *server;
 	struct nls_table *nls_codepage;
-	int retries;
 
 	/*
 	 * SMBs NegProt, SessSetup, uLogoff do not have tcon yet so check for
@@ -102,45 +101,9 @@ cifs_reconnect_tcon(struct cifs_tcon *tcon, int smb_command)
 	}
 	spin_unlock(&tcon->tc_lock);
 
-	retries = server->nr_targets;
-
-	/*
-	 * Give demultiplex thread up to 10 seconds to each target available for
-	 * reconnect -- should be greater than cifs socket timeout which is 7
-	 * seconds.
-	 */
-	while (server->tcpStatus == CifsNeedReconnect) {
-		rc = wait_event_interruptible_timeout(server->response_q,
-						      (server->tcpStatus != CifsNeedReconnect),
-						      10 * HZ);
-		if (rc < 0) {
-			cifs_dbg(FYI, "%s: aborting reconnect due to a received signal by the process\n",
-				 __func__);
-			return -ERESTARTSYS;
-		}
-
-		/* are we still trying to reconnect? */
-		spin_lock(&server->srv_lock);
-		if (server->tcpStatus != CifsNeedReconnect) {
-			spin_unlock(&server->srv_lock);
-			break;
-		}
-		spin_unlock(&server->srv_lock);
-
-		if (retries && --retries)
-			continue;
-
-		/*
-		 * on "soft" mounts we wait once. Hard mounts keep
-		 * retrying until process is killed or server comes
-		 * back on-line
-		 */
-		if (!tcon->retry) {
-			cifs_dbg(FYI, "gave up waiting on reconnect in smb_init\n");
-			return -EHOSTDOWN;
-		}
-		retries = server->nr_targets;
-	}
+	rc = cifs_wait_for_server_reconnect(server, tcon->retry);
+	if (rc)
+		return rc;
 
 	spin_lock(&ses->chan_lock);
 	if (!cifs_chan_needs_reconnect(ses, server) && !tcon->need_reconnect) {
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index ec020d860be3..5233f14f0636 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -1294,7 +1294,8 @@ cifs_match_ipaddr(struct sockaddr *srcaddr, struct sockaddr *rhs)
 	case AF_INET6: {
 		struct sockaddr_in6 *saddr6 = (struct sockaddr_in6 *)srcaddr;
 		struct sockaddr_in6 *vaddr6 = (struct sockaddr_in6 *)rhs;
-		return ipv6_addr_equal(&saddr6->sin6_addr, &vaddr6->sin6_addr);
+		return (ipv6_addr_equal(&saddr6->sin6_addr, &vaddr6->sin6_addr)
+			&& saddr6->sin6_scope_id == vaddr6->sin6_scope_id);
 	}
 	default:
 		WARN_ON(1);
@@ -1343,32 +1344,8 @@ match_port(struct TCP_Server_Info *server, struct sockaddr *addr)
 
 static bool match_server_address(struct TCP_Server_Info *server, struct sockaddr *addr)
 {
-	switch (addr->sa_family) {
-	case AF_INET: {
-		struct sockaddr_in *addr4 = (struct sockaddr_in *)addr;
-		struct sockaddr_in *srv_addr4 =
-					(struct sockaddr_in *)&server->dstaddr;
-
-		if (addr4->sin_addr.s_addr != srv_addr4->sin_addr.s_addr)
-			return false;
-		break;
-	}
-	case AF_INET6: {
-		struct sockaddr_in6 *addr6 = (struct sockaddr_in6 *)addr;
-		struct sockaddr_in6 *srv_addr6 =
-					(struct sockaddr_in6 *)&server->dstaddr;
-
-		if (!ipv6_addr_equal(&addr6->sin6_addr,
-				     &srv_addr6->sin6_addr))
-			return false;
-		if (addr6->sin6_scope_id != srv_addr6->sin6_scope_id)
-			return false;
-		break;
-	}
-	default:
-		WARN_ON(1);
-		return false; /* don't expect to be here */
-	}
+	if (!cifs_match_ipaddr(addr, (struct sockaddr *)&server->dstaddr))
+		return false;
 
 	return true;
 }
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index ebfcaae8c437..4d4a2d82636d 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -52,6 +52,8 @@ static void cifs_undirty_folios(struct inode *inode, loff_t start, unsigned int
 
 	end = (start + len - 1) / PAGE_SIZE;
 	xas_for_each_marked(&xas, folio, end, PAGECACHE_TAG_DIRTY) {
+		if (xas_retry(&xas, folio))
+			continue;
 		xas_pause(&xas);
 		rcu_read_unlock();
 		folio_lock(folio);
@@ -81,6 +83,8 @@ void cifs_pages_written_back(struct inode *inode, loff_t start, unsigned int len
 
 	end = (start + len - 1) / PAGE_SIZE;
 	xas_for_each(&xas, folio, end) {
+		if (xas_retry(&xas, folio))
+			continue;
 		if (!folio_test_writeback(folio)) {
 			WARN_ONCE(1, "bad %x @%llx page %lx %lx\n",
 				  len, start, folio_index(folio), end);
@@ -112,6 +116,8 @@ void cifs_pages_write_failed(struct inode *inode, loff_t start, unsigned int len
 
 	end = (start + len - 1) / PAGE_SIZE;
 	xas_for_each(&xas, folio, end) {
+		if (xas_retry(&xas, folio))
+			continue;
 		if (!folio_test_writeback(folio)) {
 			WARN_ONCE(1, "bad %x @%llx page %lx %lx\n",
 				  len, start, folio_index(folio), end);
@@ -2839,6 +2845,7 @@ err_xid:
 	free_xid(xid);
 	if (rc == 0) {
 		wbc->nr_to_write = count;
+		rc = len;
 	} else if (is_retryable_error(rc)) {
 		cifs_pages_write_redirty(inode, start, len);
 	} else {
@@ -3605,7 +3612,7 @@ static ssize_t __cifs_writev(
 
 		ctx->nr_pinned_pages = rc;
 		ctx->bv = (void *)ctx->iter.bvec;
-		ctx->bv_need_unpin = iov_iter_extract_will_pin(&ctx->iter);
+		ctx->bv_need_unpin = iov_iter_extract_will_pin(from);
 	} else if ((iov_iter_is_bvec(from) || iov_iter_is_kvec(from)) &&
 		   !is_sync_kiocb(iocb)) {
 		/*
@@ -4141,7 +4148,7 @@ static ssize_t __cifs_readv(
 
 		ctx->nr_pinned_pages = rc;
 		ctx->bv = (void *)ctx->iter.bvec;
-		ctx->bv_need_unpin = iov_iter_extract_will_pin(&ctx->iter);
+		ctx->bv_need_unpin = iov_iter_extract_will_pin(to);
 		ctx->should_dirty = true;
 	} else if ((iov_iter_is_bvec(to) || iov_iter_is_kvec(to)) &&
 		   !is_sync_kiocb(iocb)) {
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index 2905734eb289..a0d286ee723d 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -21,6 +21,7 @@
 #include "cifsfs.h"
 #ifdef CONFIG_CIFS_DFS_UPCALL
 #include "dns_resolve.h"
+#include "dfs_cache.h"
 #endif
 #include "fs_context.h"
 #include "cached_dir.h"
@@ -1198,4 +1199,114 @@ int cifs_update_super_prepath(struct cifs_sb_info *cifs_sb, char *prefix)
 	cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_USE_PREFIX_PATH;
 	return 0;
 }
+
+/*
+ * Handle weird Windows SMB server behaviour. It responds with
+ * STATUS_OBJECT_NAME_INVALID code to SMB2 QUERY_INFO request for
+ * "\<server>\<dfsname>\<linkpath>" DFS reference, where <dfsname> contains
+ * non-ASCII unicode symbols.
+ */
+int cifs_inval_name_dfs_link_error(const unsigned int xid,
+				   struct cifs_tcon *tcon,
+				   struct cifs_sb_info *cifs_sb,
+				   const char *full_path,
+				   bool *islink)
+{
+	struct cifs_ses *ses = tcon->ses;
+	size_t len;
+	char *path;
+	char *ref_path;
+
+	*islink = false;
+
+	/*
+	 * Fast path - skip check when @full_path doesn't have a prefix path to
+	 * look up or tcon is not DFS.
+	 */
+	if (strlen(full_path) < 2 || !cifs_sb ||
+	    (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS) ||
+	    !is_tcon_dfs(tcon) || !ses->server->origin_fullpath)
+		return 0;
+
+	/*
+	 * Slow path - tcon is DFS and @full_path has prefix path, so attempt
+	 * to get a referral to figure out whether it is an DFS link.
+	 */
+	len = strnlen(tcon->tree_name, MAX_TREE_SIZE + 1) + strlen(full_path) + 1;
+	path = kmalloc(len, GFP_KERNEL);
+	if (!path)
+		return -ENOMEM;
+
+	scnprintf(path, len, "%s%s", tcon->tree_name, full_path);
+	ref_path = dfs_cache_canonical_path(path + 1, cifs_sb->local_nls,
+					    cifs_remap(cifs_sb));
+	kfree(path);
+
+	if (IS_ERR(ref_path)) {
+		if (PTR_ERR(ref_path) != -EINVAL)
+			return PTR_ERR(ref_path);
+	} else {
+		struct dfs_info3_param *refs = NULL;
+		int num_refs = 0;
+
+		/*
+		 * XXX: we are not using dfs_cache_find() here because we might
+		 * end filling all the DFS cache and thus potentially
+		 * removing cached DFS targets that the client would eventually
+		 * need during failover.
+		 */
+		if (ses->server->ops->get_dfs_refer &&
+		    !ses->server->ops->get_dfs_refer(xid, ses, ref_path, &refs,
+						     &num_refs, cifs_sb->local_nls,
+						     cifs_remap(cifs_sb)))
+			*islink = refs[0].server_type == DFS_TYPE_LINK;
+		free_dfs_info_array(refs, num_refs);
+		kfree(ref_path);
+	}
+	return 0;
+}
 #endif
+
+int cifs_wait_for_server_reconnect(struct TCP_Server_Info *server, bool retry)
+{
+	int timeout = 10;
+	int rc;
+
+	spin_lock(&server->srv_lock);
+	if (server->tcpStatus != CifsNeedReconnect) {
+		spin_unlock(&server->srv_lock);
+		return 0;
+	}
+	timeout *= server->nr_targets;
+	spin_unlock(&server->srv_lock);
+
+	/*
+	 * Give demultiplex thread up to 10 seconds to each target available for
+	 * reconnect -- should be greater than cifs socket timeout which is 7
+	 * seconds.
+	 *
+	 * On "soft" mounts we wait once. Hard mounts keep retrying until
+	 * process is killed or server comes back on-line.
+	 */
+	do {
+		rc = wait_event_interruptible_timeout(server->response_q,
+						      (server->tcpStatus != CifsNeedReconnect),
+						      timeout * HZ);
+		if (rc < 0) {
+			cifs_dbg(FYI, "%s: aborting reconnect due to received signal\n",
+				 __func__);
+			return -ERESTARTSYS;
+		}
+
+		/* are we still trying to reconnect? */
+		spin_lock(&server->srv_lock);
+		if (server->tcpStatus != CifsNeedReconnect) {
+			spin_unlock(&server->srv_lock);
+			return 0;
+		}
+		spin_unlock(&server->srv_lock);
+	} while (retry);
+
+	cifs_dbg(FYI, "%s: gave up waiting on reconnect\n", __func__);
+	return -EHOSTDOWN;
+}
diff --git a/fs/cifs/smb2inode.c b/fs/cifs/smb2inode.c
index 37b4cd59245d..9b956294e864 100644
--- a/fs/cifs/smb2inode.c
+++ b/fs/cifs/smb2inode.c
@@ -527,12 +527,13 @@ int smb2_query_path_info(const unsigned int xid, struct cifs_tcon *tcon,
 			 struct cifs_sb_info *cifs_sb, const char *full_path,
 			 struct cifs_open_info_data *data, bool *adjust_tz, bool *reparse)
 {
-	int rc;
 	__u32 create_options = 0;
 	struct cifsFileInfo *cfile;
 	struct cached_fid *cfid = NULL;
 	struct kvec err_iov[3] = {};
 	int err_buftype[3] = {};
+	bool islink;
+	int rc, rc2;
 
 	*adjust_tz = false;
 	*reparse = false;
@@ -580,15 +581,15 @@ int smb2_query_path_info(const unsigned int xid, struct cifs_tcon *tcon,
 					      SMB2_OP_QUERY_INFO, cfile, NULL, NULL,
 					      NULL, NULL);
 			goto out;
-		} else if (rc != -EREMOTE && IS_ENABLED(CONFIG_CIFS_DFS_UPCALL) &&
-			   hdr->Status == STATUS_OBJECT_NAME_INVALID) {
-			/*
-			 * Handle weird Windows SMB server behaviour. It responds with
-			 * STATUS_OBJECT_NAME_INVALID code to SMB2 QUERY_INFO request
-			 * for "\<server>\<dfsname>\<linkpath>" DFS reference,
-			 * where <dfsname> contains non-ASCII unicode symbols.
-			 */
-			rc = -EREMOTE;
+		} else if (rc != -EREMOTE && hdr->Status == STATUS_OBJECT_NAME_INVALID) {
+			rc2 = cifs_inval_name_dfs_link_error(xid, tcon, cifs_sb,
+							     full_path, &islink);
+			if (rc2) {
+				rc = rc2;
+				goto out;
+			}
+			if (islink)
+				rc = -EREMOTE;
 		}
 		if (rc == -EREMOTE && IS_ENABLED(CONFIG_CIFS_DFS_UPCALL) && cifs_sb &&
 		    (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS))
diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
index f79b075f2992..6dfb865ee9d7 100644
--- a/fs/cifs/smb2ops.c
+++ b/fs/cifs/smb2ops.c
@@ -796,7 +796,6 @@ static int
 smb2_is_path_accessible(const unsigned int xid, struct cifs_tcon *tcon,
 			struct cifs_sb_info *cifs_sb, const char *full_path)
 {
-	int rc;
 	__le16 *utf16_path;
 	__u8 oplock = SMB2_OPLOCK_LEVEL_NONE;
 	int err_buftype = CIFS_NO_BUFFER;
@@ -804,6 +803,8 @@ smb2_is_path_accessible(const unsigned int xid, struct cifs_tcon *tcon,
 	struct kvec err_iov = {};
 	struct cifs_fid fid;
 	struct cached_fid *cfid;
+	bool islink;
+	int rc, rc2;
 
 	rc = open_cached_dir(xid, tcon, full_path, cifs_sb, true, &cfid);
 	if (!rc) {
@@ -833,15 +834,17 @@ smb2_is_path_accessible(const unsigned int xid, struct cifs_tcon *tcon,
 
 		if (unlikely(!hdr || err_buftype == CIFS_NO_BUFFER))
 			goto out;
-		/*
-		 * Handle weird Windows SMB server behaviour. It responds with
-		 * STATUS_OBJECT_NAME_INVALID code to SMB2 QUERY_INFO request
-		 * for "\<server>\<dfsname>\<linkpath>" DFS reference,
-		 * where <dfsname> contains non-ASCII unicode symbols.
-		 */
-		if (rc != -EREMOTE && IS_ENABLED(CONFIG_CIFS_DFS_UPCALL) &&
-		    hdr->Status == STATUS_OBJECT_NAME_INVALID)
-			rc = -EREMOTE;
+
+		if (rc != -EREMOTE && hdr->Status == STATUS_OBJECT_NAME_INVALID) {
+			rc2 = cifs_inval_name_dfs_link_error(xid, tcon, cifs_sb,
+							     full_path, &islink);
+			if (rc2) {
+				rc = rc2;
+				goto out;
+			}
+			if (islink)
+				rc = -EREMOTE;
+		}
 		if (rc == -EREMOTE && IS_ENABLED(CONFIG_CIFS_DFS_UPCALL) && cifs_sb &&
 		    (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_DFS))
 			rc = -EOPNOTSUPP;
diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c
index ca9d7110ddcb..0e53265e1462 100644
--- a/fs/cifs/smb2pdu.c
+++ b/fs/cifs/smb2pdu.c
@@ -139,66 +139,6 @@ out:
 	return;
 }
 
-static int wait_for_server_reconnect(struct TCP_Server_Info *server,
-				     __le16 smb2_command, bool retry)
-{
-	int timeout = 10;
-	int rc;
-
-	spin_lock(&server->srv_lock);
-	if (server->tcpStatus != CifsNeedReconnect) {
-		spin_unlock(&server->srv_lock);
-		return 0;
-	}
-	timeout *= server->nr_targets;
-	spin_unlock(&server->srv_lock);
-
-	/*
-	 * Return to caller for TREE_DISCONNECT and LOGOFF and CLOSE
-	 * here since they are implicitly done when session drops.
-	 */
-	switch (smb2_command) {
-	/*
-	 * BB Should we keep oplock break and add flush to exceptions?
-	 */
-	case SMB2_TREE_DISCONNECT:
-	case SMB2_CANCEL:
-	case SMB2_CLOSE:
-	case SMB2_OPLOCK_BREAK:
-		return -EAGAIN;
-	}
-
-	/*
-	 * Give demultiplex thread up to 10 seconds to each target available for
-	 * reconnect -- should be greater than cifs socket timeout which is 7
-	 * seconds.
-	 *
-	 * On "soft" mounts we wait once. Hard mounts keep retrying until
-	 * process is killed or server comes back on-line.
-	 */
-	do {
-		rc = wait_event_interruptible_timeout(server->response_q,
-						      (server->tcpStatus != CifsNeedReconnect),
-						      timeout * HZ);
-		if (rc < 0) {
-			cifs_dbg(FYI, "%s: aborting reconnect due to received signal\n",
-				 __func__);
-			return -ERESTARTSYS;
-		}
-
-		/* are we still trying to reconnect? */
-		spin_lock(&server->srv_lock);
-		if (server->tcpStatus != CifsNeedReconnect) {
-			spin_unlock(&server->srv_lock);
-			return 0;
-		}
-		spin_unlock(&server->srv_lock);
-	} while (retry);
-
-	cifs_dbg(FYI, "%s: gave up waiting on reconnect\n", __func__);
-	return -EHOSTDOWN;
-}
-
 static int
 smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	       struct TCP_Server_Info *server)
@@ -243,7 +183,27 @@ smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	    (!tcon->ses->server) || !server)
 		return -EIO;
 
-	rc = wait_for_server_reconnect(server, smb2_command, tcon->retry);
+	spin_lock(&server->srv_lock);
+	if (server->tcpStatus == CifsNeedReconnect) {
+		/*
+		 * Return to caller for TREE_DISCONNECT and LOGOFF and CLOSE
+		 * here since they are implicitly done when session drops.
+		 */
+		switch (smb2_command) {
+		/*
+		 * BB Should we keep oplock break and add flush to exceptions?
+		 */
+		case SMB2_TREE_DISCONNECT:
+		case SMB2_CANCEL:
+		case SMB2_CLOSE:
+		case SMB2_OPLOCK_BREAK:
+			spin_unlock(&server->srv_lock);
+			return -EAGAIN;
+		}
+	}
+	spin_unlock(&server->srv_lock);
+
+	rc = cifs_wait_for_server_reconnect(server, tcon->retry);
 	if (rc)
 		return rc;
 
diff --git a/fs/cifs/smbdirect.c b/fs/cifs/smbdirect.c
index 55b6e319a61d..0362ebd4fa0f 100644
--- a/fs/cifs/smbdirect.c
+++ b/fs/cifs/smbdirect.c
@@ -837,7 +837,7 @@ static int smbd_post_send_iter(struct smbd_connection *info,
 	int data_length;
 	struct smbd_request *request;
 	struct smbd_data_transfer *packet;
-	int new_credits;
+	int new_credits = 0;
 
 wait_credit:
 	/* Wait for send credits. A SMBD packet needs one credit */
diff --git a/fs/netfs/iterator.c b/fs/netfs/iterator.c
index f00d43b8ac0a..e9a45dea748a 100644
--- a/fs/netfs/iterator.c
+++ b/fs/netfs/iterator.c
@@ -134,7 +134,7 @@ static ssize_t netfs_extract_user_to_sg(struct iov_iter *iter,
 		npages = DIV_ROUND_UP(off + len, PAGE_SIZE);
 		sg_max -= npages;
 
-		for (; npages < 0; npages--) {
+		for (; npages > 0; npages--) {
 			struct page *page = *pages;
 			size_t seg = min_t(size_t, PAGE_SIZE - off, len);