Merge tag '4.18-rc1-more-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "Misc SMB3 fixes, including particularly important ones for signing,
  some minor documentation and debug improvements and another posix
  smb3.11 fix"

* tag '4.18-rc1-more-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: Fix invalid check in __cifs_calc_signature()
  cifs: Use correct packet length in SMB2_TRANSFORM header
  smb3: fix corrupt path in subdirs on smb311 with posix
  smb3: do not display empty interface list
  smb3: Fix mode on mkdir on smb311 mounts
  cifs: Fix kernel oops when traceSMB is enabled
  CIFS: dump every session iface info
  CIFS: parse and store info on iface queries
  CIFS: add iface info to struct cifs_ses
  CIFS: complete PDU definitions for interface queries
  CIFS: move default port definitions to cifsglob.h
  cifs: Fix encryption/signing
  cifs: update __smb_send_rqst() to take an array of requests
  cifs: remove smb2_send_recv()
  cifs: push rfc1002 generation down the stack
  smb3: increase initial number of credits requested to allow write
  cifs: minor documentation updates
  cifs: add lease tracking to the cached root fid
  smb3: note that smb3.11 posix extensions mount option is experimental

20 files changed, 753 insertions, 279 deletions

diff --git a/Documentation/filesystems/cifs/AUTHORS b/Documentation/filesystems/cifs/AUTHORS
index 9f4f87e16240..75865da2ce14 100644
--- a/Documentation/filesystems/cifs/AUTHORS
+++ b/Documentation/filesystems/cifs/AUTHORS
@@ -42,9 +42,11 @@ Jeff Layton (many, many fixes, as well as great work on the cifs Kerberos code)
 Scott Lovenberg
 Pavel Shilovsky (for great work adding SMB2 support, and various SMB3 features)
 Aurelien Aptel (for DFS SMB3 work and some key bug fixes)
-Ronnie Sahlberg (for SMB3 xattr work and bug fixes)
+Ronnie Sahlberg (for SMB3 xattr work, bug fixes, and lots of great work on compounding)
 Shirish Pargaonkar (for many ACL patches over the years)
 Sachin Prabhu (many bug fixes, including for reconnect, copy offload and security)
+Paulo Alcantara
+Long Li (some great work on RDMA, SMB Direct)
 
 
 Test case and Bug Report contributors
@@ -58,5 +60,4 @@ mention to the Stanford Checker (SWAT) which pointed out many minor
 bugs in error paths.  Valuable suggestions also have come from Al Viro
 and Dave Miller.
 
-And thanks to the IBM LTC and Power test teams and SuSE testers for
-finding multiple bugs during excellent stress test runs.
+And thanks to the IBM LTC and Power test teams and SuSE and Citrix and RedHat testers for finding multiple bugs during excellent stress test runs.
diff --git a/Documentation/filesystems/cifs/CHANGES b/Documentation/filesystems/cifs/CHANGES
index bc0025cdd1c9..455e1cc494a9 100644
--- a/Documentation/filesystems/cifs/CHANGES
+++ b/Documentation/filesystems/cifs/CHANGES
@@ -1,3 +1,6 @@
+See https://wiki.samba.org/index.php/LinuxCIFSKernel for
+more current information.
+
 Version 1.62
 ------------
 Add sockopt=TCP_NODELAY mount option. EA (xattr) routines hardened
diff --git a/Documentation/filesystems/cifs/TODO b/Documentation/filesystems/cifs/TODO
index c5adf149b57f..852499aed64b 100644
--- a/Documentation/filesystems/cifs/TODO
+++ b/Documentation/filesystems/cifs/TODO
@@ -9,14 +9,14 @@ is a partial list of the known problems and missing features:
 
 a) SMB3 (and SMB3.02) missing optional features:
    - multichannel (started), integration with RDMA
-   - directory leases (improved metadata caching)
-   - T10 copy offload (copy chunk, and "Duplicate Extents" ioctl
+   - directory leases (improved metadata caching), started (root dir only)
+   - T10 copy offload ie "ODX" (copy chunk, and "Duplicate Extents" ioctl
      currently the only two server side copy mechanisms supported)
 
 b) improved sparse file support
 
 c) Directory entry caching relies on a 1 second timer, rather than
-using Directory Leases
+using Directory Leases, currently only the root file handle is cached longer
 
 d) quota support (needs minor kernel change since quota calls
 to make it to network filesystems or deviceless filesystems)
@@ -42,6 +42,8 @@ mount or a per server basis to client UIDs or nobody if no mapping
 exists. Also better integration with winbind for resolving SID owners
 
 k) Add tools to take advantage of more smb3 specific ioctls and features
+(passthrough ioctl/fsctl for sending various SMB3 fsctls to the server
+is in progress)
 
 l) encrypted file support
 
@@ -71,9 +73,8 @@ t) split cifs and smb3 support into separate modules so legacy (and less
 secure) CIFS dialect can be disabled in environments that don't need it
 and simplify the code.
 
-u) Finish up SMB3.1.1 dialect support
-
-v) POSIX Extensions for SMB3.1.1
+v) POSIX Extensions for SMB3.1.1 (started, create and mkdir support added
+so far).
 
 KNOWN BUGS
 ====================================
@@ -92,8 +93,8 @@ Misc testing to do
 1) check out max path names and max path name components against various server
 types. Try nested symlinks (8 deep). Return max path name in stat -f information
 
-2) Improve xfstest's cifs enablement and adapt xfstests where needed to test
-cifs better
+2) Improve xfstest's cifs/smb3 enablement and adapt xfstests where needed to test
+cifs/smb3 better
 
 3) Additional performance testing and optimization using iozone and similar - 
 there are some easy changes that can be done to parallelize sequential writes,
diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c
index 116146022aa1..bfe999505815 100644
--- a/fs/cifs/cifs_debug.c
+++ b/fs/cifs/cifs_debug.c
@@ -126,6 +126,25 @@ static void cifs_debug_tcon(struct seq_file *m, struct cifs_tcon *tcon)
 	seq_putc(m, '\n');
 }
 
+static void
+cifs_dump_iface(struct seq_file *m, struct cifs_server_iface *iface)
+{
+	struct sockaddr_in *ipv4 = (struct sockaddr_in *)&iface->sockaddr;
+	struct sockaddr_in6 *ipv6 = (struct sockaddr_in6 *)&iface->sockaddr;
+
+	seq_printf(m, "\t\tSpeed: %zu bps\n", iface->speed);
+	seq_puts(m, "\t\tCapabilities: ");
+	if (iface->rdma_capable)
+		seq_puts(m, "rdma ");
+	if (iface->rss_capable)
+		seq_puts(m, "rss ");
+	seq_putc(m, '\n');
+	if (iface->sockaddr.ss_family == AF_INET)
+		seq_printf(m, "\t\tIPv4: %pI4\n", &ipv4->sin_addr);
+	else if (iface->sockaddr.ss_family == AF_INET6)
+		seq_printf(m, "\t\tIPv6: %pI6\n", &ipv6->sin6_addr);
+}
+
 static int cifs_debug_data_proc_show(struct seq_file *m, void *v)
 {
 	struct list_head *tmp1, *tmp2, *tmp3;
@@ -312,6 +331,16 @@ skip_rdma:
 					      mid_entry->mid);
 			}
 			spin_unlock(&GlobalMid_Lock);
+
+			spin_lock(&ses->iface_lock);
+			if (ses->iface_count)
+				seq_printf(m, "\n\tServer interfaces: %zu\n",
+					   ses->iface_count);
+			for (j = 0; j < ses->iface_count; j++) {
+				seq_printf(m, "\t%d)\n", j);
+				cifs_dump_iface(m, &ses->iface_list[j]);
+			}
+			spin_unlock(&ses->iface_lock);
 		}
 	}
 	spin_unlock(&cifs_tcp_ses_lock);
diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
index 937251cc61c0..ee2a8ec70056 100644
--- a/fs/cifs/cifsencrypt.c
+++ b/fs/cifs/cifsencrypt.c
@@ -37,7 +37,6 @@
 #include <crypto/aead.h>
 
 int __cifs_calc_signature(struct smb_rqst *rqst,
-			int start,
 			struct TCP_Server_Info *server, char *signature,
 			struct shash_desc *shash)
 {
@@ -45,16 +44,27 @@ int __cifs_calc_signature(struct smb_rqst *rqst,
 	int rc;
 	struct kvec *iov = rqst->rq_iov;
 	int n_vec = rqst->rq_nvec;
+	int is_smb2 = server->vals->header_preamble_size == 0;
 
-	for (i = start; i < n_vec; i++) {
+	/* iov[0] is actual data and not the rfc1002 length for SMB2+ */
+	if (is_smb2) {
+		if (iov[0].iov_len <= 4)
+			return -EIO;
+		i = 0;
+	} else {
+		if (n_vec < 2 || iov[0].iov_len != 4)
+			return -EIO;
+		i = 1; /* skip rfc1002 length */
+	}
+
+	for (; i < n_vec; i++) {
 		if (iov[i].iov_len == 0)
 			continue;
 		if (iov[i].iov_base == NULL) {
 			cifs_dbg(VFS, "null iovec entry\n");
 			return -EIO;
 		}
-		if (i == 1 && iov[1].iov_len <= 4)
-			break; /* nothing to sign or corrupt header */
+
 		rc = crypto_shash_update(shash,
 					 iov[i].iov_base, iov[i].iov_len);
 		if (rc) {
@@ -118,7 +128,7 @@ static int cifs_calc_signature(struct smb_rqst *rqst,
 		return rc;
 	}
 
-	return __cifs_calc_signature(rqst, 1, server, signature,
+	return __cifs_calc_signature(rqst, server, signature,
 				     &server->secmech.sdescmd5->shash);
 }
 
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 1efa2e65bc1a..bd78da59a4fd 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -33,6 +33,9 @@
 
 #define CIFS_MAGIC_NUMBER 0xFF534D42      /* the first four bytes of SMB PDUs */
 
+#define CIFS_PORT 445
+#define RFC1001_PORT 139
+
 /*
  * The sizes of various internal tables and strings
  */
@@ -312,6 +315,10 @@ struct smb_version_operations {
 	/* send echo request */
 	int (*echo)(struct TCP_Server_Info *);
 	/* create directory */
+	int (*posix_mkdir)(const unsigned int xid, struct inode *inode,
+			umode_t mode, struct cifs_tcon *tcon,
+			const char *full_path,
+			struct cifs_sb_info *cifs_sb);
 	int (*mkdir)(const unsigned int, struct cifs_tcon *, const char *,
 		     struct cifs_sb_info *);
 	/* set info on created directory */
@@ -838,6 +845,13 @@ static inline void cifs_set_net_ns(struct TCP_Server_Info *srv, struct net *net)
 
 #endif
 
+struct cifs_server_iface {
+	size_t speed;
+	unsigned int rdma_capable : 1;
+	unsigned int rss_capable : 1;
+	struct sockaddr_storage sockaddr;
+};
+
 /*
  * Session structure.  One of these for each uid session with a particular host
  */
@@ -875,6 +889,20 @@ struct cifs_ses {
 #ifdef CONFIG_CIFS_SMB311
 	__u8 preauth_sha_hash[SMB2_PREAUTH_HASH_SIZE];
 #endif /* 3.1.1 */
+
+	/*
+	 * Network interfaces available on the server this session is
+	 * connected to.
+	 *
+	 * Other channels can be opened by connecting and binding this
+	 * session to interfaces from this list.
+	 *
+	 * iface_lock should be taken when accessing any of these fields
+	 */
+	spinlock_t iface_lock;
+	struct cifs_server_iface *iface_list;
+	size_t iface_count;
+	unsigned long iface_last_update; /* jiffies */
 };
 
 static inline bool
@@ -883,6 +911,14 @@ cap_unix(struct cifs_ses *ses)
 	return ses->server->vals->cap_unix & ses->capabilities;
 }
 
+struct cached_fid {
+	bool is_valid:1;	/* Do we have a useable root fid */
+	struct cifs_fid *fid;
+	struct mutex fid_mutex;
+	struct cifs_tcon *tcon;
+	struct work_struct lease_break;
+};
+
 /*
  * there is one of these for each connection to a resource on a particular
  * session
@@ -987,9 +1023,7 @@ struct cifs_tcon {
 	struct fscache_cookie *fscache;	/* cookie for share */
 #endif
 	struct list_head pending_opens;	/* list of incomplete opens */
-	bool valid_root_fid:1;	/* Do we have a useable root fid */
-	struct mutex prfid_mutex; /* prevents reopen race after dead ses*/
-	struct cifs_fid *prfid;	/* handle to the directory at top of share */
+	struct cached_fid crfid; /* Cached root fid */
 	/* BB add field for back pointer to sb struct(s)? */
 };
 
diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h
index 4e0d183c3d10..03018be17283 100644
--- a/fs/cifs/cifsproto.h
+++ b/fs/cifs/cifsproto.h
@@ -112,10 +112,6 @@ extern int SendReceive2(const unsigned int /* xid */ , struct cifs_ses *,
 			struct kvec *, int /* nvec to send */,
 			int * /* type of buf returned */, const int flags,
 			struct kvec * /* resp vec */);
-extern int smb2_send_recv(const unsigned int xid, struct cifs_ses *pses,
-			  struct kvec *pkvec, int nvec_to_send,
-			  int *pbuftype, const int flags,
-			  struct kvec *presp);
 extern int SendReceiveBlockingLock(const unsigned int xid,
 			struct cifs_tcon *ptcon,
 			struct smb_hdr *in_buf ,
@@ -544,7 +540,7 @@ int cifs_create_mf_symlink(unsigned int xid, struct cifs_tcon *tcon,
 			   struct cifs_sb_info *cifs_sb,
 			   const unsigned char *path, char *pbuf,
 			   unsigned int *pbytes_written);
-int __cifs_calc_signature(struct smb_rqst *rqst, int start,
+int __cifs_calc_signature(struct smb_rqst *rqst,
 			struct TCP_Server_Info *server, char *signature,
 			struct shash_desc *shash);
 enum securityEnum cifs_select_sectype(struct TCP_Server_Info *,
@@ -552,6 +548,7 @@ enum securityEnum cifs_select_sectype(struct TCP_Server_Info *,
 struct cifs_aio_ctx *cifs_aio_ctx_alloc(void);
 void cifs_aio_ctx_release(struct kref *refcount);
 int setup_aio_ctx_iter(struct cifs_aio_ctx *ctx, struct iov_iter *iter, int rw);
+void smb2_cached_lease_break(struct work_struct *work);
 
 int cifs_alloc_hash(const char *name, struct crypto_shash **shash,
 		    struct sdesc **sdesc);
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index 42329b25877d..d352da325de3 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -107,10 +107,10 @@ cifs_mark_open_files_invalid(struct cifs_tcon *tcon)
 	}
 	spin_unlock(&tcon->open_file_lock);
 
-	mutex_lock(&tcon->prfid_mutex);
-	tcon->valid_root_fid = false;
-	memset(tcon->prfid, 0, sizeof(struct cifs_fid));
-	mutex_unlock(&tcon->prfid_mutex);
+	mutex_lock(&tcon->crfid.fid_mutex);
+	tcon->crfid.is_valid = false;
+	memset(tcon->crfid.fid, 0, sizeof(struct cifs_fid));
+	mutex_unlock(&tcon->crfid.fid_mutex);
 
 	/*
 	 * BB Add call to invalidate_inodes(sb) for all superblocks mounted
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 96645a7d8f27..a57da1b88bdf 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -57,9 +57,6 @@
 #include "smb2proto.h"
 #include "smbdirect.h"
 
-#define CIFS_PORT 445
-#define RFC1001_PORT 139
-
 extern mempool_t *cifs_req_poolp;
 extern bool disable_legacy_dialects;
 
@@ -3029,8 +3026,11 @@ cifs_get_tcon(struct cifs_ses *ses, struct smb_vol *volume_info)
 
 #ifdef CONFIG_CIFS_SMB311
 	if ((volume_info->linux_ext) && (ses->server->posix_ext_supported)) {
-		if (ses->server->vals->protocol_id == SMB311_PROT_ID)
+		if (ses->server->vals->protocol_id == SMB311_PROT_ID) {
 			tcon->posix_extensions = true;
+			printk_once(KERN_WARNING
+				"SMB3.11 POSIX Extensions are experimental\n");
+		}
 	}
 #endif /* 311 */
 
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index f4697f548a39..a2cfb33e85c1 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -1575,6 +1575,17 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, umode_t mode)
 		goto mkdir_out;
 	}
 
+	server = tcon->ses->server;
+
+#ifdef CONFIG_CIFS_SMB311
+	if ((server->ops->posix_mkdir) && (tcon->posix_extensions)) {
+		rc = server->ops->posix_mkdir(xid, inode, mode, tcon, full_path,
+					      cifs_sb);
+		d_drop(direntry); /* for time being always refresh inode info */
+		goto mkdir_out;
+	}
+#endif /* SMB311 */
+
 	if (cap_unix(tcon->ses) && (CIFS_UNIX_POSIX_PATH_OPS_CAP &
 				le64_to_cpu(tcon->fsUnixInfo.Capability))) {
 		rc = cifs_posix_mkdir(inode, direntry, mode, full_path, cifs_sb,
@@ -1583,8 +1594,6 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, umode_t mode)
 			goto mkdir_out;
 	}
 
-	server = tcon->ses->server;
-
 	if (!server->ops->mkdir) {
 		rc = -ENOSYS;
 		goto mkdir_out;
diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c
index af29ade195c0..53e8362cbc4a 100644
--- a/fs/cifs/misc.c
+++ b/fs/cifs/misc.c
@@ -82,6 +82,7 @@ sesInfoAlloc(void)
 		INIT_LIST_HEAD(&ret_buf->smb_ses_list);
 		INIT_LIST_HEAD(&ret_buf->tcon_list);
 		mutex_init(&ret_buf->session_mutex);
+		spin_lock_init(&ret_buf->iface_lock);
 	}
 	return ret_buf;
 }
@@ -102,6 +103,7 @@ sesInfoFree(struct cifs_ses *buf_to_free)
 	kfree(buf_to_free->user_name);
 	kfree(buf_to_free->domainName);
 	kzfree(buf_to_free->auth_key.response);
+	kfree(buf_to_free->iface_list);
 	kzfree(buf_to_free);
 }
 
@@ -117,8 +119,9 @@ tconInfoAlloc(void)
 		INIT_LIST_HEAD(&ret_buf->openFileList);
 		INIT_LIST_HEAD(&ret_buf->tcon_list);
 		spin_lock_init(&ret_buf->open_file_lock);
-		mutex_init(&ret_buf->prfid_mutex);
-		ret_buf->prfid = kzalloc(sizeof(struct cifs_fid), GFP_KERNEL);
+		mutex_init(&ret_buf->crfid.fid_mutex);
+		ret_buf->crfid.fid = kzalloc(sizeof(struct cifs_fid),
+					     GFP_KERNEL);
 #ifdef CONFIG_CIFS_STATS
 		spin_lock_init(&ret_buf->stat_lock);
 #endif
@@ -136,7 +139,7 @@ tconInfoFree(struct cifs_tcon *buf_to_free)
 	atomic_dec(&tconInfoAllocCount);
 	kfree(buf_to_free->nativeFileSystem);
 	kzfree(buf_to_free->password);
-	kfree(buf_to_free->prfid);
+	kfree(buf_to_free->crfid.fid);
 	kfree(buf_to_free);
 }
 
diff --git a/fs/cifs/smb2misc.c b/fs/cifs/smb2misc.c
index e2bec47c6845..3ff7cec2da81 100644
--- a/fs/cifs/smb2misc.c
+++ b/fs/cifs/smb2misc.c
@@ -454,7 +454,8 @@ cifs_convert_path_to_utf16(const char *from, struct cifs_sb_info *cifs_sb)
 #ifdef CONFIG_CIFS_SMB311
 	/* SMB311 POSIX extensions paths do not include leading slash */
 	else if (cifs_sb_master_tlink(cifs_sb) &&
-		 cifs_sb_master_tcon(cifs_sb)->posix_extensions) {
+		 cifs_sb_master_tcon(cifs_sb)->posix_extensions &&
+		 (from[0] == '/')) {
 		start_of_path = from + 1;
 	}
 #endif /* 311 */
@@ -492,10 +493,11 @@ cifs_ses_oplock_break(struct work_struct *work)
 {
 	struct smb2_lease_break_work *lw = container_of(work,
 				struct smb2_lease_break_work, lease_break);
-	int rc;
+	int rc = 0;
 
 	rc = SMB2_lease_break(0, tlink_tcon(lw->tlink), lw->lease_key,
 			      lw->lease_state);
+
 	cifs_dbg(FYI, "Lease release rc %d\n", rc);
 	cifs_put_tlink(lw->tlink);
 	kfree(lw);
@@ -561,6 +563,7 @@ smb2_tcon_has_lease(struct cifs_tcon *tcon, struct smb2_lease_break *rsp,
 
 		open->oplock = lease_state;
 	}
+
 	return found;
 }
 
@@ -603,6 +606,18 @@ smb2_is_valid_lease_break(char *buffer)
 					return true;
 				}
 				spin_unlock(&tcon->open_file_lock);
+
+				if (tcon->crfid.is_valid &&
+				    !memcmp(rsp->LeaseKey,
+					    tcon->crfid.fid->lease_key,
+					    SMB2_LEASE_KEY_SIZE)) {
+					INIT_WORK(&tcon->crfid.lease_break,
+						  smb2_cached_lease_break);
+					queue_work(cifsiod_wq,
+						   &tcon->crfid.lease_break);
+					spin_unlock(&cifs_tcp_ses_lock);
+					return true;
+				}
 			}
 		}
 	}
diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
index b15f5957d645..0356b5559c71 100644
--- a/fs/cifs/smb2ops.c
+++ b/fs/cifs/smb2ops.c
@@ -294,34 +294,191 @@ smb2_negotiate_rsize(struct cifs_tcon *tcon, struct smb_vol *volume_info)
 	return rsize;
 }
 
-#ifdef CONFIG_CIFS_STATS2
+
+static int
+parse_server_interfaces(struct network_interface_info_ioctl_rsp *buf,
+			size_t buf_len,
+			struct cifs_server_iface **iface_list,
+			size_t *iface_count)
+{
+	struct network_interface_info_ioctl_rsp *p;
+	struct sockaddr_in *addr4;
+	struct sockaddr_in6 *addr6;
+	struct iface_info_ipv4 *p4;
+	struct iface_info_ipv6 *p6;
+	struct cifs_server_iface *info;
+	ssize_t bytes_left;
+	size_t next = 0;
+	int nb_iface = 0;
+	int rc = 0;
+
+	*iface_list = NULL;
+	*iface_count = 0;
+
+	/*
+	 * Fist pass: count and sanity check
+	 */
+
+	bytes_left = buf_len;
+	p = buf;
+	while (bytes_left >= sizeof(*p)) {
+		nb_iface++;
+		next = le32_to_cpu(p->Next);
+		if (!next) {
+			bytes_left -= sizeof(*p);
+			break;
+		}
+		p = (struct network_interface_info_ioctl_rsp *)((u8 *)p+next);
+		bytes_left -= next;
+	}
+
+	if (!nb_iface) {
+		cifs_dbg(VFS, "%s: malformed interface info\n", __func__);
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (bytes_left || p->Next)
+		cifs_dbg(VFS, "%s: incomplete interface info\n", __func__);
+
+
+	/*
+	 * Second pass: extract info to internal structure
+	 */
+
+	*iface_list = kcalloc(nb_iface, sizeof(**iface_list), GFP_KERNEL);
+	if (!*iface_list) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	info = *iface_list;
+	bytes_left = buf_len;
+	p = buf;
+	while (bytes_left >= sizeof(*p)) {
+		info->speed = le64_to_cpu(p->LinkSpeed);
+		info->rdma_capable = le32_to_cpu(p->Capability & RDMA_CAPABLE);
+		info->rss_capable = le32_to_cpu(p->Capability & RSS_CAPABLE);
+
+		cifs_dbg(FYI, "%s: adding iface %zu\n", __func__, *iface_count);
+		cifs_dbg(FYI, "%s: speed %zu bps\n", __func__, info->speed);
+		cifs_dbg(FYI, "%s: capabilities 0x%08x\n", __func__,
+			 le32_to_cpu(p->Capability));
+
+		switch (p->Family) {
+		/*
+		 * The kernel and wire socket structures have the same
+		 * layout and use network byte order but make the
+		 * conversion explicit in case either one changes.
+		 */
+		case INTERNETWORK:
+			addr4 = (struct sockaddr_in *)&info->sockaddr;
+			p4 = (struct iface_info_ipv4 *)p->Buffer;
+			addr4->sin_family = AF_INET;
+			memcpy(&addr4->sin_addr, &p4->IPv4Address, 4);
+
+			/* [MS-SMB2] 2.2.32.5.1.1 Clients MUST ignore these */
+			addr4->sin_port = cpu_to_be16(CIFS_PORT);
+
+			cifs_dbg(FYI, "%s: ipv4 %pI4\n", __func__,
+				 &addr4->sin_addr);
+			break;
+		case INTERNETWORKV6:
+			addr6 =	(struct sockaddr_in6 *)&info->sockaddr;
+			p6 = (struct iface_info_ipv6 *)p->Buffer;
+			addr6->sin6_family = AF_INET6;
+			memcpy(&addr6->sin6_addr, &p6->IPv6Address, 16);
+
+			/* [MS-SMB2] 2.2.32.5.1.2 Clients MUST ignore these */
+			addr6->sin6_flowinfo = 0;
+			addr6->sin6_scope_id = 0;
+			addr6->sin6_port = cpu_to_be16(CIFS_PORT);
+
+			cifs_dbg(FYI, "%s: ipv6 %pI6\n", __func__,
+				 &addr6->sin6_addr);
+			break;
+		default:
+			cifs_dbg(VFS,
+				 "%s: skipping unsupported socket family\n",
+				 __func__);
+			goto next_iface;
+		}
+
+		(*iface_count)++;
+		info++;
+next_iface:
+		next = le32_to_cpu(p->Next);
+		if (!next)
+			break;
+		p = (struct network_interface_info_ioctl_rsp *)((u8 *)p+next);
+		bytes_left -= next;
+	}
+
+	if (!*iface_count) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+out:
+	if (rc) {
+		kfree(*iface_list);
+		*iface_count = 0;
+		*iface_list = NULL;
+	}
+	return rc;
+}
+
+
 static int
 SMB3_request_interfaces(const unsigned int xid, struct cifs_tcon *tcon)
 {
 	int rc;
 	unsigned int ret_data_len = 0;
-	struct network_interface_info_ioctl_rsp *out_buf;
+	struct network_interface_info_ioctl_rsp *out_buf = NULL;
+	struct cifs_server_iface *iface_list;
+	size_t iface_count;
+	struct cifs_ses *ses = tcon->ses;
 
 	rc = SMB2_ioctl(xid, tcon, NO_FILE_ID, NO_FILE_ID,
 			FSCTL_QUERY_NETWORK_INTERFACE_INFO, true /* is_fsctl */,
 			NULL /* no data input */, 0 /* no data input */,
 			(char **)&out_buf, &ret_data_len);
-	if (rc != 0)
+	if (rc != 0) {
 		cifs_dbg(VFS, "error %d on ioctl to get interface list\n", rc);
-	else if (ret_data_len < sizeof(struct network_interface_info_ioctl_rsp)) {
-		cifs_dbg(VFS, "server returned bad net interface info buf\n");
-		rc = -EINVAL;
-	} else {
-		/* Dump info on first interface */
-		cifs_dbg(FYI, "Adapter Capability 0x%x\t",
-			le32_to_cpu(out_buf->Capability));
-		cifs_dbg(FYI, "Link Speed %lld\n",
-			le64_to_cpu(out_buf->LinkSpeed));
+		goto out;
 	}
+
+	rc = parse_server_interfaces(out_buf, ret_data_len,
+				     &iface_list, &iface_count);
+	if (rc)
+		goto out;
+
+	spin_lock(&ses->iface_lock);
+	kfree(ses->iface_list);
+	ses->iface_list = iface_list;
+	ses->iface_count = iface_count;
+	ses->iface_last_update = jiffies;
+	spin_unlock(&ses->iface_lock);
+
+out:
 	kfree(out_buf);
 	return rc;
 }
-#endif /* STATS2 */
+
+void
+smb2_cached_lease_break(struct work_struct *work)
+{
+	struct cached_fid *cfid = container_of(work,
+				struct cached_fid, lease_break);
+	mutex_lock(&cfid->fid_mutex);
+	if (cfid->is_valid) {
+		cifs_dbg(FYI, "clear cached root file handle\n");
+		SMB2_close(0, cfid->tcon, cfid->fid->persistent_fid,
+			   cfid->fid->volatile_fid);
+		cfid->is_valid = false;
+	}
+	mutex_unlock(&cfid->fid_mutex);
+}
 
 /*
  * Open the directory at the root of a share
@@ -331,13 +488,13 @@ int open_shroot(unsigned int xid, struct cifs_tcon *tcon, struct cifs_fid *pfid)
 	struct cifs_open_parms oparams;
 	int rc;
 	__le16 srch_path = 0; /* Null - since an open of top of share */
-	u8 oplock = SMB2_OPLOCK_LEVEL_NONE;
+	u8 oplock = SMB2_OPLOCK_LEVEL_II;
 
-	mutex_lock(&tcon->prfid_mutex);
-	if (tcon->valid_root_fid) {
+	mutex_lock(&tcon->crfid.fid_mutex);
+	if (tcon->crfid.is_valid) {
 		cifs_dbg(FYI, "found a cached root file handle\n");
-		memcpy(pfid, tcon->prfid, sizeof(struct cifs_fid));
-		mutex_unlock(&tcon->prfid_mutex);
+		memcpy(pfid, tcon->crfid.fid, sizeof(struct cifs_fid));
+		mutex_unlock(&tcon->crfid.fid_mutex);
 		return 0;
 	}
 
@@ -350,10 +507,11 @@ int open_shroot(unsigned int xid, struct cifs_tcon *tcon, struct cifs_fid *pfid)
 
 	rc = SMB2_open(xid, &oparams, &srch_path, &oplock, NULL, NULL, NULL);
 	if (rc == 0) {
-		memcpy(tcon->prfid, pfid, sizeof(struct cifs_fid));
-		tcon->valid_root_fid = true;
+		memcpy(tcon->crfid.fid, pfid, sizeof(struct cifs_fid));
+		tcon->crfid.tcon = tcon;
+		tcon->crfid.is_valid = true;
 	}
-	mutex_unlock(&tcon->prfid_mutex);
+	mutex_unlock(&tcon->crfid.fid_mutex);
 	return rc;
 }
 
@@ -383,9 +541,7 @@ smb3_qfs_tcon(const unsigned int xid, struct cifs_tcon *tcon)
 	if (rc)
 		return;
 
-#ifdef CONFIG_CIFS_STATS2
 	SMB3_request_interfaces(xid, tcon);
-#endif /* STATS2 */
 
 	SMB2_QFS_attr(xid, tcon, fid.persistent_fid, fid.volatile_fid,
 			FS_ATTRIBUTE_INFORMATION);
@@ -436,7 +592,7 @@ smb2_is_path_accessible(const unsigned int xid, struct cifs_tcon *tcon,
 	struct cifs_open_parms oparms;
 	struct cifs_fid fid;
 
-	if ((*full_path == 0) && tcon->valid_root_fid)
+	if ((*full_path == 0) && tcon->crfid.is_valid)
 		return 0;
 
 	utf16_path = cifs_convert_path_to_utf16(full_path, cifs_sb);
@@ -2151,7 +2307,7 @@ fill_transform_hdr(struct smb2_transform_hdr *tr_hdr, unsigned int orig_len,
 		   struct smb_rqst *old_rq)
 {
 	struct smb2_sync_hdr *shdr =
-			(struct smb2_sync_hdr *)old_rq->rq_iov[1].iov_base;
+			(struct smb2_sync_hdr *)old_rq->rq_iov[0].iov_base;
 
 	memset(tr_hdr, 0, sizeof(struct smb2_transform_hdr));
 	tr_hdr->ProtocolId = SMB2_TRANSFORM_PROTO_NUM;
@@ -2171,14 +2327,13 @@ static inline void smb2_sg_set_buf(struct scatterlist *sg, const void *buf,
 }
 
 /* Assumes:
- * rqst->rq_iov[0]  is rfc1002 length
- * rqst->rq_iov[1]  is tranform header
- * rqst->rq_iov[2+] data to be encrypted/decrypted
+ * rqst->rq_iov[0]  is transform header
+ * rqst->rq_iov[1+] data to be encrypted/decrypted
  */
 static struct scatterlist *
 init_sg(struct smb_rqst *rqst, u8 *sign)
 {
-	unsigned int sg_len = rqst->rq_nvec + rqst->rq_npages;
+	unsigned int sg_len = rqst->rq_nvec + rqst->rq_npages + 1;
 	unsigned int assoc_data_len = sizeof(struct smb2_transform_hdr) - 20;
 	struct scatterlist *sg;
 	unsigned int i;
@@ -2189,10 +2344,10 @@ init_sg(struct smb_rqst *rqst, u8 *sign)
 		return NULL;
 
 	sg_init_table(sg, sg_len);
-	smb2_sg_set_buf(&sg[0], rqst->rq_iov[1].iov_base + 20, assoc_data_len);
-	for (i = 1; i < rqst->rq_nvec - 1; i++)
-		smb2_sg_set_buf(&sg[i], rqst->rq_iov[i+1].iov_base,
-						rqst->rq_iov[i+1].iov_len);
+	smb2_sg_set_buf(&sg[0], rqst->rq_iov[0].iov_base + 20, assoc_data_len);
+	for (i = 1; i < rqst->rq_nvec; i++)
+		smb2_sg_set_buf(&sg[i], rqst->rq_iov[i].iov_base,
+						rqst->rq_iov[i].iov_len);
 	for (j = 0; i < sg_len - 1; i++, j++) {
 		unsigned int len, offset;
 
@@ -2224,18 +2379,17 @@ smb2_get_enc_key(struct TCP_Server_Info *server, __u64 ses_id, int enc, u8 *key)
 	return 1;
 }
 /*
- * Encrypt or decrypt @rqst message. @rqst has the following format:
- * iov[0] - rfc1002 length
- * iov[1] - transform header (associate data),
- * iov[2-N] and pages - data to encrypt.
- * On success return encrypted data in iov[2-N] and pages, leave iov[0-1]
+ * Encrypt or decrypt @rqst message. @rqst[0] has the following format:
+ * iov[0]   - transform header (associate data),
+ * iov[1-N] - SMB2 header and pages - data to encrypt.
+ * On success return encrypted data in iov[1-N] and pages, leave iov[0]
  * untouched.
  */
 static int
 crypt_message(struct TCP_Server_Info *server, struct smb_rqst *rqst, int enc)
 {
 	struct smb2_transform_hdr *tr_hdr =
-			(struct smb2_transform_hdr *)rqst->rq_iov[1].iov_base;
+			(struct smb2_transform_hdr *)rqst->rq_iov[0].iov_base;
 	unsigned int assoc_data_len = sizeof(struct smb2_transform_hdr) - 20;
 	int rc = 0;
 	struct scatterlist *sg;
@@ -2323,10 +2477,6 @@ free_req:
 	return rc;
 }
 
-/*
- * This is called from smb_send_rqst. At this point we have the rfc1002
- * header as the first element in the vector.
- */
 static int
 smb3_init_transform_rq(struct TCP_Server_Info *server, struct smb_rqst *new_rq,
 		       struct smb_rqst *old_rq)
@@ -2335,7 +2485,7 @@ smb3_init_transform_rq(struct TCP_Server_Info *server, struct smb_rqst *new_rq,
 	struct page **pages;
 	struct smb2_transform_hdr *tr_hdr;
 	unsigned int npages = old_rq->rq_npages;
-	unsigned int orig_len = get_rfc1002_length(old_rq->rq_iov[0].iov_base);
+	unsigned int orig_len;
 	int i;
 	int rc = -ENOMEM;
 
@@ -2355,18 +2505,14 @@ smb3_init_transform_rq(struct TCP_Server_Info *server, struct smb_rqst *new_rq,
 			goto err_free_pages;
 	}
 
-	/* Make space for one extra iov to hold the transform header */
 	iov = kmalloc_array(old_rq->rq_nvec + 1, sizeof(struct kvec),
 			    GFP_KERNEL);
 	if (!iov)
 		goto err_free_pages;
 
-	/* copy all iovs from the old except the 1st one (rfc1002 length) */
-	memcpy(&iov[2], &old_rq->rq_iov[1],
-				sizeof(struct kvec) * (old_rq->rq_nvec - 1));