[PATCH AUTOSEL 5.13 38/59] cifs: improve fallocate emulation

[PATCH AUTOSEL 5.13 38/59] cifs: improve fallocate emulation

[Sasha Levin]

From: Ronnie Sahlberg <lsahlber@redhat.com>

[ Upstream commit 966a3cb7c7db786452a87afdc3b48858fc4d4d6b ]

RHBZ: 1866684

We don't have a real fallocate in the SMB2 protocol so we used to emulate fallocate
by simply switching the file to become non-sparse. But as that could potantially consume
a lot more data than we intended to fallocate (large sparse file and fallocating a thin
slice in the middle) we would only do this IFF the fallocate request was for virtually
the entire file.

This patch improves this and starts allowing us to fallocate smaller chunks of a file by
overwriting the region with 0, for the parts that are unallocated.

The method used is to first query the server for FSCTL_QUERY_ALLOCATED_RANGES to find what
is unallocated in the fallocate range and then to only overwrite-with-zero the unallocated
ranges to fill in the holes.

As overwriting-with-zero is different from just allocating blocks, and potentially much
more expensive, we limit this to only allow fallocate ranges up to 1Mb in size.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
Acked-by: Aurelien Aptel <aaptel@suse.com>
Acked-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/cifs/smb2ops.c | 133 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 133 insertions(+)

diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
index 21ef51d338e0..b68ba92893b6 100644
--- a/fs/cifs/smb2ops.c
+++ b/fs/cifs/smb2ops.c
@@ -3601,6 +3601,119 @@ static long smb3_punch_hole(struct file *file, struct cifs_tcon *tcon,
 	return rc;
 }
 
+static int smb3_simple_fallocate_write_range(unsigned int xid,
+					     struct cifs_tcon *tcon,
+					     struct cifsFileInfo *cfile,
+					     loff_t off, loff_t len,
+					     char *buf)
+{
+	struct cifs_io_parms io_parms = {0};
+	int nbytes;
+	struct kvec iov[2];
+
+	io_parms.netfid = cfile->fid.netfid;
+	io_parms.pid = current->tgid;
+	io_parms.tcon = tcon;
+	io_parms.persistent_fid = cfile->fid.persistent_fid;
+	io_parms.volatile_fid = cfile->fid.volatile_fid;
+	io_parms.offset = off;
+	io_parms.length = len;
+
+	/* iov[0] is reserved for smb header */
+	iov[1].iov_base = buf;
+	iov[1].iov_len = io_parms.length;
+	return SMB2_write(xid, &io_parms, &nbytes, iov, 1);
+}
+
+static int smb3_simple_fallocate_range(unsigned int xid,
+				       struct cifs_tcon *tcon,
+				       struct cifsFileInfo *cfile,
+				       loff_t off, loff_t len)
+{
+	struct file_allocated_range_buffer in_data, *out_data = NULL, *tmp_data;
+	u32 out_data_len;
+	char *buf = NULL;
+	loff_t l;
+	int rc;
+
+	in_data.file_offset = cpu_to_le64(off);
+	in_data.length = cpu_to_le64(len);
+	rc = SMB2_ioctl(xid, tcon, cfile->fid.persistent_fid,
+			cfile->fid.volatile_fid,
+			FSCTL_QUERY_ALLOCATED_RANGES, true,
+			(char *)&in_data, sizeof(in_data),
+			1024 * sizeof(struct file_allocated_range_buffer),
+			(char **)&out_data, &out_data_len);
+	if (rc)
+		goto out;
+	/*
+	 * It is already all allocated
+	 */
+	if (out_data_len == 0)
+		goto out;
+
+	buf = kzalloc(1024 * 1024, GFP_KERNEL);
+	if (buf == NULL) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	tmp_data = out_data;
+	while (len) {
+		/*
+		 * The rest of the region is unmapped so write it all.
+		 */
+		if (out_data_len == 0) {
+			rc = smb3_simple_fallocate_write_range(xid, tcon,
+					       cfile, off, len, buf);
+			goto out;
+		}
+
+		if (out_data_len < sizeof(struct file_allocated_range_buffer)) {
+			rc = -EINVAL;
+			goto out;
+		}
+
+		if (off < le64_to_cpu(tmp_data->file_offset)) {
+			/*
+			 * We are at a hole. Write until the end of the region
+			 * or until the next allocated data,
+			 * whichever comes next.
+			 */
+			l = le64_to_cpu(tmp_data->file_offset) - off;
+			if (len < l)
+				l = len;
+			rc = smb3_simple_fallocate_write_range(xid, tcon,
+					       cfile, off, l, buf);
+			if (rc)
+				goto out;
+			off = off + l;
+			len = len - l;
+			if (len == 0)
+				goto out;
+		}
+		/*
+		 * We are at a section of allocated data, just skip forward
+		 * until the end of the data or the end of the region
+		 * we are supposed to fallocate, whichever comes first.
+		 */
+		l = le64_to_cpu(tmp_data->length);
+		if (len < l)
+			l = len;
+		off += l;
+		len -= l;
+
+		tmp_data = &tmp_data[1];
+		out_data_len -= sizeof(struct file_allocated_range_buffer);
+	}
+
+ out:
+	kfree(out_data);
+	kfree(buf);
+	return rc;
+}
+
+
 static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
 			    loff_t off, loff_t len, bool keep_size)
 {
@@ -3661,6 +3774,26 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
 	}
 
 	if ((keep_size == true) || (i_size_read(inode) >= off + len)) {
+		/*
+		 * At this point, we are trying to fallocate an internal
+		 * regions of a sparse file. Since smb2 does not have a
+		 * fallocate command we have two otions on how to emulate this.
+		 * We can either turn the entire file to become non-sparse
+		 * which we only do if the fallocate is for virtually
+		 * the whole file,  or we can overwrite the region with zeroes
+		 * using SMB2_write, which could be prohibitevly expensive
+		 * if len is large.
+		 */
+		/*
+		 * We are only trying to fallocate a small region so
+		 * just write it with zero.
+		 */
+		if (len <= 1024 * 1024) {
+			rc = smb3_simple_fallocate_range(xid, tcon, cfile,
+							 off, len);
+			goto out;
+		}
+
 		/*
 		 * Check if falloc starts within first few pages of file
 		 * and ends within a few pages of the end of file to
-- 
2.30.2

[PATCH AUTOSEL 5.13 39/59] cifs: fix check of dfs interlinks

[Sasha Levin]

From: Paulo Alcantara <pc@cjr.nz>

[ Upstream commit 889c2a700799f3b6f82210925e1faf4a9b833c4a ]

Interlink is a special type of DFS link that resolves to a different
DFS domain-based namespace.  To determine whether it is an interlink
or not, check if ReferralServers and StorageServers bits are set to 1
and 0 respectively in ReferralHeaderFlags, as specified in MS-DFSC
3.1.5.4.5 Determining Whether a Referral Response is an Interlink.

Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Reviewed-by: Aurelien Aptel <aaptel@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/cifs/dfs_cache.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/cifs/dfs_cache.c b/fs/cifs/dfs_cache.c
index b1fa30fefe1f..8e16ee1e5fd1 100644
--- a/fs/cifs/dfs_cache.c
+++ b/fs/cifs/dfs_cache.c
@@ -25,8 +25,7 @@
 #define CACHE_HTABLE_SIZE 32
 #define CACHE_MAX_ENTRIES 64
 
-#define IS_INTERLINK_SET(v) ((v) & (DFSREF_REFERRAL_SERVER | \
-				    DFSREF_STORAGE_SERVER))
+#define IS_DFS_INTERLINK(v) (((v) & DFSREF_REFERRAL_SERVER) && !((v) & DFSREF_STORAGE_SERVER))
 
 struct cache_dfs_tgt {
 	char *name;
@@ -171,7 +170,7 @@ static int dfscache_proc_show(struct seq_file *m, void *v)
 				   "cache entry: path=%s,type=%s,ttl=%d,etime=%ld,hdr_flags=0x%x,ref_flags=0x%x,interlink=%s,path_consumed=%d,expired=%s\n",
 				   ce->path, ce->srvtype == DFS_TYPE_ROOT ? "root" : "link",
 				   ce->ttl, ce->etime.tv_nsec, ce->ref_flags, ce->hdr_flags,
-				   IS_INTERLINK_SET(ce->hdr_flags) ? "yes" : "no",
+				   IS_DFS_INTERLINK(ce->hdr_flags) ? "yes" : "no",
 				   ce->path_consumed, cache_entry_expired(ce) ? "yes" : "no");
 
 			list_for_each_entry(t, &ce->tlist, list) {
@@ -240,7 +239,7 @@ static inline void dump_ce(const struct cache_entry *ce)
 		 ce->srvtype == DFS_TYPE_ROOT ? "root" : "link", ce->ttl,
 		 ce->etime.tv_nsec,
 		 ce->hdr_flags, ce->ref_flags,
-		 IS_INTERLINK_SET(ce->hdr_flags) ? "yes" : "no",
+		 IS_DFS_INTERLINK(ce->hdr_flags) ? "yes" : "no",
 		 ce->path_consumed,
 		 cache_entry_expired(ce) ? "yes" : "no");
 	dump_tgts(ce);
-- 
2.30.2

[PATCH AUTOSEL 5.13 40/59] cifs: retry lookup and readdir when EAGAIN is returned.

[Sasha Levin]

From: Thiago Rafael Becker <trbecker@gmail.com>

[ Upstream commit 6efa994e35a402ae4ae2161b6439c94b64816cee ]

According to the investigation performed by Jacob Shivers at Red Hat,
cifs_lookup and cifs_readdir leak EAGAIN when the user session is
deleted on the server. Fix this issue by implementing a retry with
limits, as is implemented in cifs_revalidate_dentry_attr.

Reproducer based on the work by Jacob Shivers:

  ~~~
  $ cat readdir-cifs-test.sh
  #!/bin/bash

  # Install and configure powershell and sshd on the windows
  #  server as descibed in
  # https://docs.microsoft.com/en-us/windows-server/administration/openssh/openssh_overview
  # This script uses expect(1)

  USER=dude
  SERVER=192.168.0.2
  RPATH=root
  PASS='password'

  function debug_funcs {
  	for line in $@ ; do
  		echo "func $line +p" > /sys/kernel/debug/dynamic_debug/control
  	done
  }

  function setup {
  	echo 1 > /proc/fs/cifs/cifsFYI
  	debug_funcs wait_for_compound_request \
                smb2_query_dir_first cifs_readdir \
                compound_send_recv cifs_reconnect_tcon \
                generic_ip_connect cifs_reconnect \
                smb2_reconnect_server smb2_reconnect \
                cifs_readv_from_socket cifs_readv_receive
  	tcpdump -i eth0 -w cifs.pcap host 192.168.2.182 & sleep 5
  	dmesg -C
  }

  function test_call {
  	if [[ $1 == 1 ]] ; then
  		tracer="strace -tt -f -s 4096 -o trace-$(date -Iseconds).txt"
  	fi
        # Change the command here to anything appropriate
  	$tracer ls $2 > /dev/null
  	res=$?
  	if [[ $1 == 1 ]] ; then
  		if [[ $res == 0 ]] ; then
  			1>&2 echo success
  		else
  			1>&2 echo "failure ($res)"
  		fi
  	fi
  }

  mountpoint /mnt > /dev/null || mount -t cifs -o username=$USER,pass=$PASS //$SERVER/$RPATH /mnt

  test_call 0 /mnt/

  /usr/bin/expect << EOF
  	set timeout 60

  	spawn ssh $USER@$SERVER

  	expect "yes/no" {
  		send "yes\r"
  		expect "*?assword" { send "$PASS\r" }
  	} "*?assword" { send "$PASS\r" }

  	expect ">" { send "powershell close-smbsession -force\r" }
  	expect ">" { send "exit\r" }
  	expect eof
  EOF

  sysctl -w vm.drop_caches=2 > /dev/null
  sysctl -w vm.drop_caches=2 > /dev/null

  setup

  test_call 1 /mnt/
  ~~~

Signed-off-by: Thiago Rafael Becker <trbecker@gmail.com>
Acked-by: Ronnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/cifs/dir.c     | 4 ++++
 fs/cifs/smb2ops.c | 5 +++++
 2 files changed, 9 insertions(+)

diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 6bcd3e8f7cda..7c641f9a3dac 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -630,6 +630,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	struct inode *newInode = NULL;
 	const char *full_path;
 	void *page;
+	int retry_count = 0;
 
 	xid = get_xid();
 
@@ -673,6 +674,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 	cifs_dbg(FYI, "Full path: %s inode = 0x%p\n",
 		 full_path, d_inode(direntry));
 
+again:
 	if (pTcon->posix_extensions)
 		rc = smb311_posix_get_inode_info(&newInode, full_path, parent_dir_inode->i_sb, xid);
 	else if (pTcon->unix_ext) {
@@ -687,6 +689,8 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 		/* since paths are not looked up by component - the parent
 		   directories are presumed to be good here */
 		renew_parental_timestamps(direntry);
+	} else if (rc == -EAGAIN && retry_count++ < 10) {
+		goto again;
 	} else if (rc == -ENOENT) {
 		cifs_set_time(direntry, jiffies);
 		newInode = NULL;
diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
index b68ba92893b6..903de7449aa3 100644
--- a/fs/cifs/smb2ops.c
+++ b/fs/cifs/smb2ops.c
@@ -2325,6 +2325,7 @@ smb2_query_dir_first(const unsigned int xid, struct cifs_tcon *tcon,
 	struct smb2_query_directory_rsp *qd_rsp = NULL;
 	struct smb2_create_rsp *op_rsp = NULL;
 	struct TCP_Server_Info *server = cifs_pick_channel(tcon->ses);
+	int retry_count = 0;
 
 	utf16_path = cifs_convert_path_to_utf16(path, cifs_sb);
 	if (!utf16_path)
@@ -2372,10 +2373,14 @@ smb2_query_dir_first(const unsigned int xid, struct cifs_tcon *tcon,
 
 	smb2_set_related(&rqst[1]);
 
+again:
 	rc = compound_send_recv(xid, tcon->ses, server,
 				flags, 2, rqst,
 				resp_buftype, rsp_iov);
 
+	if (rc == -EAGAIN && retry_count++ < 10)
+		goto again;
+
 	/* If the open failed there is nothing to do */
 	op_rsp = (struct smb2_create_rsp *)rsp_iov[0].iov_base;
 	if (op_rsp == NULL || op_rsp->sync_hdr.Status != STATUS_SUCCESS) {
-- 
2.30.2

[PATCH AUTOSEL 5.13 41/59] smb3: fix uninitialized value for port in witness protocol move

[Sasha Levin]

From: Steve French <stfrench@microsoft.com>

[ Upstream commit ff93b71a3eff25fe9d4364ef13b6e01d935600c6 ]

Although in practice this can not occur (since IPv4 and IPv6 are the
only two cases currently supported), it is cleaner to avoid uninitialized
variable warnings.

Addresses smatch warning:
  fs/cifs/cifs_swn.c:468 cifs_swn_store_swn_addr() error: uninitialized symbol 'port'.

Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
CC: Samuel Cabrero <scabrero@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/cifs/cifs_swn.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/fs/cifs/cifs_swn.c b/fs/cifs/cifs_swn.c
index d829b8bf833e..93b47818c6c2 100644
--- a/fs/cifs/cifs_swn.c
+++ b/fs/cifs/cifs_swn.c
@@ -447,15 +447,13 @@ static int cifs_swn_store_swn_addr(const struct sockaddr_storage *new,
 				   const struct sockaddr_storage *old,
 				   struct sockaddr_storage *dst)
 {
-	__be16 port;
+	__be16 port = cpu_to_be16(CIFS_PORT);
 
 	if (old->ss_family == AF_INET) {
 		struct sockaddr_in *ipv4 = (struct sockaddr_in *)old;
 
 		port = ipv4->sin_port;
-	}
-
-	if (old->ss_family == AF_INET6) {
+	} else if (old->ss_family == AF_INET6) {
 		struct sockaddr_in6 *ipv6 = (struct sockaddr_in6 *)old;
 
 		port = ipv6->sin6_port;
@@ -465,9 +463,7 @@ static int cifs_swn_store_swn_addr(const struct sockaddr_storage *new,
 		struct sockaddr_in *ipv4 = (struct sockaddr_in *)new;
 
 		ipv4->sin_port = port;
-	}
-
-	if (new->ss_family == AF_INET6) {
+	} else if (new->ss_family == AF_INET6) {
 		struct sockaddr_in6 *ipv6 = (struct sockaddr_in6 *)new;
 
 		ipv6->sin6_port = port;
-- 
2.30.2

[PATCH AUTOSEL 5.13 42/59] cifs: fix SMB1 error path in cifs_get_file_info_unix

[Sasha Levin]

From: Steve French <stfrench@microsoft.com>

[ Upstream commit e39df24169a2ceb0d359eb3a05ff982711f2eb32 ]

We were trying to fill in uninitialized file attributes in the error case.

Addresses-Coverity: 139689 ("Uninitialized variables")
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/cifs/inode.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 1dfa57982522..f60f068d33e8 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -367,9 +367,12 @@ cifs_get_file_info_unix(struct file *filp)
 	} else if (rc == -EREMOTE) {
 		cifs_create_dfs_fattr(&fattr, inode->i_sb);
 		rc = 0;
-	}
+	} else
+		goto cifs_gfiunix_out;
 
 	rc = cifs_fattr_to_inode(inode, &fattr);
+
+cifs_gfiunix_out:
 	free_xid(xid);
 	return rc;
 }
-- 
2.30.2