With protocol version 2.0 mounts we have seen crashes with corrupt mid
entries. Either the server->pending_mid_q list becomes corrupt with a
cyclic reference in one element or a mid object fetched by the
demultiplexer thread becomes overwritten during use.

Code review identified a race between the demultiplexer thread and the
request issuing thread. The demultiplexer thread seems to be written
with the assumption that it is the sole user of the mid object until
it calls the mid callback which either wakes the issuer task or
deletes the mid.

This assumption is not true because the issuer task can be woken up
earlier by a signal. If the demultiplexer thread has proceeded as far
as setting the mid_state to MID_RESPONSE_RECEIVED then the issuer
thread will happily end up calling cifs_delete_mid while the
demultiplexer thread still is using the mid object.

Inserting a delay in the cifs demultiplexer thread widens the race
window and makes reproduction of the race very easy:

		if (server->large_buf)
			buf = server->bigbuf;

+		usleep_range(500, 4000);

		server->lstrp = jiffies;

To resolve this I think the proper solution involves putting a
reference count on the mid object. This patch makes sure that the
demultiplexer thread holds a reference until it has finished
processing the transaction.

Cc: stable@vger.kernel.org
Signed-off-by: NLars Persson <larper@axis.com>
Acked-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

In smb3_init_transform_rq(), 'orig_len' was only counting the request
length, but forgot to count any data pages in the request.

Writing or creating files with the 'seal' mount option was broken.

In addition, do some code refactoring by exporting smb2_rqst_len() to
calculate the appropriate packet size and avoid duplicating the same
calculation all over the code.

The start of the io vector is either the rfc1002 length (4 bytes) or a
SMB2 header which is always > 4. Use this fact to check and skip the
rfc1002 length if requested.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

mkdir was not passing the mode on smb3.11 mounts with posix extensions
Signed-off-by: NSteve French <stfrench@microsoft.com>

Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Since the rfc1002 generation was moved down to __smb_send_rqst(),
the transform header is now in rqst->rq_iov[0].

Correctly assign the transform header pointer in crypt_message().
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Move the generation of the 4 byte length field down the stack and
generate it immediately before we start writing the data to the socket.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

Use a read lease for the cached root fid so that we can detect
when the content of the directory changes (via a break) at which time
we close the handle. On next access to the root the handle will be reopened
and cached again.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This leak was introduced in 91cb74f5 and caused us
to leak one small buffer for every symlink query.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

The smb2 hdr is now in iov 1
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Reviewed-by: NPaulo Alcantara <palcantara@suse.de>
Acked-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Encryption function needs to read data starting page offset from input
buffer.

This doesn't affect decryption path since it allocates its own page
buffers.
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

With offset defined in rdata, transport functions need to look at this
offset when reading data into the correct places in pages.
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

Since header_preamble_size is 0 for SMB2+ we can remove it in those
code paths that are only invoked from SMB2.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

struct smb2_hdr is now just a wrapper for smb2_sync_hdr.
We can thus get rid of smb2_hdr completely and access the sync header directly.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

CIFS: 511c54a2 adds a check for session expiry, status STATUS_NETWORK_SESSION_EXPIRED, however the server can also respond with STATUS_USER_SESSION_DELETED in cases where the session has been idle for some time and the server reaps the session to recover resources.

Handle this additional status in the same way as SESSION_EXPIRED.
Signed-off-by: NMark Syms <mark.syms@citrix.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
CC: Stable <stable@vger.kernel.org>

Separate out all the 4 byte rfc1002 headers so that they are no longer
part of the SMB2 header structures to prepare for future work to add
compounding support.

Update the smb3 transform header processing that we no longer have
a rfc1002 header at the start of this structure.

Update smb2_readv_callback to accommodate that the first iovector in the
response is no the smb2 header and no longer a rfc1002 header.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

When loooking at the logs for the new trace-cmd tracepoints for cifs,
it would help to know which tid is for which share (UNC name) so
update /proc/fs/cifs/DebugData to display the tid.
Also display Maximal Access which was missing as well.

Now the entry for typical entry for a tcon (in proc/fs/cifs/) looks
like:

1) \\localhost\test Mounts: 1 DevInfo: 0x20 Attributes: 0x1006f
	PathComponentMax: 255 Status: 1 type: DISK
	Share Capabilities: None Aligned, Partition Aligned,	Share Flags: 0x0
	tid: 0xe0632a55	Optimal sector size: 0x200	Maximal Access: 0x1f01ff
Signed-off-by: NSteve French <smfrench@gmail.com>

Fix a few cases where we were not freeing the xid which led to
active requests being non-zero at unmount time.
Signed-off-by: NSteve French <smfrench@gmail.com>
CC: Stable <stable@vger.kernel.org>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

In SMB2_open(), if we got a lease we need to store this in the fid structure
or else we will never be able to map a lease break back to which file/fid
it applies to.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

Previous patches "cifs: update calc_size to take a server argument"
and
  "cifs: add server argument to the dump_detail method"
were broken if CONFIG_CIFS_DEBUG2 enabled
Signed-off-by: NSteve French <smfrench@gmail.com>
CC: Ronnie Sahlberg <lsahlber@redhat.com>

We need a struct TCP_Server_Info *server to this method as it calls
calc_size. The calc_size method will soon be changed to also
take a server argument.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

In SMB2/SMB3 unlike in cifs we unnecessarily open the root of the share
over and over again in various places during mount and path revalidation
and also in statfs.  This patch cuts redundant traffic (opens and closes)
by simply keeping the directory handle for the root around (and reopening
it as needed on reconnect), so query calls don't require three round
trips to copmlete - just one, and eases load on network, client and
server (on mount alone, cuts network traffic by more than a third).

Also add a new cifs mount parm "nohandlecache" to allow users whose
servers might have resource constraints (eg in case they have a server
with so many users connecting to it that this extra handle per mount
could possibly be a resource concern).
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

As per listxattr(2):

       On success, a nonnegative number is returned indicating the size
       of the extended attribute name list.  On failure, -1 is returned
       and errno  is set appropriately.

In SMB1, when the server returns an empty EA list through a listxattr(),
it will correctly return 0 as there are no EAs for the given file.

However, in SMB2+, it returns -ENODATA in listxattr() which is wrong since
the request and response were sent successfully, although there's no actual
EA for the given file.

This patch fixes listxattr() for SMB2+ by returning 0 in cifs_listxattr()
when the server returns an empty list of EAs.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

SMB server will not sign data transferred through RDMA read/write. When
signing is used, it's a good idea to have all the data signed.

In this case, use RDMA send/recv for all data transfers. This will degrade
performance as this is not generally configured in RDMA environemnt. So
warn the user on signing and RDMA send/recv.
Signed-off-by: NLong Li <longli@microsoft.com>
Acked-by: NRonnie Sahlberg <lsahlber@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: NSteve French <smfrench@gmail.com>

The current code null checks variable err_buf, which is always null
when it is checked, hence utf16_path is free'd and the function
returns -ENOENT everytime it is called, making it impossible for the
execution path to reach the following code:

err_buf = err_iov.iov_base;

Fix this by null checking err_iov.iov_base instead of err_buf. Also,
notice that err_buf no longer needs to be initialized to NULL.

Addresses-Coverity-ID: 1467876 ("Logically dead code")
Fixes: 2d636199e400 ("cifs: Change SMB2_open to return an iov for the error parameter")
Signed-off-by: NGustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

and get rid of some get_rfc1002_length() in smb2
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

This variable is set to 4 for all protocol versions and replaces
the hardcoded constant 4 throughought the code.
This will later be updated to reflect whether a response packet
has a 4 byte length preamble or not once we start removing this
field from the SMB2+ dialects.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

STATUS_FS_DRIVER_REQUIRED is expected when DFS is not turned
on on the server.  Do not log it on DFS referral response.
It clutters the dmesg log unnecessarily at mount time.
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: Aurelien Aptel <aaptel@suse.com
Reviewed-by: NRonnie sahlberg <lsahlber@redhat.com>

We can not use the standard sg_set_buf() fucntion since when
CONFIG_DEBUG_SG=y this adds a check that will BUG_ON for cifs.ko
when we pass it an object from the stack.

Create a new wrapper smb2_sg_set_buf() which avoids doing that particular check
and use it for smb3 encryption instead.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
CC: Stable <stable@vger.kernel.org>

Since IPC now has a tcon object, the caller can just pass it. This
allows domain-based DFS requests to work with smb2+.

Link: https://bugzilla.samba.org/show_bug.cgi?id=12917
Fixes: 9d49640a ("CIFS: implement get_dfs_refer for SMB2+")
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

This patch is for preparing upper layer doing SMB read via RDMA write.

When RDMA write is used for SMB read, the returned data length is in
DataRemaining in the response packet. Reading it properly by adding a
parameter to specifiy where the returned data length is.

Add the defition for memory registration to wdata and return the correct
length based on if RDMA write is used.
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

When connecting over SMB Direct, the transport negotiates its maximum I/O sizes
with the server and determines how to choose to do RDMA send/recv vs
read/write. Expose these maximum I/O sizes to upper layer so we will get the
correct sized payloads.
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

cifs.ko makes DFS queries regardless of the type of the server and
non-DFS servers are common. This often results in superfluous logging of
non-critical errors.
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

cifs starts an async. crypto op and waits for their completion.
Move it over to generic code doing the same.
Signed-off-by: NGilad Ben-Yossef <gilad@benyossef.com>
Acked-by: NPavel Shilovsky <pshilov@microsoft.com>
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>

Update reading the EA using increasingly larger buffer sizes
until the response will fit in the buffer, or we exceed the
(arbitrary) maximum set to 64kb.

Without this change, a user is able to add more and more EAs using
setfattr until the point where the total space of all EAs exceed 2kb
at which point the user can no longer list the EAs at all
and getfattr will abort with an error.

The same issue still exists for EAs in SMB1.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Reported-by: NXiaoli Feng <xifeng@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>

The newly added SMB2+ attribute support causes unused function
warnings when CONFIG_CIFS_XATTR is disabled:

fs/cifs/smb2ops.c:563:1: error: 'smb2_set_ea' defined but not used [-Werror=unused-function]
 smb2_set_ea(const unsigned int xid, struct cifs_tcon *tcon,
fs/cifs/smb2ops.c:513:1: error: 'smb2_query_eas' defined but not used [-Werror=unused-function]
 smb2_query_eas(const unsigned int xid, struct cifs_tcon *tcon,

This adds another #ifdef around the affected functions.

Fixes: 5517554e ("cifs: Add support for writing attributes on SMB2+")
Fixes: 95907fea ("cifs: Add support for reading attributes on SMB2+")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NSteve French <smfrench@gmail.com>

With the need to discourage use of less secure dialect, SMB1 (CIFS),
we temporarily upgraded the dialect to SMB3 in 4.13, but since there
are various servers which only support SMB2.1 (2.1 is more secure
than CIFS/SMB1) but not optimal for a default dialect - add support
for multidialect negotiation.  cifs.ko will now request SMB2.1
or later (ie SMB2.1 or SMB3.0, SMB3.02) and the server will
pick the latest most secure one it can support.

In addition since we are sending multidialect negotiate, add
support for secure negotiate to validate that a man in the
middle didn't downgrade us.
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
CC: Stable <stable@vger.kernel.org> # 4.13+

This adds support for writing extended attributes on SMB2+ shares.
Attributes can be written using the setfattr command.

RH-bz: 1110709
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

SMB1 already has support to read attributes. This adds similar support
to SMB2+.

With this patch, tools such as 'getfattr' will now work with SMB2+ shares.

RH-bz: 1110709
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <smfrench@gmail.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>