We can not append small padding buffers as separate iovs when encryption is
used. For this case we must flatten the request into a single buffer
containing both the data from all the iovs as well as the padding bytes.

This is at least needed for 4.20 as well due to compounding changes.

CC: Stable <stable@vger.kernel.org>
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Clarify the use of the CONFIG_DFS_UPCALL for DNS name resolution
when server ip addresses change (e.g. on long running mounts)
Signed-off-by: NSteve French <stfrench@microsoft.com>

In case a hostname resolves to a different IP address (e.g. long
running mounts), make sure to resolve it every time prior to calling
generic_ip_connect() in reconnect.
Suggested-by: NSteve French <stfrench@microsoft.com>
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Signed-off-by: NSteve French <stfrench@microsoft.com>

After a successful failover, the cifs_reconnect_tcon() function will
make sure to reconnect every tcon to new target server.

Same as previous commit but for SMB1 codepath.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

After a successful failover in cifs_reconnect(), the smb2_reconnect()
function will make sure to reconnect every tcon to new target server.

For SMB2+.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Fix potential NULL ptr deref when DFS target list is empty.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Start the DFS cache refresh worker per volume during cifs mount.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.de>
Signed-off-by: NSteve French <stfrench@microsoft.com>

A spin lock is held before kstrndup, it may sleep with holding
the spinlock, so we should use GFP_ATOMIC instead.

Fixes: e58c31d5e387 ("cifs: Add support for failover in cifs_reconnect()")
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NPaulo Alcantara <palcantara@suse.de>

After failing to reconnect to original target, it will retry any
target available from DFS cache.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This patch adds support for failover when failing to connect in
cifs_mount().
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Fixes gcc '-Wunused-but-set-variable' warning:

fs/cifs/cifs_dfs_ref.c: In function 'cifs_dfs_do_automount':
fs/cifs/cifs_dfs_ref.c:309:7: warning:
 variable 'sep' set but not used [-Wunused-but-set-variable]

It never used since introdution in commit 0f56b277073c ("cifs: Make use
of DFS cache to get new DFS referrals")
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Reviewed-by: NPaulo Alcantara <palcantara@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This patch will make use of DFS cache routines where appropriate and
do not always request a new referral from server.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Update cifs "TODO" file.
Signed-off-by: NSteve French <stfrench@microsoft.com>

kzalloc can return NULL so an additional check is needed. While there
is a check for ret_buf there is no check for the allocation of
ret_buf->crfid.fid - this check is thus added. Both call-sites
of tconInfoAlloc() check for NULL return of tconInfoAlloc()
so returning NULL on failure of kzalloc() here seems appropriate.
As the kzalloc() is the only thing here that can fail it is
moved to the beginning so as not to initialize other resources
on failure of kzalloc.

Fixes: 3d4ef9a1 ("smb3: fix redundant opens on root")
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Fixes gcc '-Wunused-but-set-variable' warning:

fs/cifs/smb2pdu.c: In function 'smb311_posix_mkdir':
fs/cifs/smb2pdu.c:2040:26: warning:
 variable 'server' set but not used [-Wunused-but-set-variable]

fs/cifs/smb2pdu.c: In function 'build_qfs_info_req':
fs/cifs/smb2pdu.c:4067:26: warning:
 variable 'server' set but not used [-Wunused-but-set-variable]

The first 'server' never used since commit bea851b8 ("smb3: Fix mode on
mkdir on smb311 mounts")
And the second not used since commit 1fc6ad2f ("cifs: remove
header_preamble_size where it is always 0")
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

We should zero out the password before we free it.

Fixes: 3d6cacbb5310 ("cifs: Add DFS cache routines")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NPaulo Alcantara <palcantara@suse.de>

memory allocated by kmem_cache_alloc() in alloc_cache_entry()
should be freed using kmem_cache_free(), not kfree().

Fixes: 34a44fb160f9 ("cifs: Add DFS cache routines")
Signed-off-by: NWei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

Fixes cifs build failure after merge of the y2038 tree

After merging the y2038 tree, today's linux-next build (x86_64
allmodconfig) failed like this:

fs/cifs/dfs_cache.c: In function 'cache_entry_expired':
fs/cifs/dfs_cache.c:106:7: error: implicit declaration of function 'current_kernel_time64'; did you mean 'core_kernel_text'? [-Werror=implicit-function-declaration]
  ts = current_kernel_time64();
       ^~~~~~~~~~~~~~~~~~~~~
       core_kernel_text
fs/cifs/dfs_cache.c:106:5: error: incompatible types when assigning to type 'struct timespec64' from type 'int'
  ts = current_kernel_time64();
     ^
fs/cifs/dfs_cache.c: In function 'get_expire_time':
fs/cifs/dfs_cache.c:342:24: error: incompatible type for argument 1 of 'timespec64_add'
  return timespec64_add(current_kernel_time64(), ts);
                        ^~~~~~~~~~~~~~~~~~~~~~~
In file included from include/linux/restart_block.h:10,
                 from include/linux/thread_info.h:13,
                 from arch/x86/include/asm/preempt.h:7,
                 from include/linux/preempt.h:78,
                 from include/linux/rcupdate.h:40,
                 from fs/cifs/dfs_cache.c:8:
include/linux/time64.h:66:66: note: expected 'struct timespec64' but argument is of type 'int'
 static inline struct timespec64 timespec64_add(struct timespec64 lhs,
                                                ~~~~~~~~~~~~~~~~~~^~~
fs/cifs/dfs_cache.c:343:1: warning: control reaches end of non-void function [-Wreturn-type]
 }
 ^

Caused by:

  commit ccea641b6742 ("timekeeping: remove obsolete time accessors")

interacting with:
  commit 34a44fb160f9 ("cifs: Add DFS cache routines")

from the cifs tree.
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: NPaulo Alcantara <palcantara@suse.de>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NSteve French <stfrench@microsoft.com>

* Add new dfs_cache.[ch] files

* Add new /proc/fs/cifs/dfscache file
  - dump current cache when read
  - clear current cache when writing "0" to it

* Add delayed_work to periodically refresh cache entries

The new interface will be used for caching DFS referrals, as well as
supporting client target failover.

The DFS cache is a hashtable that maps UNC paths to cache entries.

A cache entry contains:
- the UNC path it is mapped on
- how much the the UNC path the entry consumes
- flags
- a Time-To-Live after which the entry expires
- a list of possible targets (linked lists of UNC paths)
- a "hint target" pointing the last known working target or the first
  target if none were tried. This hint lets cifs.ko remember and try
  working targets first.

* Looking for an entry in the cache is done with dfs_cache_find()
  - if no valid entries are found, a DFS query is made, stored in the
    cache and returned
  - the full target list can be copied and returned to avoid race
    conditions and looped on with the help with the
    dfs_cache_tgt_iterator

* Updating the target hint to the next target is done with
  dfs_cache_update_tgthint()

These functions have a dfs_cache_noreq_XXX() version that doesn't
fetches referrals if no entries are found. These versions don't
require the tcp/ses/tcon/cifs_sb parameters as a result.

Expired entries cannot be used and since they have a pretty short TTL
[1] in order for them to be useful for failover the DFS cache adds a
delayed work called periodically to keep them fresh.

Since we might not have available connections to issue the referral
request when refreshing we need to store volume_info structs with
credentials and other needed info to be able to connect to the right
server.

1: Windows defaults: 5mn for domain-based referrals, 30mn for regular
links
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This will be needed by DFS cache.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Different servers have different set of file ids.

After failover, unique IDs will be different so we can't validate
them.
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Reviewed-by: NPaulo Alcantara <palcantara@suse.de>
Signed-off-by: NSteve French <stfrench@microsoft.com>

If we only want to get the mount options strings, do not return the
devname.

For DFS failover, we'll be passing the DFS full path down to
cifs_mount() rather than the devname.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

When extracting hostname from UNC, check for leading backslashes
before trying to remove them.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

* Split and refactor the very large function cifs_mount() in multiple
  functions:

- tcp, ses and tcon setup to mount_get_conns()
- tcp, ses and tcon cleanup in mount_put_conns()
- tcon tlink setup to mount_setup_tlink()
- remote path checking to is_path_remote()

* Implement 2 version of cifs_mount() for DFS-enabled builds and
  non-DFS-enabled builds (CONFIG_CIFS_DFS_UPCALL).

In preparation for DFS failover support.
Signed-off-by: NPaulo Alcantara <palcantara@suse.de>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

While resolving a bug with locks on samba shares found a strange behavior.
When a file locked by one node and we trying to lock it from another node
it fail with errno 5 (EIO) but in that case errno must be set to
(EACCES | EAGAIN).
This isn't happening when we try to lock file second time on same node.
In this case it returns EACCES as expected.
Also this issue not reproduces when we use SMB1 protocol (vers=1.0 in
mount options).

Further investigation showed that the mapping from status_to_posix_error
is different for SMB1 and SMB2+ implementations.
For SMB1 mapping is [NT_STATUS_LOCK_NOT_GRANTED to ERRlock]
(See fs/cifs/netmisc.c line 66)
but for SMB2+ mapping is [STATUS_LOCK_NOT_GRANTED to -EIO]
(see fs/cifs/smb2maperror.c line 383)

Quick changes in SMB2+ mapping from EIO to EACCES has fixed issue.

BUG: https://bugzilla.kernel.org/show_bug.cgi?id=201971Signed-off-by: NGeorgy A Bystrenin <gkot@altlinux.org>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>
CC: Stable <stable@vger.kernel.org>
Signed-off-by: NSteve French <stfrench@microsoft.com>

When pinning memory failed, we should return the correct error code and
rewind the SMB credits.
Reported-by: NMurphy Zhou <jencce.kernel@gmail.com>
Signed-off-by: NLong Li <longli@microsoft.com>
Cc: stable@vger.kernel.org
Cc: Murphy Zhou <jencce.kernel@gmail.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

The current code attempts to pin memory using the largest possible wsize
based on the currect SMB credits. This doesn't cause kernel oops but this
is not optimal as we may pin more pages then actually needed.

Fix this by only pinning what are needed for doing this write I/O.
Signed-off-by: NLong Li <longli@microsoft.com>
Cc: stable@vger.kernel.org
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NJoey Pabalinas <joeypabalinas@gmail.com>

RHBZ: 1021460

There is an issue where when multiple threads open/close the same directory
ntwrk_buf_start might end up being NULL, causing the call to smbCalcSize
later to oops with a NULL deref.

The real bug is why this happens and why this can become NULL for an
open cfile, which should not be allowed.
This patch tries to avoid a oops until the time when we fix the underlying
issue.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

password_with_pad is a fixed size buffer of 16 bytes, it contains a
password string, to be padded with \0 if shorter than 16 bytes
but is just truncated if longer.
It is not, and we do not depend on it to be, nul terminated.

As such, do not use strncpy() to populate this buffer since
the str* prefix suggests that this is a string, which it is not,
and it also confuses coverity causing a false warning.

Detected by CoverityScan CID#113743 ("Buffer not null terminated")
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Fixes gcc '-Wunused-but-set-variable' warning:

fs/cifs/sess.c: In function '_sess_auth_rawntlmssp_assemble_req':
fs/cifs/sess.c:1157:18: warning:
 variable 'smb_buf' set but not used [-Wunused-but-set-variable]

It never used since commit cc87c47d ("cifs: Separate rawntlmssp auth
from CIFS_SessSetup()")
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

To avoid the warning:

     warning: this statement may fall through [-Wimplicit-fallthrough=]
Signed-off-by: NGustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Reducing the number of network roundtrips improves the performance
of query xattrs
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Technically 3.02 is not the dialect name although that is more familiar to
many, so we should also accept the official dialect name (3.0.2 vs. 3.02)
in vers=
Signed-off-by: NKenneth D'souza <kdsouza@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

and convert statfs to use it.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This is not actually a bug but as Coverity points out we shouldn't
be doing an "|=" on a value which hasn't been set (although technically
it was memset to zero so isn't a bug) and so might as well change
"|=" to "=" in this line

Detected by CoverityScan, CID#728535 ("Unitialized scalar variable")
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

As Coverity points out le16_to_cpu(midEntry->Command) can not be
less than zero.

Detected by CoverityScan, CID#1438650 ("Macro compares unsigned to 0")
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

Improve performance by reducing number of network round trips
for set xattr.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Trivial fix to clean up indentation, replace spaces with tab
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Pull vfs fixes from Al Viro:
 "A couple of fixes - no common topic ;-)"

[ The aio spectre patch also came in from Jens, so now we have that
  doubly fixed .. ]

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  proc/sysctl: don't return ENOMEM on lookup when a table is unregistering
  aio: fix spectre gadget in lookup_ioctx