RHBZ 1484130

Update cifs_find_fid_lock_conflict() to recognize that
ODF locks do not conflict with eachother.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

It is not necessary to deregister a memory registration after it has been
successfully invalidated.
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

When issuing SMB1 read/write, pass the page offset to transport.
Signed-off-by: NLong Li <longli@microsoft.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

In some error conditions, resp_buftype can be passed uninitialised to
free_rsp_buf(), potentially resulting in a spurious debug message.
If resp_buftype randomly had the value 1 (CIFS_SMALL_BUFFER) then this
would log a debug message.
The rsp pointer is initialised to NULL so there is no other side-effect.

Detected by CoverityScan, CID 1438585 ("Uninitialized scalar variable")
Detected by CoverityScan, CID 1438667 ("Uninitialized scalar variable")
Detected by CoverityScan, CID 1438764 ("Uninitialized scalar variable")
Signed-off-by: NGarry McNulty <garrmcnu@gmail.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

Be able to log a ftrace message on success and/or failure of
sending a lease break response to the server.

Example output:

           TASK-PID   CPU#  ||||    TIMESTAMP  FUNCTION
             | |       |   ||||       |         |
     kworker/1:1-5681  [001] .... 11123.530457: smb3_lease_done: sid=0x291e3e0f tid=0x8ba43071 lease_key=0x1852ca0d3ecd9b55847750a86716fde lease_state=0x0
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

In network file system it is fairly easy for server and client
atime vs. mtime to get confused (and atime updated less frequently)
which we noticed broke some apps which expect atime >= mtime

Also ignore relatime mount option (rather than error on it) since
relatime is basically what some network server fs are doing
(relatime).
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

Modern servers often support 8MB as maximum i/o size, and we see some
performance benefits (my testing showed 1 to 13% on write paths,
and 1 to 3% on read paths for increasing the default to 4MB). If server
doesn't support larger i/o size, during negotiate protocol it is already
set correctly to the server's maximum if lower than 4MB.
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

As we reset credits later in the reconnect path, useful
to have optional (cifsFYI) debug message.
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

tcon->Flags is only used by SMB1 code and changing it is not permanent
(you lose the setting on tcon reconnect).

* Move the setting to superblock flags (per mount-points).
* Make automount callback exit early when flag present
* Make dfs resolving happening in mount syscall exit early if flag present
Signed-off-by: NAurelien Aptel <aaptel@suse.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Acked-by: NPavel Shilovsky <pshilov@microsoft.com>

Each time we reconnect to the same server, bump an instance
counter (and display in /proc/fs/cifs/DebugData) to make it
easier to debug.
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>
Reviewed-by: NPavel Shilovsky <pshilov@microsoft.com>

Previously reserved dpen response field changed in smb3
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

/proc/fs/cifs/Stats when CONFIG_CIFS_STATS2 is enabled logs 'slow'
responses, but depending on the server you are debugging a
one second timeout may be too fast, so allow setting it to
a larger number of seconds via new module parameter

/sys/module/cifs/parameters/slow_rsp_threshold

or via modprobe:

slow_rsp_threshold:Amount of time (in seconds) to wait before
logging that a response is delayed.
Default: 1 (if set to 0 disables msg). (uint)

Recommended values are 0 (disabled) to 32767 (9 hours) with
the default remaining as 1 second.
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

note smb3 (and common more modern servers) in the module description
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

For a network file system we generally prefer large i/o, but
if the server returns invalid file system block/sector sizes
in cifs (vers=1.0) QFSInfo then set block size to a default
of a reasonable minimum (4K).
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

Currently, "echo 0 > /proc/fs/cifs/Stats" resets all of the stats
except the session and share reconnect counts.  Fix it to
reset those as well.

CC: Stable <stable@vger.kernel.org>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NAurelien Aptel <aaptel@suse.com>

When "backup intent" is requested on the mount (e.g. backupuid or
backupgid mount options), the corresponding flag was missing from
some of the new compounding operations as well (now that
open_query_close is gone).

Related to kernel bugzilla #200953

Reported-and-tested-by: <whh@rubrik.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

So we don't overflow the io vector arrays accidentally
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Get rid of smb2_open_op_close() as all operations are now migrated
to smb2_compound_op().
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

We never pass is_falloc==true here anyway and if we ever need to support
is_falloc in the future, SMB2_set_eof is such a trivial wrapper around
send_set_info() that we can/should just create a differently named wrapper
for that new functionality.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Cuts number of network roundtrips significantly for some common syscalls
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This changes SMB2_OP_SET_EOF to use compounding in some situations.
This is part of the path based API to truncate a file.
Most of the time this will however not be invoked for SMB2 since
cifs_set_file_size() will as far as I can tell almost always just
open the file synchronously and switch to the handle based truncate
code path, thus bypassing the compounding we add here.

Rewriting cifs_set_file_size() and make that whole pile of code more
compounding friendly, and also easier to read and understand, is a
different project though and not for this patch.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This and previous patches drop the number of roundtrips we need for rmdir()
from 6 to 2.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

so that we can use these later for compounded set-info calls.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This,and previous patches, drops the number of roundtrips from five to two
for unlink()
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This with the previous patch changes mkdir() from needing 6 roundtrips to
just 3.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

This turns most open/query-info/close patterns in cifs.ko
to become compounds.

This changes stat from using 3 roundtrips to just a single one.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

When processing the mids for compounds we would only add credits based on
the last successful mid in the compound which would leak credits and
eventually triggering a re-connect.

Fix this by splitting the mid processing part into two loops instead of one
where the first loop just waits for all mids and then counts how many
credits we were granted for the whole compound.
Signed-off-by: NRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Add tracepoint to catch potential cases where a pending operation overlapping a
reconnect could fail and incorrectly refund its credits causing the client
to think it has more credits available than the server thinks it does.
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

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

fs/cifs/ioctl.c: In function 'cifs_ioctl':
fs/cifs/ioctl.c:164:23: warning:
 variable 'cifs_sb' set but not used [-Wunused-but-set-variable]
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Use kmemdup rather than duplicating its implementation
Signed-off-by: NYueHaibing <yuehaibing@huawei.com>
Signed-off-by: NSteve French <stfrench@microsoft.com>

Some servers (e.g. Azure) return "STATUS_NOT_IMPLEMENTED" rather
than "STATUS_INVALID_DEVICE_REQUEST" on query network interface
info at mount.  This shouldn't cause us to log a warning message
automatically.  Don't log this unless noisier cifsFYI is enabled.
Signed-off-by: NSteve French <stfrench@microsoft.com>
Reviewed-by: NRonnie Sahlberg <lsahlber@redhat.com>

fscache_set_key() can incur an out-of-bounds read, reported by KASAN:

 BUG: KASAN: slab-out-of-bounds in fscache_alloc_cookie+0x5b3/0x680 [fscache]
 Read of size 4 at addr ffff88084ff056d4 by task mount.nfs/32615

and also reported by syzbot at https://lkml.org/lkml/2018/7/8/236

  BUG: KASAN: slab-out-of-bounds in fscache_set_key fs/fscache/cookie.c:120 [inline]
  BUG: KASAN: slab-out-of-bounds in fscache_alloc_cookie+0x7a9/0x880 fs/fscache/cookie.c:171
  Read of size 4 at addr ffff8801d3cc8bb4 by task syz-executor907/4466

This happens for any index_key_len which is not divisible by 4 and is
larger than the size of the inline key, because the code allocates exactly
index_key_len for the key buffer, but the hashing loop is stepping through
it 4 bytes (u32) at a time in the buf[] array.

Fix this by calculating how many u32 buffers we'll need by using
DIV_ROUND_UP, and then using kcalloc() to allocate a precleared allocation
buffer to hold the index_key, then using that same count as the hashing
index limit.

Fixes: ec0328e4 ("fscache: Maintain a catalogue of allocated cookies")
Reported-by: syzbot+a95b989b2dde8e806af8@syzkaller.appspotmail.com
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The inline key in struct rxrpc_cookie is insufficiently initialized,
zeroing only 3 of the 4 slots, therefore an index_key_len between 13 and 15
bytes will end up hashing uninitialized memory because the memcpy only
partially fills the last buf[] element.

Fix this by clearing fscache_cookie objects on allocation rather than using
the slab constructor to initialise them.  We're going to pretty much fill
in the entire struct anyway, so bringing it into our dcache writably
shouldn't incur much overhead.

This removes the need to do clearance in fscache_set_key() (where we aren't
doing it correctly anyway).

Also, we don't need to set cookie->key_len in fscache_set_key() as we
already did it in the only caller, so remove that.

Fixes: ec0328e4 ("fscache: Maintain a catalogue of allocated cookies")
Reported-by: syzbot+a95b989b2dde8e806af8@syzkaller.appspotmail.com
Reported-by: NEric Sandeen <sandeen@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

the victim might've been rmdir'ed just before the lock_rename();
unlike the normal callers, we do not look the source up after the
parents are locked - we know it beforehand and just recheck that it's
still the child of what used to be its parent.  Unfortunately,
the check is too weak - we don't spot a dead directory since its
->d_parent is unchanged, dentry is positive, etc.  So we sail all
the way to ->rename(), with hosting filesystems _not_ expecting
to be asked renaming an rmdir'ed subdirectory.

The fix is easy, fortunately - the lock on parent is sufficient for
making IS_DEADDIR() on child safe.

Cc: stable@vger.kernel.org
Fixes: 9ae326a6 (CacheFiles: A cache that backs onto a mounted filesystem)
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When an invalid mount option is passed to jffs2, jffs2_parse_options()
will fail and jffs2_sb_info will be freed, but then jffs2_sb_info will
be used (use-after-free) and freeed (double-free) in jffs2_kill_sb().

Fix it by removing the buggy invocation of kfree() when getting invalid
mount options.

Fixes: 92abc475 ("jffs2: implement mount option parsing and compression overriding")
Cc: stable@kernel.org
Signed-off-by: NHou Tao <houtao1@huawei.com>
Reviewed-by: NRichard Weinberger <richard@nod.at>
Signed-off-by: NBoris Brezillon <boris.brezillon@bootlin.com>

The recent patch to fix the afs_server struct leak didn't actually fix the
bug, but rather fixed some of the symptoms.  The problem is that an
asynchronous call that holds a resource pointed to by call->reply[0] will
find the pointer cleared in the call destructor, thereby preventing the
resource from being cleaned up.

In the case of the server record leak, the afs_fs_get_capabilities()
function in devel code sets up a call with reply[0] pointing at the server
record that should be altered when the result is obtained, but this was
being cleared before the destructor was called, so the put in the
destructor does nothing and the record is leaked.

Commit f014ffb0 removed the additional ref obtained by
afs_install_server(), but the removal of this ref is actually used by the
garbage collector to mark a server record as being defunct after the record
has expired through lack of use.

The offending clearance of call->reply[0] upon completion in
afs_process_async_call() has been there from the origin of the code, but
none of the asynchronous calls actually use that pointer currently, so it
should be safe to remove (note that synchronous calls don't involve this
function).

Fix this by the following means:

 (1) Revert commit f014ffb0.

 (2) Remove the clearance of reply[0] from afs_process_async_call().

Without this, afs_manage_servers() will suffer an assertion failure if it
sees a server record that didn't get used because the usage count is not 1.

Fixes: f014ffb0 ("afs: Fix afs_server struct leak")
Fixes: 08e0e7c8 ("[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC.")
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

ubifs_assert() is not WARN_ON(), so we have to invert
the checks.
Randy faced this warning with UBIFS being a module, since
most users use UBIFS as builtin because UBIFS is the rootfs
nobody noticed so far. :-(
Including me.
Reported-by: NRandy Dunlap <rdunlap@infradead.org>
Fixes: 54169ddd ("ubifs: Turn two ubifs_assert() into a WARN_ON()")
Signed-off-by: NRichard Weinberger <richard@nod.at>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

On non-preempt kernels this loop can take a long time (more than 50 ticks)
processing through entries.

Link: http://lkml.kernel.org/r/20181010172623.57033-1-khazhy@google.comSigned-off-by: NKhazhismel Kumykov <khazhy@google.com>
Acked-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Reviewed-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix the following compile warning:

fs/ocfs2/dlmglue.c:99:30: warning: ‘lockdep_keys’ defined but not used [-Wunused-variable]
 static struct lock_class_key lockdep_keys[OCFS2_NUM_LOCK_TYPES];

Link: http://lkml.kernel.org/r/1536938148-32110-1-git-send-email-zhongjiang@huawei.comSigned-off-by: Nzhong jiang <zhongjiang@huawei.com>
Reviewed-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix a leak of afs_server structs.  The routine that installs them in the
various lookup lists and trees gets a ref on leaving the function, whether
it added the server or a server already exists.  It shouldn't increment
the refcount if it added the server.

The effect of this that "rmmod kafs" will hang waiting for the leaked
server to become unused.

Fixes: d2ddc776 ("afs: Overhaul volume and server record caching and fileserver rotation")
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>