This adds a HMAC covering the super block node and adds the logic that
decides if a filesystem shall be mounted unauthenticated or
authenticated.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
During creation of the default filesystem on an empty flash the default
LPT is created. With this patch a hash over the default LPT is
calculated which can be added to the default filesystems master node.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
The master node contains hashes over the root index node and the LPT.
This patch adds a HMAC to authenticate the master node itself.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
The LPT needs to be authenticated aswell. Since the LPT is only written
during commit it is enough to authenticate the whole LPT with a single
hash which is stored in the master node. Only the leaf nodes (pnodes)
are hashed which makes the implementation much simpler than it would be
to hash the complete LPT.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Make sure that during replay all buds can be authenticated. To do
this we calculate the hash chain until we find an authentication
node and check the HMAC in that node against the current status
of the hash chain.
After a power cut it can happen that some nodes have been written, but
not yet the authentication node for them. These nodes have to be
discarded during replay.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
To be able to authenticate the garbage collector journal head add
authentication nodes to the buds the garbage collector creates.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Nodes that are written to flash can only be authenticated through the
index after the next commit. When a journal replay is necessary the
nodes are not yet referenced by the index and thus can't be
authenticated.
This patch overcomes this situation by creating a hash over all nodes
beginning from the commit start node over the reference node(s) and
the buds themselves. From
time to time we insert authentication nodes. Authentication nodes
contain a HMAC from the current hash state, so that they can be
used to authenticate a journal replay up to the point where the
authentication node is. The hash is continued afterwards
so that theoretically we would only have to check the HMAC of
the last authentication node we find.
Overall we get this picture:
,,,,,,,,
,......,...........................................
,. CS , hash1.----. hash2.----.
,. | , . |hmac . |hmac
,. v , . v . v
,.REF#0,-> bud -> bud -> bud.-> auth -> bud -> bud.-> auth ...
,..|...,...........................................
, | ,
, | ,,,,,,,,,,,,,,,
. | hash3,----.
, | , |hmac
, v , v
, REF#1 -> bud -> bud,-> auth ...
,,,|,,,,,,,,,,,,,,,,,,
v
REF#2 -> ...
|
V
...
Note how hash3 covers CS, REF#0 and REF#1 so that it is not possible to
exchange or skip any reference nodes. Unlike the picture suggests the
auth nodes themselves are not hashed.
With this it is possible for an offline attacker to cut each journal
head or to drop the last reference node(s), but not to skip any journal
heads or to reorder any operations.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
With this patch the hashes over the index nodes stored in the tree node
cache are written to flash and are checked when read back from flash.
The hash of the root index node is stored in the master node.
During journal replay the hashes are regenerated from the read nodes
and stored in the tree node cache. This means the nodes must previously
be authenticated by other means. This is done in a later patch.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
As part of the UBIFS authentication support every branch in the index
gets a hash covering the referenced node. To make that happen the tree
node cache needs hashes over the nodes. This patch adds a hash argument
to ubifs_tnc_add() and ubifs_tnc_add_nm(). The hashes are calculated
from the callers of these functions which actually prepare the nodes.
With this patch all the leaf nodes of the index tree get hashes, but
currently nothing is done with these hashes, this is left for a later
patch.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
With authentication support some nodes (master node, super block node)
get a HMAC embedded into them. This patch adds functions to prepare and
write such a node.
The difficulty is that besides the HMAC the nodes also have a CRC which
must stay valid. This means we first have to initialize all fields in
the node, then calculate the HMAC (not covering the CRC) and finally
calculate the CRC.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
This patch adds the various helper functions needed for authentication
support. We need functions to hash nodes, to embed HMACs into a node and
to compare hashes and HMACs. Most functions first check if this
filesystem is authenticated and bail out early if not, which makes the
functions safe to be called with disabled authentication.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
When adding authentication support we will embed a HMAC into some
nodes. To prepare these nodes we have to first initialize the nodes,
then add a HMAC and finally add a CRC. To accomplish this add separate
ubifs_init_node/ubifs_crc_node functions.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
This patch adds the changes to the on disk format needed for
authentication support. We'll add:
* a HMAC covering super block node
* a HMAC covering the master node
* a hash over the root index node to the master node
* a hash over the LPT to the master node
* a flag to the filesystem flag indicating the filesystem is
authenticated
* an authentication node necessary to authenticate the nodes written
to the journal heads while they are written.
* a HMAC of a well known message to the super block node to be able
to check if the correct key is provided
And finally, not visible in this patch, nevertheless explained here:
* hashes over the referenced child nodes in each branch of a index node
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
The superblock node is read/modified/written several times throughout
the UBIFS code. Instead of reading it from the device each time just
keep a copy in memory and write back the modified copy when necessary.
This patch helps for authentication support, here we not only have to
read the superblock node, but also have to authenticate it, which
is easier if we do it once during initialization.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
write_node() is used only once and can easily be replaced with calls
to ubifs_prepare_node()/write_head() which makes the code a bit shorter.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
ubifs_lpt_lookup() starts by looking up the nth pnode in the LPT. We
already have this functionality in ubifs_pnode_lookup(). Use this
function rather than open coding its functionality.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
ubifs_lpt_lookup could be implemented using pnode_lookup. To make that
possible move pnode_lookup from lpt.c to lpt_commit.c. Rename it to
ubifs_pnode_lookup since it's now exported.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
read_znode() takes len, lnum and offs arguments which the caller all
extracts from the same struct ubifs_zbranch *. When adding authentication
support we would have to add a pointer to a hash to the arguments which
is also part of struct ubifs_zbranch. Pass the ubifs_zbranch * instead
so that we do not have to add another argument.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
try_read_node() takes len, lnum and offs arguments which the caller all
extracts from the same struct ubifs_zbranch *. When adding authentication
support we would have to add a pointer to a hash to the arguments which
is also part of struct ubifs_zbranch. Pass the ubifs_zbranch * instead
so that we do not have to add another argument.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
create_default_filesystem() allocates memory for a node, writes that
node and frees the memory directly afterwards. With this patch we
allocate memory for all nodes at the beginning of the function and
free the memory at the end. This makes it easier to implement
authentication support since with authentication support we'll need
the contents of some nodes when creating other nodes.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
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: ec0328e46d ("fscache: Maintain a catalogue of allocated cookies")
Reported-by: syzbot+a95b989b2dde8e806af8@syzkaller.appspotmail.com
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Greg 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: ec0328e46d ("fscache: Maintain a catalogue of allocated cookies")
Reported-by: syzbot+a95b989b2dde8e806af8@syzkaller.appspotmail.com
Reported-by: Eric Sandeen <sandeen@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Greg 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: 9ae326a690 (CacheFiles: A cache that backs onto a mounted filesystem)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
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 f014ffb025 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 f014ffb025.
(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: f014ffb025 ("afs: Fix afs_server struct leak")
Fixes: 08e0e7c82e ("[AF_RXRPC]: Make the in-kernel AFS filesystem use AF_RXRPC.")
Signed-off-by: David Howells <dhowells@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
* Fix a livelock in dax_layout_busy_page() present since v4.18. The
lockup triggers when truncating an actively mapped huge page out of a
mapping pinned for direct-I/O.
* Fix mprotect() clobbers of _PAGE_DEVMAP. Broken since v4.5 mprotect()
clears this flag that is needed to communicate the liveness of device
pages to the get_user_pages() path.
-----BEGIN PGP SIGNATURE-----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=VXnU
-----END PGP SIGNATURE-----
Merge tag 'libnvdimm-fixes-4.19-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm
Dan writes:
"libnvdimm/dax 4.19-rc8
* Fix a livelock in dax_layout_busy_page() present since v4.18. The
lockup triggers when truncating an actively mapped huge page out of
a mapping pinned for direct-I/O.
* Fix mprotect() clobbers of _PAGE_DEVMAP. Broken since v4.5
mprotect() clears this flag that is needed to communicate the
liveness of device pages to the get_user_pages() path."
* tag 'libnvdimm-fixes-4.19-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
mm: Preserve _PAGE_DEVMAP across mprotect() calls
filesystem-dax: Fix dax_layout_busy_page() livelock
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: Randy Dunlap <rdunlap@infradead.org>
Fixes: 54169ddd38 ("ubifs: Turn two ubifs_assert() into a WARN_ON()")
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg 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.com
Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
Acked-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg 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.com
Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg 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: d2ddc776a4 ("afs: Overhaul volume and server record caching and fileserver rotation")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It turns out that the fix in commit 6636c3cc56 is bad; the assertion
that the iomap code no longer creates buffer heads is incorrect for
filesystems that set the IOMAP_F_BUFFER_HEAD flag.
Instead, what's happening is that gfs2_iomap_begin_write treats all
files that have the jdata flag set as journaled files, which is
incorrect as long as those files are inline ("stuffed"). We're handling
stuffed files directly via the page cache, which is why we ended up with
pages without buffer heads in gfs2_page_add_databufs.
Fix this by handling stuffed journaled files correctly in
gfs2_iomap_begin_write.
This reverts commit 6636c3cc5690c11631e6366cf9a28fb99c8b25bb.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Access to the list of cells by /proc/net/afs/cells has a couple of
problems:
(1) It should be checking against SEQ_START_TOKEN for the keying the
header line.
(2) It's only holding the RCU read lock, so it can't just walk over the
list without following the proper RCU methods.
Fix these by using an hlist instead of an ordinary list and using the
appropriate accessor functions to follow it with RCU.
Since the code that adds a cell to the list must also necessarily change,
sort the list on insertion whilst we're at it.
Fixes: 989782dcdc ("afs: Overhaul cell database management")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Update for 4.19-rc7 to fix numerous file clone and deduplication issues.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=yj1P
-----END PGP SIGNATURE-----
Merge tag 'xfs-fixes-for-4.19-rc7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Dave writes:
"xfs: fixes for 4.19-rc7
Update for 4.19-rc7 to fix numerous file clone and deduplication issues."
* tag 'xfs-fixes-for-4.19-rc7' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
xfs: fix data corruption w/ unaligned reflink ranges
xfs: fix data corruption w/ unaligned dedupe ranges
xfs: update ctime and remove suid before cloning files
xfs: zero posteof blocks when cloning above eof
xfs: refactor clonerange preparation into a separate helper
Commit 64bc06bb32 broke buffered writes to journaled files (chattr
+j): we'll try to journal the buffer heads of the page being written to
in gfs2_iomap_journaled_page_done. However, the iomap code no longer
creates buffer heads, so we'll BUG() in gfs2_page_add_databufs. Fix
that by creating buffer heads ourself when needed.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
In the presence of multi-order entries the typical
pagevec_lookup_entries() pattern may loop forever:
while (index < end && pagevec_lookup_entries(&pvec, mapping, index,
min(end - index, (pgoff_t)PAGEVEC_SIZE),
indices)) {
...
for (i = 0; i < pagevec_count(&pvec); i++) {
index = indices[i];
...
}
index++; /* BUG */
}
The loop updates 'index' for each index found and then increments to the
next possible page to continue the lookup. However, if the last entry in
the pagevec is multi-order then the next possible page index is more
than 1 page away. Fix this locally for the filesystem-dax case by
checking for dax-multi-order entries. Going forward new users of
multi-order entries need to be similarly careful, or we need a generic
way to report the page increment in the radix iterator.
Fixes: 5fac7408d8 ("mm, fs, dax: handle layout changes to pinned dax...")
Cc: <stable@vger.kernel.org>
Cc: Ross Zwisler <zwisler@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
When reflinking sub-file ranges, a data corruption can occur when
the source file range includes a partial EOF block. This shares the
unknown data beyond EOF into the second file at a position inside
EOF, exposing stale data in the second file.
XFS only supports whole block sharing, but we still need to
support whole file reflink correctly. Hence if the reflink
request includes the last block of the souce file, only proceed with
the reflink operation if it lands at or past the destination file's
current EOF. If it lands within the destination file EOF, reject the
entire request with -EINVAL and make the caller go the hard way.
This avoids the data corruption vector, but also avoids disruption
of returning EINVAL to userspace for the common case of whole file
cloning.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
A deduplication data corruption is Exposed by fstests generic/505 on
XFS. It is caused by extending the block match range to include the
partial EOF block, but then allowing unknown data beyond EOF to be
considered a "match" to data in the destination file because the
comparison is only made to the end of the source file. This corrupts
the destination file when the source extent is shared with it.
XFS only supports whole block dedupe, but we still need to appear to
support whole file dedupe correctly. Hence if the dedupe request
includes the last block of the souce file, don't include it in the
actual XFS dedupe operation. If the rest of the range dedupes
successfully, then report the partial last block as deduped, too, so
that userspace sees it as a successful dedupe rather than return
EINVAL because we can't dedupe unaligned blocks.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
In dlm_init_lockres() we access and modify res->tracking and
dlm->tracking_list without holding dlm->track_lock. This can cause list
corruptions and can end up in kernel panic.
Fix this by locking res->tracking and dlm->tracking_list with
dlm->track_lock instead of dlm->spinlock.
Link: http://lkml.kernel.org/r/1529951192-4686-1-git-send-email-ashish.samant@oracle.com
Signed-off-by: Ashish Samant <ashish.samant@oracle.com>
Reviewed-by: Changwei Ge <ge.changwei@h3c.com>
Acked-by: Joseph Qi <jiangqi903@gmail.com>
Acked-by: Jun Piao <piaojun@huawei.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <ge.changwei@h3c.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently, you can use /proc/self/task/*/stack to cause a stack walk on
a task you control while it is running on another CPU. That means that
the stack can change under the stack walker. The stack walker does
have guards against going completely off the rails and into random
kernel memory, but it can interpret random data from your kernel stack
as instruction pointers and stack pointers. This can cause exposure of
kernel stack contents to userspace.
Restrict the ability to inspect kernel stacks of arbitrary tasks to root
in order to prevent a local attacker from exploiting racy stack unwinding
to leak kernel task stack contents. See the added comment for a longer
rationale.
There don't seem to be any users of this userspace API that can't
gracefully bail out if reading from the file fails. Therefore, I believe
that this change is unlikely to break things. In the case that this patch
does end up needing a revert, the next-best solution might be to fake a
single-entry stack based on wchan.
Link: http://lkml.kernel.org/r/20180927153316.200286-1-jannh@google.com
Fixes: 2ec220e27f ("proc: add /proc/*/stack")
Signed-off-by: Jann Horn <jannh@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Ken Chen <kenchen@google.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Laura Abbott <labbott@redhat.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H . Peter Anvin" <hpa@zytor.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ocfs2_duplicate_clusters_by_page() may crash if one of the extent's pages
is dirty. When a page has not been written back, it is still in dirty
state. If ocfs2_duplicate_clusters_by_page() is called against the dirty
page, the crash happens.
To fix this bug, we can just unlock the page and wait until the page until
its not dirty.
The following is the backtrace:
kernel BUG at /root/code/ocfs2/refcounttree.c:2961!
[exception RIP: ocfs2_duplicate_clusters_by_page+822]
__ocfs2_move_extent+0x80/0x450 [ocfs2]
? __ocfs2_claim_clusters+0x130/0x250 [ocfs2]
ocfs2_defrag_extent+0x5b8/0x5e0 [ocfs2]
__ocfs2_move_extents_range+0x2a4/0x470 [ocfs2]
ocfs2_move_extents+0x180/0x3b0 [ocfs2]
? ocfs2_wait_for_recovery+0x13/0x70 [ocfs2]
ocfs2_ioctl_move_extents+0x133/0x2d0 [ocfs2]
ocfs2_ioctl+0x253/0x640 [ocfs2]
do_vfs_ioctl+0x90/0x5f0
SyS_ioctl+0x74/0x80
do_syscall_64+0x74/0x140
entry_SYSCALL_64_after_hwframe+0x3d/0xa2
Once we find the page is dirty, we do not wait until it's clean, rather we
use write_one_page() to write it back
Link: http://lkml.kernel.org/r/20180829074740.9438-1-lchen@suse.com
[lchen@suse.com: update comments]
Link: http://lkml.kernel.org/r/20180830075041.14879-1-lchen@suse.com
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Larry Chen <lchen@suse.com>
Acked-by: Changwei Ge <ge.changwei@h3c.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Joseph Qi <jiangqi903@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----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=4PHD
-----END PGP SIGNATURE-----
Merge tag '4.19-rc6-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6
Steve writes:
"SMB3 fixes
four small SMB3 fixes: one for stable, the others to address a more
recent regression"
* tag '4.19-rc6-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb3: fix lease break problem introduced by compounding
cifs: only wake the thread for the very last PDU in a compound
cifs: add a warning if we try to to dequeue a deleted mid
smb2: fix missing files in root share directory listing
Before cloning into a file, update the ctime and remove sensitive
attributes like suid, just like we'd do for a regular file write.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
When we're reflinking between two files and the destination file range
is well beyond the destination file's EOF marker, zero any posteof
speculative preallocations in the destination file so that we don't
expose stale disk contents. The previous strategy of trying to clear
the preallocations does not work if the destination file has the
PREALLOC flag set.
Uncovered by shared/010.
Reported-by: Zorro Lang <zlang@redhat.com>
Bugzilla-id: https://bugzilla.kernel.org/show_bug.cgi?id=201259
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Refactor all the reflink preparation steps into a separate helper
that we'll use to land all the upcoming fixes for insufficient input
checks.
This rework also moves the invalidation of the destination range to
the prep function so that it is done before the range is remapped.
This ensures that nobody can access the data in range being remapped
until the remap is complete.
[dgc: fix xfs_reflink_remap_prep() return value and caller check to
handle vfs_clone_file_prep_inodes() returning 0 to mean "nothing to
do". ]
[dgc: make sure length changed by vfs_clone_file_prep_inodes() gets
propagated back to XFS code that does the remapping. ]
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQSQHSd0lITzzeNWNm3h3BK/laaZPAUCW7YOCQAKCRDh3BK/laaZ
PAmGAQC1/3+mkfHXqyDx+zv4f7A348ZULW26EWFWwMwlEljWowD+PlRFBgUu5v5c
194mCApZU7hl6o0u07aijjz6m81e6QE=
=dM89
-----END PGP SIGNATURE-----
Merge tag 'ovl-fixes-4.19-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs
Miklos writes:
"overlayfs fixes for 4.19-rc7
This update fixes a couple of regressions in the stacked file update
added in this cycle, as well as some older bugs uncovered by
syzkaller.
There's also one trivial naming change that touches other parts of
the fs subsystem."
* tag 'ovl-fixes-4.19-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
ovl: fix format of setxattr debug
ovl: fix access beyond unterminated strings
ovl: make symbol 'ovl_aops' static
vfs: swap names of {do,vfs}_clone_file_range()
ovl: fix freeze protection bypass in ovl_clone_file_range()
ovl: fix freeze protection bypass in ovl_write_iter()
ovl: fix memory leak on unlink of indexed file
Accumlated regression and bug fixes for 4.19-rc6, including:
o make iomap correctly mark dirty pages for sub-page block sizes
o fix regression in handling extent-to-btree format conversion errors
o fix torn log wrap detection for new logs
o various corrupt inode detection fixes
o various delalloc state fixes
o cleanup all the missed transaction cancel cases missed from changes merged
in 4.19-rc1
o fix lockdep false positive on transaction allocation
o fix locking and reference counting on buffer log items
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQIcBAABAgAGBQJbtVyHAAoJEK3oKUf0dfodc8YP/1hT+TdXZDBVPo9kXQwmrnra
8P1J8IEuGj851PwYQobUhifdDh4eQ+PpcYEIfqTSGhtjTg7cqyQOvTo4uUKHLn50
8mFXQb6JrAFYnjGjorOCde+lpoQrtj3oKASscs7RaL/W1RNffY74UGdiE0yOJnJs
9IC7TpJT4ZzAgYBgC61whfYWmszLmRuKlVZWgXrM8hkMqliHdfrXeZk7MXTeiflz
Q5+9eVnCCgTtC0TWgVAkFMvlZs7UtNMWIIn6zt+HQLB7Vms6q4ArpIT+fF45njwG
94tyTcFT0AcIJ8OIi5i85fOXcDGjTFFf554Yf80PlWGlk3SbXPHFQao/ItDA4S1Z
eCl2eurUOXlNXKuyzWoV5KjOBiiBicbOwl6eX0K996cGehhEdFOvBihvAGjJy5rm
c4plTTy6bhYKZWr3JLuaPDlNWDMr/P/aUkiq9tFXyaMmVKNeyxd6UKzIokl5uUfi
Ycik4Ik8zRgpFEUx5Clvb2W5qH9pqpR0WcFhrcj0mDnJa10TpBWesA0g2F+hM0Jc
0mSuGxfQeJk7tuVcvsBpPNzgVEPKabvnYbOlGK7HvCSOPRkg0Fhmj+iJK721ccVl
Ltiz0Y485eAGLKn9kOWKr47A5GAAFpuK29OrD2a8XizHPZ3Sr2CG5X1jI0gtMb8n
BiBAxO6dojRQATDuTI6w
=5ylA
-----END PGP SIGNATURE-----
Merge tag 'xfs-fixes-for-4.19-rc6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Dave writes:
"XFS fixes for 4.19-rc6
Accumlated regression and bug fixes for 4.19-rc6, including:
o make iomap correctly mark dirty pages for sub-page block sizes
o fix regression in handling extent-to-btree format conversion errors
o fix torn log wrap detection for new logs
o various corrupt inode detection fixes
o various delalloc state fixes
o cleanup all the missed transaction cancel cases missed from changes merged
in 4.19-rc1
o fix lockdep false positive on transaction allocation
o fix locking and reference counting on buffer log items"
* tag 'xfs-fixes-for-4.19-rc6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
xfs: fix error handling in xfs_bmap_extents_to_btree
iomap: set page dirty after partial delalloc on mkwrite
xfs: remove invalid log recovery first/last cycle check
xfs: validate inode di_forkoff
xfs: skip delalloc COW blocks in xfs_reflink_end_cow
xfs: don't treat unknown di_flags2 as corruption in scrub
xfs: remove duplicated include from alloc.c
xfs: don't bring in extents in xfs_bmap_punch_delalloc_range
xfs: fix transaction leak in xfs_reflink_allocate_cow()
xfs: avoid lockdep false positives in xfs_trans_alloc
xfs: refactor xfs_buf_log_item reference count handling
xfs: clean up xfs_trans_brelse()
xfs: don't unlock invalidated buf on aborted tx commit
xfs: remove last of unnecessary xfs_defer_cancel() callers
xfs: don't crash the vfs on a garbage inline symlink
Format has a typo: it was meant to be "%.*s", not "%*s". But at some point
callers grew nonprintable values as well, so use "%*pE" instead with a
maximized length.
Reported-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Fixes: 3a1e819b4e ("ovl: store file handle of lower inode on copy up")
Cc: <stable@vger.kernel.org> # v4.12
Al writes:
"xattrs regression fix from Andreas; sat in -next for quite a while."
* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
sysfs: Do not return POSIX ACL xattrs via listxattr
