platform/kernel/linux-amlogic.git
7 years agoxfs: disallow marking previously dirty buffers as ordered
Brian Foster [Sun, 17 Sep 2017 21:07:06 +0000 (14:07 -0700)]
xfs: disallow marking previously dirty buffers as ordered

commit a5814bceea48ee1c57c4db2bd54b0c0246daf54a upstream.

Ordered buffers are used in situations where the buffer is not
physically logged but must pass through the transaction/logging
pipeline for a particular transaction. As a result, ordered buffers
are not unpinned and written back until the transaction commits to
the log. Ordered buffers have a strict requirement that the target
buffer must not be currently dirty and resident in the log pipeline
at the time it is marked ordered. If a dirty+ordered buffer is
committed, the buffer is reinserted to the AIL but not physically
relogged at the LSN of the associated checkpoint. The buffer log
item is assigned the LSN of the latest checkpoint and the AIL
effectively releases the previously logged buffer content from the
active log before the buffer has been written back. If the tail
pushes forward and a filesystem crash occurs while in this state, an
inconsistent filesystem could result.

It is currently the caller responsibility to ensure an ordered
buffer is not already dirty from a previous modification. This is
unclear and error prone when not used in situations where it is
guaranteed a buffer has not been previously modified (such as new
metadata allocations).

To facilitate general purpose use of ordered buffers, update
xfs_trans_ordered_buf() to conditionally order the buffer based on
state of the log item and return the status of the result. If the
bli is dirty, do not order the buffer and return false. The caller
must either physically log the buffer (having acquired the
appropriate log reservation) or push it from the AIL to clean it
before it can be marked ordered in the current transaction.

Note that ordered buffers are currently only used in two situations:
1.) inode chunk allocation where previously logged buffers are not
possible and 2.) extent swap which will be updated to handle ordered
buffer failures in a separate patch.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: move bmbt owner change to last step of extent swap
Brian Foster [Sun, 17 Sep 2017 21:07:05 +0000 (14:07 -0700)]
xfs: move bmbt owner change to last step of extent swap

commit 6fb10d6d22094bc4062f92b9ccbcee2f54033d04 upstream.

The extent swap operation currently resets bmbt block owners before
the inode forks are swapped. The bmbt buffers are marked as ordered
so they do not have to be physically logged in the transaction.

This use of ordered buffers is not safe as bmbt buffers may have
been previously physically logged. The bmbt owner change algorithm
needs to be updated to physically log buffers that are already dirty
when/if they are encountered. This means that an extent swap will
eventually require multiple rolling transactions to handle large
btrees. In addition, all inode related changes must be logged before
the bmbt owner change scan begins and can roll the transaction for
the first time to preserve fs consistency via log recovery.

In preparation for such fixes to the bmbt owner change algorithm,
refactor the bmbt scan out of the extent fork swap code to the last
operation before the transaction is committed. Update
xfs_swap_extent_forks() to only set the inode log flags when an
owner change scan is necessary. Update xfs_swap_extents() to trigger
the owner change based on the inode log flags. Note that since the
owner change now occurs after the extent fork swap, the inode btrees
must be fixed up with the inode number of the current inode (similar
to log recovery).

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: skip bmbt block ino validation during owner change
Brian Foster [Sun, 17 Sep 2017 21:07:04 +0000 (14:07 -0700)]
xfs: skip bmbt block ino validation during owner change

commit 99c794c639a65cc7b74f30a674048fd100fe9ac8 upstream.

Extent swap uses xfs_btree_visit_blocks() to fix up bmbt block
owners on v5 (!rmapbt) filesystems. The bmbt scan uses
xfs_btree_lookup_get_block() to read bmbt blocks which verifies the
current owner of the block against the parent inode of the bmbt.
This works during extent swap because the bmbt owners are updated to
the opposite inode number before the inode extent forks are swapped.

The modified bmbt blocks are marked as ordered buffers which allows
everything to commit in a single transaction. If the transaction
commits to the log and the system crashes such that recovery of the
extent swap is required, log recovery restarts the bmbt scan to fix
up any bmbt blocks that may have not been written back before the
crash. The log recovery bmbt scan occurs after the inode forks have
been swapped, however. This causes the bmbt block owner verification
to fail, leads to log recovery failure and requires xfs_repair to
zap the log to recover.

Define a new invalid inode owner flag to inform the btree block
lookup mechanism that the current inode may be invalid with respect
to the current owner of the bmbt block. Set this flag on the cursor
used for change owner scans to allow this operation to work at
runtime and during log recovery.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Fixes: bb3be7e7c ("xfs: check for bogus values in btree block headers")
Cc: stable@vger.kernel.org
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: don't log dirty ranges for ordered buffers
Brian Foster [Sun, 17 Sep 2017 21:07:03 +0000 (14:07 -0700)]
xfs: don't log dirty ranges for ordered buffers

commit 8dc518dfa7dbd079581269e51074b3c55a65a880 upstream.

Ordered buffers are attached to transactions and pushed through the
logging infrastructure just like normal buffers with the exception
that they are not actually written to the log. Therefore, we don't
need to log dirty ranges of ordered buffers. xfs_trans_log_buf() is
called on ordered buffers to set up all of the dirty state on the
transaction, buffer and log item and prepare the buffer for I/O.

Now that xfs_trans_dirty_buf() is available, call it from
xfs_trans_ordered_buf() so the latter is now mutually exclusive with
xfs_trans_log_buf(). This reflects the implementation of ordered
buffers and helps eliminate confusion over the need to log ranges of
ordered buffers just to set up internal log state.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: refactor buffer logging into buffer dirtying helper
Brian Foster [Sun, 17 Sep 2017 21:07:02 +0000 (14:07 -0700)]
xfs: refactor buffer logging into buffer dirtying helper

commit 9684010d38eccda733b61106765e9357cf436f65 upstream.

xfs_trans_log_buf() is responsible for logging the dirty segments of
a buffer along with setting all of the necessary state on the
transaction, buffer, bli, etc., to ensure that the associated items
are marked as dirty and prepared for I/O. We have a couple use cases
that need to to dirty a buffer in a transaction without actually
logging dirty ranges of the buffer.  One existing use case is
ordered buffers, which are currently logged with arbitrary ranges to
accomplish this even though the content of ordered buffers is never
written to the log. Another pending use case is to relog an already
dirty buffer across rolled transactions within the deferred
operations infrastructure. This is required to prevent a held
(XFS_BLI_HOLD) buffer from pinning the tail of the log.

Refactor xfs_trans_log_buf() into a new function that contains all
of the logic responsible to dirty the transaction, lidp, buffer and
bli. This new function can be used in the future for the use cases
outlined above. This patch does not introduce functional changes.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: ordered buffer log items are never formatted
Brian Foster [Sun, 17 Sep 2017 21:07:01 +0000 (14:07 -0700)]
xfs: ordered buffer log items are never formatted

commit e9385cc6fb7edf23702de33a2dc82965d92d9392 upstream.

Ordered buffers pass through the logging infrastructure without ever
being written to the log. The way this works is that the ordered
buffer status is transferred to the log vector at commit time via
the ->iop_size() callback. In xlog_cil_insert_format_items(),
ordered log vectors bypass ->iop_format() processing altogether.

Therefore it is unnecessary for xfs_buf_item_format() to handle
ordered buffers. Remove the unnecessary logic and assert that an
ordered buffer never reaches this point.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: remove unnecessary dirty bli format check for ordered bufs
Brian Foster [Sun, 17 Sep 2017 21:07:00 +0000 (14:07 -0700)]
xfs: remove unnecessary dirty bli format check for ordered bufs

commit 6453c65d3576bc3e602abb5add15f112755c08ca upstream.

xfs_buf_item_unlock() historically checked the dirty state of the
buffer by manually checking the buffer log formats for dirty
segments. The introduction of ordered buffers invalidated this check
because ordered buffers have dirty bli's but no dirty (logged)
segments. The check was updated to accommodate ordered buffers by
looking at the bli state first and considering the blf only if the
bli is clean.

This logic is safe but unnecessary. There is no valid case where the
bli is clean yet the blf has dirty segments. The bli is set dirty
whenever the blf is logged (via xfs_trans_log_buf()) and the blf is
cleared in the only place BLI_DIRTY is cleared (xfs_trans_binval()).

Remove the conditional blf dirty checks and replace with an assert
that should catch any discrepencies between bli and blf dirty
states. Refactor the old blf dirty check into a helper function to
be used by the assert.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: open-code xfs_buf_item_dirty()
Brian Foster [Sun, 17 Sep 2017 21:06:59 +0000 (14:06 -0700)]
xfs: open-code xfs_buf_item_dirty()

commit a4f6cf6b2b6b60ec2a05a33a32e65caa4149aa2b upstream.

It checks a single flag and has one caller. It probably isn't worth
its own function.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: check for race with xfs_reclaim_inode() in xfs_ifree_cluster()
Omar Sandoval [Sun, 17 Sep 2017 21:06:58 +0000 (14:06 -0700)]
xfs: check for race with xfs_reclaim_inode() in xfs_ifree_cluster()

commit f2e9ad212def50bcf4c098c6288779dd97fff0f0 upstream.

After xfs_ifree_cluster() finds an inode in the radix tree and verifies
that the inode number is what it expected, xfs_reclaim_inode() can swoop
in and free it. xfs_ifree_cluster() will then happily continue working
on the freed inode. Most importantly, it will mark the inode stale,
which will probably be overwritten when the inode slab object is
reallocated, but if it has already been reallocated then we can end up
with an inode spuriously marked stale.

In 8a17d7ddedb4 ("xfs: mark reclaimed inodes invalid earlier") we added
a second check to xfs_iflush_cluster() to detect this race, but the
similar RCU lookup in xfs_ifree_cluster() needs the same treatment.

Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: evict all inodes involved with log redo item
Darrick J. Wong [Sun, 17 Sep 2017 21:06:57 +0000 (14:06 -0700)]
xfs: evict all inodes involved with log redo item

commit 799ea9e9c59949008770aab4e1da87f10e99dbe4 upstream.

When we introduced the bmap redo log items, we set MS_ACTIVE on the
mountpoint and XFS_IRECOVERY on the inode to prevent unlinked inodes
from being truncated prematurely during log recovery.  This also had the
effect of putting linked inodes on the lru instead of evicting them.

Unfortunately, we neglected to find all those unreferenced lru inodes
and evict them after finishing log recovery, which means that we leak
them if anything goes wrong in the rest of xfs_mountfs, because the lru
is only cleaned out on unmount.

Therefore, evict unreferenced inodes in the lru list immediately
after clearing MS_ACTIVE.

Fixes: 17c12bcd30 ("xfs: when replaying bmap operations, don't let unlinked inodes get reaped")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Cc: viro@ZenIV.linux.org.uk
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: stop searching for free slots in an inode chunk when there are none
Carlos Maiolino [Sun, 17 Sep 2017 21:06:56 +0000 (14:06 -0700)]
xfs: stop searching for free slots in an inode chunk when there are none

commit 2d32311cf19bfb8c1d2b4601974ddd951f9cfd0b upstream.

In a filesystem without finobt, the Space manager selects an AG to alloc a new
inode, where xfs_dialloc_ag_inobt() will search the AG for the free slot chunk.

When the new inode is in the same AG as its parent, the btree will be searched
starting on the parent's record, and then retried from the top if no slot is
available beyond the parent's record.

To exit this loop though, xfs_dialloc_ag_inobt() relies on the fact that the
btree must have a free slot available, once its callers relied on the
agi->freecount when deciding how/where to allocate this new inode.

In the case when the agi->freecount is corrupted, showing available inodes in an
AG, when in fact there is none, this becomes an infinite loop.

Add a way to stop the loop when a free slot is not found in the btree, making
the function to fall into the whole AG scan which will then, be able to detect
the corruption and shut the filesystem down.

As pointed by Brian, this might impact performance, giving the fact we
don't reset the search distance anymore when we reach the end of the
tree, giving it fewer tries before falling back to the whole AG search, but
it will only affect searches that start within 10 records to the end of the tree.

Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: add log recovery tracepoint for head/tail
Brian Foster [Sun, 17 Sep 2017 21:06:55 +0000 (14:06 -0700)]
xfs: add log recovery tracepoint for head/tail

commit e67d3d4246e5fbb0c7c700426d11241ca9c6f473 upstream.

Torn write detection and tail overwrite detection can shift the log
head and tail respectively in the event of CRC mismatch or
corruption errors. Add a high-level log recovery tracepoint to dump
the final log head/tail and make those values easily attainable in
debug/diagnostic situations.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: handle -EFSCORRUPTED during head/tail verification
Brian Foster [Sun, 17 Sep 2017 21:06:54 +0000 (14:06 -0700)]
xfs: handle -EFSCORRUPTED during head/tail verification

commit a4c9b34d6a17081005ec459b57b8effc08f4c731 upstream.

Torn write and tail overwrite detection both trigger only on
-EFSBADCRC errors. While this is the most likely failure scenario
for each condition, -EFSCORRUPTED is still possible in certain cases
depending on what ends up on disk when a torn write or partial tail
overwrite occurs. For example, an invalid log record h_len can lead
to an -EFSCORRUPTED error when running the log recovery CRC pass.

Therefore, update log head and tail verification to trigger the
associated head/tail fixups in the event of -EFSCORRUPTED errors
along with -EFSBADCRC. Also, -EFSCORRUPTED can currently be returned
from xlog_do_recovery_pass() before rhead_blk is initialized if the
first record encountered happens to be corrupted. This leads to an
incorrect 'first_bad' return value. Initialize rhead_blk earlier in
the function to address that problem as well.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: fix log recovery corruption error due to tail overwrite
Brian Foster [Sun, 17 Sep 2017 21:06:53 +0000 (14:06 -0700)]
xfs: fix log recovery corruption error due to tail overwrite

commit 4a4f66eac4681378996a1837ad1ffec3a2e2981f upstream.

If we consider the case where the tail (T) of the log is pinned long
enough for the head (H) to push and block behind the tail, we can
end up blocked in the following state without enough free space (f)
in the log to satisfy a transaction reservation:

0 phys. log N
[-------HffT---H'--T'---]

The last good record in the log (before H) refers to T. The tail
eventually pushes forward (T') leaving more free space in the log
for writes to H. At this point, suppose space frees up in the log
for the maximum of 8 in-core log buffers to start flushing out to
the log. If this pushes the head from H to H', these next writes
overwrite the previous tail T. This is safe because the items logged
from T to T' have been written back and removed from the AIL.

If the next log writes (H -> H') happen to fail and result in
partial records in the log, the filesystem shuts down having
overwritten T with invalid data. Log recovery correctly locates H on
the subsequent mount, but H still refers to the now corrupted tail
T. This results in log corruption errors and recovery failure.

Since the tail overwrite results from otherwise correct runtime
behavior, it is up to log recovery to try and deal with this
situation. Update log recovery tail verification to run a CRC pass
from the first record past the tail to the head. This facilitates
error detection at T and moves the recovery tail to the first good
record past H' (similar to truncating the head on torn write
detection). If corruption is detected beyond the range possibly
affected by the max number of iclogs, the log is legitimately
corrupted and log recovery failure is expected.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: always verify the log tail during recovery
Brian Foster [Sun, 17 Sep 2017 21:06:52 +0000 (14:06 -0700)]
xfs: always verify the log tail during recovery

commit 5297ac1f6d7cbf45464a49b9558831f271dfc559 upstream.

Log tail verification currently only occurs when torn writes are
detected at the head of the log. This was introduced because a
change in the head block due to torn writes can lead to a change in
the tail block (each log record header references the current tail)
and the tail block should be verified before log recovery proceeds.

Tail corruption is possible outside of torn write scenarios,
however. For example, partial log writes can be detected and cleared
during the initial head/tail block discovery process. If the partial
write coincides with a tail overwrite, the log tail is corrupted and
recovery fails.

To facilitate correct handling of log tail overwites, update log
recovery to always perform tail verification. This is necessary to
detect potential tail overwrite conditions when torn writes may not
have occurred. This changes normal (i.e., no torn writes) recovery
behavior slightly to detect and return CRC related errors near the
tail before actual recovery starts.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: fix recovery failure when log record header wraps log end
Brian Foster [Sun, 17 Sep 2017 21:06:51 +0000 (14:06 -0700)]
xfs: fix recovery failure when log record header wraps log end

commit 284f1c2c9bebf871861184b0e2c40fa921dd380b upstream.

The high-level log recovery algorithm consists of two loops that
walk the physical log and process log records from the tail to the
head. The first loop handles the case where the tail is beyond the
head and processes records up to the end of the physical log. The
subsequent loop processes records from the beginning of the physical
log to the head.

Because log records can wrap around the end of the physical log, the
first loop mentioned above must handle this case appropriately.
Records are processed from in-core buffers, which means that this
algorithm must split the reads of such records into two partial
I/Os: 1.) from the beginning of the record to the end of the log and
2.) from the beginning of the log to the end of the record. This is
further complicated by the fact that the log record header and log
record data are read into independent buffers.

The current handling of each buffer correctly splits the reads when
either the header or data starts before the end of the log and wraps
around the end. The data read does not correctly handle the case
where the prior header read wrapped or ends on the physical log end
boundary. blk_no is incremented to or beyond the log end after the
header read to point to the record data, but the split data read
logic triggers, attempts to read from an invalid log block and
ultimately causes log recovery to fail. This can be reproduced
fairly reliably via xfstests tests generic/047 and generic/388 with
large iclog sizes (256k) and small (10M) logs.

If the record header read has pushed beyond the end of the physical
log, the subsequent data read is actually contiguous. Update the
data read logic to detect the case where blk_no has wrapped, mod it
against the log size to read from the correct address and issue one
contiguous read for the log data buffer. The log record is processed
as normal from the buffer(s), the loop exits after the current
iteration and the subsequent loop picks up with the first new record
after the start of the log.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: Properly retry failed inode items in case of error during buffer writeback
Carlos Maiolino [Sun, 17 Sep 2017 21:06:50 +0000 (14:06 -0700)]
xfs: Properly retry failed inode items in case of error during buffer writeback

commit d3a304b6292168b83b45d624784f973fdc1ca674 upstream.

When a buffer has been failed during writeback, the inode items into it
are kept flush locked, and are never resubmitted due the flush lock, so,
if any buffer fails to be written, the items in AIL are never written to
disk and never unlocked.

This causes unmount operation to hang due these items flush locked in AIL,
but this also causes the items in AIL to never be written back, even when
the IO device comes back to normal.

I've been testing this patch with a DM-thin device, creating a
filesystem larger than the real device.

When writing enough data to fill the DM-thin device, XFS receives ENOSPC
errors from the device, and keep spinning on xfsaild (when 'retry
forever' configuration is set).

At this point, the filesystem can not be unmounted because of the flush locked
items in AIL, but worse, the items in AIL are never retried at all
(once xfs_inode_item_push() will skip the items that are flush locked),
even if the underlying DM-thin device is expanded to the proper size.

This patch fixes both cases, retrying any item that has been failed
previously, using the infra-structure provided by the previous patch.

Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: Add infrastructure needed for error propagation during buffer IO failure
Carlos Maiolino [Sun, 17 Sep 2017 21:06:49 +0000 (14:06 -0700)]
xfs: Add infrastructure needed for error propagation during buffer IO failure

commit 0b80ae6ed13169bd3a244e71169f2cc020b0c57a upstream.

With the current code, XFS never re-submit a failed buffer for IO,
because the failed item in the buffer is kept in the flush locked state
forever.

To be able to resubmit an log item for IO, we need a way to mark an item
as failed, if, for any reason the buffer which the item belonged to
failed during writeback.

Add a new log item callback to be used after an IO completion failure
and make the needed clean ups.

Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: remove xfs_trans_ail_delete_bulk
Christoph Hellwig [Sun, 17 Sep 2017 21:06:48 +0000 (14:06 -0700)]
xfs: remove xfs_trans_ail_delete_bulk

commit 27af1bbf524459962d1477a38ac6e0b7f79aaecc upstream.

xfs_iflush_done uses an on-stack variable length array to pass the log
items to be deleted to xfs_trans_ail_delete_bulk.  On-stack VLAs are a
nasty gcc extension that can lead to unbounded stack allocations, but
fortunately we can easily avoid them by simply open coding
xfs_trans_ail_delete_bulk in xfs_iflush_done, which is the only caller
of it except for the single-item xfs_trans_ail_delete.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: toggle readonly state around xfs_log_mount_finish
Eric Sandeen [Sun, 17 Sep 2017 21:06:47 +0000 (14:06 -0700)]
xfs: toggle readonly state around xfs_log_mount_finish

commit 6f4a1eefdd0ad4561543270a7fceadabcca075dd upstream.

When we do log recovery on a readonly mount, unlinked inode
processing does not happen due to the readonly checks in
xfs_inactive(), which are trying to prevent any I/O on a
readonly mount.

This is misguided - we do I/O on readonly mounts all the time,
for consistency; for example, log recovery.  So do the same
RDONLY flag twiddling around xfs_log_mount_finish() as we
do around xfs_log_mount(), for the same reason.

This all cries out for a big rework but for now this is a
simple fix to an obvious problem.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: write unmount record for ro mounts
Eric Sandeen [Sun, 17 Sep 2017 21:06:46 +0000 (14:06 -0700)]
xfs: write unmount record for ro mounts

commit 757a69ef6cf2bf839bd4088e5609ddddd663b0c4 upstream.

There are dueling comments in the xfs code about intent
for log writes when unmounting a readonly filesystem.

In xfs_mountfs, we see the intent:

/*
 * Now the log is fully replayed, we can transition to full read-only
 * mode for read-only mounts. This will sync all the metadata and clean
 * the log so that the recovery we just performed does not have to be
 * replayed again on the next mount.
 */

and it calls xfs_quiesce_attr(), but by the time we get to
xfs_log_unmount_write(), it returns early for a RDONLY mount:

 * Don't write out unmount record on read-only mounts.

Because of this, sequential ro mounts of a filesystem with
a dirty log will replay the log each time, which seems odd.

Fix this by writing an unmount record even for RO mounts, as long
as norecovery wasn't specified (don't write a clean log record
if a dirty log may still be there!) and the log device is
writable.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoiomap: fix integer truncation issues in the zeroing and dirtying helpers
Christoph Hellwig [Sun, 17 Sep 2017 21:06:45 +0000 (14:06 -0700)]
iomap: fix integer truncation issues in the zeroing and dirtying helpers

commit e28ae8e428fefe2facd72cea9f29906ecb9c861d upstream.

Fix the min_t calls in the zeroing and dirtying helpers to perform the
comparisms on 64-bit types, which prevents them from incorrectly
being truncated, and larger zeroing operations being stuck in a never
ending loop.

Special thanks to Markus Stockhausen for spotting the bug.

Reported-by: Paul Menzel <pmenzel@molgen.mpg.de>
Tested-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: don't leak quotacheck dquots when cow recovery
Darrick J. Wong [Sun, 17 Sep 2017 21:06:44 +0000 (14:06 -0700)]
xfs: don't leak quotacheck dquots when cow recovery

commit 77aff8c76425c8f49b50d0b9009915066739e7d2 upstream.

If we fail a mount on account of cow recovery errors, it's possible that
a previous quotacheck left some dquots in memory.  The bailout clause of
xfs_mountfs forgets to purge these, and so we leak them.  Fix that.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: clear MS_ACTIVE after finishing log recovery
Darrick J. Wong [Sun, 17 Sep 2017 21:06:43 +0000 (14:06 -0700)]
xfs: clear MS_ACTIVE after finishing log recovery

commit 8204f8ddaafafcae074746fcf2a05a45e6827603 upstream.

Way back when we established inode block-map redo log items, it was
discovered that we needed to prevent the VFS from evicting inodes during
log recovery because any given inode might be have bmap redo items to
replay even if the inode has no link count and is ultimately deleted,
and any eviction of an unlinked inode causes the inode to be truncated
and freed too early.

To make this possible, we set MS_ACTIVE so that inodes would not be torn
down immediately upon release.  Unfortunately, this also results in the
quota inodes not being released at all if a later part of the mount
process should fail, because we never reclaim the inodes.  So, set
MS_ACTIVE right before we do the last part of log recovery and clear it
immediately after we finish the log recovery so that everything
will be torn down properly if we abort the mount.

Fixes: 17c12bcd30 ("xfs: when replaying bmap operations, don't let unlinked inodes get reaped")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: fix inobt inode allocation search optimization
Omar Sandoval [Sun, 17 Sep 2017 21:06:42 +0000 (14:06 -0700)]
xfs: fix inobt inode allocation search optimization

commit c44245b3d5435f533ca8346ece65918f84c057f9 upstream.

When we try to allocate a free inode by searching the inobt, we try to
find the inode nearest the parent inode by searching chunks both left
and right of the chunk containing the parent. As an optimization, we
cache the leftmost and rightmost records that we previously searched; if
we do another allocation with the same parent inode, we'll pick up the
search where it last left off.

There's a bug in the case where we found a free inode to the left of the
parent's chunk: we need to update the cached left and right records, but
because we already reassigned the right record to point to the left, we
end up assigning the left record to both the cached left and right
records.

This isn't a correctness problem strictly, but it can result in the next
allocation rechecking chunks unnecessarily or allocating inodes further
away from the parent than it needs to. Fix it by swapping the record
pointer after we update the cached left and right records.

Fixes: bd169565993b ("xfs: speed up free inode search")
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: Fix per-inode DAX flag inheritance
Lukas Czerner [Sun, 17 Sep 2017 21:06:41 +0000 (14:06 -0700)]
xfs: Fix per-inode DAX flag inheritance

commit 56bdf855e676f1f2ed7033f288f57dfd315725ba upstream.

According to the commit that implemented per-inode DAX flag:
commit 58f88ca2df72 ("xfs: introduce per-inode DAX enablement")
the flag is supposed to act as "inherit flag".

Currently this only works in the situations where parent directory
already has a flag in di_flags set, otherwise inheritance does not
work. This is because setting the XFS_DIFLAG2_DAX flag is done in a
wrong branch designated for di_flags, not di_flags2.

Fix this by moving the code to branch designated for setting di_flags2,
which does test for flags in di_flags2.

Fixes: 58f88ca2df72 ("xfs: introduce per-inode DAX enablement")
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: fix multi-AG deadlock in xfs_bunmapi
Christoph Hellwig [Sun, 17 Sep 2017 21:06:40 +0000 (14:06 -0700)]
xfs: fix multi-AG deadlock in xfs_bunmapi

commit 5b094d6dac0451ad89b1dc088395c7b399b7e9e8 upstream.

Just like in the allocator we must avoid touching multiple AGs out of
order when freeing blocks, as freeing still locks the AGF and can cause
the same AB-BA deadlocks as in the allocation path.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reported-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: fix quotacheck dquot id overflow infinite loop
Brian Foster [Sun, 17 Sep 2017 21:06:39 +0000 (14:06 -0700)]
xfs: fix quotacheck dquot id overflow infinite loop

commit cfaf2d034360166e569a4929dd83ae9698bed856 upstream.

If a dquot has an id of U32_MAX, the next lookup index increment
overflows the uint32_t back to 0. This starts the lookup sequence
over from the beginning, repeats indefinitely and results in a
livelock.

Update xfs_qm_dquot_walk() to explicitly check for the lookup
overflow and exit the loop.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: check _alloc_read_agf buffer pointer before using
Darrick J. Wong [Sun, 17 Sep 2017 21:06:38 +0000 (14:06 -0700)]
xfs: check _alloc_read_agf buffer pointer before using

commit 10479e2dea83d4c421ad05dfc55d918aa8dfc0cd upstream.

In some circumstances, _alloc_read_agf can return an error code of zero
but also a null AGF buffer pointer.  Check for this and jump out.

Fixes-coverity-id: 1415250
Fixes-coverity-id: 1415320
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: set firstfsb to NULLFSBLOCK before feeding it to _bmapi_write
Darrick J. Wong [Sun, 17 Sep 2017 21:06:37 +0000 (14:06 -0700)]
xfs: set firstfsb to NULLFSBLOCK before feeding it to _bmapi_write

commit 4c1a67bd3606540b9b42caff34a1d5cd94b1cf65 upstream.

We must initialize the firstfsb parameter to _bmapi_write so that it
doesn't incorrectly treat stack garbage as a restriction on which AGs
it can search for free space.

Fixes-coverity-id: 1402025
Fixes-coverity-id: 1415167
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: check _btree_check_block value
Darrick J. Wong [Sun, 17 Sep 2017 21:06:36 +0000 (14:06 -0700)]
xfs: check _btree_check_block value

commit 1e86eabe73b73c82e1110c746ed3ec6d5e1c0a0d upstream.

Check the _btree_check_block return value for the firstrec and lastrec
functions, since we have the ability to signal that the repositioning
did not succeed.

Fixes-coverity-id: 114067
Fixes-coverity-id: 114068
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: don't crash on unexpected holes in dir/attr btrees
Darrick J. Wong [Sun, 17 Sep 2017 21:06:35 +0000 (14:06 -0700)]
xfs: don't crash on unexpected holes in dir/attr btrees

commit cd87d867920155911d0d2e6485b769d853547750 upstream.

In quite a few places we call xfs_da_read_buf with a mappedbno that we
don't control, then assume that the function passes back either an error
code or a buffer pointer.  Unfortunately, if mappedbno == -2 and bno
maps to a hole, we get a return code of zero and a NULL buffer, which
means that we crash if we actually try to use that buffer pointer.  This
happens immediately when we set the buffer type for transaction context.

Therefore, check that we have no error code and a non-NULL bp before
trying to use bp.  This patch is a follow-up to an incomplete fix in
96a3aefb8ffde231 ("xfs: don't crash if reading a directory results in an
unexpected hole").

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: free cowblocks and retry on buffered write ENOSPC
Brian Foster [Sun, 17 Sep 2017 21:06:34 +0000 (14:06 -0700)]
xfs: free cowblocks and retry on buffered write ENOSPC

commit cf2cb7845d6e101cb17bd62f8aa08cd514fc8988 upstream.

XFS runs an eofblocks reclaim scan before returning an ENOSPC error to
userspace for buffered writes. This facilitates aggressive speculative
preallocation without causing user visible side effects such as
premature ENOSPC.

Run a cowblocks scan in the same situation to reclaim lingering COW fork
preallocation throughout the filesystem.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: free uncommitted transactions during log recovery
Brian Foster [Sun, 17 Sep 2017 21:06:33 +0000 (14:06 -0700)]
xfs: free uncommitted transactions during log recovery

commit 39775431f82f890f4aaa08860a30883d081bffc7 upstream.

Log recovery allocates in-core transaction and member item data
structures on-demand as it processes the on-disk log. Transactions
are allocated on first encounter on-disk and stored in a hash table
structure where they are easily accessible for subsequent lookups.
Transaction items are also allocated on demand and are attached to
the associated transactions.

When a commit record is encountered in the log, the transaction is
committed to the fs and the in-core structures are freed. If a
filesystem crashes or shuts down before all in-core log buffers are
flushed to the log, however, not all transactions may have commit
records in the log. As expected, the modifications in such an
incomplete transaction are not replayed to the fs. The in-core data
structures for the partial transaction are never freed, however,
resulting in a memory leak.

Update xlog_do_recovery_pass() to first correctly initialize the
hash table array so empty lists can be distinguished from populated
lists on function exit. Update xlog_recover_free_trans() to always
remove the transaction from the list prior to freeing the associated
memory. Finally, walk the hash table of transaction lists as the
last step before it goes out of scope and free any transactions that
may remain on the lists. This prevents a memory leak of partial
transactions in the log.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: don't allow bmap on rt files
Darrick J. Wong [Mon, 19 Jun 2017 20:19:08 +0000 (13:19 -0700)]
xfs: don't allow bmap on rt files

commit 61d819e7bcb7f33da710bf3f5dcb2bcf1e48203c upstream.

bmap returns a dumb LBA address but not the block device that goes with
that LBA.  Swapfiles don't care about this and will blindly assume that
the data volume is the correct blockdev, which is totally bogus for
files on the rt subvolume.  This results in the swap code doing IOs to
arbitrary locations on the data device(!) if the passed in mapping is a
realtime file, so just turn off bmap for rt files.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: remove bli from AIL before release on transaction abort
Brian Foster [Thu, 15 Jun 2017 04:35:35 +0000 (21:35 -0700)]
xfs: remove bli from AIL before release on transaction abort

commit 3d4b4a3e30ae7a949c31e1e10268a3da4723d290 upstream.

When a buffer is modified, logged and committed, it ultimately ends
up sitting on the AIL with a dirty bli waiting for metadata
writeback. If another transaction locks and invalidates the buffer
(freeing an inode chunk, for example) in the meantime, the bli is
flagged as stale, the dirty state is cleared and the bli remains in
the AIL.

If a shutdown occurs before the transaction that has invalidated the
buffer is committed, the transaction is ultimately aborted. The log
items are flagged as such and ->iop_unlock() handles the aborted
items. Because the bli is clean (due to the invalidation),
->iop_unlock() unconditionally releases it. The log item may still
reside in the AIL, however, which means the I/O completion handler
may still run and attempt to access it. This results in assert
failure due to the release of the bli while still present in the AIL
and a subsequent NULL dereference and panic in the buffer I/O
completion handling. This can be reproduced by running generic/388
in repetition.

To avoid this problem, update xfs_buf_item_unlock() to first check
whether the bli is aborted and if so, remove it from the AIL before
it is released. This ensures that the bli is no longer accessed
during the shutdown sequence after it has been freed.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: release bli from transaction properly on fs shutdown
Brian Foster [Thu, 15 Jun 2017 04:35:35 +0000 (21:35 -0700)]
xfs: release bli from transaction properly on fs shutdown

commit 79e641ce29cfae5b8fc55fb77ac62d11d2d849c0 upstream.

If a filesystem shutdown occurs with a buffer log item in the CIL
and a log force occurs, the ->iop_unpin() handler is generally
expected to tear down the bli properly. This entails freeing the bli
memory and releasing the associated hold on the buffer so it can be
released and the filesystem unmounted.

If this sequence occurs while ->bli_refcount is elevated (i.e.,
another transaction is open and attempting to modify the buffer),
however, ->iop_unpin() may not be responsible for releasing the bli.
Instead, the transaction may release the final ->bli_refcount
reference and thus xfs_trans_brelse() is responsible for tearing
down the bli.

While xfs_trans_brelse() does drop the reference count, it only
attempts to release the bli if it is clean (i.e., not in the
CIL/AIL). If the filesystem is shutdown and the bli is sitting dirty
in the CIL as noted above, this ends up skipping the last
opportunity to release the bli. In turn, this leaves the hold on the
buffer and causes an unmount hang. This can be reproduced by running
generic/388 in repetition.

Update xfs_trans_brelse() to handle this shutdown corner case
correctly. If the final bli reference is dropped and the filesystem
is shutdown, remove the bli from the AIL (if necessary) and release
the bli to drop the buffer hold and ensure an unmount does not hang.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: try to avoid blowing out the transaction reservation when bunmaping a shared...
Darrick J. Wong [Thu, 15 Jun 2017 04:25:57 +0000 (21:25 -0700)]
xfs: try to avoid blowing out the transaction reservation when bunmaping a shared extent

commit e1a4e37cc7b665b6804fba812aca2f4d7402c249 upstream.

In a pathological scenario where we are trying to bunmapi a single
extent in which every other block is shared, it's possible that trying
to unmap the entire large extent in a single transaction can generate so
many EFIs that we overflow the transaction reservation.

Therefore, use a heuristic to guess at the number of blocks we can
safely unmap from a reflink file's data fork in an single transaction.
This should prevent problems such as the log head slamming into the tail
and ASSERTs that trigger because we've exceeded the transaction
reservation.

Note that since bunmapi can fail to unmap the entire range, we must also
teach the deferred unmap code to roll into a new transaction whenever we
get low on reservation.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[hch: random edits, all bugs are my fault]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: push buffer of flush locked dquot to avoid quotacheck deadlock
Brian Foster [Thu, 15 Jun 2017 04:21:45 +0000 (21:21 -0700)]
xfs: push buffer of flush locked dquot to avoid quotacheck deadlock

commit 7912e7fef2aebe577f0b46d3cba261f2783c5695 upstream.

Reclaim during quotacheck can lead to deadlocks on the dquot flush
lock:

 - Quotacheck populates a local delwri queue with the physical dquot
   buffers.
 - Quotacheck performs the xfs_qm_dqusage_adjust() bulkstat and
   dirties all of the dquots.
 - Reclaim kicks in and attempts to flush a dquot whose buffer is
   already queud on the quotacheck queue. The flush succeeds but
   queueing to the reclaim delwri queue fails as the backing buffer is
   already queued. The flush unlock is now deferred to I/O completion
   of the buffer from the quotacheck queue.
 - The dqadjust bulkstat continues and dirties the recently flushed
   dquot once again.
 - Quotacheck proceeds to the xfs_qm_flush_one() walk which requires
   the flush lock to update the backing buffers with the in-core
   recalculated values. It deadlocks on the redirtied dquot as the
   flush lock was already acquired by reclaim, but the buffer resides
   on the local delwri queue which isn't submitted until the end of
   quotacheck.

This is reproduced by running quotacheck on a filesystem with a
couple million inodes in low memory (512MB-1GB) situations. This is
a regression as of commit 43ff2122e6 ("xfs: on-stack delayed write
buffer lists"), which removed a trylock and buffer I/O submission
from the quotacheck dquot flush sequence.

Quotacheck first resets and collects the physical dquot buffers in a
delwri queue. Then, it traverses the filesystem inodes via bulkstat,
updates the in-core dquots, flushes the corrected dquots to the
backing buffers and finally submits the delwri queue for I/O. Since
the backing buffers are queued across the entire quotacheck
operation, dquot reclaim cannot possibly complete a dquot flush
before quotacheck completes.

Therefore, quotacheck must submit the buffer for I/O in order to
cycle the flush lock and flush the dirty in-core dquot to the
buffer. Add a delwri queue buffer push mechanism to submit an
individual buffer for I/O without losing the delwri queue status and
use it from quotacheck to avoid the deadlock. This restores
quotacheck behavior to as before the regression was introduced.

Reported-by: Martin Svec <martin.svec@zoner.cz>
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: fix spurious spin_is_locked() assert failures on non-smp kernels
Brian Foster [Thu, 8 Jun 2017 15:23:07 +0000 (08:23 -0700)]
xfs: fix spurious spin_is_locked() assert failures on non-smp kernels

commit 95989c46d2a156365867b1d795fdefce71bce378 upstream.

The 0-day kernel test robot reports assertion failures on
!CONFIG_SMP kernels due to failed spin_is_locked() checks. As it
turns out, spin_is_locked() is hardcoded to return zero on
!CONFIG_SMP kernels and so this function cannot be relied on to
verify spinlock state in this configuration.

To avoid this problem, replace the associated asserts with lockdep
variants that do the right thing regardless of kernel configuration.
Drop the one assert that checks for an unlocked lock as there is no
suitable lockdep variant for that case. This moves the spinlock
checks from XFS debug code to lockdep, but generally provides the
same level of protection.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoxfs: Move handling of missing page into one place in xfs_find_get_desired_pgoff()
Jan Kara [Thu, 18 May 2017 23:36:24 +0000 (16:36 -0700)]
xfs: Move handling of missing page into one place in xfs_find_get_desired_pgoff()

commit a54fba8f5a0dc36161cacdf2aa90f007f702ec1a upstream.

Currently several places in xfs_find_get_desired_pgoff() handle the case
of a missing page. Make them all handled in one place after the loop has
terminated.

Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agox86/switch_to/64: Rewrite FS/GS switching yet again to fix AMD CPUs
Andy Lutomirski [Tue, 1 Aug 2017 14:11:37 +0000 (07:11 -0700)]
x86/switch_to/64: Rewrite FS/GS switching yet again to fix AMD CPUs

commit e137a4d8f4dd2e277e355495b6b2cb241a8693c3 upstream.

Switching FS and GS is a mess, and the current code is still subtly
wrong: it assumes that "Loading a nonzero value into FS sets the
index and base", which is false on AMD CPUs if the value being
loaded is 1, 2, or 3.

(The current code came from commit 3e2b68d752c9 ("x86/asm,
sched/x86: Rewrite the FS and GS context switch code"), which made
it better but didn't fully fix it.)

Rewrite it to be much simpler and more obviously correct.  This
should fix it fully on AMD CPUs and shouldn't adversely affect
performance.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Borislav Petkov <bpetkov@suse.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Chang Seok <chang.seok.bae@intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agox86/fsgsbase/64: Report FSBASE and GSBASE correctly in core dumps
Andy Lutomirski [Tue, 1 Aug 2017 14:11:35 +0000 (07:11 -0700)]
x86/fsgsbase/64: Report FSBASE and GSBASE correctly in core dumps

commit 9584d98bed7a7a904d0702ad06bbcc94703cb5b4 upstream.

In ELF_COPY_CORE_REGS, we're copying from the current task, so
accessing thread.fsbase and thread.gsbase makes no sense.  Just read
the values from the CPU registers.

In practice, the old code would have been correct most of the time
simply because thread.fsbase and thread.gsbase usually matched the
CPU registers.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Borislav Petkov <bpetkov@suse.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Chang Seok <chang.seok.bae@intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agox86/fsgsbase/64: Fully initialize FS and GS state in start_thread_common
Andy Lutomirski [Tue, 1 Aug 2017 14:11:34 +0000 (07:11 -0700)]
x86/fsgsbase/64: Fully initialize FS and GS state in start_thread_common

commit 767d035d838f4fd6b5a5bbd7a3f6d293b7f65a49 upstream.

execve used to leak FSBASE and GSBASE on AMD CPUs.  Fix it.

The security impact of this bug is small but not quite zero -- it
could weaken ASLR when a privileged task execs a less privileged
program, but only if program changed bitness across the exec, or the
child binary was highly unusual or actively malicious.  A child
program that was compromised after the exec would not have access to
the leaked base.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Borislav Petkov <bpetkov@suse.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Chang Seok <chang.seok.bae@intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agof2fs: check hot_data for roll-forward recovery
Jaegeuk Kim [Sun, 13 Aug 2017 04:33:23 +0000 (21:33 -0700)]
f2fs: check hot_data for roll-forward recovery

commit 125c9fb1ccb53eb2ea9380df40f3c743f3fb2fed upstream.

We need to check HOT_DATA to truncate any previous data block when doing
roll-forward recovery.

Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agof2fs: let fill_super handle roll-forward errors
Jaegeuk Kim [Fri, 11 Aug 2017 00:35:04 +0000 (17:35 -0700)]
f2fs: let fill_super handle roll-forward errors

commit afd2b4da40b3b567ef8d8e6881479345a2312a03 upstream.

If we set CP_ERROR_FLAG in roll-forward error, f2fs is no longer to proceed
any IOs due to f2fs_cp_error(). But, for example, if some stale data is involved
on roll-forward process, we're able to get -ENOENT, getting fs stuck.
If we get any error, let fill_super set SBI_NEED_FSCK and try to recover back
to stable point.

Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoip_tunnel: fix setting ttl and tos value in collect_md mode
Haishuang Yan [Thu, 7 Sep 2017 06:08:34 +0000 (14:08 +0800)]
ip_tunnel: fix setting ttl and tos value in collect_md mode

[ Upstream commit 0f693f1995cf002432b70f43ce73f79bf8d0b6c9 ]

ttl and tos variables are declared and assigned, but are not used in
iptunnel_xmit() function.

Fixes: cfc7381b3002 ("ip_tunnel: add collect_md mode to IPIP tunnel")
Cc: Alexei Starovoitov <ast@fb.com>
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agosctp: fix missing wake ups in some situations
Marcelo Ricardo Leitner [Fri, 8 Sep 2017 14:35:21 +0000 (11:35 -0300)]
sctp: fix missing wake ups in some situations

[ Upstream commit 7906b00f5cd1cd484fced7fcda892176e3202c8a ]

Commit fb586f25300f ("sctp: delay calls to sk_data_ready() as much as
possible") minimized the number of wake ups that are triggered in case
the association receives a packet with multiple data chunks on it and/or
when io_events are enabled and then commit 0970f5b36659 ("sctp: signal
sk_data_ready earlier on data chunks reception") moved the wake up to as
soon as possible. It thus relies on the state machine running later to
clean the flag that the event was already generated.

The issue is that there are 2 call paths that calls
sctp_ulpq_tail_event() outside of the state machine, causing the flag to
linger and possibly omitting a needed wake up in the sequence.

One of the call paths is when enabling SCTP_SENDER_DRY_EVENTS via
setsockopt(SCTP_EVENTS), as noticed by Harald Welte. The other is when
partial reliability triggers removal of chunks from the send queue when
the application calls sendmsg().

This commit fixes it by not setting the flag in case the socket is not
owned by the user, as it won't be cleaned later. This works for
user-initiated calls and also for rx path processing.

Fixes: fb586f25300f ("sctp: delay calls to sk_data_ready() as much as possible")
Reported-by: Harald Welte <laforge@gnumonks.org>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoipv6: fix typo in fib6_net_exit()
Eric Dumazet [Fri, 8 Sep 2017 22:48:47 +0000 (15:48 -0700)]
ipv6: fix typo in fib6_net_exit()

[ Upstream commit 32a805baf0fb70b6dbedefcd7249ac7f580f9e3b ]

IPv6 FIB should use FIB6_TABLE_HASHSZ, not FIB_TABLE_HASHSZ.

Fixes: ba1cc08d9488 ("ipv6: fix memory leak with multiple tables during netns destruction")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoipv6: fix memory leak with multiple tables during netns destruction
Sabrina Dubroca [Fri, 8 Sep 2017 08:26:19 +0000 (10:26 +0200)]
ipv6: fix memory leak with multiple tables during netns destruction

[ Upstream commit ba1cc08d9488c94cb8d94f545305688b72a2a300 ]

fib6_net_exit only frees the main and local tables. If another table was
created with fib6_alloc_table, we leak it when the netns is destroyed.

Fix this in the same way ip_fib_net_exit cleans up tables, by walking
through the whole hashtable of fib6_table's. We can get rid of the
special cases for local and main, since they're also part of the
hashtable.

Reproducer:
    ip netns add x
    ip -net x -6 rule add from 6003:1::/64 table 100
    ip netns del x

Reported-by: Jianlin Shi <jishi@redhat.com>
Fixes: 58f09b78b730 ("[NETNS][IPV6] ip6_fib - make it per network namespace")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoip6_gre: update mtu properly in ip6gre_err
Xin Long [Tue, 5 Sep 2017 09:26:33 +0000 (17:26 +0800)]
ip6_gre: update mtu properly in ip6gre_err

[ Upstream commit 5c25f30c93fdc5bf25e62101aeaae7a4f9b421b3 ]

Now when probessing ICMPV6_PKT_TOOBIG, ip6gre_err only subtracts the
offset of gre header from mtu info. The expected mtu of gre device
should also subtract gre header. Otherwise, the next packets still
can't be sent out.

Jianlin found this issue when using the topo:
  client(ip6gre)<---->(nic1)route(nic2)<----->(ip6gre)server

and reducing nic2's mtu, then both tcp and sctp's performance with
big size data became 0.

This patch is to fix it by also subtracting grehdr (tun->tun_hlen)
from mtu info when updating gre device's mtu in ip6gre_err(). It
also needs to subtract ETH_HLEN if gre dev'type is ARPHRD_ETHER.

Reported-by: Jianlin Shi <jishi@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agovhost_net: correctly check tx avail during rx busy polling
Jason Wang [Tue, 5 Sep 2017 01:22:05 +0000 (09:22 +0800)]
vhost_net: correctly check tx avail during rx busy polling

[ Upstream commit 8b949bef9172ca69d918e93509a4ecb03d0355e0 ]

We check tx avail through vhost_enable_notify() in the past which is
wrong since it only checks whether or not guest has filled more
available buffer since last avail idx synchronization which was just
done by vhost_vq_avail_empty() before. What we really want is checking
pending buffers in the avail ring. Fix this by calling
vhost_vq_avail_empty() instead.

This issue could be noticed by doing netperf TCP_RR benchmark as
client from guest (but not host). With this fix, TCP_RR from guest to
localhost restores from 1375.91 trans per sec to 55235.28 trans per
sec on my laptop (Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz).

Fixes: 030881372460 ("vhost_net: basic polling support")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agogianfar: Fix Tx flow control deactivation
Claudiu Manoil [Mon, 4 Sep 2017 07:45:28 +0000 (10:45 +0300)]
gianfar: Fix Tx flow control deactivation

[ Upstream commit 5d621672bc1a1e5090c1ac5432a18c79e0e13e03 ]

The wrong register is checked for the Tx flow control bit,
it should have been maccfg1 not maccfg2.
This went unnoticed for so long probably because the impact is
hardly visible, not to mention the tangled code from adjust_link().
First, link flow control (i.e. handling of Rx/Tx link level pause frames)
is disabled by default (needs to be enabled via 'ethtool -A').
Secondly, maccfg2 always returns 0 for tx_flow_oldval (except for a few
old boards), which results in Tx flow control remaining always on
once activated.

Fixes: 45b679c9a3ccd9e34f28e6ec677b812a860eb8eb ("gianfar: Implement PAUSE frame generation support")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoRevert "net: fix percpu memory leaks"
Jesper Dangaard Brouer [Fri, 1 Sep 2017 09:26:13 +0000 (11:26 +0200)]
Revert "net: fix percpu memory leaks"

[ Upstream commit 5a63643e583b6a9789d7a225ae076fb4e603991c ]

This reverts commit 1d6119baf0610f813eb9d9580eb4fd16de5b4ceb.

After reverting commit 6d7b857d541e ("net: use lib/percpu_counter API
for fragmentation mem accounting") then here is no need for this
fix-up patch.  As percpu_counter is no longer used, it cannot
memory leak it any-longer.

Fixes: 6d7b857d541e ("net: use lib/percpu_counter API for fragmentation mem accounting")
Fixes: 1d6119baf061 ("net: fix percpu memory leaks")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoRevert "net: use lib/percpu_counter API for fragmentation mem accounting"
Jesper Dangaard Brouer [Fri, 1 Sep 2017 09:26:08 +0000 (11:26 +0200)]
Revert "net: use lib/percpu_counter API for fragmentation mem accounting"

[ Upstream commit fb452a1aa3fd4034d7999e309c5466ff2d7005aa ]

This reverts commit 6d7b857d541ecd1d9bd997c97242d4ef94b19de2.

There is a bug in fragmentation codes use of the percpu_counter API,
that can cause issues on systems with many CPUs.

The frag_mem_limit() just reads the global counter (fbc->count),
without considering other CPUs can have upto batch size (130K) that
haven't been subtracted yet.  Due to the 3MBytes lower thresh limit,
this become dangerous at >=24 CPUs (3*1024*1024/130000=24).

The correct API usage would be to use __percpu_counter_compare() which
does the right thing, and takes into account the number of (online)
CPUs and batch size, to account for this and call __percpu_counter_sum()
when needed.

We choose to revert the use of the lib/percpu_counter API for frag
memory accounting for several reasons:

1) On systems with CPUs > 24, the heavier fully locked
   __percpu_counter_sum() is always invoked, which will be more
   expensive than the atomic_t that is reverted to.

Given systems with more than 24 CPUs are becoming common this doesn't
seem like a good option.  To mitigate this, the batch size could be
decreased and thresh be increased.

2) The add_frag_mem_limit+sub_frag_mem_limit pairs happen on the RX
   CPU, before SKBs are pushed into sockets on remote CPUs.  Given
   NICs can only hash on L2 part of the IP-header, the NIC-RXq's will
   likely be limited.  Thus, a fair chance that atomic add+dec happen
   on the same CPU.

Revert note that commit 1d6119baf061 ("net: fix percpu memory leaks")
removed init_frag_mem_limit() and instead use inet_frags_init_net().
After this revert, inet_frags_uninit_net() becomes empty.

Fixes: 6d7b857d541e ("net: use lib/percpu_counter API for fragmentation mem accounting")
Fixes: 1d6119baf061 ("net: fix percpu memory leaks")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agobridge: switchdev: Clear forward mark when transmitting packet
Ido Schimmel [Fri, 1 Sep 2017 09:22:25 +0000 (12:22 +0300)]
bridge: switchdev: Clear forward mark when transmitting packet

[ Upstream commit 79e99bdd60b484af9afe0147e85a13e66d5c1cdb ]

Commit 6bc506b4fb06 ("bridge: switchdev: Add forward mark support for
stacked devices") added the 'offload_fwd_mark' bit to the skb in order
to allow drivers to indicate to the bridge driver that they already
forwarded the packet in L2.

In case the bit is set, before transmitting the packet from each port,
the port's mark is compared with the mark stored in the skb's control
block. If both marks are equal, we know the packet arrived from a switch
device that already forwarded the packet and it's not re-transmitted.

However, if the packet is transmitted from the bridge device itself
(e.g., br0), we should clear the 'offload_fwd_mark' bit as the mark
stored in the skb's control block isn't valid.

This scenario can happen in rare cases where a packet was trapped during
L3 forwarding and forwarded by the kernel to a bridge device.

Fixes: 6bc506b4fb06 ("bridge: switchdev: Add forward mark support for stacked devices")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reported-by: Yotam Gigi <yotamg@mellanox.com>
Tested-by: Yotam Gigi <yotamg@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomlxsw: spectrum: Forbid linking to devices that have uppers
Ido Schimmel [Fri, 1 Sep 2017 08:52:31 +0000 (10:52 +0200)]
mlxsw: spectrum: Forbid linking to devices that have uppers

[ Upstream commit 25cc72a33835ed8a6f53180a822cadab855852ac ]

The mlxsw driver relies on NETDEV_CHANGEUPPER events to configure the
device in case a port is enslaved to a master netdev such as bridge or
bond.

Since the driver ignores events unrelated to its ports and their
uppers, it's possible to engineer situations in which the device's data
path differs from the kernel's.

One example to such a situation is when a port is enslaved to a bond
that is already enslaved to a bridge. When the bond was enslaved the
driver ignored the event - as the bond wasn't one of its uppers - and
therefore a bridge port instance isn't created in the device.

Until such configurations are supported forbid them by checking that the
upper device doesn't have uppers of its own.

Fixes: 0d65fc13042f ("mlxsw: spectrum: Implement LAG port join/leave")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reported-by: Nogah Frankel <nogahf@mellanox.com>
Tested-by: Nogah Frankel <nogahf@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agotcp: initialize rcv_mss to TCP_MIN_MSS instead of 0
Wei Wang [Thu, 18 May 2017 18:22:33 +0000 (11:22 -0700)]
tcp: initialize rcv_mss to TCP_MIN_MSS instead of 0

[ Upstream commit 499350a5a6e7512d9ed369ed63a4244b6536f4f8 ]

When tcp_disconnect() is called, inet_csk_delack_init() sets
icsk->icsk_ack.rcv_mss to 0.
This could potentially cause tcp_recvmsg() => tcp_cleanup_rbuf() =>
__tcp_select_window() call path to have division by 0 issue.
So this patch initializes rcv_mss to TCP_MIN_MSS instead of 0.

Reported-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Wei Wang <weiwan@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoRevert "net: phy: Correctly process PHY_HALTED in phy_stop_machine()"
Florian Fainelli [Thu, 31 Aug 2017 00:49:29 +0000 (17:49 -0700)]
Revert "net: phy: Correctly process PHY_HALTED in phy_stop_machine()"

[ Upstream commit ebc8254aeae34226d0bc8fda309fd9790d4dccfe ]

This reverts commit 7ad813f208533cebfcc32d3d7474dc1677d1b09a ("net: phy:
Correctly process PHY_HALTED in phy_stop_machine()") because it is
creating the possibility for a NULL pointer dereference.

David Daney provide the following call trace and diagram of events:

When ndo_stop() is called we call:

 phy_disconnect()
    +---> phy_stop_interrupts() implies: phydev->irq = PHY_POLL;
    +---> phy_stop_machine()
    |      +---> phy_state_machine()
    |              +----> queue_delayed_work(): Work queued.
    +--->phy_detach() implies: phydev->attached_dev = NULL;

Now at a later time the queued work does:

 phy_state_machine()
    +---->netif_carrier_off(phydev->attached_dev): Oh no! It is NULL:

 CPU 12 Unable to handle kernel paging request at virtual address
0000000000000048, epc == ffffffff80de37ec, ra == ffffffff80c7c
Oops[#1]:
CPU: 12 PID: 1502 Comm: kworker/12:1 Not tainted 4.9.43-Cavium-Octeon+ #1
Workqueue: events_power_efficient phy_state_machine
task: 80000004021ed100 task.stack: 8000000409d70000
$ 0   : 0000000000000000 ffffffff84720060 0000000000000048 0000000000000004
$ 4   : 0000000000000000 0000000000000001 0000000000000004 0000000000000000
$ 8   : 0000000000000000 0000000000000000 00000000ffff98f3 0000000000000000
$12   : 8000000409d73fe0 0000000000009c00 ffffffff846547c8 000000000000af3b
$16   : 80000004096bab68 80000004096babd0 0000000000000000 80000004096ba800
$20   : 0000000000000000 0000000000000000 ffffffff81090000 0000000000000008
$24   : 0000000000000061 ffffffff808637b0
$28   : 8000000409d70000 8000000409d73cf0 80000000271bd300 ffffffff80c7804c
Hi    : 000000000000002a
Lo    : 000000000000003f
epc   : ffffffff80de37ec netif_carrier_off+0xc/0x58
ra    : ffffffff80c7804c phy_state_machine+0x48c/0x4f8
Status: 14009ce3        KX SX UX KERNEL EXL IE
Cause : 00800008 (ExcCode 02)
BadVA : 0000000000000048
PrId  : 000d9501 (Cavium Octeon III)
Modules linked in:
Process kworker/12:1 (pid: 1502, threadinfo=8000000409d70000,
task=80000004021ed100, tls=0000000000000000)
Stack : 8000000409a54000 80000004096bab68 80000000271bd300 80000000271c1e00
        0000000000000000 ffffffff808a1708 8000000409a54000 80000000271bd300
        80000000271bd320 8000000409a54030 ffffffff80ff0f00 0000000000000001
        ffffffff81090000 ffffffff808a1ac0 8000000402182080 ffffffff84650000
        8000000402182080 ffffffff84650000 ffffffff80ff0000 8000000409a54000
        ffffffff808a1970 0000000000000000 80000004099e8000 8000000402099240
        0000000000000000 ffffffff808a8598 0000000000000000 8000000408eeeb00
        8000000409a54000 00000000810a1d00 0000000000000000 8000000409d73de8
        8000000409d73de8 0000000000000088 000000000c009c00 8000000409d73e08
        8000000409d73e08 8000000402182080 ffffffff808a84d0 8000000402182080
        ...
Call Trace:
[<ffffffff80de37ec>] netif_carrier_off+0xc/0x58
[<ffffffff80c7804c>] phy_state_machine+0x48c/0x4f8
[<ffffffff808a1708>] process_one_work+0x158/0x368
[<ffffffff808a1ac0>] worker_thread+0x150/0x4c0
[<ffffffff808a8598>] kthread+0xc8/0xe0
[<ffffffff808617f0>] ret_from_kernel_thread+0x14/0x1c

The original motivation for this change originated from Marc Gonzales
indicating that his network driver did not have its adjust_link callback
executing with phydev->link = 0 while he was expecting it.

PHYLIB has never made any such guarantees ever because phy_stop() merely just
tells the workqueue to move into PHY_HALTED state which will happen
asynchronously.

Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reported-by: David Daney <ddaney.cavm@gmail.com>
Fixes: 7ad813f20853 ("net: phy: Correctly process PHY_HALTED in phy_stop_machine()")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agokcm: do not attach PF_KCM sockets to avoid deadlock
Eric Dumazet [Wed, 30 Aug 2017 16:29:31 +0000 (09:29 -0700)]
kcm: do not attach PF_KCM sockets to avoid deadlock

[ Upstream commit 351050ecd6523374b370341cc29fe61e2201556b ]

syzkaller had no problem to trigger a deadlock, attaching a KCM socket
to another one (or itself). (original syzkaller report was a very
confusing lockdep splat during a sendmsg())

It seems KCM claims to only support TCP, but no enforcement is done,
so we might need to add additional checks.

Fixes: ab7ac4eb9832 ("kcm: Kernel Connection Multiplexor module")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Tom Herbert <tom@quantonium.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agopacket: Don't write vnet header beyond end of buffer
Benjamin Poirier [Mon, 28 Aug 2017 18:29:41 +0000 (14:29 -0400)]
packet: Don't write vnet header beyond end of buffer

[ Upstream commit edbd58be15a957f6a760c4a514cd475217eb97fd ]

... which may happen with certain values of tp_reserve and maclen.

Fixes: 58d19b19cd99 ("packet: vnet_hdr support for tpacket_rcv")
Signed-off-by: Benjamin Poirier <bpoirier@suse.com>
Cc: Willem de Bruijn <willemb@google.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agocxgb4: Fix stack out-of-bounds read due to wrong size to t4_record_mbox()
Stefano Brivio [Fri, 25 Aug 2017 20:48:48 +0000 (22:48 +0200)]
cxgb4: Fix stack out-of-bounds read due to wrong size to t4_record_mbox()

[ Upstream commit 0f3086868e8889a823a6e0f3d299102aa895d947 ]

Passing commands for logging to t4_record_mbox() with size
MBOX_LEN, when the actual command size is actually smaller,
causes out-of-bounds stack accesses in t4_record_mbox() while
copying command words here:

for (i = 0; i < size / 8; i++)
entry->cmd[i] = be64_to_cpu(cmd[i]);

Up to 48 bytes from the stack are then leaked to debugfs.

This happens whenever we send (and log) commands described by
structs fw_sched_cmd (32 bytes leaked), fw_vi_rxmode_cmd (48),
fw_hello_cmd (48), fw_bye_cmd (48), fw_initialize_cmd (48),
fw_reset_cmd (48), fw_pfvf_cmd (32), fw_eq_eth_cmd (16),
fw_eq_ctrl_cmd (32), fw_eq_ofld_cmd (32), fw_acl_mac_cmd(16),
fw_rss_glb_config_cmd(32), fw_rss_vi_config_cmd(32),
fw_devlog_cmd(32), fw_vi_enable_cmd(48), fw_port_cmd(32),
fw_sched_cmd(32), fw_devlog_cmd(32).

The cxgb4vf driver got this right instead.

When we call t4_record_mbox() to log a command reply, a MBOX_LEN
size can be used though, as get_mbox_rpl() will fill cmd_rpl up
completely.

Fixes: 7f080c3f2ff0 ("cxgb4: Add support to enable logging of firmware mailbox commands")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agonetvsc: fix deadlock betwen link status and removal
stephen hemminger [Thu, 24 Aug 2017 23:49:16 +0000 (16:49 -0700)]
netvsc: fix deadlock betwen link status and removal

[ Upstream commit 9b4e946ce14e20d7addbfb7d9139e604f9fda107 ]

There is a deadlock possible when canceling the link status
delayed work queue. The removal process is run with RTNL held,
and the link status callback is acquring RTNL.

Resolve the issue by using trylock and rescheduling.
If cancel is in process, that block it from happening.

Fixes: 122a5f6410f4 ("staging: hv: use delayed_work for netvsc_send_garp()")
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoqlge: avoid memcpy buffer overflow
Arnd Bergmann [Wed, 23 Aug 2017 13:59:49 +0000 (15:59 +0200)]
qlge: avoid memcpy buffer overflow

[ Upstream commit e58f95831e7468d25eb6e41f234842ecfe6f014f ]

gcc-8.0.0 (snapshot) points out that we copy a variable-length string
into a fixed length field using memcpy() with the destination length,
and that ends up copying whatever follows the string:

    inlined from 'ql_core_dump' at drivers/net/ethernet/qlogic/qlge/qlge_dbg.c:1106:2:
drivers/net/ethernet/qlogic/qlge/qlge_dbg.c:708:2: error: 'memcpy' reading 15 bytes from a region of size 14 [-Werror=stringop-overflow=]
  memcpy(seg_hdr->description, desc, (sizeof(seg_hdr->description)) - 1);

Changing it to use strncpy() will instead zero-pad the destination,
which seems to be the right thing to do here.

The bug is probably harmless, but it seems like a good idea to address
it in stable kernels as well, if only for the purpose of building with
gcc-8 without warnings.

Fixes: a61f80261306 ("qlge: Add ethtool register dump function.")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agosctp: Avoid out-of-bounds reads from address storage
Stefano Brivio [Wed, 23 Aug 2017 11:27:13 +0000 (13:27 +0200)]
sctp: Avoid out-of-bounds reads from address storage

[ Upstream commit ee6c88bb754e3d363e568da78086adfedb692447 ]

inet_diag_msg_sctp{,l}addr_fill() and sctp_get_sctp_info() copy
sizeof(sockaddr_storage) bytes to fill in sockaddr structs used
to export diagnostic information to userspace.

However, the memory allocated to store sockaddr information is
smaller than that and depends on the address family, so we leak
up to 100 uninitialized bytes to userspace. Just use the size of
the source structs instead, in all the three cases this is what
userspace expects. Zero out the remaining memory.

Unused bytes (i.e. when IPv4 addresses are used) in source
structs sctp_sockaddr_entry and sctp_transport are already
cleared by sctp_add_bind_addr() and sctp_transport_new(),
respectively.

Noticed while testing KASAN-enabled kernel with 'ss':

[ 2326.885243] BUG: KASAN: slab-out-of-bounds in inet_sctp_diag_fill+0x42c/0x6c0 [sctp_diag] at addr ffff881be8779800
[ 2326.896800] Read of size 128 by task ss/9527
[ 2326.901564] CPU: 0 PID: 9527 Comm: ss Not tainted 4.11.0-22.el7a.x86_64 #1
[ 2326.909236] Hardware name: Dell Inc. PowerEdge R730/072T6D, BIOS 2.4.3 01/17/2017
[ 2326.917585] Call Trace:
[ 2326.920312]  dump_stack+0x63/0x8d
[ 2326.924014]  kasan_object_err+0x21/0x70
[ 2326.928295]  kasan_report+0x288/0x540
[ 2326.932380]  ? inet_sctp_diag_fill+0x42c/0x6c0 [sctp_diag]
[ 2326.938500]  ? skb_put+0x8b/0xd0
[ 2326.942098]  ? memset+0x31/0x40
[ 2326.945599]  check_memory_region+0x13c/0x1a0
[ 2326.950362]  memcpy+0x23/0x50
[ 2326.953669]  inet_sctp_diag_fill+0x42c/0x6c0 [sctp_diag]
[ 2326.959596]  ? inet_diag_msg_sctpasoc_fill+0x460/0x460 [sctp_diag]
[ 2326.966495]  ? __lock_sock+0x102/0x150
[ 2326.970671]  ? sock_def_wakeup+0x60/0x60
[ 2326.975048]  ? remove_wait_queue+0xc0/0xc0
[ 2326.979619]  sctp_diag_dump+0x44a/0x760 [sctp_diag]
[ 2326.985063]  ? sctp_ep_dump+0x280/0x280 [sctp_diag]
[ 2326.990504]  ? memset+0x31/0x40
[ 2326.994007]  ? mutex_lock+0x12/0x40
[ 2326.997900]  __inet_diag_dump+0x57/0xb0 [inet_diag]
[ 2327.003340]  ? __sys_sendmsg+0x150/0x150
[ 2327.007715]  inet_diag_dump+0x4d/0x80 [inet_diag]
[ 2327.012979]  netlink_dump+0x1e6/0x490
[ 2327.017064]  __netlink_dump_start+0x28e/0x2c0
[ 2327.021924]  inet_diag_handler_cmd+0x189/0x1a0 [inet_diag]
[ 2327.028045]  ? inet_diag_rcv_msg_compat+0x1b0/0x1b0 [inet_diag]
[ 2327.034651]  ? inet_diag_dump_compat+0x190/0x190 [inet_diag]
[ 2327.040965]  ? __netlink_lookup+0x1b9/0x260
[ 2327.045631]  sock_diag_rcv_msg+0x18b/0x1e0
[ 2327.050199]  netlink_rcv_skb+0x14b/0x180
[ 2327.054574]  ? sock_diag_bind+0x60/0x60
[ 2327.058850]  sock_diag_rcv+0x28/0x40
[ 2327.062837]  netlink_unicast+0x2e7/0x3b0
[ 2327.067212]  ? netlink_attachskb+0x330/0x330
[ 2327.071975]  ? kasan_check_write+0x14/0x20
[ 2327.076544]  netlink_sendmsg+0x5be/0x730
[ 2327.080918]  ? netlink_unicast+0x3b0/0x3b0
[ 2327.085486]  ? kasan_check_write+0x14/0x20
[ 2327.090057]  ? selinux_socket_sendmsg+0x24/0x30
[ 2327.095109]  ? netlink_unicast+0x3b0/0x3b0
[ 2327.099678]  sock_sendmsg+0x74/0x80
[ 2327.103567]  ___sys_sendmsg+0x520/0x530
[ 2327.107844]  ? __get_locked_pte+0x178/0x200
[ 2327.112510]  ? copy_msghdr_from_user+0x270/0x270
[ 2327.117660]  ? vm_insert_page+0x360/0x360
[ 2327.122133]  ? vm_insert_pfn_prot+0xb4/0x150
[ 2327.126895]  ? vm_insert_pfn+0x32/0x40
[ 2327.131077]  ? vvar_fault+0x71/0xd0
[ 2327.134968]  ? special_mapping_fault+0x69/0x110
[ 2327.140022]  ? __do_fault+0x42/0x120
[ 2327.144008]  ? __handle_mm_fault+0x1062/0x17a0
[ 2327.148965]  ? __fget_light+0xa7/0xc0
[ 2327.153049]  __sys_sendmsg+0xcb/0x150
[ 2327.157133]  ? __sys_sendmsg+0xcb/0x150
[ 2327.161409]  ? SyS_shutdown+0x140/0x140
[ 2327.165688]  ? exit_to_usermode_loop+0xd0/0xd0
[ 2327.170646]  ? __do_page_fault+0x55d/0x620
[ 2327.175216]  ? __sys_sendmsg+0x150/0x150
[ 2327.179591]  SyS_sendmsg+0x12/0x20
[ 2327.183384]  do_syscall_64+0xe3/0x230
[ 2327.187471]  entry_SYSCALL64_slow_path+0x25/0x25
[ 2327.192622] RIP: 0033:0x7f41d18fa3b0
[ 2327.196608] RSP: 002b:00007ffc3b731218 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
[ 2327.205055] RAX: ffffffffffffffda RBX: 00007ffc3b731380 RCX: 00007f41d18fa3b0
[ 2327.213017] RDX: 0000000000000000 RSI: 00007ffc3b731340 RDI: 0000000000000003
[ 2327.220978] RBP: 0000000000000002 R08: 0000000000000004 R09: 0000000000000040
[ 2327.228939] R10: 00007ffc3b730f30 R11: 0000000000000246 R12: 0000000000000003
[ 2327.236901] R13: 00007ffc3b731340 R14: 00007ffc3b7313d0 R15: 0000000000000084
[ 2327.244865] Object at ffff881be87797e0, in cache kmalloc-64 size: 64
[ 2327.251953] Allocated:
[ 2327.254581] PID = 9484
[ 2327.257215]  save_stack_trace+0x1b/0x20
[ 2327.261485]  save_stack+0x46/0xd0
[ 2327.265179]  kasan_kmalloc+0xad/0xe0
[ 2327.269165]  kmem_cache_alloc_trace+0xe6/0x1d0
[ 2327.274138]  sctp_add_bind_addr+0x58/0x180 [sctp]
[ 2327.279400]  sctp_do_bind+0x208/0x310 [sctp]
[ 2327.284176]  sctp_bind+0x61/0xa0 [sctp]
[ 2327.288455]  inet_bind+0x5f/0x3a0
[ 2327.292151]  SYSC_bind+0x1a4/0x1e0
[ 2327.295944]  SyS_bind+0xe/0x10
[ 2327.299349]  do_syscall_64+0xe3/0x230
[ 2327.303433]  return_from_SYSCALL_64+0x0/0x6a
[ 2327.308194] Freed:
[ 2327.310434] PID = 4131
[ 2327.313065]  save_stack_trace+0x1b/0x20
[ 2327.317344]  save_stack+0x46/0xd0
[ 2327.321040]  kasan_slab_free+0x73/0xc0
[ 2327.325220]  kfree+0x96/0x1a0
[ 2327.328530]  dynamic_kobj_release+0x15/0x40
[ 2327.333195]  kobject_release+0x99/0x1e0
[ 2327.337472]  kobject_put+0x38/0x70
[ 2327.341266]  free_notes_attrs+0x66/0x80
[ 2327.345545]  mod_sysfs_teardown+0x1a5/0x270
[ 2327.350211]  free_module+0x20/0x2a0
[ 2327.354099]  SyS_delete_module+0x2cb/0x2f0
[ 2327.358667]  do_syscall_64+0xe3/0x230
[ 2327.362750]  return_from_SYSCALL_64+0x0/0x6a
[ 2327.367510] Memory state around the buggy address:
[ 2327.372855]  ffff881be8779700: fc fc fc fc 00 00 00 00 00 00 00 00 fc fc fc fc
[ 2327.380914]  ffff881be8779780: fb fb fb fb fb fb fb fb fc fc fc fc 00 00 00 00
[ 2327.388972] >ffff881be8779800: 00 00 00 00 fc fc fc fc fb fb fb fb fb fb fb fb
[ 2327.397031]                                ^
[ 2327.401792]  ffff881be8779880: fc fc fc fc fb fb fb fb fb fb fb fb fc fc fc fc
[ 2327.409850]  ffff881be8779900: 00 00 00 00 00 04 fc fc fc fc fc fc 00 00 00 00
[ 2327.417907] ==================================================================

This fixes CVE-2017-7558.

References: https://bugzilla.redhat.com/show_bug.cgi?id=1480266
Fixes: 8f840e47f190 ("sctp: add the sctp_diag.c file")
Cc: Xin Long <lucien.xin@gmail.com>
Cc: Vlad Yasevich <vyasevich@gmail.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agofsl/man: Inherit parent device and of_node
Florian Fainelli [Tue, 22 Aug 2017 22:24:47 +0000 (15:24 -0700)]
fsl/man: Inherit parent device and of_node

[ Upstream commit a1a50c8e4c241a505b7270e1a3c6e50d94e794b1 ]

Junote Cai reported that he was not able to get a DSA setup involving the
Freescale DPAA/FMAN driver to work and narrowed it down to
of_find_net_device_by_node(). This function requires the network device's
device reference to be correctly set which is the case here, though we have
lost any device_node association there.

The problem is that dpaa_eth_add_device() allocates a "dpaa-ethernet" platform
device, and later on dpaa_eth_probe() is called but SET_NETDEV_DEV() won't be
propagating &pdev->dev.of_node properly. Fix this by inherenting both the parent
device and the of_node when dpaa_eth_add_device() creates the platform device.

Fixes: 3933961682a3 ("fsl/fman: Add FMan MAC driver")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoudp: on peeking bad csum, drop packets even if not at head
Eric Dumazet [Tue, 22 Aug 2017 16:39:28 +0000 (09:39 -0700)]
udp: on peeking bad csum, drop packets even if not at head

[ Upstream commit fd6055a806edc4019be1b9fb7d25262599bca5b1 ]

When peeking, if a bad csum is discovered, the skb is unlinked from
the queue with __sk_queue_drop_skb and the peek operation restarted.

__sk_queue_drop_skb only drops packets that match the queue head.

This fails if the skb was found after the head, using SO_PEEK_OFF
socket option. This causes an infinite loop.

We MUST drop this problematic skb, and we can simply check if skb was
already removed by another thread, by looking at skb->next :

This pointer is set to NULL by the  __skb_unlink() operation, that might
have happened only under the spinlock protection.

Many thanks to syzkaller team (and particularly Dmitry Vyukov who
provided us nice C reproducers exhibiting the lockup) and Willem de
Bruijn who provided first version for this patch and a test program.

Fixes: 627d2d6b5500 ("udp: enable MSG_PEEK at non-zero offset")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomacsec: add genl family module alias
Sabrina Dubroca [Tue, 22 Aug 2017 13:36:08 +0000 (15:36 +0200)]
macsec: add genl family module alias

[ Upstream commit 78362998f58c7c271e2719dcd0aaced435c801f9 ]

This helps tools such as wpa_supplicant can start even if the macsec
module isn't loaded yet.

Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoipv6: fix sparse warning on rt6i_node
Wei Wang [Fri, 25 Aug 2017 22:03:10 +0000 (15:03 -0700)]
ipv6: fix sparse warning on rt6i_node

[ Upstream commit 4e587ea71bf924f7dac621f1351653bd41e446cb ]

Commit c5cff8561d2d adds rcu grace period before freeing fib6_node. This
generates a new sparse warning on rt->rt6i_node related code:
  net/ipv6/route.c:1394:30: error: incompatible types in comparison
  expression (different address spaces)
  ./include/net/ip6_fib.h:187:14: error: incompatible types in comparison
  expression (different address spaces)

This commit adds "__rcu" tag for rt6i_node and makes sure corresponding
rcu API is used for it.
After this fix, sparse no longer generates the above warning.

Fixes: c5cff8561d2d ("ipv6: add rcu grace period before freeing fib6_node")
Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoipv6: add rcu grace period before freeing fib6_node
Wei Wang [Mon, 21 Aug 2017 16:47:10 +0000 (09:47 -0700)]
ipv6: add rcu grace period before freeing fib6_node

[ Upstream commit c5cff8561d2d0006e972bd114afd51f082fee77c ]

We currently keep rt->rt6i_node pointing to the fib6_node for the route.
And some functions make use of this pointer to dereference the fib6_node
from rt structure, e.g. rt6_check(). However, as there is neither
refcount nor rcu taken when dereferencing rt->rt6i_node, it could
potentially cause crashes as rt->rt6i_node could be set to NULL by other
CPUs when doing a route deletion.
This patch introduces an rcu grace period before freeing fib6_node and
makes sure the functions that dereference it takes rcu_read_lock().

Note: there is no "Fixes" tag because this bug was there in a very
early stage.

Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoipv6: accept 64k - 1 packet length in ip6_find_1stfragopt()
Stefano Brivio [Fri, 18 Aug 2017 12:40:53 +0000 (14:40 +0200)]
ipv6: accept 64k - 1 packet length in ip6_find_1stfragopt()

[ Upstream commit 3de33e1ba0506723ab25734e098cf280ecc34756 ]

A packet length of exactly IPV6_MAXPLEN is allowed, we should
refuse parsing options only if the size is 64KiB or more.

While at it, remove one extra variable and one assignment which
were also introduced by the commit that introduced the size
check. Checking the sum 'offset + len' and only later adding
'len' to 'offset' doesn't provide any advantage over directly
summing to 'offset' and checking it.

Fixes: 6399f1fae4ec ("ipv6: avoid overflow of offset in ip6_find_1stfragopt")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoLinux 4.9.50 v4.9.50
Greg Kroah-Hartman [Wed, 13 Sep 2017 21:13:54 +0000 (14:13 -0700)]
Linux 4.9.50

7 years agoxfs: XFS_IS_REALTIME_INODE() should be false if no rt device present
Richard Wareing [Tue, 12 Sep 2017 23:09:35 +0000 (09:09 +1000)]
xfs: XFS_IS_REALTIME_INODE() should be false if no rt device present

commit b31ff3cdf540110da4572e3e29bd172087af65cc upstream.

If using a kernel with CONFIG_XFS_RT=y and we set the RHINHERIT flag on
a directory in a filesystem that does not have a realtime device and
create a new file in that directory, it gets marked as a real time file.
When data is written and a fsync is issued, the filesystem attempts to
flush a non-existent rt device during the fsync process.

This results in a crash dereferencing a null buftarg pointer in
xfs_blkdev_issue_flush():

  BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
  IP: xfs_blkdev_issue_flush+0xd/0x20
  .....
  Call Trace:
    xfs_file_fsync+0x188/0x1c0
    vfs_fsync_range+0x3b/0xa0
    do_fsync+0x3d/0x70
    SyS_fsync+0x10/0x20
    do_syscall_64+0x4d/0xb0
    entry_SYSCALL64_slow_path+0x25/0x25

Setting RT inode flags does not require special privileges so any
unprivileged user can cause this oops to occur.  To reproduce, confirm
kernel is compiled with CONFIG_XFS_RT=y and run:

  # mkfs.xfs -f /dev/pmem0
  # mount /dev/pmem0 /mnt/test
  # mkdir /mnt/test/foo
  # xfs_io -c 'chattr +t' /mnt/test/foo
  # xfs_io -f -c 'pwrite 0 5m' -c fsync /mnt/test/foo/bar

Or just run xfstests with MKFS_OPTIONS="-d rtinherit=1" and wait.

Kernels built with CONFIG_XFS_RT=n are not exposed to this bug.

Fixes: f538d4da8d52 ("[XFS] write barrier support")
Signed-off-by: Richard Wareing <rwareing@fb.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoNFS: Sync the correct byte range during synchronous writes
tarangg@amazon.com [Thu, 7 Sep 2017 13:29:23 +0000 (09:29 -0400)]
NFS: Sync the correct byte range during synchronous writes

commit e973b1a5999e57da677ab50da5f5479fdc0f0c31 upstream.

Since commit 18290650b1c8 ("NFS: Move buffered I/O locking into
nfs_file_write()") nfs_file_write() has not flushed the correct byte
range during synchronous writes.  generic_write_sync() expects that
iocb->ki_pos points to the right edge of the range rather than the
left edge.

To replicate the problem, open a file with O_DSYNC, have the client
write at increasing offsets, and then print the successful offsets.
Block port 2049 partway through that sequence, and observe that the
client application indicates successful writes in advance of what the
server received.

Fixes: 18290650b1c8 ("NFS: Move buffered I/O locking into nfs_file_write()")
Signed-off-by: Jacob Strauss <jsstraus@amazon.com>
Signed-off-by: Tarang Gupta <tarangg@amazon.com>
Tested-by: Tarang Gupta <tarangg@amazon.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoNFS: Fix 2 use after free issues in the I/O code
Trond Myklebust [Sat, 9 Sep 2017 01:28:11 +0000 (21:28 -0400)]
NFS: Fix 2 use after free issues in the I/O code

commit 196639ebbe63a037fe9a80669140bd292d8bcd80 upstream.

The writeback code wants to send a commit after processing the pages,
which is why we want to delay releasing the struct path until after
that's done.

Also, the layout code expects that we do not free the inode before
we've put the layout segments in pnfs_writehdr_free() and
pnfs_readhdr_free()

Fixes: 919e3bd9a875 ("NFS: Ensure we commit after writeback is complete")
Fixes: 4714fb51fd03 ("nfs: remove pgio_header refcount, related cleanup")
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoARM: 8692/1: mm: abort uaccess retries upon fatal signal
Mark Rutland [Tue, 22 Aug 2017 10:36:17 +0000 (11:36 +0100)]
ARM: 8692/1: mm: abort uaccess retries upon fatal signal

commit 746a272e44141af24a02f6c9b0f65f4c4598ed42 upstream.

When there's a fatal signal pending, arm's do_page_fault()
implementation returns 0. The intent is that we'll return to the
faulting userspace instruction, delivering the signal on the way.

However, if we take a fatal signal during fixing up a uaccess, this
results in a return to the faulting kernel instruction, which will be
instantly retried, resulting in the same fault being taken forever. As
the task never reaches userspace, the signal is not delivered, and the
task is left unkillable. While the task is stuck in this state, it can
inhibit the forward progress of the system.

To avoid this, we must ensure that when a fatal signal is pending, we
apply any necessary fixup for a faulting kernel instruction. Thus we
will return to an error path, and it is up to that code to make forward
progress towards delivering the fatal signal.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Steve Capper <steve.capper@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoARM64: dts: marvell: armada-37xx: Fix GIC maintenance interrupt
Marc Zyngier [Sat, 1 Jul 2017 14:16:34 +0000 (15:16 +0100)]
ARM64: dts: marvell: armada-37xx: Fix GIC maintenance interrupt

commit 95696d292e204073433ed2ef3ff4d3d8f42a8248 upstream.

The GIC-500 integrated in the Armada-37xx SoCs is compliant with
the GICv3 architecture, and thus provides a maintenance interrupt
that is required for hypervisors to function correctly.

With the interrupt provided in the DT, KVM now works as it should.
Tested on an Espressobin system.

Fixes: adbc3695d9e4 ("arm64: dts: add the Marvell Armada 3700 family and a development board")
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoBluetooth: Properly check L2CAP config option output buffer length
Ben Seri [Sat, 9 Sep 2017 21:15:59 +0000 (23:15 +0200)]
Bluetooth: Properly check L2CAP config option output buffer length

commit e860d2c904d1a9f38a24eb44c9f34b8f915a6ea3 upstream.

Validate the output buffer length for L2CAP config requests and responses
to avoid overflowing the stack buffer used for building the option blocks.

Signed-off-by: Ben Seri <ben@armis.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoALSA: msnd: Optimize / harden DSP and MIDI loops
Takashi Iwai [Thu, 6 Jul 2017 10:34:40 +0000 (12:34 +0200)]
ALSA: msnd: Optimize / harden DSP and MIDI loops

commit 20e2b791796bd68816fa115f12be5320de2b8021 upstream.

The ISA msnd drivers have loops fetching the ring-buffer head, tail
and size values inside the loops.  Such codes are inefficient and
fragile.

This patch optimizes it, and also adds the sanity check to avoid the
endless loops.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=196131
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=196133
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: grygorii tertychnyi <gtertych@cisco.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agolocktorture: Fix potential memory leak with rw lock test
Yang Shi [Thu, 10 Nov 2016 21:06:39 +0000 (13:06 -0800)]
locktorture: Fix potential memory leak with rw lock test

commit f4dbba591945dc301c302672adefba9e2ec08dc5 upstream.

When running locktorture module with the below commands with kmemleak enabled:

$ modprobe locktorture torture_type=rw_lock_irq
$ rmmod locktorture

The below kmemleak got caught:

root@10:~# echo scan > /sys/kernel/debug/kmemleak
[  323.197029] kmemleak: 2 new suspected memory leaks (see /sys/kernel/debug/kmemleak)
root@10:~# cat /sys/kernel/debug/kmemleak
unreferenced object 0xffffffc07592d500 (size 128):
  comm "modprobe", pid 368, jiffies 4294924118 (age 205.824s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 c3 7b 02 00 00 00 00 00  .........{......
    00 00 00 00 00 00 00 00 d7 9b 02 00 00 00 00 00  ................
  backtrace:
    [<ffffff80081e5a88>] create_object+0x110/0x288
    [<ffffff80086c6078>] kmemleak_alloc+0x58/0xa0
    [<ffffff80081d5acc>] __kmalloc+0x234/0x318
    [<ffffff80006fa130>] 0xffffff80006fa130
    [<ffffff8008083ae4>] do_one_initcall+0x44/0x138
    [<ffffff800817e28c>] do_init_module+0x68/0x1cc
    [<ffffff800811c848>] load_module+0x1a68/0x22e0
    [<ffffff800811d340>] SyS_finit_module+0xe0/0xf0
    [<ffffff80080836f0>] el0_svc_naked+0x24/0x28
    [<ffffffffffffffff>] 0xffffffffffffffff
unreferenced object 0xffffffc07592d480 (size 128):
  comm "modprobe", pid 368, jiffies 4294924118 (age 205.824s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 3b 6f 01 00 00 00 00 00  ........;o......
    00 00 00 00 00 00 00 00 23 6a 01 00 00 00 00 00  ........#j......
  backtrace:
    [<ffffff80081e5a88>] create_object+0x110/0x288
    [<ffffff80086c6078>] kmemleak_alloc+0x58/0xa0
    [<ffffff80081d5acc>] __kmalloc+0x234/0x318
    [<ffffff80006fa22c>] 0xffffff80006fa22c
    [<ffffff8008083ae4>] do_one_initcall+0x44/0x138
    [<ffffff800817e28c>] do_init_module+0x68/0x1cc
    [<ffffff800811c848>] load_module+0x1a68/0x22e0
    [<ffffff800811d340>] SyS_finit_module+0xe0/0xf0
    [<ffffff80080836f0>] el0_svc_naked+0x24/0x28
    [<ffffffffffffffff>] 0xffffffffffffffff

It is because cxt.lwsa and cxt.lrsa don't get freed in module_exit, so free
them in lock_torture_cleanup() and free writer_tasks if reader_tasks is
failed at memory allocation.

Signed-off-by: Yang Shi <yang.shi@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Cc: 石洋 <yang.s@alibaba-inc.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomm/memory.c: fix mem_cgroup_oom_disable() call missing
Laurent Dufour [Fri, 8 Sep 2017 23:13:12 +0000 (16:13 -0700)]
mm/memory.c: fix mem_cgroup_oom_disable() call missing

commit de0c799bba2610a8e1e9a50d76a28614520a4cd4 upstream.

Seen while reading the code, in handle_mm_fault(), in the case
arch_vma_access_permitted() is failing the call to
mem_cgroup_oom_disable() is not made.

To fix that, move the call to mem_cgroup_oom_enable() after calling
arch_vma_access_permitted() as it should not have entered the memcg OOM.

Link: http://lkml.kernel.org/r/1504625439-31313-1-git-send-email-ldufour@linux.vnet.ibm.com
Fixes: bae473a423f6 ("mm: introduce fault_env")
Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Acked-by: Kirill A. Shutemov <kirill@shutemov.name>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoselftests/x86/fsgsbase: Test selectors 1, 2, and 3
Andy Lutomirski [Tue, 1 Aug 2017 14:11:36 +0000 (07:11 -0700)]
selftests/x86/fsgsbase: Test selectors 1, 2, and 3

commit 23d98c204386a98d9ef9f9e744f41443ece4929f upstream.

Those are funny cases.  Make sure they work.

(Something is screwy with signal handling if a selector is 1, 2, or 3.
Anyone who wants to dive into that rabbit hole is welcome to do so.)

Signed-off-by: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Borislav Petkov <bpetkov@suse.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Chang Seok <chang.seok.bae@intel.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agobtrfs: resume qgroup rescan on rw remount
Aleksa Sarai [Tue, 4 Jul 2017 11:49:06 +0000 (21:49 +1000)]
btrfs: resume qgroup rescan on rw remount

commit 6c6b5a39c4bf3dbd8cf629c9f5450e983c19dbb9 upstream.

Several distributions mount the "proper root" as ro during initrd and
then remount it as rw before pivot_root(2). Thus, if a rescan had been
aborted by a previous shutdown, the rescan would never be resumed.

This issue would manifest itself as several btrfs ioctl(2)s causing the
entire machine to hang when btrfs_qgroup_wait_for_completion was hit
(due to the fs_info->qgroup_rescan_running flag being set but the rescan
itself not being resumed). Notably, Docker's btrfs storage driver makes
regular use of BTRFS_QUOTA_CTL_DISABLE and BTRFS_IOC_QUOTA_RESCAN_WAIT
(causing this problem to be manifested on boot for some machines).

Cc: Jeff Mahoney <jeffm@suse.com>
Fixes: b382a324b60f ("Btrfs: fix qgroup rescan resume on mount")
Signed-off-by: Aleksa Sarai <asarai@suse.de>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Tested-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agonvme-fabrics: generate spec-compliant UUID NQNs
Daniel Verkamp [Wed, 30 Aug 2017 22:18:19 +0000 (15:18 -0700)]
nvme-fabrics: generate spec-compliant UUID NQNs

commit 40a5fce495715c48c2e02668144e68a507ac5a30 upstream.

The default host NQN, which is generated based on the host's UUID,
does not follow the UUID-based NQN format laid out in the NVMe 1.3
specification.  Remove the "NVMf:" portion of the NQN to match the spec.

Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomtd: nand: qcom: fix config error for BCH
Abhishek Sahu [Thu, 3 Aug 2017 15:56:39 +0000 (17:56 +0200)]
mtd: nand: qcom: fix config error for BCH

commit 10777de570016471fd929869c7830a7772893e39 upstream.

The configuration for BCH is not correct in the current driver.
The ECC_CFG_ECC_DISABLE bit defines whether to enable or disable the
BCH ECC in which

0x1 : BCH_DISABLED
0x0 : BCH_ENABLED

But currently host->bch_enabled is being assigned to BCH_DISABLED.

Fixes: c76b78d8ec05a ("mtd: nand: Qualcomm NAND controller driver")
Signed-off-by: Abhishek Sahu <absahu@codeaurora.org>
Reviewed-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomtd: nand: qcom: fix read failure without complete bootchain
Abhishek Sahu [Fri, 11 Aug 2017 11:39:16 +0000 (17:09 +0530)]
mtd: nand: qcom: fix read failure without complete bootchain

commit d8a9b320a26c1ea28e51e4f3ecfb593d5aac2910 upstream.

The NAND page read fails without complete boot chain since
NAND_DEV_CMD_VLD value is not proper. The default power on reset
value for this register is

    0xe - ERASE_START_VALID | WRITE_START_VALID | READ_STOP_VALID

The READ_START_VALID should be enabled for sending PAGE_READ
command. READ_STOP_VALID should be cleared since normal NAND
page read does not require READ_STOP command.

Fixes: c76b78d8ec05a ("mtd: nand: Qualcomm NAND controller driver")
Reviewed-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Abhishek Sahu <absahu@codeaurora.org>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomtd: nand: mxc: Fix mxc_v1 ooblayout
Boris Brezillon [Fri, 25 Nov 2016 10:32:32 +0000 (11:32 +0100)]
mtd: nand: mxc: Fix mxc_v1 ooblayout

commit 3bff08dffe3115a25ce04b95ea75f6d868572c60 upstream.

Commit a894cf6c5a82 ("mtd: nand: mxc: switch to mtd_ooblayout_ops")
introduced a bug in the OOB layout description. Even if the driver claims
that 3 ECC bytes are reserved to protect 512 bytes of data, it's actually
5 ECC bytes to protect 512+6 bytes of data (some OOB bytes are also
protected using extra ECC bytes).

Fix the mxc_v1_ooblayout_{free,ecc}() functions to reflect this behavior.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Fixes: a894cf6c5a82 ("mtd: nand: mxc: switch to mtd_ooblayout_ops")
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoLinux 4.9.49 v4.9.49
Greg Kroah-Hartman [Sun, 10 Sep 2017 05:49:19 +0000 (07:49 +0200)]
Linux 4.9.49

7 years agodrm/bridge: adv7511: Switch to using drm_kms_helper_hotplug_event()
John Stultz [Tue, 17 Jan 2017 00:52:48 +0000 (16:52 -0800)]
drm/bridge: adv7511: Switch to using drm_kms_helper_hotplug_event()

commit 6d5104c5a6b56385426e15047050584794bb6254 upstream.

In chasing down a previous issue with EDID probing from calling
drm_helper_hpd_irq_event() from irq context, Laurent noticed
that the DRM documentation suggests that
drm_kms_helper_hotplug_event() should be used instead.

Thus this patch replaces drm_helper_hpd_irq_event() with
drm_kms_helper_hotplug_event(), which requires we update the
connector.status entry and only call _hotplug_event() when the
status changes.

Cc: David Airlie <airlied@linux.ie>
Cc: Archit Taneja <architt@codeaurora.org>
Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Link: http://patchwork.freedesktop.org/patch/msgid/1484614372-15342-3-git-send-email-john.stultz@linaro.org
Signed-off-by: Thong Ho <thong.ho.px@rvc.renesas.com>
Signed-off-by: Nhan Nguyen <nhan.nguyen.yb@renesas.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agodrm/bridge: adv7511: Use work_struct to defer hotplug handing to out of irq context
John Stultz [Tue, 17 Jan 2017 00:52:47 +0000 (16:52 -0800)]
drm/bridge: adv7511: Use work_struct to defer hotplug handing to out of irq context

commit 518cb7057a59b9441336d2e88a396d52b6ab0cce upstream.

I was recently seeing issues with EDID probing, where
the logic to wait for the EDID read bit to be set by the
IRQ wasn't happening and the code would time out and fail.

Digging deeper, I found this was due to the fact that
IRQs were disabled as we were running in IRQ context from
the HPD signal.

Thus this patch changes the logic to handle the HPD signal
via a work_struct so we can be out of irq context.

With this patch, the EDID probing on hotplug does not time
out.

Cc: David Airlie <airlied@linux.ie>
Cc: Archit Taneja <architt@codeaurora.org>
Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Link: http://patchwork.freedesktop.org/patch/msgid/1484614372-15342-2-git-send-email-john.stultz@linaro.org
Signed-off-by: Thong Ho <thong.ho.px@rvc.renesas.com>
Signed-off-by: Nhan Nguyen <nhan.nguyen.yb@renesas.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoscsi: sg: recheck MMAP_IO request length with lock held
Todd Poynor [Wed, 16 Aug 2017 04:48:43 +0000 (21:48 -0700)]
scsi: sg: recheck MMAP_IO request length with lock held

commit 8d26f491116feaa0b16de370b6a7ba40a40fa0b4 upstream.

Commit 1bc0eb044615 ("scsi: sg: protect accesses to 'reserved' page
array") adds needed concurrency protection for the "reserve" buffer.
Some checks that are initially made outside the lock are replicated once
the lock is taken to ensure the checks and resulting decisions are made
using consistent state.

The check that a request with flag SG_FLAG_MMAP_IO set fits in the
reserve buffer also needs to be performed again under the lock to ensure
the reserve buffer length compared against matches the value in effect
when the request is linked to the reserve buffer.  An -ENOMEM should be
returned in this case, instead of switching over to an indirect buffer
as for non-MMAP_IO requests.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoscsi: sg: protect against races between mmap() and SG_SET_RESERVED_SIZE
Todd Poynor [Wed, 16 Aug 2017 05:41:08 +0000 (22:41 -0700)]
scsi: sg: protect against races between mmap() and SG_SET_RESERVED_SIZE

commit 6a8dadcca81fceff9976e8828cceb072873b7bd5 upstream.

Take f_mutex around mmap() processing to protect against races with the
SG_SET_RESERVED_SIZE ioctl.  Ensure the reserve buffer length remains
consistent during the mapping operation, and set the "mmap called" flag
to prevent further changes to the reserved buffer size as an atomic
operation with the mapping.

[mkp: fixed whitespace]

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agocs5536: add support for IDE controller variant
Andrey Korolyov [Thu, 10 Aug 2017 10:21:14 +0000 (13:21 +0300)]
cs5536: add support for IDE controller variant

commit 591b6bb605785c12a21e8b07a08a277065b655a5 upstream.

Several legacy devices such as Geode-based Cisco ASA appliances
and DB800 development board do possess CS5536 IDE controller
with different PCI id than existing one. Using pata_generic is
not always feasible as at least DB800 requires MSR quirk from
pata_cs5536 to be used with vendor firmware.

Signed-off-by: Andrey Korolyov <andrey@xdel.ru>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoworkqueue: Fix flag collision
Ben Hutchings [Sun, 3 Sep 2017 00:18:41 +0000 (01:18 +0100)]
workqueue: Fix flag collision

commit fbf1c41fc0f4d3574ac2377245efd666c1fa3075 upstream.

Commit 0a94efb5acbb ("workqueue: implicit ordered attribute should be
overridable") introduced a __WQ_ORDERED_EXPLICIT flag but gave it the
same value as __WQ_LEGACY.  I don't believe these were intended to
mean the same thing, so renumber __WQ_ORDERED_EXPLICIT.

Fixes: 0a94efb5acbb ("workqueue: implicit ordered attribute should be ...")
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agodrm/nouveau/pci/msi: disable MSI on big-endian platforms by default
Ilia Mirkin [Thu, 10 Aug 2017 16:13:40 +0000 (12:13 -0400)]
drm/nouveau/pci/msi: disable MSI on big-endian platforms by default

commit bc60c90f472b6e762ea96ef384072145adc8d4af upstream.

It appears that MSI does not work on either G5 PPC nor on a E5500-based
platform, where other hardware is reported to work fine with MSI.

Both tests were conducted with NV4x hardware, so perhaps other (or even
this) hardware can be made to work. It's still possible to force-enable
with config=NvMSI=1 on load.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agos390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs
Christian Borntraeger [Thu, 24 Aug 2017 10:55:08 +0000 (12:55 +0200)]
s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs

commit fa41ba0d08de7c975c3e94d0067553f9b934221f upstream.

Right now there is a potential hang situation for postcopy migrations,
if the guest is enabling storage keys on the target system during the
postcopy process.

For storage key virtualization, we have to forbid the empty zero page as
the storage key is a property of the physical page frame.  As we enable
storage key handling lazily we then drop all mappings for empty zero
pages for lazy refaulting later on.

This does not work with the postcopy migration, which relies on the
empty zero page never triggering a fault again in the future. The reason
is that postcopy migration will simply read a page on the target system
if that page is a known zero page to fault in an empty zero page.  At
the same time postcopy remembers that this page was already transferred
- so any future userfault on that page will NOT be retransmitted again
to avoid races.

If now the guest enters the storage key mode while in postcopy, we will
break this assumption of postcopy.

The solution is to disable the empty zero page for KVM guests early on
and not during storage key enablement. With this change, the postcopy
migration process is guaranteed to start after no zero pages are left.

As guest pages are very likely not empty zero pages anyway the memory
overhead is also pretty small.

While at it this also adds proper page table locking to the zero page
removal.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Janosch Frank <frankja@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoMCB: add support for SC31 to mcb-lpc
Michael Moese [Tue, 29 Aug 2017 12:47:24 +0000 (14:47 +0200)]
MCB: add support for SC31 to mcb-lpc

commit acf5e051ac44d5dc60b21bc4734ef1b844d55551 upstream.

This patch adds the resources and DMI ID's for the MEN SC31,
which uses a different address region to map the LPC bus than
the one used for the existing SC24.

Signed-off-by: Michael Moese <michael.moese@men.de>
[jth add stable tag]
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agomwifiex: correct channel stat buffer overflows
Brian Norris [Fri, 30 Jun 2017 01:23:54 +0000 (18:23 -0700)]
mwifiex: correct channel stat buffer overflows

commit 4b5dde2d6234ff5bc68e97e6901d1f2a0a7f3749 upstream.

mwifiex records information about various channels as it receives scan
information. It does this by appending to a buffer that was sized
to the max number of supported channels on any band, but there are
numerous problems:

(a) scans can return info from more than one band (e.g., both 2.4 and 5
    GHz), so the determined "max" is not large enough
(b) some firmware appears to return multiple results for a given
    channel, so the max *really* isn't large enough
(c) there is no bounds checking when stashing these stats, so problems
    (a) and (b) can easily lead to buffer overflows

Let's patch this by setting a slightly-more-correct max (that accounts
for a combination of both 2.4G and 5G bands) and adding a bounds check
when writing to our statistics buffer.

Due to problem (b), we still might not properly report all known survey
information (e.g., with "iw <dev> survey dump"), since duplicate results
(or otherwise "larger than expected" results) will cause some
truncation. But that's a problem for a future bugfix.

(And because of this known deficiency, only log the excess at the WARN
level, since that isn't visible by default in this driver and would
otherwise be a bit too noisy.)

Fixes: bf35443314ac ("mwifiex: channel statistics support for mwifiex")
Cc: Avinash Patil <patila@marvell.com>
Cc: Xinming Hu <huxm@marvell.com>
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Reviewed-by: Ganapathi Bhat <gbhat@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agodlm: avoid double-free on error path in dlm_device_{register,unregister}
Edwin Török [Thu, 3 Aug 2017 09:30:06 +0000 (10:30 +0100)]
dlm: avoid double-free on error path in dlm_device_{register,unregister}

commit 55acdd926f6b21a5cdba23da98a48aedf19ac9c3 upstream.

Can be reproduced when running dlm_controld (tested on 4.4.x, 4.12.4):
 # seq 1 100 | xargs -P0 -n1 dlm_tool join
 # seq 1 100 | xargs -P0 -n1 dlm_tool leave

misc_register fails due to duplicate sysfs entry, which causes
dlm_device_register to free ls->ls_device.name.
In dlm_device_deregister the name was freed again, causing memory
corruption.

According to the comment in dlm_device_deregister the name should've been
set to NULL when registration fails,
so this patch does that.

sysfs: cannot create duplicate filename '/dev/char/10:1'
------------[ cut here ]------------
warning: cpu: 1 pid: 4450 at fs/sysfs/dir.c:31 sysfs_warn_dup+0x56/0x70
modules linked in: msr rfcomm dlm ccm bnep dm_crypt uvcvideo
videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 videobuf2_core videodev
btusb media btrtl btbcm btintel bluetooth ecdh_generic intel_rapl
x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm
snd_hda_codec_hdmi irqbypass crct10dif_pclmul crc32_pclmul
ghash_clmulni_intel thinkpad_acpi pcbc nvram snd_seq_midi
snd_seq_midi_event aesni_intel snd_hda_codec_realtek snd_hda_codec_generic
snd_rawmidi aes_x86_64 crypto_simd glue_helper snd_hda_intel snd_hda_codec
cryptd intel_cstate arc4 snd_hda_core snd_seq snd_seq_device snd_hwdep
iwldvm intel_rapl_perf mac80211 joydev input_leds iwlwifi serio_raw
cfg80211 snd_pcm shpchp snd_timer snd mac_hid mei_me lpc_ich mei soundcore
sunrpc parport_pc ppdev lp parport autofs4 i915 psmouse
 e1000e ahci libahci i2c_algo_bit sdhci_pci ptp drm_kms_helper sdhci
pps_core syscopyarea sysfillrect sysimgblt fb_sys_fops drm wmi video
cpu: 1 pid: 4450 comm: dlm_test.exe not tainted 4.12.4-041204-generic
hardware name: lenovo 232425u/232425u, bios g2et82ww (2.02 ) 09/11/2012
task: ffff96b0cbabe140 task.stack: ffffb199027d0000
rip: 0010:sysfs_warn_dup+0x56/0x70
rsp: 0018:ffffb199027d3c58 eflags: 00010282
rax: 0000000000000038 rbx: ffff96b0e2c49158 rcx: 0000000000000006
rdx: 0000000000000000 rsi: 0000000000000086 rdi: ffff96b15e24dcc0
rbp: ffffb199027d3c70 r08: 0000000000000001 r09: 0000000000000721
r10: ffffb199027d3c00 r11: 0000000000000721 r12: ffffb199027d3cd1
r13: ffff96b1592088f0 r14: 0000000000000001 r15: ffffffffffffffef
fs:  00007f78069c0700(0000) gs:ffff96b15e240000(0000)
knlgs:0000000000000000
cs:  0010 ds: 0000 es: 0000 cr0: 0000000080050033
cr2: 000000178625ed28 cr3: 0000000091d3e000 cr4: 00000000001406e0
call trace:
 sysfs_do_create_link_sd.isra.2+0x9e/0xb0
 sysfs_create_link+0x25/0x40
 device_add+0x5a9/0x640
 device_create_groups_vargs+0xe0/0xf0
 device_create_with_groups+0x3f/0x60
 ? snprintf+0x45/0x70
 misc_register+0x140/0x180
 device_write+0x6a8/0x790 [dlm]
 __vfs_write+0x37/0x160
 ? apparmor_file_permission+0x1a/0x20
 ? security_file_permission+0x3b/0xc0
 vfs_write+0xb5/0x1a0
 sys_write+0x55/0xc0
 ? sys_fcntl+0x5d/0xb0
 entry_syscall_64_fastpath+0x1e/0xa9
rip: 0033:0x7f78083454bd
rsp: 002b:00007f78069bbd30 eflags: 00000293 orig_rax: 0000000000000001
rax: ffffffffffffffda rbx: 0000000000000006 rcx: 00007f78083454bd
rdx: 000000000000009c rsi: 00007f78069bee00 rdi: 0000000000000005
rbp: 00007f77f8000a20 r08: 000000000000fcf0 r09: 0000000000000032
r10: 0000000000000024 r11: 0000000000000293 r12: 00007f78069bde00
r13: 00007f78069bee00 r14: 000000000000000a r15: 00007f78069bbd70
code: 85 c0 48 89 c3 74 12 b9 00 10 00 00 48 89 c2 31 f6 4c 89 ef e8 2c c8
ff ff 4c 89 e2 48 89 de 48 c7 c7 b0 8e 0c a8 e8 41 e8 ed ff <0f> ff 48 89
df e8 00 d5 f4 ff 5b 41 5c 41 5d 5d c3 66 0f 1f 84
---[ end trace 40412246357cc9e0 ]---

dlm: 59f24629-ae39-44e2-9030-397ebc2eda26: leaving the lockspace group...
bug: unable to handle kernel null pointer dereference at 0000000000000001
ip: [<ffffffff811a3b4a>] kmem_cache_alloc+0x7a/0x140
pgd 0
oops: 0000 [#1] smp
modules linked in: dlm 8021q garp mrp stp llc openvswitch nf_defrag_ipv6
nf_conntrack libcrc32c iptable_filter dm_multipath crc32_pclmul dm_mod
aesni_intel psmouse aes_x86_64 sg ablk_helper cryptd lrw gf128mul
glue_helper i2c_piix4 nls_utf8 tpm_tis tpm isofs nfsd auth_rpcgss
oid_registry nfs_acl lockd grace sunrpc xen_wdt ip_tables x_tables autofs4
hid_generic usbhid hid sr_mod cdrom sd_mod ata_generic pata_acpi 8139too
serio_raw ata_piix 8139cp mii uhci_hcd ehci_pci ehci_hcd libata
scsi_dh_rdac scsi_dh_hp_sw scsi_dh_emc scsi_dh_alua scsi_mod ipv6
cpu: 0 pid: 394 comm: systemd-udevd tainted: g w 4.4.0+0 #1
hardware name: xen hvm domu, bios 4.7.2-2.2 05/11/2017
task: ffff880002410000 ti: ffff88000243c000 task.ti: ffff88000243c000
rip: e030:[<ffffffff811a3b4a>] [<ffffffff811a3b4a>]
kmem_cache_alloc+0x7a/0x140
rsp: e02b:ffff88000243fd90 eflags: 00010202
rax: 0000000000000000 rbx: ffff8800029864d0 rcx: 000000000007b36c
rdx: 000000000007b36b rsi: 00000000024000c0 rdi: ffff880036801c00
rbp: ffff88000243fdc0 r08: 0000000000018880 r09: 0000000000000054
r10: 000000000000004a r11: ffff880034ace6c0 r12: 00000000024000c0
r13: ffff880036801c00 r14: 0000000000000001 r15: ffffffff8118dcc2
fs: 00007f0ab77548c0(0000) gs:ffff880036e00000(0000) knlgs:0000000000000000
cs: e033 ds: 0000 es: 0000 cr0: 0000000080050033
cr2: 0000000000000001 cr3: 000000000332d000 cr4: 0000000000040660
stack:
ffffffff8118dc90 ffff8800029864d0 0000000000000000 ffff88003430b0b0
ffff880034b78320 ffff88003430b0b0 ffff88000243fdf8 ffffffff8118dcc2
ffff8800349c6700 ffff8800029864d0 000000000000000b 00007f0ab7754b90
call trace:
[<ffffffff8118dc90>] ? anon_vma_fork+0x60/0x140
[<ffffffff8118dcc2>] anon_vma_fork+0x92/0x140
[<ffffffff8107033e>] copy_process+0xcae/0x1a80
[<ffffffff8107128b>] _do_fork+0x8b/0x2d0
[<ffffffff81071579>] sys_clone+0x19/0x20
[<ffffffff815a30ae>] entry_syscall_64_fastpath+0x12/0x71
] code: f6 75 1c 4c 89 fa 44 89 e6 4c 89 ef e8 a7 e4 00 00 41 f7 c4 00 80
00 00 49 89 c6 74 47 eb 32 49 63 45 20 48 8d 4a 01 4d 8b 45 00 <49> 8b 1c
06 4c 89 f0 65 49 0f c7 08 0f 94 c0 84 c0 74 ac 49 63
rip [<ffffffff811a3b4a>] kmem_cache_alloc+0x7a/0x140
rsp <ffff88000243fd90>
cr2: 0000000000000001
--[ end trace 70cb9fd1b164a0e8 ]--

Signed-off-by: Edwin Török <edvin.torok@citrix.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
7 years agoiwlwifi: pci: add new PCI ID for 7265D
Luca Coelho [Wed, 16 Aug 2017 05:47:38 +0000 (08:47 +0300)]
iwlwifi: pci: add new PCI ID for 7265D

commit 3f7a5e13e85026b6e460bbd6e87f87379421d272 upstream.

We have a new PCI subsystem ID for 7265D.  Add it to the list.

Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>