The CCID3 code fails to initialize the trailing padding bytes of struct
tfrc_tx_info added for alignment on 64 bit architectures. It that for
potentially leaks four bytes kernel stack via the getsockopt() syscall.
Add an explicit memset(0) before filling the structure to avoid the
info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Gerrit Renker <gerrit@erg.abdn.ac.uk> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The LLC code wrongly returns 0, i.e. "success", when the socket is
zapped. Together with the uninitialized uaddrlen pointer argument from
sys_getsockname this leads to an arbitrary memory leak of up to 128
bytes kernel stack via the getsockname() syscall.
Return an error instead when the socket is zapped to prevent the info
leak. Also remove the unnecessary memset(0). We don't directly write to
the memory pointed by uaddr but memcpy() a local structure at the end of
the function that is properly initialized.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The L2CAP code fails to initialize the l2_bdaddr_type member of struct
sockaddr_l2 and the padding byte added for alignment. It that for leaks
two bytes kernel stack via the getsockname() syscall. Add an explicit
memset(0) before filling the structure to avoid the info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Marcel Holtmann <marcel@holtmann.org> Cc: Gustavo Padovan <gustavo@padovan.org> Cc: Johan Hedberg <johan.hedberg@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The RFCOMM code fails to initialize the trailing padding byte of struct
sockaddr_rc added for alignment. It that for leaks one byte kernel stack
via the getsockname() syscall. Add an explicit memset(0) before filling
the structure to avoid the info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Marcel Holtmann <marcel@holtmann.org> Cc: Gustavo Padovan <gustavo@padovan.org> Cc: Johan Hedberg <johan.hedberg@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The RFCOMM code fails to initialize the two padding bytes of struct
rfcomm_dev_list_req inserted for alignment before copying it to
userland. Additionally there are two padding bytes in each instance of
struct rfcomm_dev_info. The ioctl() that for disclosures two bytes plus
dev_num times two bytes uninitialized kernel heap memory.
Allocate the memory using kzalloc() to fix this issue.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Marcel Holtmann <marcel@holtmann.org> Cc: Gustavo Padovan <gustavo@padovan.org> Cc: Johan Hedberg <johan.hedberg@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The HCI code fails to initialize the hci_channel member of struct
sockaddr_hci and that for leaks two bytes kernel stack via the
getsockname() syscall. Initialize hci_channel with 0 to avoid the
info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Marcel Holtmann <marcel@holtmann.org> Cc: Gustavo Padovan <gustavo@padovan.org> Cc: Johan Hedberg <johan.hedberg@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The HCI code fails to initialize the two padding bytes of struct
hci_ufilter before copying it to userland -- that for leaking two
bytes kernel stack. Add an explicit memset(0) before filling the
structure to avoid the info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Marcel Holtmann <marcel@holtmann.org> Cc: Gustavo Padovan <gustavo@padovan.org> Cc: Johan Hedberg <johan.hedberg@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ATM code fails to initialize the two padding bytes of struct
sockaddr_atmpvc inserted for alignment. Add an explicit memset(0)
before filling the structure to avoid the info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ATM code fails to initialize the two padding bytes of struct
sockaddr_atmpvc inserted for alignment. Add an explicit memset(0)
before filling the structure to avoid the info leak.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
addrconf_forward_change() uses RCU iteration over the netdev list,
which is unnecessary since it already holds the RTNL lock. We also
cannot reasonably require netdevice notifier functions not to sleep.
Reported-by: Cong Wang <amwang@redhat.com> Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Here's a quote of the comment about the BUG macro from asm-generic/bug.h:
Don't use BUG() or BUG_ON() unless there's really no way out; one
example might be detecting data structure corruption in the middle
of an operation that can't be backed out of. If the (sub)system
can somehow continue operating, perhaps with reduced functionality,
it's probably not BUG-worthy.
If you're tempted to BUG(), think again: is completely giving up
really the *only* solution? There are usually better options, where
users don't need to reboot ASAP and can mostly shut down cleanly.
In our case, the status flag of a ring buffer slot is managed from both sides,
the kernel space and the user space. This means that even though the kernel
side might work as expected, the user space screws up and changes this flag
right between the send(2) is triggered when the flag is changed to
TP_STATUS_SENDING and a given skb is destructed after some time. Then, this
will hit the BUG macro. As David suggested, the best solution is to simply
remove this statement since it cannot be used for kernel side internal
consistency checks. I've tested it and the system still behaves /stable/ in
this case, so in accordance with the above comment, we should rather remove it.
Signed-off-by: Daniel Borkmann <daniel.borkmann@tik.ee.ethz.ch> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Do not leak memory by updating pointer with potentially NULL realloc return value.
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
gact_rand array is accessed by gact->tcfg_ptype whose value
is assumed to less than MAX_RAND, but any range checks are
not performed.
So add a check in tcf_gact_init(). And in tcf_gact(), we can
reduce a branch.
Signed-off-by: Hiroaki SHIMODA <shimoda.hiroaki@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cache the device gso_max_segs in sock::sk_gso_max_segs and use it to
limit the size of TSO skbs. This avoids the need to fall back to
software GSO for local TCP senders.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently an skb requiring TSO may not fit within a minimum-size TX
queue. The TX queue selected for the skb may stall and trigger the TX
watchdog repeatedly (since the problem skb will be retried after the
TX reset). This issue is designated as CVE-2012-3412.
Set the maximum number of TSO segments for our devices to 100. This
should make no difference to behaviour unless the actual MSS is less
than about 700. Increase the minimum TX queue size accordingly to
allow for 2 worst-case skbs, so that there will definitely be space
to add an skb after we wake a queue.
To avoid invalidating existing configurations, change
efx_ethtool_set_ringparam() to fix up values that are too small rather
than returning -EINVAL.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A peer (or local user) may cause TCP to use a nominal MSS of as little
as 88 (actual MSS of 76 with timestamps). Given that we have a
sufficiently prodigious local sender and the peer ACKs quickly enough,
it is nevertheless possible to grow the window for such a connection
to the point that we will try to send just under 64K at once. This
results in a single skb that expands to 861 segments.
In some drivers with TSO support, such an skb will require hundreds of
DMA descriptors; a substantial fraction of a TX ring or even more than
a full ring. The TX queue selected for the skb may stall and trigger
the TX watchdog repeatedly (since the problem skb will be retried
after the TX reset). This particularly affects sfc, for which the
issue is designated as CVE-2012-3412.
Therefore:
1. Add the field net_device::gso_max_segs holding the device-specific
limit.
2. In netif_skb_features(), if the number of segments is too high then
mask out GSO features to force fall back to software GSO.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The old interface is bugged and reads the wrong sensor when retrieving
the reading for the chassis fan (it reads the CPU sensor); the new
interface works fine.
ccid_hc_rx_getsockopt() and ccid_hc_tx_getsockopt() might be called with
a NULL ccid pointer leading to a NULL pointer dereference. This could
lead to a privilege escalation if the attacker is able to map page 0 and
prepare it with a fake ccid_ops pointer.
Signed-off-by: Mathias Krause <minipli@googlemail.com> Cc: Gerrit Renker <gerrit@erg.abdn.ac.uk> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The following build error occured during a parisc build with
swap-over-NFS patches applied.
net/core/sock.c:274:36: error: initializer element is not constant
net/core/sock.c:274:36: error: (near initialization for 'memalloc_socks')
net/core/sock.c:274:36: error: initializer element is not constant
Dave Anglin says:
> Here is the line in sock.i:
>
> struct static_key memalloc_socks = ((struct static_key) { .enabled =
> ((atomic_t) { (0) }) });
The above line contains two compound literals. It also uses a designated
initializer to initialize the field enabled. A compound literal is not a
constant expression.
The location of the above statement isn't fully clear, but if a compound
literal occurs outside the body of a function, the initializer list must
consist of constant expressions.
They all define their chassis type as "Other" and therefore are not
categorized as "laptops" by the driver, which tries to perform AUX IRQ
delivery test which fails and causes touchpad not working.
Code tracking when transaction needs to be committed on fdatasync(2) forgets
to handle a situation when only inode's i_size is changed. Thus in such
situations fdatasync(2) doesn't force transaction with new i_size to disk
and that can result in wrong i_size after a crash.
Fix the issue by updating inode's i_datasync_tid whenever its size is
updated.
Reported-by: Kristian Nielsen <knielsen@knielsen-hq.org> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When a file is stored in ICB (inode), we overwrite part of the file, and
the page containing file's data is not in page cache, we end up corrupting
file's data by overwriting them with zeros. The problem is we use
simple_write_begin() which simply zeroes parts of the page which are not
written to. The problem has been introduced by be021ee4 (udf: convert to
new aops).
Fix the problem by providing a ->write_begin function which makes the page
properly uptodate.
Reported-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is a particularly nasty SCSI ATA Translation Layer (SATL) problem.
SAT-2 says (section 8.12.2)
if the device is in the stopped state as the result of
processing a START STOP UNIT command (see 9.11), then the SATL
shall terminate the TEST UNIT READY command with CHECK CONDITION
status with the sense key set to NOT READY and the additional
sense code of LOGICAL UNIT NOT READY, INITIALIZING COMMAND
REQUIRED;
mpt2sas internal SATL seems to implement this. The result is very confusing
standby behaviour (using hdparm -y). If you suspend a drive and then send
another command, usually it wakes up. However, if the next command is a TEST
UNIT READY, the SATL sees that the drive is suspended and proceeds to follow
the SATL rules for this, returning NOT READY to all subsequent commands. This
means that the ordering of TEST UNIT READY is crucial: if you send TUR and
then a command, you get a NOT READY to both back. If you send a command and
then a TUR, you get GOOD status because the preceeding command woke the drive.
block: flush MEDIA_CHANGE from drivers on close(2)
Changed our ordering on TEST UNIT READY commands meaning that SATA drives
connected to an mpt2sas now suspend and refuse to wake (because the mpt2sas
SATL sees the suspend *before* the drives get awoken by the next ATA command)
resulting in lots of failed commands.
The standard is completely nuts forcing this inconsistent behaviour, but we
have to work around it.
The fix for this is twofold:
1. Set the allow_restart flag so we wake the drive when we see it has been
suspended
2. Return all TEST UNIT READY status directly to the mid layer without any
further error handling which prevents us causing error handling which
may offline the device just because of a media check TUR.
If the specified max_queue_depth setting is less than the
expected number of internal commands, then driver will calculate
the queue depth size to a negitive number. This negitive number
is actually a very large number because variable is unsigned
16bit integer. So, the driver will ask for a very large amount of
memory for message frames and resulting into oops as memory
allocation routines will not able to handle such a large request.
So, in order to limit this kind of oops, The driver need to set
the max_queue_depth to a scsi mid layer's can_queue value. Then
the overall message frames required for IO is minimum of either
(max_queue_depth plus internal commands) or the IOC global
credits.
The following patch moves the poll_aen_lock initializer from
megasas_probe_one() to megasas_init(). This prevents a crash when a user
loads the driver and tries to issue a poll() system call on the ioctl
interface with no adapters present.
Signed-off-by: Kashyap Desai <Kashyap.Desai@lsi.com> Signed-off-by: Adam Radford <aradford@gmail.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
During a context switch we always restore the per thread DSCR value.
If we aren't doing explicit DSCR management
(ie thread.dscr_inherit == 0) and the default DSCR changed while
the process has been sleeping we end up with the wrong value.
Check thread.dscr_inherit and select the default DSCR or per thread
DSCR as required.
This was found with the following test case, when running with
more threads than CPUs (ie forcing context switching):
If the default DSCR is non zero we set thread.dscr_inherit in
copy_thread() meaning the new thread and all its children will ignore
future updates to the default DSCR. This is not intended and is
a change in behaviour that a number of our users have hit.
We just need to inherit thread.dscr and thread.dscr_inherit from
the parent which ends up being much simpler.
If a device specifies zero endpoints in its interface descriptor,
the kernel oopses in acm_probe(). Even though that's clearly an
invalid descriptor, we should test wether we have all endpoints.
This is especially bad as this oops can be triggered by just
plugging a USB device in.
Signed-off-by: Sven Schnelle <svens@stackframe.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Reported-by: Bjørn Mork <bjorn@mork.no> CC: Herton Ronaldo Krzesinski <herton@canonical.com> CC: Hin-Tak Leung <htl10@users.sourceforge.net> CC: Larry Finger <Larry.Finger@lwfinger.net> CC: "John W. Linville" <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
Commit 91f68c89d8f3 ("block: fix infinite loop in __getblk_slow")
is not good: a successful call to grow_buffers() cannot guarantee
that the page won't be reclaimed before the immediate next call to
__find_get_block(), which is why there was always a loop there.
Yesterday I got "EXT4-fs error (device loop0): __ext4_get_inode_loc:3595:
inode #19278: block 664: comm cc1: unable to read itable block" on console,
which pointed to this commit.
I've been trying to bisect for weeks, why kbuild-on-ext4-on-loop-on-tmpfs
sometimes fails from a missing header file, under memory pressure on
ppc G5. I've never seen this on x86, and I've never seen it on 3.5-rc7
itself, despite that commit being in there: bisection pointed to an
irrelevant pinctrl merge, but hard to tell when failure takes between
18 minutes and 38 hours (but so far it's happened quicker on 3.6-rc2).
(I've since found such __ext4_get_inode_loc errors in /var/log/messages
from previous weeks: why the message never appeared on console until
yesterday morning is a mystery for another day.)
Revert 91f68c89d8f3, restoring __getblk_slow() to how it was (plus
a checkpatch nitfix). Simplify the interface between grow_buffers()
and grow_dev_page(), and avoid the infinite loop beyond end of device
by instead checking init_page_buffers()'s end_block there (I presume
that's more efficient than a repeated call to blkdev_max_block()),
returning -ENXIO to __getblk_slow() in that case.
And remove akpm's ten-year-old "__getblk() cannot fail ... weird"
comment, but that is worrying: are all users of __getblk() really
now prepared for a NULL bh beyond end of device, or will some oops??
Commit dbf0e4c (PCI: EHCI: fix crash during suspend on ASUS
computers) added a workaround for an ASUS suspend issue related to
USB EHCI and a bug in a number of ASUS BIOSes that attempt to shut
down the EHCI controller during system suspend if its PCI command
register doesn't contain 0 at that time.
It turns out that the same workaround is necessary in the analogous
hibernation code path, so add it.
References: https://bugzilla.kernel.org/show_bug.cgi?id=45811 Reported-and-tested-by: Oleksij Rempel <bug-track@fisher-privat.net> Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ath_rx_tasklet() calls ath9k_rx_skb_preprocess() and ath9k_rx_skb_postprocess()
in a loop over the received frames. The decrypt_error flag is
initialized to false
just outside ath_rx_tasklet() loop. ath9k_rx_accept(), called by
ath9k_rx_skb_preprocess(),
only sets decrypt_error to true and never to false.
Then ath_rx_tasklet() calls ath9k_rx_skb_postprocess() and passes
decrypt_error to it.
So, after a decryption error, in ath9k_rx_skb_postprocess(), we can
have a leftover value
from another processed frame. In that case, the frame will not be marked with
RX_FLAG_DECRYPTED even if it is decrypted correctly.
When using CCMP encryption this issue can lead to connection stuck
because of CCMP
PN corruption and a waste of CPU time since mac80211 tries to decrypt an already
deciphered frame with ieee80211_aes_ccm_decrypt.
Fix the issue initializing decrypt_error flag at the begging of the
ath_rx_tasklet() loop.
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Delete code which sets SCSI status incorrectly as it's already been set
correctly above this incorrect code. The bug was introduced in 2009 by
commit b0e15f6db111 ("cciss: fix typo that causes scsi status to be
lost.")
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com> Reported-by: Roel van Meer <roel.vanmeer@bokxing.nl> Tested-by: Roel van Meer <roel.vanmeer@bokxing.nl> Cc: Jens Axboe <axboe@kernel.dk> 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>
svc_tcp_sendto sets XPT_CLOSE if we fail to transmit the entire reply.
However, the XPT_CLOSE won't be acted on immediately. Meanwhile other
threads could send further replies before the socket is really shut
down. This can manifest as data corruption: for example, if a truncated
read reply is followed by another rpc reply, that second reply will look
to the client like further read data.
Symptoms were data corruption preceded by svc_tcp_sendto logging
something like
kernel: rpc-srv/tcp: nfsd: sent only 963696 when sending 1048708 bytes - shutting down socket
Reported-by: Malahal Naineni <malahal@us.ibm.com> Tested-by: Malahal Naineni <malahal@us.ibm.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The rpc server tries to ensure that there will be room to send a reply
before it receives a request.
It does this by tracking, in xpt_reserved, an upper bound on the total
size of the replies that is has already committed to for the socket.
Currently it is adding in the estimate for a new reply *before* it
checks whether there is space available. If it finds that there is not
space, it then subtracts the estimate back out.
This may lead the subsequent svc_xprt_enqueue to decide that there is
space after all.
The results is a svc_recv() that will repeatedly return -EAGAIN, causing
server threads to loop without doing any actual work.
Reported-by: Michael Tokarev <mjt@tls.msk.ru> Tested-by: Michael Tokarev <mjt@tls.msk.ru> Signed-off-by: J. Bruce Fields <bfields@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Examination of svc_tcp_clear_pages shows that it assumes sk_tcplen is
consistent with sk_pages[] (in particular, sk_pages[n] can't be NULL if
sk_tcplen would lead us to expect n pages of data).
svc_tcp_restore_pages zeroes out sk_pages[] while leaving sk_tcplen.
This is OK, since both functions are serialized by XPT_BUSY. However,
that means the inconsistency must be repaired before dropping XPT_BUSY.
Therefore we should be ensuring that svc_tcp_save_pages repairs the
problem before exiting svc_tcp_recv_record on error.
Symptoms were a BUG() in svc_tcp_clear_pages.
Signed-off-by: J. Bruce Fields <bfields@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This was reported by various people as triggering Oops when stopping auditd.
We could just remove the put_mark from audit_tree_freeing_mark() but that would
break freeing via inode destruction. So this patch simply omits a put_mark
after calling destroy_mark or adds a get_mark before.
The additional get_mark is necessary where there's no other put_mark after
fsnotify_destroy_mark() since it assumes that the caller is holding a reference
(or the inode is keeping the mark pinned, not the case here AFAICS).
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Reported-by: Valentin Avram <aval13@gmail.com> Reported-by: Peter Moody <pmoody@google.com> Acked-by: Eric Paris <eparis@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ever since commit 0a57cdac3f (NFSv4.1 send layoutreturn to fence
disconnected data server) we've been sending layoutreturn calls
while there is potentially still outstanding I/O to the data
servers. The reason we do this is to avoid races between replayed
writes to the MDS and the original writes to the DS.
When this happens, the BUG_ON() in nfs4_layoutreturn_done can
be triggered because it assumes that we would never call
layoutreturn without knowing that all I/O to the DS is
finished. The fix is to remove the BUG_ON() now that the
assumptions behind the test are obsolete.
If the rpc call to NFS3PROC_FSINFO fails, then we need to report that
error so that the mount fails. Otherwise we can end up with a
superblock with completely unusable values for block sizes, maxfilesize,
etc.
Each page mapped in a process's address space must be correctly
accounted for in _mapcount. Normally the rules for this are
straightforward but hugetlbfs page table sharing is different. The page
table pages at the PMD level are reference counted while the mapcount
remains the same.
If this accounting is wrong, it causes bugs like this one reported by
Larry Woodman:
kernel BUG at mm/filemap.c:135!
invalid opcode: 0000 [#1] SMP
CPU 22
Modules linked in: bridge stp llc sunrpc binfmt_misc dcdbas microcode pcspkr acpi_pad acpi]
Pid: 18001, comm: mpitest Tainted: G W 3.3.0+ #4 Dell Inc. PowerEdge R620/07NDJ2
RIP: 0010:[<ffffffff8112cfed>] [<ffffffff8112cfed>] __delete_from_page_cache+0x15d/0x170
Process mpitest (pid: 18001, threadinfo ffff880428972000, task ffff880428b5cc20)
Call Trace:
delete_from_page_cache+0x40/0x80
truncate_hugepages+0x115/0x1f0
hugetlbfs_evict_inode+0x18/0x30
evict+0x9f/0x1b0
iput_final+0xe3/0x1e0
iput+0x3e/0x50
d_kill+0xf8/0x110
dput+0xe2/0x1b0
__fput+0x162/0x240
During fork(), copy_hugetlb_page_range() detects if huge_pte_alloc()
shared page tables with the check dst_pte == src_pte. The logic is if
the PMD page is the same, they must be shared. This assumes that the
sharing is between the parent and child. However, if the sharing is
with a different process entirely then this check fails as in this
diagram:
For this situation to occur, it must be possible for Parent and Other to
have faulted and failed to share page tables with each other. This is
possible due to the following style of race.
PROC A PROC B
copy_hugetlb_page_range copy_hugetlb_page_range
src_pte == huge_pte_offset src_pte == huge_pte_offset
!src_pte so no sharing !src_pte so no sharing
These two processes are not poing to the same data page but are not
sharing page tables because the opportunity was missed. When either
process later forks, the src_pte == dst pte is potentially insufficient.
As the check falls through, the wrong PTE information is copied in
(harmless but wrong) and the mapcount is bumped for a page mapped by a
shared page table leading to the BUG_ON.
This patch addresses the issue by moving pmd_alloc into huge_pmd_share
which guarantees that the shared pud is populated in the same critical
section as pmd. This also means that huge_pte_offset test in
huge_pmd_share is serialized correctly now which in turn means that the
success of the sharing will be higher as the racing tasks see the pud
and pmd populated together.
Race identified and changelog written mostly by Mel Gorman.
{akpm@linux-foundation.org: attempt to make the huge_pmd_share() comment comprehensible, clean up coding style] Reported-by: Larry Woodman <lwoodman@redhat.com> Tested-by: Larry Woodman <lwoodman@redhat.com> Reviewed-by: Mel Gorman <mgorman@suse.de> Signed-off-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Rik van Riel <riel@redhat.com> Cc: David Gibson <david@gibson.dropbear.id.au> Cc: Ken Chen <kenchen@google.com> Cc: Cong Wang <xiyou.wangcong@gmail.com> Cc: Hillf Danton <dhillf@gmail.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>
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
Currently we export SOCK_NONBLOCK to user space but that conflicts with
the definition from glibc leading to compilation errors in user programs
(e.g. see Debian bug #658460).
The generic socket.h restricts the definition of SOCK_NONBLOCK to the
kernel, as does the MIPS specific socket.h, so let's do the same on
Alpha.
Signed-off-by: Michael Cree <mcree@orcon.net.nz> Acked-by: Matt Turner <mattst88@gmail.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the microphone input source is not selectable as while there is
a DAPM widget it's not connected to anything so it won't be properly
instantiated. Add something more correct for the input structure to get
things going, even though it's not hooked into the rest of the routing
map and so won't actually achieve anything except allowing the relevant
register bits to be written.
Reported-by: Christop Fritz <chf.fritz@googlemail.com> Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The i.MX cpufreq implementation uses the CPU_FREQ_TABLE helpers,
so it needs to select that code to be built. This problem has
apparently existed since the i.MX cpufreq code was first merged
in v2.6.37.
Building IMX without CPU_FREQ_TABLE results in:
arch/arm/plat-mxc/built-in.o: In function `mxc_cpufreq_exit':
arch/arm/plat-mxc/cpufreq.c:173: undefined reference to `cpufreq_frequency_table_put_attr'
arch/arm/plat-mxc/built-in.o: In function `mxc_set_target':
arch/arm/plat-mxc/cpufreq.c:84: undefined reference to `cpufreq_frequency_table_target'
arch/arm/plat-mxc/built-in.o: In function `mxc_verify_speed':
arch/arm/plat-mxc/cpufreq.c:65: undefined reference to `cpufreq_frequency_table_verify'
arch/arm/plat-mxc/built-in.o: In function `mxc_cpufreq_init':
arch/arm/plat-mxc/cpufreq.c:154: undefined reference to `cpufreq_frequency_table_cpuinfo'
arch/arm/plat-mxc/cpufreq.c:162: undefined reference to `cpufreq_frequency_table_get_attr'
Commit d670ac019f60 (ARM: SAMSUNG: DMA Cleanup as per sparse) changed the
prototype of the s3c2410_dma_* functions to use the enum dma_ch instead
of an generic unsigned int.
In the s3c24xx dma.c s3c2410_dma_enqueue seems to have been forgotten,
the other functions there were changed correctly.
Commit 5a783cbc4836 ("ARM: 7478/1: errata: extend workaround for erratum
#720789") added workarounds for erratum #720789 to the range TLB
invalidation functions with the observation that the erratum only
affects SMP platforms. However, when running an SMP_ON_UP kernel on a
uniprocessor platform we must take care to preserve the ASID as the
workaround is not required.
This patch ensures that we don't set the ASID to 0 when flushing the TLB
on such a system, preserving the original behaviour with the workaround
disabled.
Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Page migration encodes the pfn in the offset field of a swp_entry_t.
For LPAE, we support physical addresses of up to 36 bits (due to
sparsemem limitations with the size of page flags), requiring 24 bits
to represent a pfn. A further 3 bits are used to encode a swp_entry into
a pte, leaving 5 bits for the type field. Furthermore, the core code
defines MAX_SWAPFILES_SHIFT as 5, so the additional type bit does not
get used.
This patch reduces the width of the type field to 5 bits, allowing us
to create up to 31 swapfiles of 64GB each.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Swap entries are encoding in ptes such that !pte_present(pte) and
pte_file(pte). The remaining bits of the descriptor are used to identify
the swapfile and offset within it to the swap entry.
When writing such a pte for a user virtual address, set_pte_at
unconditionally sets the nG bit, which (in the case of LPAE) will
corrupt the swapfile offset and lead to a BUG:
[ 140.494067] swap_free: Unused swap offset entry 000763b4
[ 140.509989] BUG: Bad page map in process rs:main Q:Reg pte:0ec76800 pmd:8f92e003
This patch fixes the problem by only setting the nG bit for user
mappings that are actually present.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Reported-by: Bjørn Mork <bjorn@mork.no> CC: Paul Gortmaker <paul.gortmaker@windriver.com> CC: Andrew Morton <akpm@linux-foundation.org> CC: Felipe Balbi <balbi@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This structure needs to always stick around, even if CONFIG_HOTPLUG
is disabled, otherwise we can oops when trying to probe a device that
was added after the structure is thrown away.
Thanks to Fengguang Wu and Bjørn Mork for tracking this issue down.
Reported-by: Fengguang Wu <fengguang.wu@intel.com> Reported-by: Bjørn Mork <bjorn@mork.no> CC: Forest Bond <forest@alittletooquiet.net> CC: Marcos Paulo de Souza <marcos.souza.org@gmail.com> CC: "David S. Miller" <davem@davemloft.net> CC: Jesper Juhl <jj@chaosbits.net> CC: Jiri Pirko <jpirko@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Avoid a crash caused by the scmnd->scsi_done(scmnd) call in
srp_process_rsp() being invoked with scsi_done == NULL. This can
happen if a reply is received during or after a command abort.
Reported-by: Joseph Glanville <joseph.glanville@orionvm.com.au>
Reference: http://marc.info/?l=linux-rdma&m=134314367801595 Acked-by: David Dillow <dillowda@ornl.gov> Signed-off-by: Bart Van Assche <bvanassche@acm.org> Signed-off-by: Roland Dreier <roland@purestorage.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Jeongdo Son <sohn9086@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
* Use the buffer content length as opposed to the total buffer size. This can
be a real problem when using the mos7840 as a usb serial-console as all
kernel output is truncated during boot.
Signed-off-by: Mark Ferrell <mferrell@uplogix.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In this patch, we add new declarations into option.c to support the new
interfaces of Huawei Data Card devices. And at the same time, remove the
redundant declarations from option.c.
A lot of Broadcom Bluetooth devices provides vendor specific interface
class and we are getting flooded by patches adding new device support.
This change will help us enable support for any other Broadcom with vendor
specific device that arrives in the future.
Only the product id changes for those devices, so this macro would be
perfect for us:
The Intel desktop boards DH77EB and DH77DF have a hardware issue that
can be worked around by BIOS. If the USB ports are switched to xHCI on
shutdown, the xHCI host will send a spurious interrupt, which will wake
the system. Some BIOS will work around this, but not all.
The bug can be avoided if the USB ports are switched back to EHCI on
shutdown. The Intel Windows driver switches the ports back to EHCI, so
change the Linux xHCI driver to do the same.
Unfortunately, we can't tell the two effected boards apart from other
working motherboards, because the vendors will change the DMI strings
for the DH77EB and DH77DF boards to their own custom names. One example
is Compulab's mini-desktop, the Intense-PC. Instead, key off the
Panther Point xHCI host PCI vendor and device ID, and switch the ports
over for all PPT xHCI hosts.
The only impact this will have on non-effected boards is to add a couple
hundred milliseconds delay on boot when the BIOS has to switch the ports
over from EHCI to xHCI.
This patch should be backported to kernels as old as 3.0, that contain
the commit 69e848c2090aebba5698a1620604c7dccb448684 "Intel xhci: Support
EHCI/xHCI port switching."
The NEC/Renesas 720201 xHCI host controller does not complete its reset
within 250 milliseconds. In fact, it takes about 9 seconds to reset the
host controller, and 1 second for the host to be ready for doorbell
rings. Extend the reset and CNR polling timeout to 10 seconds each.
This patch should be backported to kernels as old as 2.6.31, that
contain the commit 66d4eadd8d067269ea8fead1a50fe87c2979a80d "USB: xhci:
BIOS handoff and HW initialization."
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com> Reported-by: Edwin Klein Mentink <e.kleinmentink@zonnet.nl> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Gary reports that with recent kernels, he notices more xHCI driver
warnings:
xhci_hcd 0000:03:00.0: WARN Successful completion on short TX: needs XHCI_TRUST_TX_LENGTH quirk?
We think his Etron xHCI host controller may have the same buggy behavior
as the Fresco Logic xHCI host. When a short transfer is received, the
host will mark the transfer as successfully completed when it should be
marking it with a short completion.
Fix this by turning on the XHCI_TRUST_TX_LENGTH quirk when the Etron
host is discovered. Note that Gary has revision 1, but if Etron fixes
this bug in future revisions, the quirk will have no effect.
This patch should be backported to kernels as old as 2.6.36, that
contain a backported version of commit 1530bbc6272d9da1e39ef8e06190d42c13a02733 "xhci: Add new short TX quirk
for Fresco Logic host."
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com> Reported-by: Gary E. Miller <gem@rellim.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 03179fe923 introduced a kmemcheck complaint in
ext4_da_get_block_prep() because we save and restore
ei->i_da_metadata_calc_last_lblock even though it is left
uninitialized in the case where i_da_metadata_calc_len is zero.
This doesn't hurt anything, but silencing the kmemcheck complaint
makes it easier for people to find real bugs.
Addresses https://bugzilla.kernel.org/show_bug.cgi?id=45631
(which is marked as a regression).
It seems we can not update the crtc scanout address. After disabling
crtc, update to base address do not take effect after crtc being
reenable leading to at least frame being scanout from the old crtc
base address. Disabling crtc display request lead to same behavior.
So after changing the vram address if we don't keep crtc disabled
we will have the GPU trying to read some random system memory address
with some iommu this will broke the crtc engine and will lead to
broken display and iommu error message.
So to avoid this, disable crtc. For flicker less boot we will need
to avoid moving the vram start address.
We may only start to set up the new register values after having
confirmed that the ring is truely off. Otherwise the hw might lose the
newly written register values. This is caught later on in the init
sequence, when we check whether the register writes have stuck.
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=50522 Tested-by: Yang Guang <guang.a.yang@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the frontend and the backend reside on the same domain, even if we
add pages to the m2p_override, these pages will never be returned by
mfn_to_pfn because the check "get_phys_to_machine(pfn) != mfn" will
always fail, so the pfn of the frontend will be returned instead
(resulting in a deadlock because the frontend pages are already locked).
The explanation is in the comment within the code:
We need to do this because the pages shared by the frontend
(xen-blkfront) can be already locked (lock_page, called by
do_read_cache_page); when the userspace backend tries to use them
with direct_IO, mfn_to_pfn returns the pfn of the frontend, so
do_blockdev_direct_IO is going to try to lock the same pages
again resulting in a deadlock.
Internally the xen-blkback uses m2p_add_override to swizzle (temporarily)
a 'struct page' to have a different MFN (so that it can point to another
guest). It also can easily find out whether another pfn corresponding
to the mfn exists in the m2p, and can set the FOREIGN bit
in the p2m, making sure that mfn_to_pfn returns the pfn of the backend.
This allows the backend to perform direct_IO on these pages, but as a
side effect prevents the frontend from using get_user_pages_fast on
them while they are being shared with the backend.
Commit 7572777eef78ebdee1ecb7c258c0ef94d35bad16 attempted to verify that
the total iovec from the client doesn't overflow iov_length() but it
only checked the first element. The iovec could still overflow by
starting with a small element. The obvious fix is to check all the
elements.
The overflow case doesn't look dangerous to the kernel as the copy is
limited by the length after the overflow. This fix restores the
intention of returning an error instead of successfully copying less
than the iovec represented.
I found this by code inspection. I built it but don't have a test case.
I'm cc:ing stable because the initial commit did as well.
The native 31 bit and the compat behaviour for the mmap system calls differ:
In native 31 bit mode the passed in address for the mmap system call will be
unmodified passed to sys_mmap_pgoff().
In compat mode however the passed in address will be modified with
compat_ptr() which masks out the most significant bit.
The result is that in native 31 bit mode each mmap request (with MAP_FIXED)
will fail where the most significat bit is set, while in compat mode it
may succeed.
This odd behaviour was introduced with d3815898 "[S390] mmap: add missing
compat_ptr conversion to both mmap compat syscalls".
To restore a consistent behaviour accross native and compat mode this
patch functionally reverts the above mentioned commit.
Bamboo One's with ID of 0x6a and 0x6b were added with correct
indication of 1024 pressure levels but the Graphire packet routine
was only looking at 9 bits. Increased to 10 bits.
This bug caused these devices to roll over to zero pressure at half
way mark.
The other devices using this routine only support 256 or 512 range
and look to fix unused bits at zero.
Found that commit d478eb44 was a bad commit.
If the link partner is transmitting codeword (even if NULL codeword),
then the RXCW.C bit will be set so check for RXCW.CW is unnecessary.
Ref: RH BZ 840642
Reported-by: Fabio Futigami <ffutigam@redhat.com> Signed-off-by: Tushar Dave <tushar.n.dave@intel.com> CC: Marcelo Ricardo Leitner <mleitner@redhat.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As part of commit 463454b5dbd8 ("cfg80211: fix interface
combinations check"), this extra check was introduced:
if ((all_iftypes & used_iftypes) != used_iftypes)
goto cont;
However, most wireless NIC drivers did not advertise ADHOC in
wiphy.iface_combinations[i].limits[] and hence we'll get -EBUSY
when we bring up a ADHOC wlan with commands similar to:
# iwconfig wlan0 mode ad-hoc && ifconfig wlan0 up
In commit 8e8b41f9d8c8e ("cfg80211: enforce lack of interface
combinations"), the change below fixes the issue:
if (total == 1)
return 0;
But it also introduces other dependencies for stable. For example,
a full cherry pick of 8e8b41f9d8c8e would introduce additional
regressions unless we also start cherry picking driver specific
fixes like the following:
9b4760e ath5k: add possible wiphy interface combinations 1ae2fc2 mac80211_hwsim: advertise interface combinations 20c8e8d ath9k: add possible wiphy interface combinations
And the purpose of the 'if (total == 1)' is to cover the specific
use case (IBSS, adhoc) that was mentioned above. So we just pick
the specific part out from 8e8b41f9d8c8e here.
Doing so gives stable kernels a way to fix the change introduced
by 463454b5dbd8, without having to make cherry picks specific to
various NIC drivers.
Signed-off-by: Liang Li <liang.li@windriver.com> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
libertas currently calls cfg80211_disconnected() when it is being
brought down. This causes an event to be allocated, but since the
wdev is already removed from the rdev by the time that the event
processing work executes, the event is never processed or freed.
http://article.gmane.org/gmane.linux.kernel.wireless.general/95666
Fix this leak, and other possible situations, by processing the event
queue when a device is being unregistered. Thanks to Johannes Berg for
the suggestion.
Signed-off-by: Daniel Drake <dsd@laptop.org> Reviewed-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The irq_to_gpio function was removed from the pxa platform
in linux-3.2, and this driver has been broken since.
There is actually no in-tree user of this driver that adds
this platform device, but the driver can and does get enabled
on some platforms.
Without this patch, building ezx_defconfig results in:
drivers/mfd/ezx-pcap.c: In function 'pcap_isr_work':
drivers/mfd/ezx-pcap.c:205:2: error: implicit declaration of function 'irq_to_gpio' [-Werror=implicit-function-declaration]
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com> Cc: Samuel Ortiz <sameo@linux.intel.com> Cc: Daniel Ribeiro <drwyrm@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The CONFIG_DEFAULT_MMAP_MIN_ADDR was set to 65536 in mxs_defconfig,
this caused severe breakage of userland applications since the upper
limit for ARM is 32768. By default CONFIG_DEFAULT_MMAP_MIN_ADDR is
set to 4096 and can also be changed via /proc/sys/vm/mmap_min_addr
if needed.
Quoting Russell King [1]:
"4096 is also fine for ARM too. There's not much point in having
defconfigs change it - that would just be pure noise in the config
files."
the CONFIG_DEFAULT_MMAP_MIN_ADDR can be removed from the defconfig
altogether.
This problem was introduced by commit cde7c41 (ARM: configs: add
defconfig for mach-mxs).
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Russell King <linux@arm.linux.org.uk> Cc: Wolfgang Denk <wd@denx.de> Signed-off-by: Shawn Guo <shawn.guo@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a process creates a large hugetlbfs mapping that is eligible for page
table sharing and forks heavily with children some of whom fault and
others which destroy the mapping then it is possible for page tables to
get corrupted. Some teardowns of the mapping encounter a "bad pmd" and
output a message to the kernel log. The final teardown will trigger a
BUG_ON in mm/filemap.c.
This was reproduced in 3.4 but is known to have existed for a long time
and goes back at least as far as 2.6.37. It was probably was introduced
in 2.6.20 by [39dde65c: shared page table for hugetlb page]. The messages
look like this;
The bug is a race and not always easy to reproduce. To reproduce it I was
doing the following on a single socket I7-based machine with 16G of RAM.
$ hugeadm --pool-pages-max DEFAULT:13G
$ echo $((18*1048576*1024)) > /proc/sys/kernel/shmmax
$ echo $((18*1048576*1024)) > /proc/sys/kernel/shmall
$ for i in `seq 1 9000`; do ./hugetlbfs-test; done
On my particular machine, it usually triggers within 10 minutes but
enabling debug options can change the timing such that it never hits.
Once the bug is triggered, the machine is in trouble and needs to be
rebooted. The machine will respond but processes accessing proc like "ps
aux" will hang due to the BUG_ON. shutdown will also hang and needs a
hard reset or a sysrq-b.
The basic problem is a race between page table sharing and teardown. For
the most part page table sharing depends on i_mmap_mutex. In some cases,
it is also taking the mm->page_table_lock for the PTE updates but with
shared page tables, it is the i_mmap_mutex that is more important.
Unfortunately it appears to be also insufficient. Consider the following
situation
In this scenario, it is possible for Process A to share page tables with
Process B that is trying to tear them down. The i_mmap_mutex on its own
does not prevent Process A walking Process B's page tables. At (1) above,
the page tables are not shared yet so it unmaps the PMDs. Process A sets
up page table sharing and at (2) faults a new entry. Process B then trips
up on it in free_pgtables.
This patch fixes the problem by adding a new function
__unmap_hugepage_range_final that is only called when the VMA is about to
be destroyed. This function clears VM_MAYSHARE during
unmap_hugepage_range() under the i_mmap_mutex. This makes the VMA
ineligible for sharing and avoids the race. Superficially this looks like
it would then be vunerable to truncate and madvise issues but hugetlbfs
has its own truncate handlers so does not use unmap_mapping_range() and
does not support madvise(DONTNEED).
This should be treated as a -stable candidate if it is merged.
Test program is as follows. The test case was mostly written by Michal
Hocko with a few minor changes to reproduce this bug.
Microcode reloading in a per-core manner is a very bad idea for both
major x86 vendors. And the thing is, we have such interface with which
we can end up with different microcode versions applied on different
cores of an otherwise homogeneous wrt (family,model,stepping) system.
So turn off the possibility of doing that per core and allow it only
system-wide.
This is a minimal fix which we'd like to see in stable too thus the
more-or-less arbitrary decision to allow system-wide reloading only on
the BSP:
Also, allowing the reload only from one CPU (the BSP in
that case) doesn't allow the reload procedure to degenerate
into an O(n^2) deal when triggering reloads from all
/sys/devices/system/cpu/cpuX/microcode/reload sysfs nodes
simultaneously.
A more generic fix will follow.
Signed-off-by: Borislav Petkov <borislav.petkov@amd.com> Cc: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1340280437-7718-2-git-send-email-bp@amd64.org Signed-off-by: H. Peter Anvin <hpa@zytor.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mix in any architectural randomness in extract_buf() instead of
xfer_secondary_buf(). This allows us to mix in more architectural
randomness, and it also makes xfer_secondary_buf() faster, moving a
tiny bit of additional CPU overhead to process which is extracting the
randomness.
[ Commit description modified by tytso to remove an extended
advertisement for the RDRAND instruction. ]
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Acked-by: Ingo Molnar <mingo@kernel.org> Cc: DJ Johnston <dj.johnston@intel.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Many platforms have per-machine instance data (serial numbers,
asset tags, etc.) squirreled away in areas that are accessed
during early system bringup. Mixing this data into the random
pools has a very high value in providing better random data,
so we should allow (and even encourage) architecture code to
call add_device_randomness() from the setup_arch() paths.
However, this limits our options for internal structure of
the random driver since random_initialize() is not called
until long after setup_arch().
Add a big fat comment to rand_initialize() spelling out
this requirement.
With the new interrupt sampling system, we are no longer using the
timer_rand_state structure in the irq descriptor, so we can stop
initializing it now.
[ Merged in fixes from Sedat to find some last missing references to
rand_initialize_irq() ]
The tamper evident features of the RTC include the "write counter" which
is a pseudo-random number regenerated whenever we set the RTC. Since this
value is unpredictable it should provide some useful seeding to the random
number generator.
Only do this on boot since the goal is to seed the pool rather than add
useful entropy.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / karo-tx-linux.git / log