Soeren tracked down a very difficult bug in ehci-hcd's DMA pool
management of iTD and siTD structures. Some background: ehci-hcd
gives each isochronous endpoint its own set of active and free itd's
(or sitd's for full-speed devices). When a new itd is needed, it is
taken from the head of the free list, if possible. However, itd's
must not be used twice in a single frame because the hardware
continues to access the data structure for the entire duration of a
frame. Therefore if the itd at the head of the free list has its
"frame" member equal to the current value of ehci->now_frame, it
cannot be reused and instead a new itd is allocated from the DMA pool.
The entries on the free list are not released back to the pool until
the endpoint is no longer in use.
The bug arises from the fact that sometimes an itd can be moved back
onto the free list before itd->frame has been set properly. In
Soeren's case, this happened because ehci-hcd can allocate one more
itd than it actually needs for an URB; the extra itd may or may not be
required depending on how the transfer aligns with a frame boundary.
For example, an URB with 8 isochronous packets will cause two itd's to
be allocated. If the URB is scheduled to start in microframe 3 of
frame N then it will require both itds: one for microframes 3 - 7 of
frame N and one for microframes 0 - 2 of frame N+1. But if the URB
had been scheduled to start in microframe 0 then it would require only
the first itd, which could cover microframes 0 - 7 of frame N. The
second itd would be returned to the end of the free list.
The itd allocation routine initializes the entire structure to 0, so
the extra itd ends up on the free list with itd->frame set to 0
instead of a meaningful value. After a while the itd reaches the head
of the list, and occasionally this happens when ehci->now_frame is
equal to 0. Then, even though it would be okay to reuse this itd, the
driver thinks it must get another itd from the DMA pool.
For as long as the isochronous endpoint remains in use, this flaw in
the mechanism causes more and more itd's to be taken slowly from the
DMA pool. Since none are released back, the pool eventually becomes
exhausted.
This reuslts in memory allocation failures, which typically show up
during a long-running audio stream. Video might suffer the same
effect.
The fix is very simple. To prevent allocations from the pool when
they aren't needed, make sure that itd's sent back to the free list
prematurely have itd->frame set to an invalid value which can never be
equal to ehci->now_frame.
This should be applied to -stable kernels going back to 3.6.
... lest we get livelocks between path_is_under() and d_path() and friends.
The thing is, wrt fairness lglocks are more similar to rwsems than to rwlocks;
it is possible to have thread B spin on attempt to take lock shared while thread
A is already holding it shared, if B is on lower-numbered CPU than A and there's
a thread C spinning on attempt to take the same lock exclusive.
As the result, we need consistent ordering between vfsmount_lock (lglock) and
rename_lock (seq_lock), even though everything that takes both is going to take
vfsmount_lock only shared.
Note that clearing NFS_INO_LAYOUTCOMMIT is tricky, since it requires
you to also clear the NFS_LSEG_LAYOUTCOMMIT bits from the layout
segments.
The only two sites that need to do this are the ones that call
pnfs_return_layout() without first doing a layout commit.
We need to clear the NFS_LSEG_LAYOUTCOMMIT bits atomically with the
NFS_INO_LAYOUTCOMMIT bit, otherwise we may end up with situations
where the two are out of sync.
The first half of the problem is to ensure that pnfs_layoutcommit_inode
clears the NFS_LSEG_LAYOUTCOMMIT bit through pnfs_list_write_lseg.
We still need to keep the reference to those segments until the RPC call
is finished, so in order to make it clear _where_ those references come
from, we add a helper pnfs_list_write_lseg_done() that cleans up after
pnfs_list_write_lseg.
Functions like nfs_map_uid_to_name() and nfs_map_gid_to_group() are
expected to return a string without any terminating NUL character.
Regression introduced by commit 57e62324e469e092ecc6c94a7a86fe4bd6ac5172
(NFS: Store the legacy idmapper result in the keyring).
when pnfs block using device mapper,if umounting later,it maybe
cause oops. we apply "1 + sizeof(bl_umount_request)" memory for
msg->data, the memory maybe overflow when we do "memcpy(&dataptr
[sizeof(bl_msg)], &bl_umount_request, sizeof(bl_umount_request))",
because the size of bl_msg is more than 1 byte.
curr_cmd points to the command that is in processing or waiting
for its command response from firmware. If the function shutdown
happens to occur at this time we should cancel the cmd timer and
put the command back to free queue.
Tested-by: Marco Cesarano <marco@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mwifiex_send_cmd_async() is called for sync as well as async
commands. (mwifiex_send_cmd_sync() internally calls it for
sync command.)
"adapter->cmd_queued" gets filled inside mwifiex_send_cmd_async()
routine for both types of commands. But it is used only for sync
commands in mwifiex_wait_queue_complete(). This could lead to a
race when two threads try to queue a sync command with another
sync/async command simultaneously.
Get rid of global variable and pass command node as a parameter
to mwifiex_wait_queue_complete() to fix the problem.
Reported-by: Daniel Drake <dsd@laptop.org> Tested-by: Daniel Drake <dsd@laptop.org> Tested-by: Marco Cesarano <marco@marvell.com> Signed-off-by: Amitkumar Karwar <akarwar@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vcs_poll_data_free() calls unregister_vt_notifier(), which calls
atomic_notifier_chain_unregister(), which calls synchronize_rcu().
Do it *after* we'd dropped ->f_lock.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It enhances the driver for FTDI-based USB serial adapters
to recognize Mitsubishi Electric Corp. USB/RS422 Converters
as FT232BM chips and support them.
https://search.meau.com/?q=FX-USB-AW
Signed-off-by: Konstantin Holoborodko <klh.kernel@gmail.com> Tested-by: Konstantin Holoborodko <klh.kernel@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
was a stop-gate to fix a GCC4.1 bug. The appropiate way
is to actually use an list instead of using an llist.
As such this patch replaces the usage of llist with an
list.
Since we always manipulate the list while holding the io_lock, there's
no need for additional locking (llist used previously is safe to use
concurrently without additional locking).
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
[v1: Redid the git commit description] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 7708992 ("xen/blkback: Seperate the bio allocation and the bio
submission") consolidated the pendcnt updates to just a single write,
neglecting the fact that the error path relied on it getting set to 1
up front (such that the decrement in __end_block_io_op() would actually
drop the count to zero, triggering the necessary cleanup actions).
Also remove a misleading and a stale (after said commit) comment.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the frontend is using a non-native protocol (e.g., a 64-bit
frontend with a 32-bit backend) and it sent an unrecognized request,
the request was not translated and the response would have the
incorrect ID. This may cause the frontend driver to behave
incorrectly or crash.
Since the ID field in the request is always in the same place,
regardless of the request type we can get the correct ID and make a
valid response (which will report BLKIF_RSP_EOPNOTSUPP).
This bug affected 64-bit SLES 11 guests when using a 32-bit backend.
This guest does a BLKIF_OP_RESERVED_1 (BLKIF_OP_PACKET in the SLES
source) and would crash in blkif_int() as the ID in the response would
be invalid.
Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For MSI-X capable devices the hypervisor wants to write protect the
MSI-X table and PBA, yet it can't assume that resources have been
assigned to their final values at device enumeration time. Thus have
pciback do that notification, as having the device controlled by it is
a prerequisite to assigning the device to guests anyway.
This is the kernel part of hypervisor side commit 4245d33 ("x86/MSI:
add mechanism to fully protect MSI-X table from PV guest accesses") on
the master branch of git://xenbits.xen.org/xen.git.
Signed-off-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a bug introduced with commit 27c2127 that causes
devices which are hot unplugged and then hot-replugged to
not have per-device dma_ops set. This causes these devices
to not function correctly. Fixed with this patch.
If there are keys left during station removal, then a
synchronize_net() will be done (for each key, I have a
patch to address this for 3.10), otherwise it won't be
done at all which causes issues because the station
could be used for TX while it's being removed from the
driver -- that might confuse the driver.
Fix this by always doing synchronize_net() if no key
was present any more.
Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Limit the hardware's reported firmware string length (max 255 bytes) to
stay inside the driver's firmware string length (32 bytes). On overflow,
truncate the formatted firmware string instead of potentially overwriting
portions of the tg3 struct.
When calculating "offset" for final RSSI calibration we're using numbers
bigger than s8 can hold. We have for example:
offset[j] = 232 - poll_results[j];
formula. If poll_results[j] is small enough (it usually is) we treat
number's bit as a sign bit. For example 232 - 1 becomes:
0xE8 - 0x1 = 0xE7, which is not 231 but -25.
Intermittently, b43 will report "Out of order TX status report on DMA ring".
When this happens, the driver must be reset before communication can resume.
The cause of the problem is believed to be an error in the closed-source
firmware; however, all versions of the firmware are affected.
This change uses the observation that the expected status is always 2 less
than the observed value, and supplies a fake status report to skip one
header/data pair.
Not all devices suffer from this problem, but it can occur several times
per second under heavy load. As each occurence kills the unmodified driver,
this patch makes if possible for the affected devices to function. The patch
logs only the first instance of the reset operation to prevent spamming
the logs.
Tested-by: Chris Vine <chris@cvine.freeserve.co.uk> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We're using "mind" variable to find the VCM that got the best polling
results. For each VCM we calculte "currd" which is compared to the
"mind". For PHY rev3+ "currd" gets values around 14k-40k. Looking for a
value smaller than 40 makes no sense, so increase the initial value.
As reported by Ben Hutchings, there was a harmless issue in
the checks being done on the lengths of the TBs while
building the TFD for a multi-TB host command.
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Acked-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The beacon and multicast-buffer queues are managed by the beacon
tasklet, and the generic tx path hang check does not help in any way
here. Running it on those queues anyway can introduce some race
conditions leading to unnecessary chip resets.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit 'ath9k_hw: fix calibration issues on chainmask that don't
include chain 0' changed the hardware chainmask to the chip chainmask
for the duration of the calibration, but the revert to user
configuration in the reset path runs too early.
That causes some issues with limiting the number of antennas (including
spurious failure in hardware-generated packets).
Fix this by reverting the chainmask after the essential parts of the
calibration that need the workaround, and before NF calibration is run.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Reported-by: Wojciech Dubowik <Wojciech.Dubowik@neratec.com> Tested-by: Wojciech Dubowik <Wojciech.Dubowik@neratec.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Thias patch fixes a define conflict between the SH architecture and the sja1000
driver:
drivers/net/can/sja1000/sja1000.h:59:0: warning:
"REG_SR" redefined [enabled by default]
arch/sh/include/asm/ptrace_32.h:25:0: note:
this is the location of the previous definition
A SJA1000_ prefix is added to the offending sja1000 define only, to make a
minimal patch suited for stable. A later patch will add a SJA1000_ prefix to
all defines in sja1000.h.
In case of 'if (filp->f_pos == 0 or 1)' of sysfs_readdir(),
the failure from filldir() isn't handled, and the reference counter
of the sysfs_dirent object pointed by filp->private_data will be
released without clearing filp->private_data, so use after free
bug will be triggered later.
This patch returns immeadiately under the situation for fixing the bug,
and it is reasonable to return from readdir() when filldir() fails.
Reported-by: Dave Jones <davej@redhat.com> Tested-by: Sasha Levin <levinsasha928@gmail.com> Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While readdir() is running, lseek() may set filp->f_pos as zero,
then may leave filp->private_data pointing to one sysfs_dirent
object without holding its reference counter, so the sysfs_dirent
object may be used after free in next readdir().
This patch holds inode->i_mutex to avoid the problem since
the lock is always held in readdir path.
Reported-by: Dave Jones <davej@redhat.com> Tested-by: Sasha Levin <levinsasha928@gmail.com> Signed-off-by: Ming Lei <ming.lei@canonical.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For the s626 driver, there is a bug in the handling of asynchronous
commands on the AI subdevice when the stop source is `TRIG_NONE`. The
command should run continuously until cancelled, but the interrupt
handler stops the command running after the first scan.
The command set-up function `s626_ai_cmd()` contains this code:
switch (cmd->stop_src) {
case TRIG_COUNT:
/* data arrives as one packet */
devpriv->ai_sample_count = cmd->stop_arg;
devpriv->ai_continous = 0;
break;
case TRIG_NONE:
/* continous acquisition */
devpriv->ai_continous = 1;
devpriv->ai_sample_count = 0;
break;
}
The interrupt handler `s626_irq_handler()` contains this code:
if (!(devpriv->ai_continous))
devpriv->ai_sample_count--;
if (devpriv->ai_sample_count <= 0) {
devpriv->ai_cmd_running = 0;
/* ... */
}
So `devpriv->ai_sample_count` is only decremented for the `TRIG_COUNT`
case, but `devpriv->ai_cmd_running` is set to 0 (and the command
stopped) regardless.
Fix this in `s626_ai_cmd()` by setting `devpriv->ai_sample_count = 1`
for the `TRIG_NONE` case. The interrupt handler will not decrement it
so it will remain greater than 0 and the check for stopping the
acquisition will fail.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
With deferred setup for SCO, it is possible that userspace closes the
socket when it is in the BT_CONNECT2 state, after the Connect Request is
received but before the Accept Synchonous Connection is sent.
If this happens the following crash was observed, when the connection is
terminated:
The current Tilera boot infrastructure now provides the initramfs
to Linux as a Tilera-hypervisor file named "initramfs", rather than
"initramfs.cpio.gz", as before. (This makes it reasonable to use
other compression techniques than gzip on the file without having to
worry about the name causing confusion.) Adapt to use the new name,
but also fall back to checking for the old name.
Cc'ing to stable so that older kernels will remain compatible with
newer Tilera boot infrastructure.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We need to be careful when testing task->tk_waitqueue in
rpc_wake_up_task_queue_locked, because it can be changed while we
are holding the queue->lock.
By adding appropriate memory barriers, we can ensure that it is safe to
test task->tk_waitqueue for equality if the RPC_TASK_QUEUED bit is set.
If CONFIG_MAC80211_MESH is not set, cfg80211 will now allow advertising
interface combinations with NL80211_IFTYPE_MESH_POINT present.
Add appropriate ifdefs to avoid running into errors.
[Backported for 3.8-stable. Removed code of simultaneous AP and mesh
mode added in 4a5fc6d 3.9-rc1.]
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Acked-by: Gertjan van Wingerde <gwingerde@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Lingzhu Xiang <lxiang@redhat.com> Reviewed-by: CAI Qian <caiqian@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Atomic pool should always be allocated from DMA zone if such zone is
available in the system to avoid issues caused by limited dma mask of
any of the devices used for making an atomic allocation.
Reported-by: Krzysztof Halasa <khc@pm.waw.pl> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The two-port driver was not updated when switching to the new icount
interface in commit 0bca1b913aff ("tty: Convert the USB drivers to the
new icount interface").
Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove bogus disconnect test introduced by 95bef012e ("USB: more serial
drivers writing after disconnect") which prevented queued data from
being freed on disconnect.
The possible IO it was supposed to prevent is long gone.
Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For large UDF filesystems with 512-byte blocks the number of necessary
bitmap blocks is larger than 2^16 so s_nr_groups in udf_bitmap overflows
(the number will overflow for filesystems larger than 128 GB with
512-byte blocks). That results in ENOSPC errors despite the filesystem
has plenty of free space.
Fix the problem by changing s_nr_groups' type to 'int'. That is enough
even for filesystems 2^32 blocks (UDF maximum) and 512-byte blocksize.
Reported-and-tested-by: v10lator@myway.de Signed-off-by: Jan Kara <jack@suse.cz> Cc: Jim Trigg <jtrigg@spamcop.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Observe that acpi_get_child() doesn't need to use the helper
struct acpi_find_child structure and change it to work without it.
Also, using acpi_get_object_info() to get the output of _ADR for the
given device is overkill, because that function does much more than
just evaluating _ADR (let alone the additional memory allocation
done by it).
Moreover, acpi_get_child() doesn't need to loop any more once it has
found a matching handle, so make it stop in that case. To prevent
the results from changing, make it use do_acpi_find_child() as
a post-order callback.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Cc: Josh Boyer <jwboyer@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We know that with some firmware implementations writing too much data to
UEFI variables can lead to bricking machines. Recent changes attempt to
address this issue, but for some it may still be prudent to avoid
writing large amounts of data until the solution has been proven on a
wide variety of hardware.
Crash dumps or other data from pstore can potentially be a large data
source. Add a pstore_module parameter to efivars to allow disabling its
use as a backend for pstore. Also add a config option,
CONFIG_EFI_VARS_PSTORE_DEFAULT_DISABLE, to allow setting the default
value of this paramter to true (i.e. disabled by default).
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Cc: Josh Boyer <jwboyer@redhat.com> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Cc: Seiji Aguchi <seiji.aguchi@hds.com> Cc: Tony Luck <tony.luck@intel.com> Signed-off-by: Matt Fleming <matt.fleming@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add a new option, CONFIG_EFI_VARS_PSTORE, which can be set to N to
avoid using efivars as a backend to pstore, as some users may want to
compile out the code completely.
Set the default to Y to maintain backwards compatability, since this
feature has always been enabled until now.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Cc: Josh Boyer <jwboyer@redhat.com> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Cc: Seiji Aguchi <seiji.aguchi@hds.com> Cc: Tony Luck <tony.luck@intel.com> Signed-off-by: Matt Fleming <matt.fleming@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make sure the interface is not released before our serial device.
Note that drivers are still not allowed to access the interface in
any way that may interfere with another driver that may have gotten
bound to the same interface after disconnect returns.
Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes an "off-by-one" bug found in 581791f (FunctionFS: enable multiple functions).
During gfs_bind/gfs_unbind the functionfs_bind/functionfs_unbind should be
called for every functionfs instance. With the "i" pre-decremented they
were not called for the zeroth instance.
Acked-by: Michal Nazarewicz <mina86@mina86.com> Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
[ balbi@ti.com : added offending commit's subject ] Signed-off-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch (as1670) fixes a regression caused by commit 6402c796d3b4205d3d7296157956c5100a05d7d6 (USB: EHCI: work around
silicon bug in Intel's EHCI controllers). The workaround goes through
two IAA cycles for each QH being unlinked. During the first cycle,
the QH is not added to the async_iaa list (because it isn't fully gone
from the hardware yet), which means that list will be empty.
Unfortunately, I forgot to update the IAA watchdog timer routine. It
thinks that an empty async_iaa list means the timer expiration was an
error, which isn't true any more. This problem didn't show up during
initial testing because the controllers being tested all had working
IAA interrupts. But not all controllers do, and when the watchdog
timer expires, the empty-list check prevents the second IAA cycle from
starting. As a result, URB unlinks never complete. The check needs
to be removed.
Among the symptoms of the regression are processes stuck in D wait
states and hangs during system shutdown.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Reported-and-tested-by: Stephen Warren <swarren@wwwdotorg.org> Reported-and-tested-by: Sven Joachim <svenjoac@gmx.de> Reported-by: Andreas Bombe <aeb@debian.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch (as1663) fixes a regression caused by commit 6e0c3339a6f19d748f16091d0a05adeb1e1f822b (USB: EHCI: unlink one async
QH at a time). In order to avoid keeping multiple QHs in an unusable
intermediate state, that commit changed unlink_empty_async() so that
it unlinks only one empty QH at a time.
However, when the EHCI root hub is suspended, _all_ async QHs need to
be unlinked. ehci_bus_suspend() used to do this by calling
unlink_empty_async(), but now this only unlinks one of the QHs, not
all of them.
The symptom is that when the root hub is resumed, USB communications
don't work for some period of time. This is because ehci-hcd doesn't
realize it needs to restart the async schedule; it assumes that
because some QHs are already on the schedule, the schedule must be
running.
The easiest way to fix the problem is add a new function that unlinks
all the async QHs when the root hub is suspended.
This patch should be applied to all kernels that have the 6e0c3339a6f1
commit.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Reported-and-tested-by: Adrian Bassett <adrian.bassett@hotmail.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Unregister tty device in disconnect as is required by the USB stack.
By deferring unregistration to when the last tty reference is dropped,
the parent interface device can get unregistered before the child
resulting in broken hotplug events being generated when the tty is
finally closed:
The driver must deal with tty callbacks after disconnect by checking the
disconnected flag. Specifically, further opens must be prevented and
this is already implemented.
Acked-by: Oliver Neukum <oneukum@suse.de> Cc: Oliver Neukum <oneukum@suse.de> Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
xhci has its own interrupt enabling routine, which will try to
use MSI-X/MSI if present. So the usb core shouldn't try to enable
legacy interrupts; on some machines the xhci legacy IRQ setting
is invalid.
v3: Be careful to not break XHCI_BROKEN_MSI workaround (by trenn)
Cc: Bjorn Helgaas <bhelgaas@google.com> Cc: Oliver Neukum <oneukum@suse.de> Cc: Thomas Renninger <trenn@suse.de> Cc: Yinghai Lu <yinghai@kernel.org> Cc: Frederik Himpe <fhimpe@vub.ac.be> Cc: David Haerdeman <david@hardeman.nu> Cc: Alan Stern <stern@rowland.harvard.edu> Acked-by: Sarah Sharp <sarah.a.sharp@linux.intel.com> Reviewed-by: Thomas Renninger <trenn@suse.de> Signed-off-by: Hannes Reinecke <hare@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
According to XHCI specification (5.5.2.1) the IP is bit 0 and IE is bit 1
of IMAN register. Previously their definitions were reversed.
Even though there are no ill effects being observed from the swapped
definitions (because IMAN_IP is RW1C and in legacy PCI case we come in
with it already set to 1 so it was clearing itself even though we were
setting IMAN_IE instead of IMAN_IP), we should still correct the values.
This patch should be backported to kernels as old as 2.6.36, that
contain the commit 4e833c0b87a30798e67f06120cecebef6ee9644c "xhci: don't
re-enable IE constantly".
The increment of "to" in copy_user_handle_tail() will have incremented
before a failure has been noted. This causes us to skip a byte in the
failure case.
Only do the increment when assured there is no failure.
Currently tick_check_broadcast_device doesn't reject clock_event_devices
with CLOCK_EVT_FEAT_DUMMY, and may select them in preference to real
hardware if they have a higher rating value. In this situation, the
dummy timer is responsible for broadcasting to itself, and the core
clockevents code may attempt to call non-existent callbacks for
programming the dummy, eventually leading to a panic.
This patch makes tick_check_broadcast_device always reject dummy timers,
preventing this problem.
Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: linux-arm-kernel@lists.infradead.org Cc: Jon Medhurst (Tixy) <tixy@linaro.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
vfs_writev() updates the offset argument - but the code then passes the
offset to vfs_fsync_range(). Since offset now points to the offset after
what was just written, this is probably not what was intended
Signed-off-by: Kent Overstreet <koverstreet@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Reviewed-by: Zach Brown <zab@redhat.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A number of problems can occur due to races between
resync/recovery and discard.
- if sync_request calls handle_stripe() while a discard is
happening on the stripe, it might call handle_stripe_clean_event
before all of the individual discard requests have completed
(so some devices are still locked, but not all).
Since commit ca64cae96037de16e4af92678814f5d4bf0c1c65
md/raid5: Make sure we clear R5_Discard when discard is finished.
this will cause R5_Discard to be cleared for the parity device,
so handle_stripe_clean_event() will not be called when the other
devices do become unlocked, so their ->written will not be cleared.
This ultimately leads to a WARN_ON in init_stripe and a lock-up.
- If handle_stripe_clean_event() does clear R5_UPTODATE at an awkward
time for resync, it can lead to s->uptodate being less than disks
in handle_parity_checks5(), which triggers a BUG (because it is
one).
So:
- keep R5_Discard on the parity device until all other devices have
completed their discard request
- make sure we don't try to have a 'discard' and a 'sync' action at
the same time.
This involves a new stripe flag to we know when a 'discard' is
happening, and the use of R5_Overlap on the parity disk so when a
discard is wanted while a sync is active, so we know to wake up
the discard at the appropriate time.
Discard support for RAID5 was added in 3.7, so this is suitable for
any -stable kernel since 3.7.
MD RAID5: Fix kernel oops when RAID4/5/6 is used via device-mapper
Commit a9add5d (v3.8-rc1) added blktrace calls to the RAID4/5/6 driver.
However, when device-mapper is used to create RAID4/5/6 arrays, the
mddev->gendisk and mddev->queue fields are not setup. Therefore, calling
things like trace_block_bio_remap will cause a kernel oops. This patch
conditionalizes those calls on whether the proper fields exist to make
the calls. (Device-mapper will call trace_block_bio_remap on its own.)
This patch is suitable for the 3.8.y stable kernel.
It has been possible for handle_stripe_dirtying to be called
when there isn't actually any work to do.
It then calls schedule_reconstruction() which will set R5_LOCKED
on the parity block(s) even when nothing else is happening.
This then causes problems in do_release_stripe().
So add checks to schedule_reconstruction() so that if it doesn't
find anything to do, it just aborts.
This bug was introduced in v3.7, so the patch is suitable
for -stable kernels since then.
A problem was found on PC's with the SB700 chipset: The PC fails to
load BIOS after running the 3.8.x kernel until the power is completely
cut off. It occurs in all 3.8.x versions and the mainline version as of
2/4. The issue does not occur with the 3.7.x builds.
There are two methods for accessing the watchdog registers.
1. Re-programming a resource address obtained by allocate_resource()
to chipset.
2. Use the direct memory-mapped IO access.
The method 1 can be used by all the chipsets (SP5100, SB7x0, SB8x0 or
later). However, experience shows that only PC with the SB8x0 (or
later) chipsets can use the method 2.
This patch removes the method 1, because the critical problem was found.
That's why the watchdog timer was able to be used on SP5100 and SB7x0
chipsets until now.
The AcpiMmioSel bit is bit 1 in the AcpiMmioEn register, but the current
sp5100_tco driver is using bit 2.
See 2.3.3 Power Management (PM) Registers page 150 of the
AMD SB800-Series Southbridges Register Reference Guide [1].
AcpiMmioEn - RW – 8/16/32 bits - [PM_Reg: 24h]
Field Name Bits Default Description
AcpiMMioDecodeEn 0 0b Set to 1 to enable AcpiMMio space.
AcpiMMIoSel 1 0b Set AcpiMMio registers to be memory-mapped or IO-mapped space.
0: Memory-mapped space
1: I/O-mapped space
The sp5100_tco driver expects zero as a value of AcpiMmioSel (bit 1).
Fortunately, no problems were caused by this typo, because the default
value of the undocumented misused bit 2 seems to be zero.
However, the sp5100_tco driver should use the correct bitmask value.
Commit f0dc117abdfa ("IPoIB: Fix TX queue lockup with mixed UD/CM
traffic") attempts to solve an issue where unprocessed UD send
completions can deadlock the netdev.
The patch doesn't fully resolve the issue because if more than half
the tx_outstanding's were UD and all of the destinations are RC
reachable, arming the CQ doesn't solve the issue.
This patch uses the IB_CQ_REPORT_MISSED_EVENTS on the
ib_req_notify_cq(). If the rc is above 0, the UD send cq completion
callback is called directly to re-arm the send completion timer.
This issue is seen in very large parallel filesystem deployments
and the patch has been shown to correct the issue.
Reviewed-by: Dean Luick <dean.luick@intel.com> Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Roland Dreier <roland@purestorage.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In data=journal mode, if we unmount the file system before a
transaction has a chance to complete, when the journal inode is being
evicted, we can end up calling into jbd2_log_wait_commit() for the
last transaction, after the journalling machinery has been shut down.
Arguably we should adjust ext4_should_journal_data() to return FALSE
for the journal inode, but the only place it matters is
ext4_evict_inode(), and so to save a bit of CPU time, and to make the
patch much more obviously correct by inspection(tm), we'll fix it by
explicitly not trying to waiting for a journal commit when we are
evicting the journal inode, since it's guaranteed to never succeed in
this case.
This can be easily replicated via:
mount -t ext4 -o data=journal /dev/vdb /vdb ; umount /vdb
Currently when converting extent to initialized, we have to decide
whether to zeroout part/all of the uninitialized extent in order to
avoid extent tree growing rapidly.
The decision is made by comparing the size of the extent with the
configurable value s_extent_max_zeroout_kb which is in kibibytes units.
However when converting it to number of blocks we currently use it as it
was in bytes. This is obviously bug and it will result in ext4 _never_
zeroout extents, but rather always split and convert parts to
initialized while leaving the rest uninitialized in default setting.
Fix this by using s_extent_max_zeroout_kb as kibibytes.
A user who was using a 8TB+ file system and with a very large flexbg
size (> 65536) could cause the atomic_t used in the struct flex_groups
to overflow. This was detected by PaX security patchset:
This commit fixes a wrong return value of the number of the allocated
blocks in ext4_split_extent. When the length of blocks we want to
allocate is greater than the length of the current extent, we return a
wrong number. Let's see what happens in the following case when we
call ext4_split_extent().
map: [48, 72]
ex: [32, 64, u]
'ex' will be split into two parts:
ex1: [32, 47, u]
ex2: [48, 64, w]
'map->m_len' is returned from this function, and the value is 24. But
the real length is 16. So it should be fixed.
Meanwhile in this commit we use right length of the allocated blocks
when get_reserved_cluster_alloc in ext4_ext_handle_uninitialized_extents
is called.
jbd2_journal_dirty_metadata() didn't get a reference to journal_head it
was working with. This is OK in most of the cases since the journal head
should be attached to a transaction but in rare occasions when we are
journalling data, __ext4_journalled_writepage() can race with
jbd2_journal_invalidatepage() stripping buffers from a page and thus
journal head can be freed under hands of jbd2_journal_dirty_metadata().
Fix the problem by getting own journal head reference in
jbd2_journal_dirty_metadata() (and also in jbd2_journal_set_triggers()
which can possibly have the same issue).
Reported-by: Zheng Liu <gnehzuil.liu@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We've had several reports of people attempting to mount Windows 8 shares
and getting failures with a return code of -EINVAL. The default sec=
mode changed recently to sec=ntlmssp. With that, we expect and parse a
SPNEGO blob from the server in the NEGOTIATE reply.
The current decode_negTokenInit function first parses all of the
mechTypes and then tries to parse the rest of the negTokenInit reply.
The parser however currently expects a mechListMIC or nothing to follow the
mechTypes, but Windows 8 puts a mechToken field there instead to carry
some info for the new NegoEx stuff.
In practice, we don't do anything with the fields after the mechTypes
anyway so I don't see any real benefit in continuing to parse them.
This patch just has the kernel ignore the fields after the mechTypes.
We'll probably need to reinstate some of this if we ever want to support
NegoEx.
Reported-by: Jason Burgess <jason@jacknife2.dns2go.com> Reported-by: Yan Li <elliot.li.tech@gmail.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cifsFileInfo objects hold references to dentries and it is possible that
these will still be around in workqueues when VFS decides to kill super
block during unmount.
Fix this by making each cifsFileInfo object hold a reference to cifs
super block, which implicitly keeps VFS super block around as well.
Signed-off-by: Mateusz Guzik <mguzik@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Reported-and-Tested-by: Ben Greear <greearb@candelatech.com> Signed-off-by: Steve French <sfrench@us.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While testing the mgag200 kms driver on the HP ProLiant Gen8, a
bug was seen. Once the bootloader would load the selected kernel,
the screen would go black. At first it was assumed that the
mgag200 kms driver was hanging. But after setting up the grub
serial output, it was seen that the driver was being loaded
properly. After trying serval monitors, one finaly displayed
the message "Frequency Out of Range". By comparing the kms pll
algorithm with the previous mgag200 xorg driver pll algorithm,
discrepencies were found. Once the kms pll algorithm was
modified, the expected pll values were produced. This fix was
tested on several monitors of varying native resolutions.
Signed-off-by: Julia Lemire <jlemire@matrox.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A deadlock was found in the prefetch code in the dm verity map
function. This patch fixes this by transferring the prefetch
to a worker thread and skipping it completely if kmalloc fails.
If generic_make_request is called recursively, it queues the I/O
request on the current->bio_list without making the I/O request
and returns. The routine making the recursive call cannot wait
for the I/O to complete.
The deadlock occurs when one thread grabs the bufio_client
mutex and waits for an I/O to complete but the I/O is queued
on another thread's current->bio_list and is waiting to get
the mutex held by the first thread.
The fix recognises that prefetching is not essential. If memory
can be allocated, it queues the prefetch request to the worker thread,
but if not, it does nothing.
Signed-off-by: Paul Taysom <taysom@chromium.org> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix a bug in dm_btree_remove that could leave leaf values with incorrect
reference counts. The effect of this was that removal of a shared block
could result in the space maps thinking the block was no longer used.
More concretely, if you have a thin device and a snapshot of it, sending
a discard to a shared region of the thin could corrupt the snapshot.
Thinp uses a 2-level nested btree to store it's mappings. This first
level is indexed by thin device, and the second level by logical
block.
Often when we're removing an entry in this mapping tree we need to
rebalance nodes, which can involve shadowing them, possibly creating a
copy if the block is shared. If we do create a copy then children of
that node need to have their reference counts incremented. In this
way reference counts percolate down the tree as shared trees diverge.
The rebalance functions were incrementing the children at the
appropriate time, but they were always assuming the children were
internal nodes. This meant the leaf values (in our case packed
block/flags entries) were not being incremented.
Signed-off-by: Joe Thornber <ejt@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch bumps the default FILEIO backend FD_MAX_SECTORS value from
1024 -> 2048 in order to allow block_size=512 to handle 1M sized I/Os.
The current default rejects I/Os larger than 512K in sbc_parse_cdb():
[12015.915146] SCSI OP 2ah with too big sectors 1347 exceeds backend
hw_max_sectors: 1024
[12015.977744] SCSI OP 2ah with too big sectors 2048 exceeds backend
hw_max_sectors: 1024
This issue is present in >= v3.5 based kernels, introduced after the
removal of se_task logic.
mnt_drop_write() must be called only if mnt_want_write() succeeded,
otherwise the mnt_writers counter will diverge.
mnt_writers counters are used to check if remounting FS as read-only is
OK, so after an extra mnt_drop_write() call, it would be impossible to
remount mqueue FS as read-only. Besides, on umount a warning would be
printed like this one:
=====================================
[ BUG: bad unlock balance detected! ]
3.9.0-rc3 #5 Not tainted
-------------------------------------
a.out/12486 is trying to release lock (sb_writers) at:
mnt_drop_write+0x1f/0x30
but there are no more locks to release!
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Doug Ledford <dledford@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Al Viro <viro@zeniv.linux.org.uk> 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>
Commit be8678149701 ("drivers/video/ep93xx-fb.c: use devm_ functions")
introduced a build error:
drivers/video/ep93xx-fb.c: In function 'ep93xxfb_probe':
drivers/video/ep93xx-fb.c:532: error: implicit declaration of function 'devm_ioremap'
drivers/video/ep93xx-fb.c:533: warning: assignment makes pointer from integer without a cast
Include <linux/io.h> to pickup the declaration of 'devm_ioremap'.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Acked-by: Ryan Mallon <rmallon@gmail.com> Cc: Damien Cassou <damien.cassou@lifl.fr> 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>
hugetlb_total_pages is used for overcommit calculations but the current
implementation considers only the default hugetlb page size (which is
either the first defined hugepage size or the one specified by
default_hugepagesz kernel boot parameter).
If the system is configured for more than one hugepage size, which is
possible since commit a137e1cc6d6e ("hugetlbfs: per mount huge page
sizes") then the overcommit estimation done by __vm_enough_memory()
(resp. shown by meminfo_proc_show) is not precise - there is an
impression of more available/allowed memory. This can lead to an
unexpected ENOMEM/EFAULT resp. SIGSEGV when memory is accounted.
Testcase:
boot: hugepagesz=1G hugepages=1
the default overcommit ratio is 50
before patch:
On some revisions of AT91 SoCs, the RTC IMR register is not working.
Instead of elaborating a workaround for that specific SoC or IP version,
we simply use a software variable to store the Interrupt Mask Register
and modify it for each enabling/disabling of an interrupt. The overhead
of this is negligible anyway.
The interrupt mask register (IMR) for the RTC is broken on the AT91SAM9x5
sub-family of SoCs (good overview of the members here:
http://www.eewiki.net/display/linuxonarm/AT91SAM9x5 ). The "user visible
effect" is the RTC doesn't work.
That sub-family is less than two years old and only has devicetree (DT)
support and came online circa lk 3.7 . The dust is yet to settle on the
DT stuff at least for AT91 SoCs (translation: lots of stuff is still
broken, so much that it is hard to know where to start).
The fix in the patch is pretty simple: just shadow the silicon IMR
register with a variable in the driver. Some older SoCs (pre-DT) use the
the rtc-at91rm9200 driver (e.g. obviously the AT91RM9200) and they should
not be impacted by the change. There shouldn't be a large volume of
interrupts associated with a RTC.
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com> Reported-by: Douglas Gilbert <dgilbert@interlog.com> Cc: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com> Cc: Ludovic Desroches <ludovic.desroches@atmel.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>
When KMS has parsed an EDID "detailed timing", it leaves the frame rate
zeroed. Consecutive (debug-) output of that mode thus yields 0 for
vsync. This simple fix also speeds up future invocations of
drm_mode_vrefresh().
While it is debatable whether this qualifies as a -stable fix I'd apply
it for consistency's sake; drm_helper_probe_single_connector_modes()
does the same thing already for all probed modes.
EDID spreads some values across multiple bytes; bit-fiddling is needed
to retrieve these. The current code to parse "detailed timings" has a
cut&paste error that results in a vsync offset of at most 15 lines
instead of 63.
See
http://en.wikipedia.org/wiki/EDID
and in the "EDID Detailed Timing Descriptor" see bytes 10+11 show why
that needs to be a left shift.
NVIDIA's Tegra SoC allows read/write of controller register only
if controller clock is enabled. System hangs if read/write happens
to registers without enabling clock.
clk_prepare_enable() can be fail due to unknown reason and hence
adding check for return value of this function. If this function
success then only access register otherwise return to caller with
error.
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com> Reviewed-by: Stephen Warren <swarren@nvidia.com> Signed-off-by: Wolfram Sang <wsa@the-dreams.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fixed a regression in 3.6-rc kernels for which we've never figured out
the exact root cause. But some further inspection of the backlight
code reveals that it's seriously lacking locking. And especially the
asle backlight update is know to get fired (through some smm magic)
when writing specific backlight control registers. So the possibility
of suffering from races is rather real.
Until those races are fixed I don't think it makes sense to try
further hacks. Which sucks a bit, but sometimes that's how it is :(
It is possible to wrap the counter used to allocate the buffer for
relocation copies. This could lead to heap writing overflows.
CVE-2013-0913
v3: collapse test, improve comment
v2: move check into validate_exec_list
Signed-off-by: Kees Cook <keescook@chromium.org> Reported-by: Pinkie Pie Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The driver was failing to clear the BSSID when a disconnect happened. That
prevented a reconnection. This problem is reported at
https://bugzilla.redhat.com/show_bug.cgi?id=789605,
https://bugzilla.redhat.com/show_bug.cgi?id=866786,
https://bugzilla.redhat.com/show_bug.cgi?id=906734, and
https://bugzilla.kernel.org/show_bug.cgi?id=46171.
Thanks to Jussi Kivilinna for making the critical observation
that led to the solution.
Reported-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Tested-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Tested-by: Alessandro Lannocca <alessandro.lannocca@gmail.com> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The cause is a synchronous read from routine rtl92cu_update_hal_rate_table().
The resulting output is not critical, thus the debug statement is
deleted.
Reported-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / karo-tx-linux.git / log