Use the v4l2-async API to support asynchronous subdevice probing,
including the CSI2 subdevice. Synchronous probing is still supported too.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Add support for asynchronous subdevice probing, using the v4l2-async API.
The legacy synchronous mode is still supported too, which allows to
gradually update drivers and platforms. The selected approach adds a
notifier for each struct soc_camera_device instance, i.e. for each video
device node, even when there are multiple such instances registered with a
single soc-camera host simultaneously.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Instead of centrally enabling and disabling subdevice master clocks in
soc-camera core, let subdevice drivers do that themselves, using the
V4L2 clock API and soc-camera convenience wrappers.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Currently bridge device drivers register devices for all subdevices
synchronously, typically, during their probing. E.g. if an I2C CMOS sensor
is attached to a video bridge device, the bridge driver will create an I2C
device and wait for the respective I2C driver to probe. This makes linking
of devices straight forward, but this approach cannot be used with
intrinsically asynchronous and unordered device registration systems like
the Flattened Device Tree. To support such systems this patch adds an
asynchronous subdevice registration framework to V4L2. To use it respective
(e.g. I2C) subdevice drivers must register themselves with the framework.
A bridge driver on the other hand must register notification callbacks,
that will be called upon various related events.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

It is often useful to have simple means to get from a subdevice to the
underlying physical device. This patch adds such a pointer to struct
v4l2_subdev and sets it accordingly in the I2C and SPI cases.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Typical video devices like camera sensors require an external clock source.
Many such devices cannot even access their hardware registers without a
running clock. These clock sources should be controlled by their consumers.
This should be performed, using the generic clock framework. Unfortunately
so far only very few systems have been ported to that framework. This patch
adds a set of temporary helpers, mimicking the generic clock API, to V4L2.
Platforms, adopting the clock API, should switch to using it. Eventually
this temporary API should be removed.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Currently soc-camera uses a single camera host callback to activate the
interface master clock and to configure the interface for a specific
client. However, during probing we might not have the information about
a client, we just need to activate the clock. Add new camera host driver
callbacks to only start and stop the clock without and client-specific
configuration.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

All soc-camera host drivers include a pointer to an soc-camera device in
their host private struct to check, that only one client is connected.
Move this common code to soc_camera.c.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

The last set of drivers still using the parent field of video_device instead
of the v4l2_dev field have been converted, so v4l2_dev is now always set.
A proper pointer to v4l2_dev is necessary these days otherwise the advanced
debugging ioctls will not work when addressing sub-devices. It also ensures
that the core can always go from a video_device struct to the top-level
v4l2_device struct.
There is still one single use case for the parent pointer: if there are
multiple busses, each being the parent of one or more video nodes, and if
they all share the same v4l2_device struct. In that case one still needs a
parent pointer since the v4l2_device struct can only refer to a single
parent device. The cx88 driver is one such case. Unfortunately, the cx88
failed to set the parent pointer since 3.6. The next patch will correct this.
In order to support this use-case the parent pointer is only renamed to
dev_parent, not removed altogether. It has been renamed to ensure that the
compiler will catch any (possibly out-of-tree) drivers that were missed during
the conversion.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@iki.fi>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

The parent field will eventually disappear to be replaced by v4l2_dev.
soc_camera does provide a v4l2_device struct but did not point to it in
struct video_device. This is now fixed.
Now the video nodes can be found under the correct platform bus, and
the advanced debug ioctls work correctly as well (the core implementation
of those ioctls requires that v4l2_dev is set correctly).
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

This has been replaced by the new and much better VIDIOC_DBG_G_CHIP_INFO.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Cc: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

This function is no longer needed since it is now the responsibility of the
v4l2 core to check if the DBG_G/S_REGISTER and DBG_G_CHIP_INFO ioctls are
called for the bridge driver or not.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

This function allows to remove all media entity's links to other
entities, leaving no references to a media entity's links array
at its remote entities.
Currently, when a driver of some entity is removed it will free its
media entities links[] array, leaving dangling pointers at other
entities that are part of same media graph. This is troublesome when
drivers of a media device entities are in separate kernel modules,
removing only some modules will leave others in an incorrect state.
This function is intended to be used when an entity is being
unregistered from a media device.
With an assumption that normally the media links should be created
between media entities registered to a media device, with the graph
mutex held.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: NAndrzej Hajda <a.hajda@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSakari Ailus <sakari.ailus@iki.fi>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

The use of current_norm to keep track of the current standard has been
deprecated for quite some time. Now that all drivers that were using it
have been converted to use g_std we can drop it from the core.
It was a bad idea to introduce this at the time: since it is a per-device
node field it didn't work for drivers that create multiple nodes, all sharing
the same tuner (e.g. video and vbi nodes, or a raw video node and a compressed
video node). In addition it was very surprising behavior that g_std was
implemented in the core. Often drivers implemented both g_std and current_norm,
because they didn't understand how it should be used.
Since the benefits were very limited (if they were there at all), it is better
to just drop it and require that drivers just implement g_std.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Replace the V4L2_IDENT_* usage with tveeprom-specific defines. This header
is deprecated, so those defines shouldn't be used anymore.
The em28xx driver is the only one that uses the tveeprom audio_processor
field, so that has been updated to use the new tveeprom AUDPROC define.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Thanks to commit f91eb62f ("init: scream bloody murder if interrupts
are enabled too early"), "bloody murder" is now being screamed.

With a MIPS OCTEON config, we use on_each_cpu() in our
irq_chip.irq_bus_sync_unlock() function.  This gets called in early as a
result of the time_init() call.  Because the !SMP version of
on_each_cpu() unconditionally enables irqs, we get:

    WARNING: at init/main.c:560 start_kernel+0x250/0x410()
    Interrupts were enabled early
    CPU: 0 PID: 0 Comm: swapper Not tainted 3.10.0-rc5-Cavium-Octeon+ #801
    Call Trace:
      show_stack+0x68/0x80
      warn_slowpath_common+0x78/0xb0
      warn_slowpath_fmt+0x38/0x48
      start_kernel+0x250/0x410

Suggested fix: Do what we already do in the SMP version of
on_each_cpu(), and use local_irq_save/local_irq_restore.  Because we
need a flags variable, make it a static inline to avoid name space
issues.

[ Change from v1: Convert on_each_cpu to a static inline function, add
  #include <linux/irqflags.h> to avoid build breakage on some files.

  on_each_cpu_mask() and on_each_cpu_cond() suffer the same problem as
  on_each_cpu(), but they are not causing !SMP bugs for me, so I will
  defer changing them to a less urgent patch. ]
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This patch removes init_enable option from pdata, the init_enable
was intended that the device should start streaming video immediately
but ideally the bridge drivers should call s_stream explicitly for such
devices to start video.
Signed-off-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

If CONFIG_NET_NS is not set then __net_init is the same as __init and
__net_exit is the same as __exit. These functions will be removed from
memory after the module loads or is removed. Functions that are exported
for use by other functions should never be labeled for removal.

Bug introduced by commit c5441932
("GRE: Refactor GRE tunneling code.")
Reported-by: NSteinar H. Gunderson <sgunderson@bigfoot.com>
Signed-off-by: NSteven Rostedt <rostedt@goodmis.org>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently the media device link_notify callback is invoked before the
actual change of state of a link when the link is being enabled, and
after the actual change of state when the link is being disabled.
This doesn't allow a media device driver to perform any operations
on a full graph before a link is disabled, as well as performing
any tasks on a modified graph right after a link's state is changed.
This patch modifies signature of the link_notify callback. This
callback is now called always before and after a link's state change.
To distinguish the notifications a 'notification' argument is added
to the link_notify callback: MEDIA_DEV_NOTIFY_PRE_LINK_CH indicates
notification before link's state change and
MEDIA_DEV_NOTIFY_POST_LINK_CH corresponds to a notification after
link flags change.

[mchehab@redhat.com: whitespace cleanups]
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSakari Ailus <sakari.ailus@iki.fi>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

This enumeration is now private to exynos4-is and the exynos5 camera
subsystem driver may have the subdevs handling designed differently.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

This patch introduces common structure for the video entities
to handle all video nodes and media pipelines associated with
them in more generic way.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

There is div64_long() to handle the s64/long division, but no mocro do
u64/ul division.  It is necessary in some scenarios, so add this
function.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: NAlex Shi <alex.shi@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When we have a page fault for the address which is backed by a hugepage
under migration, the kernel can't wait correctly and do busy looping on
hugepage fault until the migration finishes.  As a result, users who try
to kick hugepage migration (via soft offlining, for example) occasionally
experience long delay or soft lockup.

This is because pte_offset_map_lock() can't get a correct migration entry
or a correct page table lock for hugepage.  This patch introduces
migration_entry_wait_huge() to solve this.
Signed-off-by: NNaoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reviewed-by: NRik van Riel <riel@redhat.com>
Reviewed-by: NWanpeng Li <liwanp@linux.vnet.ibm.com>
Reviewed-by: NMichal Hocko <mhocko@suse.cz>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: <stable@vger.kernel.org>	[2.6.35+]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The dmesg_restrict sysctl currently covers the syslog method for access
dmesg, however /dev/kmsg isn't covered by the same protections.  Most
people haven't noticed because util-linux dmesg(1) defaults to using the
syslog method for access in older versions.  With util-linux dmesg(1)
defaults to reading directly from /dev/kmsg.

To fix /dev/kmsg, let's compare the existing interfaces and what they
allow:

 - /proc/kmsg allows:
  - open (SYSLOG_ACTION_OPEN) if CAP_SYSLOG since it uses a destructive
    single-reader interface (SYSLOG_ACTION_READ).
  - everything, after an open.

 - syslog syscall allows:
  - anything, if CAP_SYSLOG.
  - SYSLOG_ACTION_READ_ALL and SYSLOG_ACTION_SIZE_BUFFER, if
    dmesg_restrict==0.
  - nothing else (EPERM).

The use-cases were:
 - dmesg(1) needs to do non-destructive SYSLOG_ACTION_READ_ALLs.
 - sysklog(1) needs to open /proc/kmsg, drop privs, and still issue the
   destructive SYSLOG_ACTION_READs.

AIUI, dmesg(1) is moving to /dev/kmsg, and systemd-journald doesn't
clear the ring buffer.

Based on the comments in devkmsg_llseek, it sounds like actions besides
reading aren't going to be supported by /dev/kmsg (i.e.
SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive
syslog syscall actions.

To this end, move the check as Josh had done, but also rename the
constants to reflect their new uses (SYSLOG_FROM_CALL becomes
SYSLOG_FROM_READER, and SYSLOG_FROM_FILE becomes SYSLOG_FROM_PROC).
SYSLOG_FROM_READER allows non-destructive actions, and SYSLOG_FROM_PROC
allows destructive actions after a capabilities-constrained
SYSLOG_ACTION_OPEN check.

 - /dev/kmsg allows:
  - open if CAP_SYSLOG or dmesg_restrict==0
  - reading/polling, after open

Addresses https://bugzilla.redhat.com/show_bug.cgi?id=903192

[akpm@linux-foundation.org: use pr_warn_once()]
Signed-off-by: NKees Cook <keescook@chromium.org>
Reported-by: NChristian Kujau <lists@nerdbynature.de>
Tested-by: NJosh Boyer <jwboyer@redhat.com>
Cc: Kay Sievers <kay@vrfy.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

There are instances in the kernel where we would like to disable CPU
hotplug (from sysfs) during some important operation.  Today the freezer
code depends on this and the code to do it was kinda tailor-made for
that.

Restructure the code and make it generic enough to be useful for other
usecases too.
Signed-off-by: NSrivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: NRobin Holt <holt@sgi.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Russ Anderson <rja@sgi.com>
Cc: Robin Holt <holt@sgi.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Shawn Guo <shawn.guo@linaro.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

If hci_dev_open fails we need to ensure that the corresponding
mgmt_set_powered command gets an appropriate response. This patch fixes
the missing response by adding a new mgmt_set_powered_failed function
that's used to indicate a power on failure to mgmt. Since a situation
with the device being rfkilled may require special handling in user
space the patch uses a new dedicated mgmt status code for this.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Cc: stable@vger.kernel.org
Acked-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

should be checked if "cur" is txable, not "port".

Introduced by commit 6e88e135 "team: use function team_port_txable()
for determing enabled and up port"
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We use 0x7000000000000000ULL as 0x6000000000000000ULL is reserved for
ARM64.
Signed-off-by: NDavid Daney <david.daney@cavium.com>
Signed-off-by: NGleb Natapov <gleb@redhat.com>

Filters need to be translated to real BPF code for userland, like SO_GETFILTER.
Signed-off-by: NNicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

__DECLARE_TRACE_RCU() currently creates an _rcuidle() tracepoint which
may safely be invoked from what RCU considers to be an idle CPU.
However, these _rcuidle() tracepoints may -not- be invoked from the
handler of an irq taken from idle, because rcu_idle_enter() zeroes
RCU's nesting-level counter, so that the rcu_irq_exit() returning to
idle will trigger a WARN_ON_ONCE().

This commit therefore substitutes rcu_irq_enter() for rcu_idle_exit()
and rcu_irq_exit() for rcu_idle_enter() in order to make the _rcuidle()
tracepoints usable from irq handlers as well as from process context.
Reported-by: NDave Jones <davej@redhat.com>
Signed-off-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>

Function media_entity_remote_source actually returns the remote pad to
the given one, regardless if this is the source or the sink pad.
Name media_entity_remote_pad is more adequate for this function.
Signed-off-by: NAndrzej Hajda <a.hajda@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NSakari Ailus <sakari.ailus@iki.fi>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Even though they are virtual widgets DAI widgets still get counted for the
DAPM context power management so we can't just use the active state to
check if they should be powered as they may not be part of a complete path.

Instead split them into input and output widgets and do the same power
checks as we perform on AIFs.
Reported-by: NStephen Warren <swarren@nvidia.com>
Tested-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

I broke them in this commit:

    commit 1be374a0
    Author: Andy Lutomirski <luto@amacapital.net>
    Date:   Wed May 22 14:07:44 2013 -0700

        net: Block MSG_CMSG_COMPAT in send(m)msg and recv(m)msg

This patch adds __sys_sendmsg and __sys_sendmsg as common helpers that accept
MSG_CMSG_COMPAT and blocks MSG_CMSG_COMPAT at the syscall entrypoints.  It
also reverts some unnecessary checks in sys_socketcall.

Apparently I was suffering from underscore blindness the first time around.
Signed-off-by: NAndy Lutomirski <luto@amacapital.net>
Tested-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since the introduction of preemptible mmu_gather TLB fast mode has been
broken. TLB fast mode relies on there being absolutely no concurrency;
it frees pages first and invalidates TLBs later.

However now we can get concurrency and stuff goes *bang*.

This patch removes all tlb_fast_mode() code; it was found the better
option vs trying to patch the hole by entangling tlb invalidation with
the scheduler.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Tony Luck <tony.luck@intel.com>
Reported-by: NMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: NPeter Zijlstra <peterz@infradead.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

According to include/uapi/linux/kvm_para.h architectures should define
kvm_para_available, so add an implementation to asm-generic/kvm_para.h
which just returns false.

This fixes intel8x0.c build failure on mips with KVM enabled.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: NGleb Natapov <gleb@redhat.com>

Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: NKamil Debski <k.debski@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

commit 56b765b7 ("htb: improved accuracy at high rates")
broke the "overhead xxx" handling, as well as the "linklayer atm"
attribute.

tc class add ... htb rate X ceil Y linklayer atm overhead 10

This patch restores the "overhead xxx" handling, for htb, tbf
and act_police

The "linklayer atm" thing needs a separate fix.
Reported-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Vimalkumar <j.vimal@gmail.com>
Cc: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Roman Gushchin discovered that udp4_lib_lookup2() was not reloading
first item in the rcu protected list, in case the loop was restarted.

This produced soft lockups as in https://lkml.org/lkml/2013/4/16/37

rcu_dereference(X)/ACCESS_ONCE(X) seem to not work as intended if X is
ptr->field :

In some cases, gcc caches the value or ptr->field in a register.

Use a barrier() to disallow such caching, as documented in
Documentation/atomic_ops.txt line 114

Thanks a lot to Roman for providing analysis and numerous patches.
Diagnosed-by: NRoman Gushchin <klamm@yandex-team.ru>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Reported-by: NBoris Zhmurov <zhmurov@yandex-team.ru>
Signed-off-by: NRoman Gushchin <klamm@yandex-team.ru>
Acked-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>