Upcoming Intel systems will have an ACPI method to control whether a USB
port can be completely powered off.  The implication of powering off a
USB port is that the device and host sees a physical disconnect, and
subsequent port connections and remote wakeups will be lost.

Add a new function, usb_acpi_power_manageable(), that can be used to
find whether the usb port has ACPI power resources that can be used to
power on and off the port on these machines. Also add a new function
called usb_acpi_set_power_state() that controls the port power via these
ACPI methods.

When the USB core calls into the xHCI hub driver to power off a port,
check whether the port can be completely powered off via this new ACPI
mechanism.  If so, call into these new ACPI methods.  Also use the ACPI
methods when the USB core asks to power on a port.
Signed-off-by: NLan Tianyu <tianyu.lan@intel.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

In the upcoming USB port power off patches, we need to know whether a
USB port can ever see a disconnect event.  Often USB ports are internal
to a system, and users can't disconnect USB devices from that port.
Sometimes those ports will remain empty, because the OEM chose not to
connect an internal USB device to that port.

According to ACPI Spec 9.13, PLD indicates whether USB port is
user visible and _UPC indicates whether a USB device can be connected to
the USB port (we'll call this "connectible").  Here's a matrix of the
possible combinations:

Visible Connectible
		Name		Example
-------------------------------------------------------------------------

Yes	No	Unknown		(Invalid state.)

Yes	Yes	Hot-plug	USB ports on the outside of a laptop.
				A user could freely connect and disconnect
				USB devices.

No	Yes	Hard-wired	A USB modem hard-wired to a port on the
				inside of a laptop.

No	No	Not used	The port is internal to the system and
				will remain empty.

Represent each of these four states with an enum usb_port_connect_type.
The four states are USB_PORT_CONNECT_TYPE_UNKNOWN,
USB_PORT_CONNECT_TYPE_HOT_PLUG, USB_PORT_CONNECT_TYPE_HARD_WIRED, and
USB_PORT_NOT_USED.  When we get the USB port's acpi_handle, store the
state in connect_type in struct usb_port.
Signed-off-by: NLan Tianyu <tianyu.lan@intel.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The usb_device structure contains an array of usb_device "children".
This array is only valid if the usb_device is a hub, so it makes no
sense to store it there.  Instead, store the usb_device child
in its parent usb_port structure.

Since usb_port is an internal USB core structure, add a new function to
get the USB device child, usb_hub_find_child().  Add a new macro,
usb_hub_get_each_child(), to iterate over all the children attached to a
particular USB hub.

Remove the printing the USB children array pointer from the usb-ip
driver, since it's really not necessary.
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NLan Tianyu <tianyu.lan@intel.com>
Signed-off-by: NSarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

DeviceRemovalbe and wHubDelay for usb3.0 hub are little-endian
and so define them as _le16.
Signed-off-by: NLan Tianyu <tianyu.lan@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The kernel's version number is used as decimal in the bcdDevice field of
the RH descriptor. For kernel version v3.12 we would see 3.0c in lsusb.
I am not sure how important it is to stick with bcd values since this is
this way since we started git history and nobody complained (however back
then we reported only 2.6).
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Apps which deal with devices which also have a kernel driver, need to do
the following:
1) Check which driver is attached, so as to not detach the wrong driver
   (ie detaching usbfs while another instance of the app is using the device)
2) Detach the kernel driver
3) Claim the interface

Where moving from one step to the next for both 1-2 and 2-3 consists of
a (small) race window. So currently such apps are racy and people just live
with it.

This patch adds a new ioctl which makes it possible for apps to do this
in a race free manner. For flexibility apps can choose to:
1) Specify the driver to disconnect
2) Specify to disconnect any driver except for the one named by the app
3) Disconnect any driver

Note that if there is no driver attached, the ioctl will just act like the
regular claim-interface ioctl, this is by design, as returning an error for
this condition would open a new bag of race-conditions.

Changes in v2:
-Fix indentation of if blocks where the condition spans multiple lines
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The bcdDevice field is defined as
|Device release number in binary-coded decimal
in the USB 2.0 specification. We use this field to distinguish the UDCs
from each other. In theory this could be used on the host side to apply
certain quirks if the "special" UDC in combination with this gadget is
used. This hasn't been done as far as I am aware. In practice it would
be better to fix the UDC driver before shipping since a later release
might not need this quirk anymore.

There are some driver in tree (on the host side) which use the bcdDevice
field to figure out special workarounds for a given firmware revision.
This seems to make sense. Therefore this patch converts all gadgets
(except a few) to use the kernel version instead a random 2 or 3 plus
the UDC number. The few that don't report kernel's version are:
- webcam
  This one reports always a version 0x10 so allow it to do so in future.
- nokia
  This one reports always 0x211. The comment says that this gadget works
  only if the UDC supports altsettings so I added a check for this.
- serial
  This one reports 0x2400 + UDC number. Since the gadget version is 2.4
  this could make sense. Therefore bcdDevice is 0x2400 here.

I also remove various gadget_is_<name> macros which are unused. The
remaining few macros should be moved to feature / bug bitfield.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Some gadgets provide custom entry here. Some may override it with an
etntry that is also created by composite if there was no value sumbitted
at all.
This patch removes all "custom manufacturer" strings which are the same
as these which are created by composite. Then it moves the creation of
the default manufacturer string to usb_composite_overwrite_options() in
case no command line argument has been used and the entry is still an
empty string.
By doing this we get rid of the global variable "composite_manufacturer"
in composite.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The struct usb_composite_driver members iProduct, iSerial and
iManufacturer can be entered directly via the string array. There is no
need for them to appear here.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch pushes the iProduct module argument from composite
into each gadget.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch pushes the iManufacturer module argument from composite into
each gadget. Once the user uses the module paramter, the string is
overwritten with the final value.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch pushes the iSerialNumber module argument from composite into
each gadget. Once the user uses the module paramter, the string is
overwritten with the final value.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The index in usb_string array is defined by the gadget. The gadget can
choose which index entry it assigns for the serial number and which the
product name. The gadget has just to ensure that the descriptor contains
the proper string id which is assigned by composite.
If the composite layer knows the index of the "default" information
which will be overwritten by module parameters, it can be used later to
overwrite it.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch moves the module options idVendor, idProduct and bcdDevice
from composite.c into each gadgets. This ensures compatibility with
current gadgets and removes the global variable which brings me step
closer towards composite.c in libcomposite
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

there is no read user of bufsiz, its content is available via
USB_COMP_EP0_BUFSIZ. Remove it.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch moves USB_BUFSIZ into global header file as
USB_COMP_EP0_BUFSIZ. There is currently only one user (f_sourcesink)
besides composite which need it. Ideally f_sourcesink would have its
own ep0 buffer. Lets keep it that way it was for now.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

epautoconf has two global variables which count the endpoint number of
last assigned endpoint.
This patch removes the global variable and keeps it as per (UDC) gadget.
While here, the ifdef is removed and now the in and outpoint are
enumerated unconditionally. The dwc3 for instance supports 32 endpoints
in total.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch removes the global variable composite in composite.c.
The private data which was saved there is now passed via an additional
argument to the bind() function in struct usb_gadget_driver.

Only the "old-style" UDC drivers have to be touched here, new style are
doing it right because this change is made in udc-core.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This partly reverts 07a18bd7 ("usb gadget: don't save bind callback in
struct usb_composite_driver") and fixes new drivers. The section missmatch
problems was solved by whitelisting bind callback in modpost.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This partly reverts 07a18bd7 ("usb gadget: don't save bind callback in
struct usb_composite_driver") and fixes new drivers. The section missmatch
problems was solved by whitelisting structs in question via __ref.

Cc: devel@driverdev.osuosl.org
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

It was moved to be an argument in 07a18bd7 ("usb gadget: don't
save bind callback in struct usb_composite_driver"). The reason was to
avoid the section missmatch. The warning was shown because ->bind is
marked as __init becuase it is a one time init. The warning can be also
suppresed by whitelisting the variable i.e. rename it to lets say _probe.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The "Low Performance USB Block driver" has been removed which a user of
libusual. Now we have only the usb-storage driver as the only driver in
tree. This makes libusual needless.
This patch removes libusal, fixes up all users. The usual-table is now
linked into usb-storage.
usb_usual.h remains in public include directory because some staging
users seem to need it.
Signed-off-by: NSebastian Andrzej Siewior <sebastian@breakpoint.cc>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

when missing USB PHY clock, kernel booting up will hang during USB
initialization. We should check USBGP[PHY_CLK_VALID] bit to avoid
CPU hanging in this case.
Signed-off-by: NShengzhou Liu <Shengzhou.Liu@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This commit removes USB_GADGET_DUALSPEED and USB_GADGET_SUPERSPEED
Kconfig options.  Since now kernel allows many UDC drivers to be
compiled, those options may turn to no longer be valid.  For
instance, if someone decides to build UDC that supports super
speed and UDC that supports high speed only, the latter will be
"assumed" to support super speed since USB_GADGET_SUPERSPEED will
be selected by the former.

The test of whether CONFIG_USB_GADGET_*SPEED was defined was just
an optimisation which removed otherwise dead code (ie. if UDC is
not dual speed, there is no need to handle cases that can happen
if speed is high).  This commit removes those checks.
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

equivalent of
	mutex_lock(mutex);
	if (!kref_put(kref, release))
		mutex_unlock(mutex);
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Jim Schutt reported a problem that pointed at compaction contending
heavily on locks.  The workload is straight-forward and in his own words;

	The systems in question have 24 SAS drives spread across 3 HBAs,
	running 24 Ceph OSD instances, one per drive.  FWIW these servers
	are dual-socket Intel 5675 Xeons w/48 GB memory.  I've got ~160
	Ceph Linux clients doing dd simultaneously to a Ceph file system
	backed by 12 of these servers.

Early in the test everything looks fine

  procs -------------------memory------------------ ---swap-- -----io---- --system-- -----cpu-------
   r  b       swpd       free       buff      cache   si   so    bi    bo   in   cs  us sy  id wa st
  31 15          0     287216        576   38606628    0    0     2  1158    2   14   1  3  95  0  0
  27 15          0     225288        576   38583384    0    0    18 2222016 203357 134876  11 56  17 15  0
  28 17          0     219256        576   38544736    0    0    11 2305932 203141 146296  11 49  23 17  0
   6 18          0     215596        576   38552872    0    0     7 2363207 215264 166502  12 45  22 20  0
  22 18          0     226984        576   38596404    0    0     3 2445741 223114 179527  12 43  23 22  0

and then it goes to pot

  procs -------------------memory------------------ ---swap-- -----io---- --system-- -----cpu-------
   r  b       swpd       free       buff      cache   si   so    bi    bo   in   cs  us sy  id wa st
  163  8          0     464308        576   36791368    0    0    11 22210  866  536   3 13  79  4  0
  207 14          0     917752        576   36181928    0    0   712 1345376 134598 47367   7 90   1  2  0
  123 12          0     685516        576   36296148    0    0   429 1386615 158494 60077   8 84   5  3  0
  123 12          0     598572        576   36333728    0    0  1107 1233281 147542 62351   7 84   5  4  0
  622  7          0     660768        576   36118264    0    0   557 1345548 151394 59353   7 85   4  3  0
  223 11          0     283960        576   36463868    0    0    46 1107160 121846 33006   6 93   1  1  0

Note that system CPU usage is very high blocks being written out has
dropped by 42%. He analysed this with perf and found

  perf record -g -a sleep 10
  perf report --sort symbol --call-graph fractal,5
    34.63%  [k] _raw_spin_lock_irqsave
            |
            |--97.30%-- isolate_freepages
            |          compaction_alloc
            |          unmap_and_move
            |          migrate_pages
            |          compact_zone
            |          compact_zone_order
            |          try_to_compact_pages
            |          __alloc_pages_direct_compact
            |          __alloc_pages_slowpath
            |          __alloc_pages_nodemask
            |          alloc_pages_vma
            |          do_huge_pmd_anonymous_page
            |          handle_mm_fault
            |          do_page_fault
            |          page_fault
            |          |
            |          |--87.39%-- skb_copy_datagram_iovec
            |          |          tcp_recvmsg
            |          |          inet_recvmsg
            |          |          sock_recvmsg
            |          |          sys_recvfrom
            |          |          system_call
            |          |          __recv
            |          |          |
            |          |           --100.00%-- (nil)
            |          |
            |           --12.61%-- memcpy
             --2.70%-- [...]

There was other data but primarily it is all showing that compaction is
contended heavily on the zone->lock and zone->lru_lock.

commit [b2eef8c0: mm: compaction: minimise the time IRQs are disabled
while isolating pages for migration] noted that it was possible for
migration to hold the lru_lock for an excessive amount of time. Very
broadly speaking this patch expands the concept.

This patch introduces compact_checklock_irqsave() to check if a lock
is contended or the process needs to be scheduled. If either condition
is true then async compaction is aborted and the caller is informed.
The page allocator will fail a THP allocation if compaction failed due
to contention. This patch also introduces compact_trylock_irqsave()
which will acquire the lock only if it is not contended and the process
does not need to schedule.
Reported-by: NJim Schutt <jaschut@sandia.gov>
Tested-by: NJim Schutt <jaschut@sandia.gov>
Signed-off-by: NMel Gorman <mgorman@suse.de>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix the following warning:

  usr/include/linux/string.h:8: userspace cannot reference function or variable defined in the kernel
Signed-off-by: NWANG Cong <xiyou.wangcong@gmail.com>
Acked-by: NAkinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Since the attribute avoid_reset_quirk is work for all devices including
those devices that can't morph, convert USB_QUIRK_RESET_MORPHS to
USB_QUIRK_RESET.
Signed-off-by: NLan Tianyu <tianyu.lan@intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

If a packet is emitted on one socket in one group of fanout sockets,
it is transmitted again. It is thus read again on one of the sockets
of the fanout group. This result in a loop for software which
generate packets when receiving one.
This retransmission is not the intended behavior: a fanout group
must behave like a single socket. The packet should not be
transmitted on a socket if it originates from a socket belonging
to the same fanout group.

This patch fixes the issue by changing the transmission check to
take fanout group info account.
Reported-by: NAleksandr Kotov <a1k@mail.ru>
Signed-off-by: NEric Leblond <eric@regit.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Macro for_each_child_of_node() makes it easy to iterate over all of the
children for a given device tree node, including those nodes that are
marked as unavailable (i.e. status = "disabled").

Introduce for_each_available_child_of_node(), which is like
for_each_child_of_node(), but it automatically skips unavailable nodes.
This also requires the introduction of helper function
of_get_next_available_child(), which returns the next available child
node.
Signed-off-by: NTimur Tabi <timur@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As struct device is used as a function argument, it should at
least be declared (device.h is not included).
Signed-off-by: NRichard Genoud <richard.genoud@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Changed the type of rsrc_start and rsrc_len from u64 to resource_size_t
so that it behaves well with resource API's like *resource_size()* on both
32bit and 64bit devices.
Signed-off-by: NKishon Vijay Abraham I <kishon@ti.com>
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

With small channel spacing values and high reference frequencies it is
possible to exceed the range of the 10-bit counter.
Workaround by checking the range and widening some constrains.

We don't use the REG1_PHASE value in this case the datasheet recommends to set
it to 1 if not used.
Signed-off-by: NMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

These functions are just boolean, let them return
bool instead of int.

Cc: David Miller <davem@davemloft.net>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Although this doesn't matter actually, because netpoll_tx_running()
doesn't use the parameter, the code will be more readable.

For team_dev_queue_xmit() we have to move it down to avoid
compile errors.

Cc: David Miller <davem@davemloft.net>
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch fixes several problems in the call path of
netpoll_send_skb_on_dev():

1. Disable IRQ's before calling netpoll_send_skb_on_dev().

2. All the callees of netpoll_send_skb_on_dev() should use
   rcu_dereference_bh() to dereference ->npinfo.

3. Rename arp_reply() to netpoll_arp_reply(), the former is too generic.

Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The logic of the code is same, just call netpoll_rx_on().

Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In __netpoll_rx(), it dereferences ->npinfo without rcu_dereference_bh(),
this patch fixes it by using the 'npinfo' passed from netpoll_rx()
where it is already dereferenced with rcu_dereference_bh().

Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Like the previous patch, slave_disable_netpoll() and __netpoll_cleanup()
may be called with read_lock() held too, so we should make them
non-block, by moving the cleanup and kfree() to call_rcu_bh() callbacks.

Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

slave_enable_netpoll() and __netpoll_setup() may be called
with read_lock() held, so should use GFP_ATOMIC to allocate
memory. Eric suggested to pass gfp flags to __netpoll_setup().

Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NCong Wang <amwang@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>