As GPIO descriptors are not going to remain unique anymore, having this
function public is not safe. Restrain its use to gpiolib since we have
no user outside of it.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

For some reason gpiochip_export() would invalidate all the descriptors
of a chip if exporting it to sysfs failed. This does not appear as
necessary. Remove that part of the code.

While we are at it, add a note about the non-safety of temporarily
releasing a spinlock in the middle of the loop that protects its
iterator, and explain why this is done.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

of_get_named_gpio_flags() has been made gpiolib-private by commit
f01d9075, but its EXPORT statement has not been removed. Fix this.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The gpio include file and the gpio documentation declare and document
GPIOF_ACTIVE_LOW as one of the flags to be passed to gpio_request_one
and related functions. However, the flag is not evaluated or used.

This can cause problems in at least two areas: First, the same API can
be used to auto-export pins to user space. The missing support for
GPIOF_ACTIVE_LOW results in unexpected behavior for such auto-exported
pins. Second, the requested gpio pin can be convered for use by
gpiod functions with gpio_to_desc(). While gpio API functions do not
support GPIOF_ACTIVE_LOW, gpiod functions do, which again results in
unexpected behavior.

Check the flag in gpio_request_one and set the gpio internal flag
FLAG_ACTIVE_LOW if it is set to address those problems.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The Zynq GPIO controller does not disable the interrupt detection when the
interrupt is masked and only disables the propagation of the interrupt. This
means when the controller detects an interrupt condition while the interrupt is
logically disabled (and masked) it will propagate the recorded interrupt event
once the interrupt is enabled. This will cause the interrupt consumer to see
spurious interrupts to prevent this first make sure that the interrupt is not
asserted and then enable it.

E.g. when a interrupt is requested with request_irq() it will be configured
according to the requested type (edge/level triggered, etc.) after that it will
be enabled. But the detection circuit might have already registered a false
interrupt before the interrupt type was correctly configured and once the
interrupt is unmasked this false interrupt will be propagated and the interrupt
handler for the just request interrupt will called.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Signed-off-by: Nabdoulaye berthe <berthe.ab@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Add support for GPIO controller used by Xilinx Zynq.

v3 changes:
 - Use linux/gpio/driver.h instead of linux/gpio.h
 - Make irq a local variable in probe

v2 changes:
 - convert to pm_runtime_force_(suspend|resume)
 - add pm_runtime_set_active in probe()
 - also (un)prepare clocks when they are dis-/enabled
 - add some missing calls to pm_runtime_get()
 - use pm_runtime_put() instead of sync variant
 - remove gpio chip in driver remove()
 - remove redundant type casts
 - directly use IO helpers
 - use BIT macro to set/clear bits
 - migrate to GPIOLIB_IRQCHIP
Signed-off-by: NHarini Katakam <harinik@xilinx.com>
Signed-off-by: NSoren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Following is the debug output (only a few examples) before and after
the patch.

$ dmesg | grep of_get_named_gpiod_flags

Before:
of_get_named_gpiod_flags: can't parse gpios property
 of node '/mmc@12220000[0]'
of_get_named_gpiod_flags exited with status 0

After:
of_get_named_gpiod_flags: can't parse 'wp-gpios' property
 of node '/mmc@12220000[0]'
of_get_named_gpiod_flags: parsed 'gpios' property of node
 '/gpio-keys/power[0]' - status (0)
Signed-off-by: NTushar Behera <tushar.b@samsung.com>
Reviewed-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The old integer GPIO interface is, in effect, a privileged user of the
gpiod interface. Reflect this fact further by moving legacy GPIO support
into its own source file. This makes the code clearer and will allow us
to disable legacy GPIO support in the (far) future.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

sysfs support is currently entangled within the core GPIO support, while
it should relly just be a (privileged) user of the integer GPIO API.
This patch is a first step towards making the gpiolib code more readable
by splitting it into logical parts.

Move all sysfs support to their own source file, and share static
members of gpiolib that need to be in the private gpiolib.h file. In
the future we will want to put some of them back into gpiolib.c, but this
first patch let us at least identify them.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Compiling out GPIO labels results in a space gain so small that it can
hardly be justified. Labels can also be useful for printing debug
messages, so always keep them around.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The GPIO OMAP driver didn't have a consistent naming scheme for
all its functions. Some of them had an omap prefix while others
didn't. There are many advantages on having a separate namespace
for driver functions so let's add an "omap" prefix to all of them.
Signed-off-by: NJavier Martinez Canillas <jmartinez@softcrates.net>
Acked-by: NKevin Hilman <khilman@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The <linux/irqchip/chained_irq.h> header is already included
when selecting GPIOLIB_IRQCHIP so there is no need to do it
in the driver. This is a left over from commit fb655f57
("gpio: omap: convert driver to use gpiolib irqchip").
Signed-off-by: NJavier Martinez Canillas <jmartinez@softcrates.net>
Acked-by: NKevin Hilman <khilman@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

GPIO irqchips assign to the cascaded IRQs their own lock class
in order to avoid warnings about lockdep recursions since that
allow the lockdep core to keep track of things.

Since commit e45d1c80 ("gpio: put GPIO IRQs into their own lock class")
there is no need to do this in a driver if it's using the GPIO
irqchip helpers since gpiolib already assigns a lockdep class.
Signed-off-by: NJavier Martinez Canillas <jmartinez@softcrates.net>
Acked-by: NKevin Hilman <khilman@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Removal of null pointer checks that could never happen

This was found using a static code analysis program called cppcheck
Signed-off-by: NRickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Removal of null pointer checks that could never happen

This was found using a static code analysis program called cppcheck
Signed-off-by: NRickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The driver was not checking the return value from gpiochip_add()
properly, so add a bail-out check.
Reported-by: NRickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

While it will be clamped to bool by gpiolib, let's make this sane
in the driver as well.
Signed-off-by: NJürg Billeter <j@bitron.ch>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

pxa_gpio_probe() has some issues supporting the gpio0 and gpio1
interrupts under device-tree - it never actually sets up the chain
handler to get interrupts on edge detect for GPIO0 and GPIO1.
Signed-off-by: NAndrew Ruder <andrew.ruder@elecsyscorp.com>
Acked-by: NHaojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Make of_device_id array const, because all OF functions handle
it as const.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Acked-by: NNeil Zhang <zhangwm@marvell.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Make of_device_id array const, because all OF functions handle
it as const.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Acked-By: NRoland Stigge <stigge@antcom.de>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

This switches the Intel MID GPIO driver over to using the gpiolib
irqchip helpers in the gpiolib core.

Cc: xinhui.pan <xinhuiX.pan@intel.com>
Acked-by: NDavid Cohen <david.a.cohen@linux.intel.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

This switches the STMPE driver to use the gpiolib irqchip
helpers.
Tested-by: NSilvio Fricke <silvio.fricke@gmail.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Call skb_pop_rcv_encapsulation and postpull_rcsum for the Ethernet
header to work properly with checksum complete.
Signed-off-by: NTom Herbert <therbert@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When we mirror packets from a vxlan tunnel to other device,
the mirror device should see the same packets (that is, without
outer header). Because vxlan tunnel sets dev->hard_header_len,
tcf_mirred() resets mac header back to outer mac, the mirror device
actually sees packets with outer headers

Vxlan tunnel should set dev->needed_headroom instead of
dev->hard_header_len, like what other ip tunnels do. This fixes
the above problem.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: stephen hemminger <stephen@networkplumber.org>
Cc: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: NCong Wang <cwang@twopensource.com>
Signed-off-by: NCong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch contains several fixes for Scsi START_STOP_UNIT. The previous
code did not account for signed vs. unsigned arithmetic which resulted
in an invalid lowest power state caculation when the device only supports
1 power state.

The code for Power Condition == 2 (Idle) was not following the spec. The
spec calls for setting the device to specific power states, depending
upon Power Condition Modifier, without accounting for the number of
power states supported by the device.

The code for Power Condition == 3 (Standby) was using a hard-coded '0'
which is replaced with the macro POWER_STATE_0.
Signed-off-by: NDan McLeran <daniel.mcleran@intel.com>
Reviewed-by: NVishal Verma <vishal.l.verma@linux.intel.com>
Signed-off-by: NMatthew Wilcox <matthew.r.wilcox@intel.com>

The nvme-scsi file defined its own Log Page constant.  Use the
newly-defined one from the header file instead.
Signed-off-by: NMatthew Wilcox <matthew.r.wilcox@intel.com>

There is a potential dead lock if a cpu event occurs during nvme probe
since it registered with hot cpu notification. This fixes the race by
having the module register with notification outside of probe rather
than have each device register.

The actual work is done in a scheduled work queue instead of in the
notifier since assigning IO queues has the potential to block if the
driver creates additional queues.
Signed-off-by: NKeith Busch <keith.busch@intel.com>
Signed-off-by: NMatthew Wilcox <matthew.r.wilcox@intel.com>

This patch fixes a NULL pointer dereference regression bug that was
introduced with:

commit 1e1110c4
Author: Mikulas Patocka <mpatocka@redhat.com>
Date:   Sat May 17 06:49:22 2014 -0400

    target: fix memory leak on XCOPY

Now that target_put_sess_cmd() -> kref_put_spinlock_irqsave() is
called with a valid se_cmd->cmd_kref, a NULL pointer dereference
is triggered because the XCOPY passthrough commands don't have
an associated se_session pointer.

To address this bug, go ahead and checking for a NULL se_sess pointer
within target_put_sess_cmd(), and call se_cmd->se_tfo->release_cmd()
to release the XCOPY's xcopy_pt_cmd memory.
Reported-by: NThomas Glanzmann <thomas@glanzmann.de>
Cc: Thomas Glanzmann <thomas@glanzmann.de>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org # 3.12+
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

Add software TSO support for FEC.
This feature allows to improve outbound throughput performance.

Tested on imx6dl sabresd board, running iperf tcp tests shows:
- 16.2% improvement comparing with FEC SG patch
- 82% improvement comparing with NO SG & TSO patch

$ ethtool -K eth0 tso on
$ iperf -c 10.192.242.167 -t 3 &
[  3] local 10.192.242.108 port 35388 connected with 10.192.242.167 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0- 3.0 sec   181 MBytes   506 Mbits/sec

During the testing, CPU loading is 30%.
Since imx6dl FEC Bandwidth is limited to SOC system bus bandwidth, the
performance with SW TSO is a milestone.

CC: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
CC: Eric Dumazet <eric.dumazet@gmail.com>
CC: David Laight <David.Laight@ACULAB.COM>
CC: Li Frank <B20596@freescale.com>
Signed-off-by: NFugang Duan <B38611@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add Scatter/gather support for FEC.
This feature allows to improve outbound throughput performance.

Tested on imx6dl sabresd board:
Running iperf tests shows a 55.4% improvement.

$ ethtool -K eth0 sg off
$ iperf -c 10.192.242.167 -t 3 &
[  3] local 10.192.242.108 port 52618 connected with 10.192.242.167 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0- 3.0 sec  99.5 MBytes   278 Mbits/sec

$ ethtool -K eth0 sg on
$ iperf -c 10.192.242.167 -t 3 &
[  3] local 10.192.242.108 port 52617 connected with 10.192.242.167 port 5001
[ ID] Interval       Transfer     Bandwidth
[  3]  0.0- 3.0 sec   154 MBytes   432 Mbits/sec

CC: Li Frank <B20596@freescale.com>
Signed-off-by: NFugang Duan <B38611@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to support SG, software TSO, let's increase BD entry number.

CC: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
CC: Eric Dumazet <eric.dumazet@gmail.com>
CC: David Laight <David.Laight@ACULAB.COM>
Signed-off-by: NFugang Duan <B38611@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to enhance the code readable, let's factorize the
feature list.
Signed-off-by: NFugang Duan <B38611@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

IP header checksum is calcalated by network layer in default.
To support software TSO, it is better to use HW calculate the
IP header checksum.

FEC hw checksum feature request the checksum field in frame
is zero, otherwise the calculative CRC is not correct.

For segmentated TCP packet, HW calculate the IP header checksum again,
it doesn't bring any impact. For SW TSO, HW calculated checksum bring
better performance.
Signed-off-by: NFugang Duan <B38611@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Make the code more readable and easy to support other features like
SG, TSO, moving the common transmit function to one api.

And the patch also factorize the getting BD index to it own function.

CC: David Laight <David.Laight@ACULAB.COM>
Signed-off-by: NFugang Duan <B38611@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

With some specific configuration (VT6105M on Soekris 5510 and depending
on the device at the other end), fragmented packets were not transmitted
when forcing 100 full-duplex with autoneg disable.

This fix now write full-duplex chips register when forcing full or
half-duplex not only when autoneg is enable.
Signed-off-by: NFrançois Cachereul <f.cachereul@alphalink.fr>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

A malicious VF might try to starve the other VFs & PF by creating
contineous doorbell floods. In order to negate this, HW has a threshold of
doorbells per client, which will stop the client doorbells from arriving
if crossed.

The threshold currently configured for VFs is too low - under extreme traffic
scenarios, it's possible for a VF to reach the threshold and thus for its
fastpath to stop working.
Signed-off-by: NAriel Elior <ariel.elior@qlogic.com>
Signed-off-by: NYuval Mintz <yuval.mintz@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If L2FW utilized by the UNDI driver has the same version number as that
of the regular FW, a driver loading after UNDI and receiving an uncommon
answer from management will mistakenly assume the loaded FW matches its
own requirement and try to exist the flow via FLR.
Signed-off-by: NYuval Mintz <yuval.mintz@qlogic.com>
Signed-off-by: NAriel Elior <ariel.elior@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Set the phy access mode even in case of link-flap avoidance.
Signed-off-by: NYaniv Rosner <yaniv.rosner@qlogic.com>
Signed-off-by: NYuval Mintz <yuval.mintz@qlogic.com>
Signed-off-by: NAriel Elior <ariel.elior@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This avoids clearing the RX polarity setting in KR mode when polarity lane
is swapped, as otherwise this will result in failed link.
Signed-off-by: NYaniv Rosner <yaniv.rosner@qlogic.com>
Signed-off-by: NYuval Mintz <yuval.mintz@qlogic.com>
Signed-off-by: NAriel Elior <ariel.elior@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>