The tx interrupt is of edge type, and in case such interrupt is triggered
while it is masked it will not be handled even after tx interrupts are
re-enabled in the end of NAPI poll.
This will cause tx network to stop in the following scenario:
 * Rx is being handled, hence interrupts are masked.
 * Tx interrupt is triggered after checking if there is some tx to handle
   and before re-enabling the interrupts.
In this situation only rx transaction will release tx requests.

In order to handle the tx that was missed( if there was one ),
a NAPI reschdule was added after enabling the interrupts.
Signed-off-by: NElad Kanfi <eladkan@mellanox.com>
Acked-by: NNoam Camus <noamca@mellanox.com>
Acked-by: NGilad Ben-Yossef <giladby@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Below is a description of a possible problematic
sequence. CPU-A is sending a frame and CPU-B handles
the interrupt that indicates the frame was sent. CPU-B
reads an invalid value of tx_packet_sent.

	CPU-A				CPU-B
	-----				-----
	nps_enet_send_frame
	.
	.
	tx_skb = skb
	tx_packet_sent = true
	order HW to start tx
	.
	.
	HW complete tx
			    ------> 	get tx complete interrupt
					.
					.
					if(tx_packet_sent == true)
						handle tx_skb

	end memory transaction
	(tx_packet_sent actually
	 written)

Furthermore there is a dependency between tx_skb and tx_packet_sent.
There is no assurance that tx_skb contains a valid pointer at CPU B
when it sees tx_packet_sent == true.

Solution:

Initialize tx_skb to NULL and use it to indicate that packet was sent,
in this way tx_packet_sent can be removed.
Add a write memory barrier after setting tx_skb in order to make sure
that it is valid before HW is informed and IRQ is fired.

Fixed sequence will be:

       CPU-A                           CPU-B
       -----                           -----

	tx_skb = skb
	wmb()
	.
	.
	order HW to start tx
	.
	.
	HW complete tx
			------>		get tx complete interrupt
					.
					.
					if(tx_skb != NULL)
						handle tx_skb

					tx_skb = NULL
Signed-off-by: NElad Kanfi <eladkan@mellanox.com>
Acked-by: NNoam Camus <noamca@mellanox.com>
Acked-by: NGilad Ben-Yossef <giladby@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Reserved fields should be set to zero to avoid exposing
bits from the kernel stack.
Signed-off-by: NHeinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When reopening the network device on ra7795/salvator-x, e.g. after a
DHCP timeout:

    IP-Config: Reopening network devices...
    genirq: Flags mismatch irq 139. 00000000 (eth0:ch24:emac) vs. 00000000 (eth0:ch24:emac)
    ravb e6800000.ethernet eth0: cannot request IRQ eth0:ch24:emac
    IP-Config: Failed to open eth0
    IP-Config: No network devices available

The "mismatch" is due to requesting an IRQ that is already in use,
while IRQF_PROBE_SHARED wasn't set.

However, the real cause is that ravb_close() doesn't release the R-Car
Gen3-specific secondary IRQ.

Add the missing free_irq() call to fix this.

Fixes: 22d4df8f ("ravb: Add support for r8a7795 SoC")
Signed-off-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Acked-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

VXLAN can be disabled at compile-time or it can be a loadable
module while mlx5 is built-in, which leads to a link error:

drivers/net/built-in.o: In function `mlx5e_create_netdev':
ntb_netdev.c:(.text+0x106de4): undefined reference to `vxlan_get_rx_port'

This avoids the link error and makes the vxlan code optional,
like the other ethernet drivers do as well.

Link: https://patchwork.ozlabs.org/patch/589296/
Fixes: b3f63c3d ("net/mlx5e: Add netdev support for VXLAN tunneling")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This reverts commit 69976fb1.

We cannot select VXLAN when IPv4 support is disabled, that just gives
us additional build errors, including:

warning: (MLX5_CORE_EN) selects VXLAN which has unmet direct dependencies (NETDEVICES && NET_CORE && INET)
In file included from ../drivers/net/vxlan.c:36:0:
include/net/udp_tunnel.h: In function 'udp_tunnel_handle_offloads':
include/net/udp_tunnel.h:112:9: error: implicit declaration of function 'iptunnel_handle_offloads' [-Werror=implicit-function-declaration]
  return iptunnel_handle_offloads(skb, type);
         ^~~~~~~~~~~~~~~~~~~~~~~~

I'm sending a proper fix for the original bug in a separate patch.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The MACsec standard mentions a key identifier for each key, but
doesn't specify anything about it, so I arbitrarily chose 64 bits.

IEEE 802.1X-2010 specifies MKA (MACsec Key Agreement), and defines the
key identifier to be 128 bits (96 bits "member identifier" + 32 bits
"key number").
Signed-off-by: NSabrina Dubroca <sd@queasysnail.net>
Acked-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If GSO packet is segmented and its segments are properly queued,
we call consume_skb() instead of kfree_skb() to be drop monitor
friendly.

Fixes: 3e4f8b78 ("macvtap: Perform GSO on forwarding path.")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Vlad Yasevich <vyasevic@redhat.com>
Reviewed-by: NShmulik Ladkani <shmulik.ladkani@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

"data_split" was never set to false.  It's just uninitialized.

Fixes: 2950219d ('qede: Add basic network device support')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The error handling is broken here.  netxen_rom_fast_read() returns zero
on success and -EIO on error.  It never returns -1.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

My static checker complains that we are using "autoneg" without
initializing it.  The problem is the ->phy_read() condition is reversed
so we only set this on error instead of success.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

My static checker complained that "v" can be used unintialized if
netxen_rom_fast_read() returns -EIO.  That function never actually
returns -1.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When we fail to set the flooding configuration for the broadcast and
unregistered multicast traffic, we should revert the flooding
configuration of the unknown unicast traffic.

Fixes: 0293038e ("mlxsw: spectrum: Add support for flood control")
Signed-off-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Make the leave procedure in the error path symmetric to the join
procedure and first remove the port from the collector before
potentially destroying the LAG.

Fixes: 0d65fc13 ("mlxsw: spectrum: Implement LAG port join/leave")
Signed-off-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

UDP tunnel segmentation code relies on the inner offsets being set for
an UDP tunnel GSO packet, but the inner *_complete() functions will
set the inner offsets only if 'encapsulation' is set before calling
them.  Currently, udp_gro_complete() sets 'encapsulation' only after
the inner *_complete() functions are done.  This causes the inner
offsets having invalid values after udp_gro_complete() returns, which
in turn will make it impossible to properly segment the packet in case
it needs to be forwarded, which would be visible to the user either as
invalid packets being sent or as packet loss.

This patch fixes this by setting skb's 'encapsulation' in
udp_gro_complete() before calling into the inner complete functions,
and by making each possible UDP tunnel gro_complete() callback set the
inner_mac_header to the beginning of the tunnel payload.
Signed-off-by: NJarno Rajahalme <jarno@ovn.org>
Reviewed-by: NAlexander Duyck <aduyck@mirantis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The setting of the UDP tunnel GSO type is already performed by
udp[46]_gro_complete().
Signed-off-by: NJarno Rajahalme <jarno@ovn.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

qede requires qed to provide enough resources to accommodate 16 combined
channels, but that upper-bound isn't actually being enforced by it.
Instead, qed inform back to qede how many channels can be opened based on
available resources - but that calculation doesn't really take into account
the resources requested by qede; Instead it considers other FW/HW available
resources.

As a result, if a user would increase the number of channels to more than
16 [e.g., using ethtool] the chip would hang.

This change increments the resources requested by qede to 64 combined
channels instead of 16; This value is an upper bound on the possible
available channels [due to other FW/HW resources].
Signed-off-by: NSudarsana Reddy Kalluru <sudarsana.kalluru@qlogic.com>
Signed-off-by: NYuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use htons instead of unconditionally byte swapping nexthdr.  On a little
endian systems shifting the byte is correct behavior, but it results in
incorrect csums on big endian architectures.

Fixes: f8c6455b ('net/mlx4_en: Extend checksum offloading by CHECKSUM COMPLETE')
Signed-off-by: NDaniel Jurgens <danielj@mellanox.com>
Reviewed-by: NCarol Soto <clsoto@us.ibm.com>
Tested-by: NCarol Soto <clsoto@us.ibm.com>
Signed-off-by: NTariq Toukan <tariqt@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix problems in uapi definitions reported by Gabriel Laskar: (see
https://lkml.org/lkml/2016/4/5/205 for details)

 - move public header file rio_mport_cdev.h to include/uapi/linux directory
 - change types in data structures passed as IOCTL parameters
 - improve parameter checking in some IOCTL service routines
Signed-off-by: NAlexandre Bounine <alexandre.bounine@idt.com>
Reported-by: NGabriel Laskar <gabriel@lse.epita.fr>
Tested-by: NBarry Wood <barry.wood@idt.com>
Cc: Gabriel Laskar <gabriel@lse.epita.fr>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Aurelien Jacquiot <a-jacquiot@ti.com>
Cc: Andre van Herk <andre.van.herk@prodrive-technologies.com>
Cc: Barry Wood <barry.wood@idt.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

iSER currently has a couple places that set max_sectors in either the host
template or SCSI host, and all of them get it wrong.

This patch instead uses a single assignment that (hopefully) gets it right:
the max_sectors value must be derived from the number of segments in the
FR or FMR structure, but actually be one lower than the page size multiplied
by the number of sectors, as it has to handle the case of non-aligned I/O.

Without this I get trivial to reproduce hangs when running xfstests
(on XFS) over iSER to Linux targets.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMax Gurtovoy <maxg@mellanox.com>
Acked-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Just checking ifdef CONFIG_USB is not enough, if the USB is compiled
as module. The same applies to PCI.

Tested with the following .config alternatives:

CONFIG_USB=m
CONFIG_MEDIA_CONTROLLER=y
CONFIG_MEDIA_SUPPORT=m
CONFIG_VIDEO_AU0828=m

CONFIG_USB=m
CONFIG_MEDIA_CONTROLLER=y
CONFIG_MEDIA_SUPPORT=y
CONFIG_VIDEO_AU0828=m

CONFIG_USB=y
CONFIG_MEDIA_CONTROLLER=y
CONFIG_MEDIA_SUPPORT=y
CONFIG_VIDEO_AU0828=m

CONFIG_USB=y
CONFIG_MEDIA_CONTROLLER=y
CONFIG_MEDIA_SUPPORT=y
CONFIG_VIDEO_AU0828=y
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

If of_node is set before calling platform_device_add, the driver core
will try to use of: modalias matching, which fails because the device
tree nodes don't have a compatible property set. This patch fixes
imx-ipuv3-crtc module autoloading by setting the of_node property only
after the platform modalias is set.

Fixes: 304e6be6 ("gpu: ipu-v3: Assign of_node of child platform devices to corresponding ports")
Reported-by: NDennis Gilmore <dennis@ausil.us>
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Tested-By: NDennis Gilmore <dennis@ausil.us>
Cc: stable@vger.kernel.org # 4.4+
Signed-off-by: NDave Airlie <airlied@redhat.com>

Regulators are optional for devices using OPPs and the OPP core
shouldn't be printing any errors for such missing regulators.

It was fine before the commit 0c717d0f, but that failed to update
this part of the code to remove an 'always true' check and an extra
unwanted print message.

Fix that now.

Fixes: 0c717d0f (PM / OPP: Initialize regulator pointer to an error value)
Reported-by: NMarc Gonzalez <marc_gonzalez@sigmadesigns.com>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: NStephen Boyd <sboyd@codeaurora.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

The multicast/all-multicast internal flags are not properly restored
after device reset.  This could lead to unreliable multicast operations
after an ethtool configuration change for example.

Call bnxt_mc_list_updated() and setup the vnic->mask in bnxt_init_chip()
to fix the issue.
Signed-off-by: NMichael Chan <michael.chan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The code determines if the next ring entry is valid before proceeding
further to read the rest of the entry.  The CPU can re-order and read
the rest of the entry first, possibly reading a stale entry, if DMA
of a new entry happens right after reading it.  This issue can be
readily seen on a ppc64 system, causing it to crash.
Signed-off-by: NMichael Chan <michael.chan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

ACPICA commit 7a3bd2d962f221809f25ddb826c9e551b916eb25

Set the mutex owner thread ID.
Original patch from: Prarit Bhargava <prarit@redhat.com>

Link: https://bugzilla.kernel.org/show_bug.cgi?id=115121
Link: https://github.com/acpica/acpica/commit/7a3bd2d9Signed-off-by: NPrarit Bhargava <prarit@redhat.com>
Tested-by: Andy Lutomirski <luto@kernel.org> # On a Dell XPS 13 9350
Signed-off-by: NBob Moore <robert.moore@intel.com>
Signed-off-by: NLv Zheng <lv.zheng@intel.com>
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

In the receive path a queue's work bit was cleared unconditionally even
if fec_enet_rx_queue only read out a part of the available packets from
the hardware. This resulted in not reading any packets in the next napi
turn and so packets were delayed or lost.

The obvious fix is to only clear a queue's bit when the queue was
emptied.

Fixes: 4d494cdc ("net: fec: change data structure to support multiqueue")
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: NLucas Stach <l.stach@pengutronix.de>
Tested-by: NFugang Duan <fugang.duan@nxp.com>
Acked-by: NFugang Duan <fugang.duan@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When probe bails out with an error, we try to unregister the
netdev before we have even registered it. Fix the goto statements
for that.
Signed-off-by: NMatthias Brugger <mbrugger@suse.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The TSO code creates A-MSDUs from a single large send. Each
A-MSDU is an skb and skb->len doesn't include the number of
bytes which need to be added for the headers being added
(subframe header, TCP header, IP header, SNAP, padding).

To be able to set the right value in the Tx command, we
put the number of bytes added by those headers in
driver_data in iwl_mvm_tx_tso and use this value in
iwl_mvm_set_tx_cmd.

The problem by setting this value in driver_data is that
it overrides the ieee80211_tx_info. The bug manifested
itself when we send P2P related frames in CCK since the
rate in ieee80211_tx_info is zero-ed. This of course is
a violation of the P2P specification.

To fix this, copy the original ieee80211_tx_info to the
stack and pass it to the functions which need it.
Assign the number of bytes added by the headers to the
driver_data inside the skb itself.

Fixes: a6d5e32f ("iwlwifi: mvm: send large SKBs to the transport")
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>

The copying of ring data was wrong for two cases: For a full ring
nothing got copied at all (as in that case the canonicalized producer
and consumer indexes are identical). And in case one or both of the
canonicalized (after the resize) indexes would point into the second
half of the buffer, the copied data ended up in the wrong (free) part
of the new buffer. In both cases uninitialized data would get passed
back to the caller.

Fix this by simply copying the old ring contents twice: Once to the
low half of the new buffer, and a second time to the high half.

This addresses the inability to boot a HVM guest with 64 or more
vCPUs.  This regression was caused by 86200154 (xen/evtchn:
dynamically grow pending event channel ring).
Reported-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: NJan Beulich <jbeulich@suse.com>
Cc: <stable@vger.kernel.org> # 4.4+
Signed-off-by: NDavid Vrabel <david.vrabel@citrix.com>

After commit 8fa520af "intel_pstate: Remove freq calculation from
intel_pstate_calc_busy()" intel_pstate_get() calls get_avg_frequency()
to compute the average frequency, which is problematic for two reasons.

First, intel_pstate_get() may be invoked before the driver reads the
CPU feedback registers for the first time and if that happens,
get_avg_frequency() will attempt to divide by zero.

Second, the get_avg_frequency() call in intel_pstate_get() is racy
with respect to intel_pstate_sample() and it may end up returning
completely meaningless values for this reason.

Moreover, after commit 7349ec04 "intel_pstate: Move
intel_pstate_calc_busy() into get_target_pstate_use_performance()"
sample.core_pct_busy is never computed on Atom, but it is used in
intel_pstate_adjust_busy_pstate() in that case too.

To address those problems notice that if sample.core_pct_busy
was used in the average frequency computation carried out by
get_avg_frequency(), both the divide by zero problem and the
race with respect to intel_pstate_sample() would be avoided.

Accordingly, move the invocation of intel_pstate_calc_busy() from
get_target_pstate_use_performance() to intel_pstate_update_util(),
which also will take care of the uninitialized sample.core_pct_busy
on Atom, and modify get_avg_frequency() to use sample.core_pct_busy
as per the above.
Reported-by: Nkernel test robot <ying.huang@linux.intel.com>
Link: http://marc.info/?l=linux-kernel&m=146226437623173&w=4
Fixes: 8fa520af "intel_pstate: Remove freq calculation from intel_pstate_calc_busy()"
Fixes: 7349ec04 "intel_pstate: Move intel_pstate_calc_busy() into get_target_pstate_use_performance()"
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

When using ACPI, id is NULL and the current code automatically
defaults name to NULL and chip id to 0. We should instead use
the data provided in the ACPI device table.

Fixes: c816d9e7 ("iio: imu: mpu6050: fix possible NULL dereferences")
Signed-off-by: NDaniel Baluta <daniel.baluta@intel.com>
Reviewed-By: NMatt Ranostay <matt.ranostay@intel.com>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

Fix possible null dereferencing of i2c and spi driver data.
Signed-off-by: NMatt Ranostay <matt.ranostay@intel.com>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

The current sequence makes us register for a network device prior to
registering and probing the MDIO bus which could lead to some unwanted
consequences, like a thread of execution calling into ndo_open before
register_netdev() returns, while the MDIO bus is not ready yet.

Rework the sequence to register for the MDIO bus, and therefore attach
to a PHY prior to calling register_netdev(), which implies reworking the
error path a bit.
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Acked-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since commit 3b9d6da6 ("cpu/hotplug: Fix rollback during error-out
in __cpu_disable()") it is ensured that callbacks of CPU_ONLINE and
CPU_DOWN_PREPARE are processed on the hotplugged CPU. Due to this SMP
function calls are no longer required.

Replace smp_call_function_single() with a direct call to
mvneta_percpu_enable() or mvneta_percpu_disable(). The functions do
not require to be called with interrupts disabled, therefore the
smp_call_function_single() calling convention is not preserved.

Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: netdev@vger.kernel.org
Signed-off-by: NAnna-Maria Gleixner <anna-maria@linutronix.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Now mdiobus_scan() returns ERR_PTR(-ENODEV) instead of NULL if the PHY
device ID was read as all ones. As this was not  an error before, this
value  should be filtered out now in this driver.

Fixes: b74766a0 ("phylib: don't return NULL from get_phy_device()")
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: NFlorian Fainelli <f.fainelli@gmail.com>
Acked-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since mdiobus_scan() returns either an error code or NULL on error, the
driver should check  for both,  not only for NULL, otherwise a crash is
imminent...
Reported-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

hw doesn't like a 0 value.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

hw doesn't like a 0 value.
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

The vxlan add/delete port NDOs are called under rcu lock.
The current mlx5e implementation can potentially block in these
calls, which is not allowed.  Move to using the mlx5e workqueue
to handle these NDOs.

Fixes: b3f63c3d ('net/mlx5e: Add netdev support for VXLAN tunneling')
Signed-off-by: NMatthew Finlay <matt@mellanox.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>