openwrt/staging/blogic.git
10 years agoixgbe: Add timeout parameter to ixgbe_host_interface_command
Don Skidmore [Sat, 29 Nov 2014 05:22:32 +0000 (05:22 +0000)]
ixgbe: Add timeout parameter to ixgbe_host_interface_command

Since on X550 we use host interface commands to read,write and erase
some commands require more time to complete. So this adds a timeout
parameter to ixgbe_host_interface_command as wells as a return_data
parameter allowing us to return with any data.

Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
10 years agoixgbe: add support for X550 extended RSS support
Don Skidmore [Tue, 18 Nov 2014 09:35:08 +0000 (09:35 +0000)]
ixgbe: add support for X550 extended RSS support

The new X550 family of MAC's will have a larger RSS hash (16 -> 64).
It will also support individual VF to have their own independent RSS
hash key.  This patch will enable this functionality

Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
10 years agoixgbe: remove CIAA/D register reads from bad VF check
Emil Tantilov [Wed, 19 Nov 2014 03:18:19 +0000 (03:18 +0000)]
ixgbe: remove CIAA/D register reads from bad VF check

Accessing the CIAA/D register can block access to the PCI config space.

This patch removes the read/write operations to the CIAA/D registers
and makes use of standard kernel functions for accessing the PCI config
space.

In addition it moves ixgbevf_check_for_bad_vf() into the watchdog subtask
which reduces the frequency of the checks.

CC: Alex Williamson <alex.williamson@redhat.com>
Reported-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
10 years agoixgbe: Look up MAC address in Open Firmware or IDPROM
Martin K Petersen [Sat, 15 Nov 2014 14:24:51 +0000 (14:24 +0000)]
ixgbe: Look up MAC address in Open Firmware or IDPROM

Attempt to look up the MAC address in Open Firmware on systems that
support it. On SPARC resort to using the IDPROM if no OF address is
found.

Signed-off-by: Martin K Petersen <martin.petersen@oracle.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
10 years agoixgbe: Remove tail write abstraction and add missing barrier
Alexander Duyck [Fri, 14 Nov 2014 00:56:35 +0000 (00:56 +0000)]
ixgbe: Remove tail write abstraction and add missing barrier

This change cleans up the tail writes for the ixgbe descriptor queues.  The
current implementation had me confused as I wasn't sure if it was still
making use of the surprise remove logic or not.

It also adds the mmiowb which is needed on ia64, mips, and a couple other
architectures in order to synchronize the MMIO writes with the Tx queue
_xmit_lock spinlock.

Cc: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
10 years agoixgbe: Clean-up page reuse code
Alexander Duyck [Fri, 14 Nov 2014 00:56:29 +0000 (00:56 +0000)]
ixgbe: Clean-up page reuse code

This patch cleans up the page reuse code getting it into a state where all
the workarounds needed are in place as well as cleaning up a few minor
oversights such as using __free_pages instead of put_page to drop a locally
allocated page.

It also cleans up how we clear the descriptor status bits.  Previously they
were zeroed as a part of clearing the hdr_addr.  However the hdr_addr is a
64 bit field and 64 bit writes can be a bit more expensive on on 32 bit
systems.  Since we are no longer using the header split feature the upper
32 bits of the address no longer need to be cleared.  As a result we can
just clear the status bits and leave the length and VLAN fields as-is which
should provide more information in debugging.

Cc: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
10 years agotun: Fix GSO meta-data handling in tun_get_user
Herbert Xu [Sun, 30 Nov 2014 10:03:31 +0000 (18:03 +0800)]
tun: Fix GSO meta-data handling in tun_get_user

When we write the GSO meta-data in tun_get_user we end up advancing
the IO vector twice, thus exhausting the user buffer before we can
finish writing the packet.

Fixes: f5ff53b4d97c ("{macvtap,tun}_get_user(): switch to iov_iter")
Reported-by: Marcelo Ricardo Leitner <mleitner@redhat.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'rocker-next'
David S. Miller [Wed, 3 Dec 2014 04:01:31 +0000 (20:01 -0800)]
Merge branch 'rocker-next'

Jiri Pirko says:

====================
introduce rocker switch driver with hardware accelerated datapath api - phase 1: bridge fdb offload

This patchset is just the first phase of switch and switch-ish device
support api in kernel. Note that the api will extend.

So what this patchset includes:
- introduce switchdev api skeleton for implementing switch drivers
- introduce rocker switch driver which implements switchdev api fdb and
  bridge set/get link ndos

As to the discussion if there is need to have specific class of device
representing the switch itself, so far we found no need to introduce that.
But we are generally ok with the idea and when the time comes and it will
be needed, it can be easily introduced without any disturbance.

This patchset introduces switch id export through rtnetlink and sysfs,
which is similar to what we have for port id in SR-IOV. I will send iproute2
patchset for showing the switch id for port netdevs once this is applied.
This applies also for the PF_BRIDGE and fdb iproute2 patches.

iproute2 patches are now available here:
https://github.com/jpirko/iproute2-rocker

For detailed description and version history, please see individual patches.

In v4 I reordered the patches leaving rocker patches on the end of the patchset.

In v5 I only fixed whitespace issues of patch #13

We have a TODO for related items we want to work on in near future:
https://etherpad.wikimedia.org/p/netdev-swdev-todo
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: Use logical operators on booleans
Thomas Graf [Fri, 28 Nov 2014 13:34:32 +0000 (14:34 +0100)]
rocker: Use logical operators on booleans

Silences various sparse warnings

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: Add proper validation of Netlink attributes
Thomas Graf [Fri, 28 Nov 2014 13:34:31 +0000 (14:34 +0100)]
rocker: Add proper validation of Netlink attributes

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: add ndo_bridge_setlink/getlink support for learning policy
Scott Feldman [Fri, 28 Nov 2014 13:34:30 +0000 (14:34 +0100)]
rocker: add ndo_bridge_setlink/getlink support for learning policy

Rocker ports will use new "swdev" hwmode for bridge port offload policy.
Current supported policy settings are BR_LEARNING and BR_LEARNING_SYNC.
User can turn on/off device port FDB learning and syncing to bridge.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: implement ndo_fdb_dump
Jiri Pirko [Fri, 28 Nov 2014 13:34:29 +0000 (14:34 +0100)]
rocker: implement ndo_fdb_dump

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: implement L2 bridge offloading
Scott Feldman [Fri, 28 Nov 2014 13:34:28 +0000 (14:34 +0100)]
rocker: implement L2 bridge offloading

Add L2 bridge offloading support to rocker driver.  Here, the Linux bridge
driver is used to collect swdev ports into a tagged (or untagged) VLAN
bridge.  The switchdev will offload from the bridge driver the following L2
bridging functions:

 - Learning of neighbor MAC addresses on VLAN X  Learned mac/vlan is
installed in bridge FDB.  (And removed when device unlearns mac/vlan).
Learning must be turned off on each bridge port to disable the feature in
the bridge driver.

- Flooding of multicast/broadcast and unknown unicast pkts to (STP)
active ports in bridge.  The bridge driver is unaware of the flooding happening
at the device level.  Flooding must be turned off on each bridge port to
disable the feature on the bridge driver.

- STP port state is pushed down to driver/device.  The bridge still processes
STP BDPUs and maintains port STP state (for all VLANs in bridge), but
the driver/device must be notified of port STP state change to program
the device.

Multiple (VLAN) bridges are supported.  The device (implemented per
the OF-DPA spec) must use a portion of the VLAN namespace for
internal VLANs.  Right now, the upper 255 VLANs (0xf00 to 0xffe) are
used as internal VLAN IDs for untagged traffic and are not available
as port VLANs.

The driver uses the following interfaces:

1. To track VLAN add/del on ports in bridge:

.ndo_vlan_rx_add_vid
.ndo_vlan_rx_kill_vid

2. To track port add/del membership in bridge:

NETDEV_CHANGEUPPER netdevice notifier

3. To catch static FDB entries installed on bridge/vlan by user using netlink:

.ndo_fdb_add
.ndo_fdb_del

4. To be notified on port STP state change:

.ndo_switch_port_stp_update

5. To notify bridge driver on learned/forgotten mac/vlans on bridge port:

br_fdb_external_learn_add
br_fdb_external_learn_del

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: implement rocker ofdpa flow table manipulation
Scott Feldman [Fri, 28 Nov 2014 13:34:27 +0000 (14:34 +0100)]
rocker: implement rocker ofdpa flow table manipulation

The rocker driver maintains 4 hash tables: flows, groups, FDB, and VLANs.

Flow and group tables track the entries installed to OF-DPA tables,
per the OF-DPA spec.  See OF-DPA spec for full description of fields
in each flow and group table.  New table entries are pushed to the
device with ADD cmd.  Updated entries are pushed to the device with
MOD cmd.  For flow table entries, a crc32 key is made from fields of
the particular field.  For group table entries, the group_id is used
as the key.

The FDB table tracks fdb entries learned by the device or manually
pushed to the bridge by the user.  A crc32 key is made from the
port/mac/vlan tuple for the fdb entry.

The VLAN table tracks the ifindex-to-internal-vlan mapping for
untagged pkts.  On ingress, an untagged pkt is inserted with an
internal VLAN ID based on the input port's current internal VLAN ID.
The input port's internal VLAN will either be referenced by the port's
ifindex, if not bridged, or the containing bridge's ifindex, if
bridged.  Since the ifindex space isn't within a fixed range, uses a
hash table (with ifindex as key) to track internal VLAN ID for a given
ifindex.  The internal VLAN ID range is fixed and currently uses the
upper 255 VLAN IDs, starting at 0xf00.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agorocker: introduce rocker switch driver
Jiri Pirko [Fri, 28 Nov 2014 13:34:26 +0000 (14:34 +0100)]
rocker: introduce rocker switch driver

This patch introduces the first driver to benefit from the switchdev
infrastructure and to implement newly introduced switch ndos. This is a
driver for emulated switch chip implemented in qemu:
https://github.com/sfeldma/qemu-rocker/

This patch is a result of joint work with Scott Feldman.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Reviewed-by: Thomas Graf <tgraf@suug.ch>
Reviewed-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: add brport flags to dflt bridge_getlink
Scott Feldman [Fri, 28 Nov 2014 13:34:25 +0000 (14:34 +0100)]
bridge: add brport flags to dflt bridge_getlink

To allow brport device to return current brport flags set on port.  Add
returned flags to nested IFLA_PROTINFO netlink msg built in dflt getlink.
With this change, netlink msg returned for bridge_getlink contains the port's
offloaded flag settings (the port's SELF settings).

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: add new hwmode swdev
Scott Feldman [Fri, 28 Nov 2014 13:34:24 +0000 (14:34 +0100)]
bridge: add new hwmode swdev

Current hwmode settings are "vepa" or "veb".  These are for NIC interfaces
with basic bridging function offloaded to HW.  Add new "swdev" for full
switch device offloads.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: add new brport flag LEARNING_SYNC
Scott Feldman [Fri, 28 Nov 2014 13:34:23 +0000 (14:34 +0100)]
bridge: add new brport flag LEARNING_SYNC

This policy flag controls syncing of learned FDB entries to bridge's FDB.  If
on, FDB entries learned on bridge port device will be synced.  If off, device
may still learn new FDB entries but they will not be synced with bridge's FDB.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: move private brport flags to if_bridge.h so port drivers can use flags
Scott Feldman [Fri, 28 Nov 2014 13:34:22 +0000 (14:34 +0100)]
bridge: move private brport flags to if_bridge.h so port drivers can use flags

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: add API to notify bridge driver of learned FBD on offloaded device
Scott Feldman [Fri, 28 Nov 2014 13:34:21 +0000 (14:34 +0100)]
bridge: add API to notify bridge driver of learned FBD on offloaded device

When the swdev device learns a new mac/vlan on a port, it sends some async
notification to the driver and the driver installs an FDB in the device.
To give a holistic system view, the learned mac/vlan should be reflected
in the bridge's FBD table, so the user, using normal iproute2 cmds, can view
what is currently learned by the device.  This API on the bridge driver gives
a way for the swdev driver to install an FBD entry in the bridge FBD table.
(And remove one).

This is equivalent to the device running these cmds:

  bridge fdb [add|del] <mac> dev <dev> vid <vlan id> master

This patch needs some extra eyeballs for review, in paricular around the
locking and contexts.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: call netdev_sw_port_stp_update when bridge port STP status changes
Scott Feldman [Fri, 28 Nov 2014 13:34:20 +0000 (14:34 +0100)]
bridge: call netdev_sw_port_stp_update when bridge port STP status changes

To notify switch driver of change in STP state of bridge port, add new
.ndo op and provide switchdev wrapper func to call ndo op. Use it in bridge
code then.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet-sysfs: expose physical switch id for particular device
Jiri Pirko [Fri, 28 Nov 2014 13:34:19 +0000 (14:34 +0100)]
net-sysfs: expose physical switch id for particular device

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Reviewed-by: Thomas Graf <tgraf@suug.ch>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agortnl: expose physical switch id for particular device
Jiri Pirko [Fri, 28 Nov 2014 13:34:18 +0000 (14:34 +0100)]
rtnl: expose physical switch id for particular device

The netdevice represents a port in a switch, it will expose
IFLA_PHYS_SWITCH_ID value via rtnl. Two netdevices with the same value
belong to one physical switch.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Reviewed-by: Thomas Graf <tgraf@suug.ch>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: introduce generic switch devices support
Jiri Pirko [Fri, 28 Nov 2014 13:34:17 +0000 (14:34 +0100)]
net: introduce generic switch devices support

The goal of this is to provide a possibility to support various switch
chips. Drivers should implement relevant ndos to do so. Now there is
only one ndo defined:
- for getting physical switch id is in place.

Note that user can use random port netdevice to access the switch.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Reviewed-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: rename netdev_phys_port_id to more generic name
Jiri Pirko [Fri, 28 Nov 2014 13:34:16 +0000 (14:34 +0100)]
net: rename netdev_phys_port_id to more generic name

So this can be reused for identification of other "items" as well.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Reviewed-by: Thomas Graf <tgraf@suug.ch>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: make vid as a parameter for ndo_fdb_add/ndo_fdb_del
Jiri Pirko [Fri, 28 Nov 2014 13:34:15 +0000 (14:34 +0100)]
net: make vid as a parameter for ndo_fdb_add/ndo_fdb_del

Do the work of parsing NDA_VLAN directly in rtnetlink code, pass simple
u16 vid to drivers from there.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: convert flags in fbd entry into bitfields
Jiri Pirko [Fri, 28 Nov 2014 13:34:14 +0000 (14:34 +0100)]
bridge: convert flags in fbd entry into bitfields

Suggested-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoneigh: sort Neighbor Cache Entry Flags
Jiri Pirko [Fri, 28 Nov 2014 13:34:13 +0000 (14:34 +0100)]
neigh: sort Neighbor Cache Entry Flags

Suggested-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: rename fdb_*_hw to fdb_*_hw_addr to avoid confusion
Jiri Pirko [Fri, 28 Nov 2014 13:34:12 +0000 (14:34 +0100)]
bridge: rename fdb_*_hw to fdb_*_hw_addr to avoid confusion

The current name might seem that this actually offloads the fdb entry to
hw. So rename it to clearly present that this for hardware address
addition/removal.

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonetpoll: delete defconfig references to obsolete NETPOLL_TRAP
Paul Gortmaker [Thu, 27 Nov 2014 15:28:16 +0000 (10:28 -0500)]
netpoll: delete defconfig references to obsolete NETPOLL_TRAP

In commit 9c62a68d13119a1ca9718381d97b0cb415ff4e9d ("netpoll:
Remove dead packet receive code (CONFIG_NETPOLL_TRAP)") this
Kconfig option was removed.  So remove references to it from
all defconfigs as well.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoipvlan: ipvlan depends on INET and IPV6
Mahesh Bandewar [Thu, 27 Nov 2014 05:13:45 +0000 (21:13 -0800)]
ipvlan: ipvlan depends on INET and IPV6

This driver uses ip_out_local() and ip6_route_output() which are
defined only if CONFIG_INET and CONFIG_IPV6 are enabled respectively.

Reported-by: Jim Davis <jim.epost@gmail.com>
Signed-off-by: Mahesh Bandewar <maheshb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Sun, 30 Nov 2014 04:47:48 +0000 (20:47 -0800)]
Merge git://git./linux/kernel/git/davem/net

10 years agomacvlan: delay the header check for dodgy packets into lower device
Jason Wang [Wed, 26 Nov 2014 09:21:14 +0000 (17:21 +0800)]
macvlan: delay the header check for dodgy packets into lower device

We do header check twice for a dodgy packet. One is done before
macvlan_start_xmit(), another is done before lower device's
ndo_start_xmit(). The first one seems redundant so this patch tries to
delay header check until a packet reaches its lower device (or macvtap)
through always enabling NETIF_F_GSO_ROBUST for macvlan device.

Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 29 Nov 2014 18:49:24 +0000 (10:49 -0800)]
Merge branch 'irq-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull irq fixes from Thomas Gleixner:
 "Three fixlets from the ARM SoC camp:
   - correct irqdomain initialization for atmel-aic
   - correct error handling for device tree parsing in bcm controllers"

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqchip: brcmstb-l2: Fix error handling of irq_of_parse_and_map
  irqchip: bcm7120-l2: Fix error handling of irq_of_parse_and_map
  irqchip: atmel-aic: Fix irqdomain initialization

10 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sat, 29 Nov 2014 18:15:31 +0000 (10:15 -0800)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is a set of ten fixes: 8 for UFS including four static checker
  warnings, a potential null deref in the voltage regulator code, a race
  on module unload, a ref counting fix on the well known LUNs which made
  it impossible to remove the ufs module and fix to correct the
  information in pwr_info.

  In addition to UFS, there's a blacklist for the Intel Multi-Flex array
  which chokes on report supported operation codes and a fix to an oops
  in bnx2fc caused by shared skbs"

[ For us non-SCSI people: "UFS" here is "Universal Flash Storage" not
  the filesystem.  - Linus ]

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  ufs: fix NULL dereference when no regulators are defined
  ufs: ensure clk gating work is finished before module unloading
  scsi: ufs: fix static checker warning in ufshcd_parse_clock_info
  scsi: ufs: fix static checker warning in __ufshcd_setup_clocks
  scsi: ufs: fix static checker warning in ufshcd_populate_vreg
  scsi: ufs: fix static checker errors in ufshcd_system_suspend
  ufs: fix power info after link start-up
  ufs: fix reference counting of W-LUs
  scsi: add Intel Multi-Flex to scsi scan blacklist
  bnx2fc: do not add shared skbs to the fcoe_rx_list

10 years agoMerge tag 'staging-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 29 Nov 2014 00:08:09 +0000 (16:08 -0800)]
Merge tag 'staging-3.18-rc7' of git://git./linux/kernel/git/gregkh/staging

Pull staging/IIO driver fixes from Greg KH:
 "Here are some staging and IIO driver fixes for 3.18-rc7 that resolve a
  number of reported issues, and a new device id for a staging wireless
  driver.

  All of these have been in linux-next"

* tag 'staging-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: r8188eu: Add new device ID for DLink GO-USB-N150
  staging: r8188eu: Fix scheduling while atomic error introduced in commit fadbe0cd
  iio: accel: bmc150: set low default thresholds
  iio: accel: bmc150: Fix iio_event_spec direction
  iio: accel: bmc150: Send x, y and z motion separately
  iio: accel: bmc150: Error handling when mode set fails
  iio: gyro: bmg160: Fix iio_event_spec direction
  iio: gyro: bmg160: Send x, y and z motion separately
  iio: gyro: bmg160: Don't let interrupt mode to be open drain
  iio: gyro: bmg160: Error handling when mode set fails
  iio: adc: men_z188_adc: Add terminating entry for men_z188_ids
  iio: accel: kxcjk-1013: Fix kxcjk10013_set_range
  iio: Fix IIO_EVENT_CODE_EXTRACT_DIR bit mask

10 years agoMerge tag 'tty-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sat, 29 Nov 2014 00:03:24 +0000 (16:03 -0800)]
Merge tag 'tty-3.18-rc7' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial fix from Greg KH:
 "Here is a single revert for the of-serial driver that resolves a
  reported issue.

  This revert has been in linux-next for a while"

* tag 'tty-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  Revert "serial: of-serial: add PM suspend/resume support"

10 years agoMerge tag 'usb-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 28 Nov 2014 23:55:14 +0000 (15:55 -0800)]
Merge tag 'usb-3.18-rc7' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some USB driver fixes and new device ids for 3.18-rc7.

  Full details are in the shortlog, and all of these have been in the
  linux-next tree for a while"

* tag 'usb-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  usb-quirks: Add reset-resume quirk for MS Wireless Laser Mouse 6000
  usb: xhci: rework root port wake bits if controller isn't allowed to wakeup
  USB: xhci: Reset a halted endpoint immediately when we encounter a stall.
  Revert "xhci: clear root port wake on bits if controller isn't wake-up capable"
  USB: xhci: don't start a halted endpoint before its new dequeue is set
  USB: uas: Add no-uas quirk for Hitachi usb-3 enclosures 4971:1012
  USB: ssu100: fix overrun-error reporting
  USB: keyspan: fix overrun-error reporting
  USB: keyspan: fix tty line-status reporting
  usb: serial: ftdi_sio: add PIDs for Matrix Orbital products
  usb: dwc3: ep0: fix for dead code
  USB: serial: cp210x: add IDs for CEL MeshConnect USB Stick

10 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/evalenti/linux...
Linus Torvalds [Fri, 28 Nov 2014 22:00:33 +0000 (14:00 -0800)]
Merge branch 'fixes' of git://git./linux/kernel/git/evalenti/linux-soc-thermal

Pull thermal fixes from Eduardo Valentin:
 "In this -rc still very minor changes:

   - Lee Jones fixes compilation warning in sti thermal driver
   - Marjus Elfring removes unnecessary checks in exynos thermal driver
     (as per coccinelle)
   - Now we always update cpufreq policies, and thus get (hopefully)
     always in sync with cpufreq, thanks to Yadwinder"

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/evalenti/linux-soc-thermal:
  thermal: Exynos: Deletion of unnecessary checks before two function calls
  thermal: sti: Ignore suspend/resume functions when !PM_SLEEP
  thermal: cpu_cooling: Update always cpufreq policy with thermal constraints

10 years agoMerge tag 'sound-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 28 Nov 2014 21:54:53 +0000 (13:54 -0800)]
Merge tag 'sound-3.18-rc7' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "No excitement, here are only minor fixes: an endian fix for the new
  DSD format we added in 3.18, a fix for HP mute LED, and a fix for
  Native Instrument quirk"

* tag 'sound-3.18-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: pcm: Add big-endian DSD sample formats and fix XMOS DSD sample format
  ALSA: hda - One more HP machine needs to change mute led quirk
  ALSA: usb-audio: Use snd_usb_ctl_msg() for Native Instruments quirk

10 years agoMerge tag 'armsoc-for-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Linus Torvalds [Fri, 28 Nov 2014 21:34:32 +0000 (13:34 -0800)]
Merge tag 'armsoc-for-rc7' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Arnd Bergmann:
 "Not much interesting going on fixes-wise for us this week, as it
  should be for an -rc7.  I'm not expecting Olof to work much over
  Thanksgiving weekend, so I decided to take over again and push these
  out to you.

  Just four simple fixes this week:

   - one missing of_node_put() on armv7 based mvebu
   - forcing the USB host into the right mode on Chromebook
     (exynos5-snow)
   - enabling two important drivers for exynos_defconfig
   - fixing a noncritical bug for tegra that would cause a regression
     with common code patches queued for 3.19"

* tag 'armsoc-for-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: tegra: irq: fix buggy usage of irq_data irq field
  ARM: exynos_defconfig: Enable max77802 rtc and clock drivers
  ARM: dts: Explicitly set dr_mode on exynos5250-snow
  ARM: mvebu: add missing of_node_put() call in coherency.c

10 years agoMerge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm
Linus Torvalds [Fri, 28 Nov 2014 21:32:47 +0000 (13:32 -0800)]
Merge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

Pull ARM fixes from Russell King:
 "Another round of relatively small ARM fixes.

  Thomas spotted that the strex backoff delay bit was a disable bit, so
  it needed to be clear for this to work.  Vladimir spotted that using a
  restart block for the cache flush operation would return -EINTR, which
  userspace was not expecting.  Dmitry spotted that the auxiliary
  control register accesses for Xscale were not correct"

* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  ARM: 8226/1: cacheflush: get rid of restarting block
  ARM: 8222/1: mvebu: enable strex backoff delay
  ARM: 8216/1: xscale: correct auxiliary register in suspend/resume

10 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Fri, 28 Nov 2014 02:32:49 +0000 (18:32 -0800)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull mips fixes from Ralf Baechle:
 "The hopefully final round of fixes for 3.18:

   - Fix a number of build errors affecting particular configurations.
   - Handle EVA correctly when flushing a signal trampoline and dcache
     lines.
   - Fix printks printing jibberish.
   - Handle 64 bit memory addresses correctly when adding memory chunk
     on 32 bit kernels.
   - Fix a race condition in the hardware tablewalker code"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  MIPS: tlbex: Fix potential HTW race on TLBL/M/S handlers
  MIPS: Fix address type used for early memory detection.
  MIPS: Kconfig: Don't allow both microMIPS and SmartMIPS to be selected.
  MIPS: kernel: cps-vec: Set ISA level to mips32r2 for the MIPS MT ASE
  MIPS: Netlogic: handle modular AHCI builds
  MIPS: Netlogic: handle modular USB case
  MIPS: Loongson: Make platform serial setup always built-in.
  MIPS: fix EVA & non-SMP non-FPU FP context signal handling
  MIPS: cpu-probe: Set the FTLB probability bit on supported cores
  MIPS: BMIPS: Fix ".previous without corresponding .section" warnings
  MIPS: uaccess.h: Fix strnlen_user comment.
  MIPS: r4kcache: Add EVA case for protected_writeback_dcache_line
  MIPS: Fix info about plat_setup in arch_mem_init comment
  MIPS: rtlx: Remove KERN_DEBUG from pr_debug() arguments in rtlx.c
  MIPS: SEAD3: Fix LED device registration.
  MIPS: Fix a copy & paste error in unistd.h

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mpe/linux
Linus Torvalds [Fri, 28 Nov 2014 02:23:41 +0000 (18:23 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mpe/linux

Pull powerpc fixes from Michael Ellerman:
 "Here are five fixes for you to pull please.

  They're all CC'ed to stable except the "Fix PE state format" one which
  went in this release"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mpe/linux:
  powerpc: 32 bit getcpu VDSO function uses 64 bit instructions
  powerpc/powernv: Replace OPAL_DEASSERT_RESET with EEH_RESET_DEACTIVATE
  powerpc/eeh: Fix PE state format
  powerpc/pseries: Fix endiannes issue in RTAS call from xmon
  powerpc/powernv: Fix the hmi event version check.

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc
Linus Torvalds [Fri, 28 Nov 2014 02:19:25 +0000 (18:19 -0800)]
Merge git://git./linux/kernel/git/davem/sparc

Pull sparc fixlet from David Miller:
 "Aparc fix to add dma_cache_sync(), even if a nop it should be provided
  if dma_{alloc,free}_noncoherent() is provided too"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc: Add NOP dma_cache_sync() implementation.

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 28 Nov 2014 02:05:05 +0000 (18:05 -0800)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:
 "Several small fixes here:

   1) Don't crash in tg3 driver when the number of tx queues has been
      configured to be different from the number of rx queues.  From
      Thadeu Lima de Souza Cascardo.

   2) VLAN filter not disabled properly in promisc mode in ixgbe driver,
      from Vlad Yasevich.

   3) Fix OOPS on dellink op in VTI tunnel driver, from Xin Long.

   4) IPV6 GRE driver WCCP code checks skb->protocol for ETH_P_IP
      instead of ETH_P_IPV6, whoops.  From Yuri Chislov.

   5) Socket matching in ping driver is buggy when packet AF does not
      match socket's AF.  Fix from Jane Zhou.

   6) Fix checksum calculation errors in VXLAN due to where the
      udp_tunnel6_xmit_skb() helper gets it's saddr/daddr from.  From
      Alexander Duyck.

   7) Fix 5G detection problem in rtlwifi driver, from Larry Finger.

   8) Fix NULL deref in tcp_v{4,6}_send_reset, from Eric Dumazet.

   9) Various missing netlink attribute verifications in bridging code,
      from Thomas Graf.

  10) tcp_recvmsg() unconditionally calls ipv4 ip_recv_error even for
      ipv6 sockets, whoops.  Fix from Willem de Bruijn"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (29 commits)
  net-timestamp: make tcp_recvmsg call ipv6_recv_error for AF_INET6 socks
  bridge: Sanitize IFLA_EXT_MASK for AF_BRIDGE:RTM_GETLINK
  bridge: Add missing policy entry for IFLA_BRPORT_FAST_LEAVE
  net: Check for presence of IFLA_AF_SPEC
  net: Validate IFLA_BRIDGE_MODE attribute length
  bridge: Validate IFLA_BRIDGE_FLAGS attribute length
  stmmac: platform: fix default values of the filter bins setting
  net/mlx4_core: Limit count field to 24 bits in qp_alloc_res
  net: dsa: bcm_sf2: reset switch prior to initialization
  net: dsa: bcm_sf2: fix unmapping registers in case of errors
  tg3: fix ring init when there are more TX than RX channels
  tcp: fix possible NULL dereference in tcp_vX_send_reset()
  rtlwifi: Change order in device startup
  rtlwifi: rtl8821ae: Fix 5G detection problem
  Revert "netfilter: conntrack: fix race in __nf_conntrack_confirm against get_next_corpse"
  vxlan: Fix boolean flip in VXLAN_F_UDP_ZERO_CSUM6_[TX|RX]
  ip6_udp_tunnel: Fix checksum calculation
  net-timestamp: Fix a documentation typo
  net/ping: handle protocol mismatching scenario
  af_packet: fix sparse warning
  ...

10 years agoMerge tag 'spi-v3.18-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Linus Torvalds [Fri, 28 Nov 2014 01:55:42 +0000 (17:55 -0800)]
Merge tag 'spi-v3.18-rc6' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "There's a couple of driver fixes here, plus one core fix for the DMA
  mapping which wasn't doing the right thing for vmalloc()ed addresses
  that hadn't been through kmap().  It's fairly rare to use vmalloc()
  with SPI and it's a subset of those users who might fail so it's
  unsurprising that this wasn't noticed sooner"

* tag 'spi-v3.18-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: sirf: fix word width configuration
  spi: Fix mapping from vmalloc-ed buffer to scatter list
  spi: dw: Fix dynamic speed change.

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Fri, 28 Nov 2014 01:51:50 +0000 (17:51 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

Pull input layer fixes from Dmitry Torokhov:
 "The main change is to fix breakage in Elantech driver introduced by
  the recent commit adding trackpoint reporting to protocol v4.  Now we
  are trusting the hardware to advertise the trackpoint properly and do
  not try to decode the data as trackpoint if firmware told us it is not
  present"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: xpad - use proper endpoint type
  Input: elantech - trust firmware about trackpoint presence
  Input: synaptics - adjust min/max on Thinkpad E540

10 years agostaging: r8188eu: Add new device ID for DLink GO-USB-N150
Larry Finger [Thu, 27 Nov 2014 16:10:21 +0000 (10:10 -0600)]
staging: r8188eu: Add new device ID for DLink GO-USB-N150

The DLink GO-USB-N150 with revision B1 uses this driver.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoMIPS: tlbex: Fix potential HTW race on TLBL/M/S handlers
Leonid Yegoshin [Thu, 27 Nov 2014 11:13:08 +0000 (11:13 +0000)]
MIPS: tlbex: Fix potential HTW race on TLBL/M/S handlers

There is a potential race when probing the TLB in TLBL/M/S exception
handlers for a matching entry. Between the time we hit a TLBL/S/M
exception and the time we get to execute the TLBP instruction, the
HTW may have replaced the TLB entry we are interested in hence the TLB
probe may fail. However, in the existing handlers, we never checked the
status of the TLBP (ie check the result in the C0/Index register). We
fix this by adding such a check when the core implements the HTW. If
we couldn't find a matching entry, we return back and try again.

Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Cc: <stable@vger.kernel.org> # v3.17+
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/8599/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
10 years agoARM: 8226/1: cacheflush: get rid of restarting block
Vladimir Murzin [Thu, 27 Nov 2014 10:39:04 +0000 (11:39 +0100)]
ARM: 8226/1: cacheflush: get rid of restarting block

We cannot restart cacheflush safely if a process provides user-defined
signal handler and signal is pending. In this case -EINTR is returned
and it is expected that process re-invokes syscall. However, there are
a few problems with that:
 * looks like nobody bothers checking return value from cacheflush
 * but if it did, we don't provide the restart address for that, so the
   process has to use the same range again
 * ...and again, what might lead to looping forever

So, remove cacheflush restarting code and terminate cache flushing
as early as fatal signal is pending.

Cc: stable@vger.kernel.org # 3.12+
Reported-by: Chanho Min <chanho.min@lge.com>
Signed-off-by: Vladimir Murzin <vladimir.murzin@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoARM: 8222/1: mvebu: enable strex backoff delay
Thomas Petazzoni [Tue, 25 Nov 2014 17:43:15 +0000 (18:43 +0100)]
ARM: 8222/1: mvebu: enable strex backoff delay

Under extremely rare conditions, in an MPCore node consisting of at
least 3 CPUs, two CPUs trying to perform a STREX to data on the same
shared cache line can enter a livelock situation.

This patch enables the HW mechanism that overcomes the bug. This fixes
the incorrect setup of the STREX backoff delay bit due to a wrong
description in the specification.

Note that enabling the STREX backoff delay mechanism is done by
leaving the bit *cleared*, while the bit was currently being set by
the proc-v7.S code.

[Thomas: adapt to latest mainline, slightly reword the commit log, add
stable markers.]

Fixes: de4901933f6d ("arm: mm: Add support for PJ4B cpu and init routines")
Cc: <stable@vger.kernel.org> # v3.8+
Signed-off-by: Nadav Haklai <nadavh@marvell.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: Jason Cooper <jason@lakedaemon.net>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
10 years agoMerge tag 'mvebu-fixes-3.18-2' of git://git.infradead.org/linux-mvebu into fixes
Arnd Bergmann [Thu, 27 Nov 2014 13:29:23 +0000 (14:29 +0100)]
Merge tag 'mvebu-fixes-3.18-2' of git://git.infradead.org/linux-mvebu into fixes

Pull "mvebu fixes for v3.18 (round 2)" frm Jason Cooper:

 - mvebu
    - coherency.c needed an of_node_put()

* tag 'mvebu-fixes-3.18-2' of git://git.infradead.org/linux-mvebu:
  ARM: mvebu: add missing of_node_put() call in coherency.c

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
10 years agoMerge tag 'samsung-defconfig-v3.18' of git://git.kernel.org/pub/scm/linux/kernel...
Arnd Bergmann [Thu, 27 Nov 2014 13:26:52 +0000 (14:26 +0100)]
Merge tag 'samsung-defconfig-v3.18' of git://git./linux/kernel/git/kgene/linux-samsung into fixes

Pull "Samsung defconfig update for v3.18" from Kukjin Kim:

- enable max77802 rtc and clock drivers for exynos_defconfig
  : enable the kernel config options to have the drivers for
    max77802 including rtc and 2-ch 32kHz clock outputs

* tag 'samsung-defconfig-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: exynos_defconfig: Enable max77802 rtc and clock drivers

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
10 years agoMerge tag 'samsung-fixes-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git...
Arnd Bergmann [Thu, 27 Nov 2014 13:24:41 +0000 (14:24 +0100)]
Merge tag 'samsung-fixes-v3.18' of git://git./linux/kernel/git/kgene/linux-samsung into fixes

Pull "Samsung fixes for v3.18" from Kukjin Kim:

- explicitly set dr_mode on exynos5250-snow
  this is required when kernel is built with USB gadget support.

* tag 'samsung-fixes-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: dts: Explicitly set dr_mode on exynos5250-snow

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
10 years agoARM: tegra: irq: fix buggy usage of irq_data irq field
Marc Zyngier [Wed, 26 Nov 2014 17:55:31 +0000 (17:55 +0000)]
ARM: tegra: irq: fix buggy usage of irq_data irq field

The crazy gic_arch_extn thing that Tegra uses contains multiple
references to the irq field in struct irq_data, and uses this
to directly poke hardware register.

But irq is the *virtual* irq number, something that has nothing
to do with the actual HW irq (stored in the hwirq field). And once
we put the stacked domain code in action, the whole thing explodes,
as these two values are *very* different:

root@bacon-fat:~# cat /proc/interrupts
            CPU0       CPU1
 16:      25801       2075       GIC  29  twd
 17:          0          0       GIC  73  timer0
112:          0          0      GPIO  58  c8000600.sdhci cd
123:          0          0      GPIO  69  c8000200.sdhci cd
279:       1126          0       GIC 122  serial
281:          0          0       GIC  70  7000c000.i2c
282:          0          0       GIC 116  7000c400.i2c
283:          0          0       GIC 124  7000c500.i2c
284:        300          0       GIC  85  7000d000.i2c
[...]

Just replacing all instances of irq with hwirq fixes the issue.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
10 years agopowerpc: 32 bit getcpu VDSO function uses 64 bit instructions
Anton Blanchard [Wed, 26 Nov 2014 21:11:28 +0000 (08:11 +1100)]
powerpc: 32 bit getcpu VDSO function uses 64 bit instructions

I used some 64 bit instructions when adding the 32 bit getcpu VDSO
function. Fix it.

Fixes: 18ad51dd342a ("powerpc: Add VDSO version of getcpu")
Cc: stable@vger.kernel.org
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
10 years agopowerpc/powernv: Replace OPAL_DEASSERT_RESET with EEH_RESET_DEACTIVATE
Gavin Shan [Mon, 24 Nov 2014 22:26:59 +0000 (09:26 +1100)]
powerpc/powernv: Replace OPAL_DEASSERT_RESET with EEH_RESET_DEACTIVATE

The flag passed to ioda_eeh_phb_reset() should be EEH_RESET_DEACTIVATE,
which is translated to OPAL_DEASSERT_RESET or something else by the
EEH backend accordingly.

The patch replaces OPAL_DEASSERT_RESET with EEH_RESET_DEACTIVATE for
ioda_eeh_phb_reset().

Cc: stable@vger.kernel.org
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
10 years agopowerpc/eeh: Fix PE state format
Gavin Shan [Mon, 24 Nov 2014 22:26:58 +0000 (09:26 +1100)]
powerpc/eeh: Fix PE state format

Obviously I had wrong format given to the PE state output from
/sys/bus/pci/devices/xxxx/eeh_pe_state with some typoes, which
was introduced by commit 2013add4ce73. The patch fixes it up.

Fixes: 2013add4ce73 ("powerpc/eeh: Show hex prefix for PE state sysfs")
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
10 years agopowerpc/pseries: Fix endiannes issue in RTAS call from xmon
Laurent Dufour [Mon, 24 Nov 2014 14:07:53 +0000 (15:07 +0100)]
powerpc/pseries: Fix endiannes issue in RTAS call from xmon

On pseries system (LPAR) xmon failed to enter when running in LE mode,
system is hunging. Inititating xmon will lead to such an output on the
console:

SysRq : Entering xmon
cpu 0x15: Vector: 0  at [c0000003f39ffb10]
    pc: c00000000007ed7c: sysrq_handle_xmon+0x5c/0x70
    lr: c00000000007ed7c: sysrq_handle_xmon+0x5c/0x70
    sp: c0000003f39ffc70
   msr: 8000000000009033
  current = 0xc0000003fafa7180
  paca    = 0xc000000007d75e80  softe: 0  irq_happened: 0x01
    pid   = 14617, comm = bash
Bad kernel stack pointer fafb4b0 at eca7cc4
cpu 0x15: Vector: 300 (Data Access) at [c000000007f07d40]
    pc: 000000000eca7cc4
    lr: 000000000eca7c44
    sp: fafb4b0
   msr: 8000000000001000
   dar: 10000000
 dsisr: 42000000
  current = 0xc0000003fafa7180
  paca    = 0xc000000007d75e80  softe: 0  irq_happened: 0x01
    pid   = 14617, comm = bash
cpu 0x15: Exception 300 (Data Access) in xmon, returning to main loop
xmon: WARNING: bad recursive fault on cpu 0x15

The root cause is that xmon is calling RTAS to turn off the surveillance
when entering xmon, and RTAS is requiring big endian parameters.

This patch is byte swapping the RTAS arguments when running in LE mode.

Cc: stable@vger.kernel.org
Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
10 years agopowerpc/powernv: Fix the hmi event version check.
Mahesh Salgaonkar [Thu, 20 Nov 2014 04:14:36 +0000 (09:44 +0530)]
powerpc/powernv: Fix the hmi event version check.

The current HMI event structure is an ABI and carries a version field to
accommodate future changes without affecting/rearranging current structure
members that are valid for previous versions.

The current version check "if (hmi_evt->version != OpalHMIEvt_V1)"
doesn't accomodate the fact that the version number may change in
future.

If firmware starts returning an HMI event with version > 1, this check
will fail and no HMI information will be printed on older kernels.

This patch fixes this issue.

Cc: stable@vger.kernel.org # 3.17+
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
[mpe: Reword changelog]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
10 years agostaging: r8188eu: Fix scheduling while atomic error introduced in commit fadbe0cd
Larry Finger [Wed, 12 Nov 2014 16:07:49 +0000 (10:07 -0600)]
staging: r8188eu: Fix scheduling while atomic error introduced in commit fadbe0cd

In commit fadbe0cd5292851608e2e01b91d9295fa287b9fe entitled "staging:
rtl8188eu:Remove rtw_zmalloc(), wrapper for kzalloc()", the author failed
to note that the original code in the wrapper tested whether the caller
could sleep, and set the flags argument to kzalloc() appropriately.
After the patch, GFP_KERNEL is used unconditionally. Unfortunately, several
of the routines may be entered from an interrupt routine and generate
a BUG splat for every such call. Routine rtw_sitesurvey_cmd() is used in the
example below:

BUG: sleeping function called from invalid context at mm/slub.c:1240
in_atomic(): 1, irqs_disabled(): 0, pid: 756, name: wpa_supplicant
INFO: lockdep is turned off.
CPU: 2 PID: 756 Comm: wpa_supplicant Tainted: G        WC O   3.18.0-rc4+ #34
Hardware name: TOSHIBA TECRA A50-A/TECRA A50-A, BIOS Version 4.20   04/17/2014
ffffc90005557000 ffff880216fafaa8 ffffffff816b0bbf 0000000000000000
ffff8800c3b58000 ffff880216fafac8 ffffffff8107af77 0000000000000001
0000000000000010 ffff880216fafb18 ffffffff811b06ce 0000000000000000
Call Trace:
 [<ffffffff816b0bbf>] dump_stack+0x4e/0x71
 [<ffffffff8107af77>] __might_sleep+0xf7/0x120
 [<ffffffff811b06ce>] kmem_cache_alloc_trace+0x4e/0x1f0
 [<ffffffffa0888226>] ? rtw_sitesurvey_cmd+0x56/0x2a0 [r8188eu]
 [<ffffffffa0888226>] rtw_sitesurvey_cmd+0x56/0x2a0 [r8188eu]
 [<ffffffffa088f00d>] rtw_do_join+0x22d/0x370 [r8188eu]
 [<ffffffffa088f6e8>] rtw_set_802_11_ssid+0x218/0x3d0 [r8188eu]
 [<ffffffffa08c3ca5>] rtw_wx_set_essid+0x1e5/0x410 [r8188eu]
 [<ffffffffa08c3ac0>] ? rtw_wx_get_rate+0x50/0x50 [r8188eu]
 [<ffffffff816938f1>] ioctl_standard_iw_point+0x151/0x3f0
 [<ffffffff81693d52>] ioctl_standard_call+0xb2/0xe0
 [<ffffffff81597df7>] ? rtnl_lock+0x17/0x20
 [<ffffffff816945a0>] ? iw_handler_get_private+0x70/0x70
 [<ffffffff81693ca0>] ? call_commit_handler+0x40/0x40
 [<ffffffff81693256>] wireless_process_ioctl+0x176/0x1c0
 [<ffffffff81693e79>] wext_handle_ioctl+0x69/0xc0
 [<ffffffff8159fe79>] dev_ioctl+0x309/0x5e0
 [<ffffffff810be9c7>] ? call_rcu+0x17/0x20
 [<ffffffff8156a472>] sock_ioctl+0x142/0x2e0
 [<ffffffff811e0c70>] do_vfs_ioctl+0x300/0x520
 [<ffffffff81101514>] ? __audit_syscall_entry+0xb4/0x110
 [<ffffffff81101514>] ? __audit_syscall_entry+0xb4/0x110
 [<ffffffff810102bc>] ? do_audit_syscall_entry+0x6c/0x70
 [<ffffffff811e0f11>] SyS_ioctl+0x81/0xa0
 [<ffffffff816ba1d2>] system_call_fastpath+0x12/0x17

Additional routines that generate this BUG are rtw_joinbss_cmd(),
rtw_dynamic_chk_wk_cmd(), rtw_lps_ctrl_wk_cmd(), rtw_rpt_timer_cfg_cmd(),
rtw_ps_cmd(), report_survey_event(), report_join_res(), survey_timer_hdl(),
and rtw_check_bcn_info().

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: navin patidar <navin.patidar@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoMerge tag 'master-2014-11-25' of git://git.kernel.org/pub/scm/linux/kernel/git/linvil...
David S. Miller [Wed, 26 Nov 2014 21:38:35 +0000 (16:38 -0500)]
Merge tag 'master-2014-11-25' of git://git./linux/kernel/git/linville/wireless

John W. Linville says:

====================
pull request: wireless 2014-11-26

Please pull this little batch of fixes intended for the 3.18 stream...

For the iwlwifi one, Emmanuel says:

"Not all the firmware know how to handle the HOT_SPOT_CMD.
Make sure that the firmware will know this command before
sending it. This avoids a firmware crash."

Along with that, Larry sends a pair of rtlwifi fixes to address some
discrepancies from moving drivers out of staging.  Larry says:

"These two patches are needed to fix a regression introduced when
driver rtl8821ae was moved from staging to the regular wireless tree."

Please let me know if there are problems!
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agosparc: Add NOP dma_cache_sync() implementation.
David S. Miller [Wed, 26 Nov 2014 21:36:41 +0000 (13:36 -0800)]
sparc: Add NOP dma_cache_sync() implementation.

This can be a NOP because we forward dma_{alloc,free}_noncoherent to
dma_{alloc,free}_coherent.

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet-timestamp: make tcp_recvmsg call ipv6_recv_error for AF_INET6 socks
Willem de Bruijn [Wed, 26 Nov 2014 19:53:02 +0000 (14:53 -0500)]
net-timestamp: make tcp_recvmsg call ipv6_recv_error for AF_INET6 socks

TCP timestamping introduced MSG_ERRQUEUE handling for TCP sockets.
If the socket is of family AF_INET6, call ipv6_recv_error instead
of ip_recv_error.

This change is more complex than a single branch due to the loadable
ipv6 module. It reuses a pre-existing indirect function call from
ping. The ping code is safe to call, because it is part of the core
ipv6 module and always present when AF_INET6 sockets are active.

Fixes: 4ed2d765 (net-timestamp: TCP timestamping)
Signed-off-by: Willem de Bruijn <willemb@google.com>
----

It may also be worthwhile to add WARN_ON_ONCE(sk->family == AF_INET6)
to ip_recv_error.
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'bridge_nl_validation'
David S. Miller [Wed, 26 Nov 2014 20:29:10 +0000 (15:29 -0500)]
Merge branch 'bridge_nl_validation'

Thomas Graf says:

====================
bridge: Fix missing Netlink message validations

Adds various missing length checks in the bridging code for Netlink
messages and corresponding attributes provided by user space.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: Sanitize IFLA_EXT_MASK for AF_BRIDGE:RTM_GETLINK
Thomas Graf [Wed, 26 Nov 2014 12:42:20 +0000 (13:42 +0100)]
bridge: Sanitize IFLA_EXT_MASK for AF_BRIDGE:RTM_GETLINK

Only search for IFLA_EXT_MASK if the message actually carries a
ifinfomsg header and validate minimal length requirements for
IFLA_EXT_MASK.

Fixes: 6cbdceeb ("bridge: Dump vlan information from a bridge port")
Cc: Vlad Yasevich <vyasevic@redhat.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: Add missing policy entry for IFLA_BRPORT_FAST_LEAVE
Thomas Graf [Wed, 26 Nov 2014 12:42:19 +0000 (13:42 +0100)]
bridge: Add missing policy entry for IFLA_BRPORT_FAST_LEAVE

Fixes: c2d3babf ("bridge: implement multicast fast leave")
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: Check for presence of IFLA_AF_SPEC
Thomas Graf [Wed, 26 Nov 2014 12:42:18 +0000 (13:42 +0100)]
net: Check for presence of IFLA_AF_SPEC

ndo_bridge_setlink() is currently only called on the slave if
IFLA_AF_SPEC is set but this is a very fragile assumption and may
change in the future.

Cc: Ajit Khaparde <ajit.khaparde@emulex.com>
Cc: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: Validate IFLA_BRIDGE_MODE attribute length
Thomas Graf [Wed, 26 Nov 2014 12:42:17 +0000 (13:42 +0100)]
net: Validate IFLA_BRIDGE_MODE attribute length

Payload is currently accessed blindly and may exceed valid message
boundaries.

Fixes: a77dcb8c8 ("be2net: set and query VEB/VEPA mode of the PF interface")
Fixes: 815cccbf1 ("ixgbe: add setlink, getlink support to ixgbe and ixgbevf")
Cc: Ajit Khaparde <ajit.khaparde@emulex.com>
Cc: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobridge: Validate IFLA_BRIDGE_FLAGS attribute length
Thomas Graf [Wed, 26 Nov 2014 12:42:16 +0000 (13:42 +0100)]
bridge: Validate IFLA_BRIDGE_FLAGS attribute length

Payload is currently accessed blindly and may exceed valid message
boundaries.

Fixes: 407af3299 ("bridge: Add netlink interface to configure vlans on bridge ports")
Cc: Vlad Yasevich <vyasevic@redhat.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agosky2: Fix crash inside sky2_rx_clean
Mirko Lindner [Wed, 26 Nov 2014 14:13:38 +0000 (15:13 +0100)]
sky2: Fix crash inside sky2_rx_clean

If sky2->tx_le = pci_alloc_consistent() or sky2->tx_ring = kcalloc() in
sky2_alloc_buffers() fails, sky2->rx_ring = kcalloc() will never be called.
In this error case handling, sky2_rx_clean() is called from within
sky2_free_buffers().

In sky2_rx_clean() we find the following:

...
   memset(sky2->rx_le, 0, RX_LE_BYTES);
...

This results in a memset using a NULL pointer and will crash the system.

Signed-off-by: Mirko Lindner <mlindner@marvell.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoipvlan: fix sparse warnings
Mahesh Bandewar [Wed, 26 Nov 2014 05:24:43 +0000 (21:24 -0800)]
ipvlan: fix sparse warnings

Fix sparse warnings reported by kbuild robot

drivers/net/ipvlan/ipvlan_main.c:172:13: warning: symbol 'ipvlan_start_xmit' was not declared. Should it be static?
drivers/net/ipvlan/ipvlan_main.c:256:33: warning: incorrect type in initializer (different address spaces)
drivers/net/ipvlan/ipvlan_main.c:256:33:    expected void const [noderef] <asn:3>*__vpp_verify
drivers/net/ipvlan/ipvlan_main.c:256:33:    got struct ipvl_pcpu_stats *<noident>
drivers/net/ipvlan/ipvlan_main.c:544:5: warning: symbol 'ipvlan_link_register' was not declared. Should it be static

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Mahesh Bandewar <maheshb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'bcmgenet_eee'
David S. Miller [Wed, 26 Nov 2014 20:08:38 +0000 (15:08 -0500)]
Merge branch 'bcmgenet_eee'

Florian Fainelli says:

====================
net: bcmgenet: EEE support

This patch series add EEE support to the Broadcom GENET driver, first patch
basically adds register definitions for the hardware, the second patch does the
actual implementation and hooks the {get,set}_eee ethtool callbacks, the last
patch adds auto-negotiation restart capability.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: bcmgenet: support restarting auto-negotiation
Florian Fainelli [Wed, 26 Nov 2014 05:16:36 +0000 (21:16 -0800)]
net: bcmgenet: support restarting auto-negotiation

Hook a nway_reset ethtool callback to allow restarting the
auto-negotiation process when asked to. We defer to the PHY library call
to do the heavy lifting.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: bcmgenet: add EEE support
Florian Fainelli [Wed, 26 Nov 2014 05:16:35 +0000 (21:16 -0800)]
net: bcmgenet: add EEE support

Allow enabling and disabling EEE using the designated ethtool getters
and setters. GENET allows controlling EEE at the UniMAC, RBUF and TBUF
levels. We also take care of restoring EEE after a suspend/resume cycle
if it was enabled prior to suspending.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: bcmgenet: add register definitions for EEE
Florian Fainelli [Wed, 26 Nov 2014 05:16:34 +0000 (21:16 -0800)]
net: bcmgenet: add register definitions for EEE

Add register definitions to control EEE in the UniMAC, RBUF and TBUF
register ranges.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Wed, 26 Nov 2014 19:16:44 +0000 (11:16 -0800)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "Last minute KVM/ARM fixes; even the generic change actually affects
  nothing but ARM"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  kvm: fix kvm_is_mmio_pfn() and rename to kvm_is_reserved_pfn()
  arm/arm64: kvm: drop inappropriate use of kvm_is_mmio_pfn()
  arm/arm64: KVM: vgic: Fix error code in kvm_vgic_create()
  arm64: KVM: Handle traps of ICC_SRE_EL1 as RAZ/WI
  arm64: KVM: fix unmapping with 48-bit VAs

10 years agoMerge tag 'iio-fixes-for-3.18c' of git://git.kernel.org/pub/scm/linux/kernel/git...
Greg Kroah-Hartman [Wed, 26 Nov 2014 19:06:36 +0000 (11:06 -0800)]
Merge tag 'iio-fixes-for-3.18c' of git://git./linux/kernel/git/jic23/iio into staging-linus

Jonathan writes:

Third set of IIO fixes for the 3.18 cycle.

Most of these are fairly standard little fixes, a bmc150 and bmg160 patch
is to make an ABI change to indicated a specific axis in an event rather
than the generic option in the original drivers.  As both of these drivers
are new in this cycle it would be ideal to push this minor change through
even though it isn't strictly a fix.  A couple of other 'fixes' change
defaults for some settings on these new drivers to more intuitive calues.
Looks like some useful feedback has been coming in for this driver
since it was applied.

* IIO_EVENT_CODE_EXTRACT_DIR bit mask was wrong and has been for a while
  0xCF clearly doesn't give a contiguous bitmask.
* kxcjk-1013 range setting was failing to mask out the previous value
  in the register and hence was 'enable only'.
* men_z188 device id table wasn't null terminated.
* bmg160 and bmc150 both failed to correctly handling an error in mode
  setting.
* bmg160 and bmc150 both had a bug in setting the event direction in the
  event spec (leads to an attribute name being incorrect)
* bmg160 defaulted to an open drain output for the interrupt - as a default
  this obviously only works with some interrupt chips - hence change the
  default to push-pull (note this is a new driver so we aren't going to
  cause any regressions with this change).
* bmc150 had an unintuitive default for the rate of change (motion detector)
  so change it to 0 (new driver so change of default won't cause any
  regressions).

10 years agoMerge remote-tracking branches 'spi/fix/dw' and 'spi/fix/sirf' into spi-linus
Mark Brown [Wed, 26 Nov 2014 19:05:25 +0000 (19:05 +0000)]
Merge remote-tracking branches 'spi/fix/dw' and 'spi/fix/sirf' into spi-linus

10 years agoMerge remote-tracking branch 'spi/fix/dma' into spi-linus
Mark Brown [Wed, 26 Nov 2014 19:05:25 +0000 (19:05 +0000)]
Merge remote-tracking branch 'spi/fix/dma' into spi-linus

10 years agoMerge branch 'tipc-next'
David S. Miller [Wed, 26 Nov 2014 17:30:23 +0000 (12:30 -0500)]
Merge branch 'tipc-next'

Ying Xue says:

====================
standardize TIPC SKB queue operations

Now the following SKB queues are created and maintained within internal
TIPC stack:

- link transmission queue
- link deferred queue
- link receive queue
- socket outgoing packet chain
- name table outgoing packet chain

In order to manage above queues, TIPC stack declares a sk_buff pointer
for each queue to record its head, and directly modifies "prev" and
"next" SKB pointers of SKB structure when inserting or deleting a SKB
to or from the queue. As these operations are pretty complex, they
easily involve fatal mistakes. If these sk_buff pointers are replaced
with sk_buff_head instances as queue heads and corresponding generic
SKB list APIs are used to manage them, the entire TIPC code would
become quite clean and readable. But before make the change, we need
to clean up below redundant functionalities:

- remove node subscribe infrastructure
- remove protocol message queue
- remove retransmission queue
- clean up process of pushing packets in link layer
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: use generic SKB list APIs to manage TIPC outgoing packet chains
Ying Xue [Wed, 26 Nov 2014 03:41:55 +0000 (11:41 +0800)]
tipc: use generic SKB list APIs to manage TIPC outgoing packet chains

Use standard SKB list APIs associated with struct sk_buff_head to
manage socket outgoing packet chain and name table outgoing packet
chain, having relevant code simpler and more readable.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: use generic SKB list APIs to manage link receive queue
Ying Xue [Wed, 26 Nov 2014 03:41:54 +0000 (11:41 +0800)]
tipc: use generic SKB list APIs to manage link receive queue

Use standard SKB list APIs associated with struct sk_buff_head to
manage link's receive queue to simplify its relevant code cemplexity.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: use generic SKB list APIs to manage deferred queue of link
Ying Xue [Wed, 26 Nov 2014 03:41:53 +0000 (11:41 +0800)]
tipc: use generic SKB list APIs to manage deferred queue of link

Use standard SKB list APIs associated with struct sk_buff_head to
manage link's deferred queue, simplifying relevant code.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: use generic SKB list APIs to manage link transmission queue
Ying Xue [Wed, 26 Nov 2014 03:41:52 +0000 (11:41 +0800)]
tipc: use generic SKB list APIs to manage link transmission queue

Use standard SKB list APIs associated with struct sk_buff_head to
manage link transmission queue, having relevant code more clean.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: use skb_queue_walk_safe marco to simplify link_prepare_wakeup routine
Ying Xue [Wed, 26 Nov 2014 03:41:51 +0000 (11:41 +0800)]
tipc: use skb_queue_walk_safe marco to simplify link_prepare_wakeup routine

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: remove unused between routine
Ying Xue [Wed, 26 Nov 2014 03:41:50 +0000 (11:41 +0800)]
tipc: remove unused between routine

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: eliminate two pseudo message types of BUNDLE_OPEN and BUNDLE_CLOSED
Ying Xue [Wed, 26 Nov 2014 03:41:49 +0000 (11:41 +0800)]
tipc: eliminate two pseudo message types of BUNDLE_OPEN and BUNDLE_CLOSED

The pseudo message types of BUNDLE_CLOSED as well as BUNDLE_OPEN are
used to flag whether or not more messages can be bundled into a data
packet in the outgoing transmission queue. Obviously, no more messages
can be appended after the packet has been sent and is waiting to be
acknowledged and deleted. These message types do in reality represent
a send-side local implementation flag, and are not defined as part of
the protocol. It is therefore safe to move it to to where it belongs,
that is, the control area (TIPC_SKB_CB) of the buffer.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: clean up the process of link pushing packets
Ying Xue [Wed, 26 Nov 2014 03:41:48 +0000 (11:41 +0800)]
tipc: clean up the process of link pushing packets

In original tipc_link_push_packet(), it pushes messages from protocol
message queue, retransmission queue and next_out queue. But as the two
first queues are removed, we can simplify its relevant code through
deleting tipc_link_push_queue().

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: remove retransmission queue
Ying Xue [Wed, 26 Nov 2014 03:41:47 +0000 (11:41 +0800)]
tipc: remove retransmission queue

TIPC retransmission queue is intended to record which messages
should be retransmitted when bearer is not congested. However,
as the retransmission queue becomes useless with the removal of
bearer congestion mechanism, it should be removed.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: remove protocol message queue
Ying Xue [Wed, 26 Nov 2014 03:41:46 +0000 (11:41 +0800)]
tipc: remove protocol message queue

TIPC protocol message queue is intended to save one protocol message
when bearer is congested so that the message stored in the queue can
be immediately transmitted when bearer congestion is released. However,
as now the protocol queue has no mission any more with the removal of
bearer congestion mechanism, it should be removed.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotipc: remove node subscription infrastructure
Ying Xue [Wed, 26 Nov 2014 03:41:45 +0000 (11:41 +0800)]
tipc: remove node subscription infrastructure

The node subscribe infrastructure represents a virtual base class, so
its users, such as struct tipc_port and struct publication, can derive
its implemented functionalities. However, after the removal of struct
tipc_port, struct publication is left as its only single user now. So
defining an abstract infrastructure for one user becomes no longer
reasonable. If corresponding new functions associated with the
infrastructure are moved to name_table.c file, the node subscription
infrastructure can be removed as well.

Signed-off-by: Ying Xue <ying.xue@windriver.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agostmmac: platform: fix default values of the filter bins setting
Huacai Chen [Wed, 26 Nov 2014 02:38:06 +0000 (10:38 +0800)]
stmmac: platform: fix default values of the filter bins setting

The commit 3b57de958e2a brought the support for a different amount of
the filter bins, but didn't update the platform driver that without
CONFIG_OF.

Fixes: 3b57de958e2a (net: stmmac: Support devicetree configs for mcast
and ucast filter entries)

Signed-off-by: Huacai Chen <chenhc@lemote.com>
Acked-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoipv6: Remove unnecessary test
zhuyj [Wed, 26 Nov 2014 02:25:58 +0000 (10:25 +0800)]
ipv6: Remove unnecessary test

The "init_net" test in function addrconf_exit_net is introduced
in commit 44a6bd29 [Create ipv6 devconf-s for namespaces] to avoid freeing
init_net. In commit c900a800 [ipv6: fix bad free of addrconf_init_net],
function addrconf_init_net will allocate memory for every net regardless of
init_net. In this case, it is unnecessary to make "init_net" test.

CC: Hong Zhiguo <honkiko@gmail.com>
CC: Octavian Purdila <opurdila@ixiacom.com>
CC: Pavel Emelyanov <xemul@openvz.org>
CC: Cong Wang <cwang@twopensource.com>
Suggested-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Zhu Yanjun <Yanjun.Zhu@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'remcsum_adjust'
David S. Miller [Wed, 26 Nov 2014 17:25:51 +0000 (12:25 -0500)]
Merge branch 'remcsum_adjust'

Tom Herbert says:

====================
gue: Generalize remote checksum offload

The remote checksum offload is generalized by creating a common
function (remcsum_adjust) that does the work of modifying the
checksum in remote checksum offload. This function can be called
from normal or GRO path. GUE was modified to use this function.

Remote checksum offload is described in
https://tools.ietf.org/html/draft-herbert-remotecsumoffload-01

Tested by running 200 TCP_STREAM connections over GUE, did not see
any problems with remote checksum offload enabled.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogue: Call remcsum_adjust
Tom Herbert [Tue, 25 Nov 2014 19:21:20 +0000 (11:21 -0800)]
gue: Call remcsum_adjust

Change remote checksum offload to call remcsum_adjust. This also
eliminates the optimization to skip an IP header as part of the
adjustment (really does not seem to be much of a win).

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: Add remcsum_adjust as common function for remote checksum offload
Tom Herbert [Tue, 25 Nov 2014 19:21:19 +0000 (11:21 -0800)]
net: Add remcsum_adjust as common function for remote checksum offload

This function does the work to update a checksum field as part of
remote checksum offload.

remcsum_adjust does the following:

1) Subtract out the calculated checksum from the beginning of the
   packet (ptr arg) to the start offset.
2) Adjust the checksum field indicated by offset based on the modified
   checksum value from above step.
3) Return the difference in the old checksum field value and the
   new one. The caller will use this to update skb->csum and NAPI csum.

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agopkt_sched: fq: increase max delay from 125 ms to one second
Eric Dumazet [Tue, 25 Nov 2014 16:57:29 +0000 (08:57 -0800)]
pkt_sched: fq: increase max delay from 125 ms to one second

FQ/pacing has a clamp of delay of 125 ms, to avoid some possible harm.

It turns out this delay is too small to allow pacing low rates :
Some ISP setup very aggressive policers as low as 16kbit.

Now TCP stack has spurious rtx prevention, it seems safe to increase
this fixed parameter, without adding a qdisc attribute.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet/mlx4_core: Limit count field to 24 bits in qp_alloc_res
Jack Morgenstein [Tue, 25 Nov 2014 09:54:31 +0000 (11:54 +0200)]
net/mlx4_core: Limit count field to 24 bits in qp_alloc_res

Some VF drivers use the upper byte of "param1" (the qp count field)
in mlx4_qp_reserve_range() to pass flags which are used to optimize
the range allocation.

Under the current code, if any of these flags are set, the 32-bit
count field yields a count greater than 2^24, which is out of range,
and this VF fails.

As these flags represent a "best-effort" allocation hint anyway, they may
safely be ignored. Therefore, the PF driver may simply mask out the bits.

Fixes: c82e9aa0a8 "mlx4_core: resource tracking for HCA resources used by guests"
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>