platform/kernel/linux-exynos.git
8 years agoi40evf: create a generic config RSS function
Helin Zhang [Tue, 27 Oct 2015 20:15:06 +0000 (16:15 -0400)]
i40evf: create a generic config RSS function

There are two ways to configure RSS, this patch adjusts those two
functions with the same input parameters, and creates a more
generic function for configuring RSS.

Change-ID: Iace73bdeba4831909979bef221011060ab327f71
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40evf: rename VF adapter specific RSS function
Helin Zhang [Mon, 26 Oct 2015 23:44:31 +0000 (19:44 -0400)]
i40evf: rename VF adapter specific RSS function

This patch renames old VF adapter specific RSS function to clarify
its scope.

Change-ID: Ie5253083a44c677ebb7709a8a3a18402ad2dc6a6
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: prefetch skb data on transmit
Jesse Brandeburg [Mon, 26 Oct 2015 23:44:30 +0000 (19:44 -0400)]
i40e/i40evf: prefetch skb data on transmit

Issue a prefetch for data early in the transmit path.
This should not be generally needed for Tx traffic, but
it helps immensely for pktgen workloads and should help
for forwarding workloads as well.

Change-ID: Iefee870c20599e0c4240e1d8637e4f16b625f83a
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: Fix RS bit update in Tx path and disable force WB workaround
Anjali Singhai Jain [Mon, 26 Oct 2015 23:44:29 +0000 (19:44 -0400)]
i40e/i40evf: Fix RS bit update in Tx path and disable force WB workaround

This patch fixes the issue of forcing WB too often causing us to not
benefit from NAPI.

Without this patch we were forcing WB/arming interrupt too often taking
away the benefits of NAPI and causing a performance impact.

With this patch we disable force WB in the clean routine for X710
and XL710 adapters. X722 adapters do not enable interrupt to force
a WB and benefit from WB_ON_ITR and hence force WB is left enabled
for those adapters.
For XL710 and X710 adapters if we have less than 4 packets pending
a software Interrupt triggered from service task will force a WB.

This patch also changes the conditions for setting RS bit as described
in code comments. This optimizes when the HW does a tail bump amd when
it does a WB. It also optimizes when we do a wmb.

Change-ID: Id831e1ae7d3e2ec3f52cd0917b41ce1d22d75d9d
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: rename rss_size to alloc_rss_size in i40e_pf
Helin Zhang [Mon, 26 Oct 2015 23:44:28 +0000 (19:44 -0400)]
i40e: rename rss_size to alloc_rss_size in i40e_pf

This patch renames rss_size to alloc_rss_size in i40e_pf, which is
clearer and avoids confusion. It also adds comments to the other
related structure members to help clarify usage.

Change-ID: Ia90090609d006ab589cb639975bb8a0af795d16f
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: add new fields to store user configuration
Helin Zhang [Mon, 26 Oct 2015 23:44:27 +0000 (19:44 -0400)]
i40e: add new fields to store user configuration

This patch adds new fields to i40e_vsi to store user configured
RSS config data and code to use it.

Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Change-ID: I73886469dca9e9f6b16d842182a87f3f4009f95d
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoisdn: remove spellcaster driver
Arnd Bergmann [Mon, 30 Nov 2015 10:34:09 +0000 (11:34 +0100)]
isdn: remove spellcaster driver

The 'sc' ISDN driver relies on using readl() to access ISA I/O memory.
This has been deprecated and produced warnings since linux-2.3.23,
disabled by default since 2.4.10 and finally removed in 2.6.5.

I found this because the compiling the driver for ARM produces
a warning:

In file included from ../drivers/isdn/sc/includes.h:8:0,
                 from ../drivers/isdn/sc/init.c:13:
../arch/arm/include/asm/io.h:115:21: note: expected 'const volatile void *' but argument is of type 'long unsigned int'

It is pretty clear that this driver has not been used for a long time
and there is no point fixing it now, so let's remove it.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agostmmac: support Reg_9 to get HW level information
Giuseppe CAVALLARO [Mon, 30 Nov 2015 10:33:10 +0000 (11:33 +0100)]
stmmac: support Reg_9 to get HW level information

For GMAC newer than 3.40a there is a new register (Reg_9) that provides the
status of all modules of the transmit and receive paths and FIFO status.
These can be exposed via ethtool.

Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'qed-ethtool-ops'
David S. Miller [Tue, 1 Dec 2015 21:02:41 +0000 (16:02 -0500)]
Merge branch 'qed-ethtool-ops'

Yuval Mintz says:

====================
qede/qed: Implement various ethtool operations

This series adds several new ethtool operations to qede:
  - {get, set}_channels
  - {get, set}_ringparam
  - set_phys_id
  - nway_reset
  - {get, set}_pauseparam
As well as extending the qed APIs to support these commands.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqede: Add support for {get, set}_pauseparam
Sudarsana Kalluru [Mon, 30 Nov 2015 10:25:06 +0000 (12:25 +0200)]
qede: Add support for {get, set}_pauseparam

Signed-off-by: Sudarsana Kalluru <Sudarsana.Kalluru@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqede: Add support for nway_reset
Sudarsana Kalluru [Mon, 30 Nov 2015 10:25:05 +0000 (12:25 +0200)]
qede: Add support for nway_reset

Signed-off-by: Sudarsana Kalluru <Sudarsana.Kalluru@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqede: Add support for set_phys_id
Sudarsana Kalluru [Mon, 30 Nov 2015 10:25:04 +0000 (12:25 +0200)]
qede: Add support for set_phys_id

Signed-off-by: Sudarsana Kalluru <Sudarsana.Kalluru@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed: Add support for changing LED state
Sudarsana Kalluru [Mon, 30 Nov 2015 10:25:03 +0000 (12:25 +0200)]
qed: Add support for changing LED state

Physical LEDs are being controlled by the management FW.
This adds the qed functionality required to request management FW to
change the LED configuration, as well as the necessary APIs for this
functionality to later be used by the protocol drivers.

Signed-off-by: Sudarsana Kalluru <Sudarsana.Kalluru@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqede: Add support for {get, set}_ringparam
Sudarsana Kalluru [Mon, 30 Nov 2015 10:25:02 +0000 (12:25 +0200)]
qede: Add support for {get, set}_ringparam

Signed-off-by: Sudarsana Kalluru <Sudarsana.Kalluru@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqede: Add support for {get, set}_channels
Sudarsana Kalluru [Mon, 30 Nov 2015 10:25:01 +0000 (12:25 +0200)]
qede: Add support for {get, set}_channels

Signed-off-by: Sudarsana Kalluru <Sudarsana.Kalluru@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'sfc-8000'
David S. Miller [Tue, 1 Dec 2015 20:46:40 +0000 (15:46 -0500)]
Merge branch 'sfc-8000'

Bert Kenward says:

====================
Basic support for Solarflare 8000 series NICs

The upcoming Solarflare 8000 series 10G/40G network card supports a
similar interface to the current 7000 series cards. This patch series
provides basic support for these cards, making no use of any new
functionality.

v2: fix indenting in ef10.c in patch 1/2.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosfc: Add PCI ID for Solarflare 8000 series 10/40G NIC
Bert Kenward [Mon, 30 Nov 2015 09:05:47 +0000 (09:05 +0000)]
sfc: Add PCI ID for Solarflare 8000 series 10/40G NIC

Also add support for 7000 series 40G NIC VF.

Signed-off-by: Bert Kenward <bkenward@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosfc: make TSO version a per-queue parameter
Bert Kenward [Mon, 30 Nov 2015 09:05:35 +0000 (09:05 +0000)]
sfc: make TSO version a per-queue parameter

The Solarflare 8000 series NIC will use a new TSO scheme. The current
driver refuses to load if the current TSO scheme is not found. Remove
that check and instead make the TSO version a per-queue parameter.

Signed-off-by: Bert Kenward <bkenward@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: add support for netdev notifier error injection
Nikolay Aleksandrov [Sat, 28 Nov 2015 12:45:28 +0000 (13:45 +0100)]
net: add support for netdev notifier error injection

This module allows to insert errors in some of netdevice's notifier
events. All network drivers use these notifiers to signal various events
and to check if they are allowed, e.g. PRECHANGEMTU and CHANGEMTU
afterwards. Until recently I had to run failure tests by injecting
a custom module, but now this infrastructure makes it trivial to test
these failure paths. Some of the recent bugs I fixed were found using
this module.
Here's an example:
 $ cd /sys/kernel/debug/notifier-error-inject/netdev
 $ echo -22 > actions/NETDEV_CHANGEMTU/error
 $ ip link set eth0 mtu 1024
 RTNETLINK answers: Invalid argument

CC: Akinobu Mita <akinobu.mita@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev <netdev@vger.kernel.org>
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agohv_netvsc: rework link status change handling
Vitaly Kuznetsov [Fri, 27 Nov 2015 10:39:55 +0000 (11:39 +0100)]
hv_netvsc: rework link status change handling

There are several issues in hv_netvsc driver with regards to link status
change handling:
- RNDIS_STATUS_NETWORK_CHANGE results in calling userspace helper doing
  '/etc/init.d/network restart' and this is inappropriate and broken for
  many reasons.
- link_watch infrastructure only sends one notification per second and
  in case of e.g. paired disconnect/connect events we get only one
  notification with last status. This makes it impossible to handle such
  situations in userspace.

Redo link status changes handling in the following way:
- Create a list of reconfig events in network device context.
- On a reconfig event add it to the list of events and schedule
  netvsc_link_change().
- In netvsc_link_change() ensure 2-second delay between link status
  changes.
- Handle RNDIS_STATUS_NETWORK_CHANGE as a paired disconnect/connect event.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agounix: use wq_has_sleeper in unix_dgram_recvmsg
Rainer Weikusat [Thu, 26 Nov 2015 19:23:15 +0000 (19:23 +0000)]
unix: use wq_has_sleeper in unix_dgram_recvmsg

The current unix_dgram_recvmsg does a wake up for every received
datagram. This seems wasteful as only SOCK_DGRAM client sockets in an
n:1 association with a server socket will ever wait because of the
associated condition. The patch below changes the function such that the
wake up only happens if wq_has_sleeper indicates that someone actually
wants to be notified. Testing with SOCK_SEQPACKET and SOCK_DGRAM socket
seems to confirm that this is an improvment.

Signed-Off-By: Rainer Weikusat <rweikusat@mobileactivedefense.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'ipmr-nl'
David S. Miller [Mon, 30 Nov 2015 20:26:23 +0000 (15:26 -0500)]
Merge branch 'ipmr-nl'

Nikolay Aleksandrov says:

====================
net: ipmr: more cleanups and mfc netlink support

This set continues with the minor cleanups in the first 6 patches and
patch 7 adds the first new feature - MFC manipulation via netlink. It
registers NEWROUTE/DELROUTE for that purpose and uses the same semantics
as the already present netlink dump. The only new attribute that is used
is RTA_PREFSRC to denote an MFC_PROXY entry. Currently the table must
exist before adding an entry, and new tables can be created only via
setsockopt, but that will be changed in the future.
This set was tested with modified iproute2 which supports NEWROUTE/DELROUTE
for RTNL_FAMILY_IPMR.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: add mfc newroute/delroute netlink support
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:50 +0000 (15:23 +0100)]
net: ipmr: add mfc newroute/delroute netlink support

This patch adds support to add and remove MFC entries. It uses the
same attributes like the already present dump support in order to be
consistent. There's one new entry - RTA_PREFSRC, it's used to denote an
MFC_PROXY entry (see MRT_ADD_MFC vs MRT_ADD_MFC_PROXY).
The already existing infrastructure is used to create and delete the
entries, the netlink message gets converted internally to a struct mfcctl
which is used with ipmr_mfc_add/delete.
The other used attributes are:
RTA_IIF - used for mfcc_parent (when adding it's required to be valid)
RTA_SRC - used for mfcc_origin
RTA_DST - used for mfcc_mcastgrp
RTA_TABLE - the MRT table id
RTA_MULTIPATH - the "oifs" ttl array

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: fix setsockopt error return
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:49 +0000 (15:23 +0100)]
net: ipmr: fix setsockopt error return

We can have both errors and we'll return the second one, fix it to
return an error at a time as it's normal. I've overlooked this in my
previous set.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: move pimsm_enabled to pim.h and rename
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:48 +0000 (15:23 +0100)]
net: ipmr: move pimsm_enabled to pim.h and rename

Move the inline pimsm_enabled() to pim.h and rename it to
ipmr_pimsm_enabled to show it's for the ipv4 ipmr code since pim.h is
used by IPv6 too.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: move struct mr_table and VIF_EXISTS to mroute.h
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:47 +0000 (15:23 +0100)]
net: ipmr: move struct mr_table and VIF_EXISTS to mroute.h

Move the definitions of VIF_EXISTS() and struct mr_table to mroute.h

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: adjust mroute.h style and drop extern
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:46 +0000 (15:23 +0100)]
net: ipmr: adjust mroute.h style and drop extern

Remove extra spaces and tabs, adjust function definitions, remove an
unnecessary ifdef (already used below, just move code) and drop extern
from the functions.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: remove unused MFC_NOTIFY flag and make the flags enum
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:45 +0000 (15:23 +0100)]
net: ipmr: remove unused MFC_NOTIFY flag and make the flags enum

MFC_NOTIFY was introduced in kernel 2.1.68 but afaik it hasn't been used
and I couldn't find any users currently so just remove it. Only
MFC_STATIC is left, so move it into an enum, add a description and use
BIT().

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: remove unnecessary mroute.h includes
Nikolay Aleksandrov [Thu, 26 Nov 2015 14:23:44 +0000 (15:23 +0100)]
net: remove unnecessary mroute.h includes

It looks like many files are including mroute.h unnecessarily, so remove
the include. Most importantly remove it from ipv6.

CC: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
CC: Steffen Klassert <steffen.klassert@secunet.com>
CC: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'mlxsw-hwmon'
David S. Miller [Mon, 30 Nov 2015 20:05:41 +0000 (15:05 -0500)]
Merge branch 'mlxsw-hwmon'

Jiri Pirko says:

====================
mlxsw: driver update

This patchset carries support for port identification, monitoring
of ASIC temperature and board fans.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: core: Implement fan control using hwmon
Jiri Pirko [Fri, 27 Nov 2015 12:45:59 +0000 (13:45 +0100)]
mlxsw: core: Implement fan control using hwmon

ASIC provides access to fans. Implement their exposure to userspace
using hwmon.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: reg: Add definition of fan management registers
Jiri Pirko [Fri, 27 Nov 2015 12:45:58 +0000 (13:45 +0100)]
mlxsw: reg: Add definition of fan management registers

Add definition of MFCR, MFSC and MFSM which provide possibility to
control and monitor fans.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: core: Implement temperature hwmon interface
Jiri Pirko [Fri, 27 Nov 2015 12:45:57 +0000 (13:45 +0100)]
mlxsw: core: Implement temperature hwmon interface

ASIC provides access to temperature sensors. Implement their exposure to
userspace using hwmon.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: reg: Add definition of temperature management registers
Jiri Pirko [Fri, 27 Nov 2015 12:45:56 +0000 (13:45 +0100)]
mlxsw: reg: Add definition of temperature management registers

Add definition of MTCAP and MTMP registers which provide access to
temperature sensors.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: spectrum: Add support for port identification
Ido Schimmel [Fri, 27 Nov 2015 12:45:55 +0000 (13:45 +0100)]
mlxsw: spectrum: Add support for port identification

Allow a user to flash the port's LED in order to identify it. This is
achieved by setting the Management LED Control Register (MLCR).

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: reg: Add Management LED Control register definition
Ido Schimmel [Fri, 27 Nov 2015 12:45:54 +0000 (13:45 +0100)]
mlxsw: reg: Add Management LED Control register definition

Add the MLCR register, which controls physical port identification LEDs.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: Generalise wq_has_sleeper helper
Herbert Xu [Thu, 26 Nov 2015 05:55:39 +0000 (13:55 +0800)]
net: Generalise wq_has_sleeper helper

The memory barrier in the helper wq_has_sleeper is needed by just
about every user of waitqueue_active.  This patch generalises it
by making it take a wait_queue_head_t directly.  The existing
helper is renamed to skwq_has_sleeper.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge tag 'wireless-drivers-next-for-davem-2015-11-25' of git://git.kernel.org/pub...
David S. Miller [Mon, 30 Nov 2015 03:32:32 +0000 (22:32 -0500)]
Merge tag 'wireless-drivers-next-for-davem-2015-11-25' of git://git./linux/kernel/git/kvalo/wireless-drivers-next

Kalle Valo says:

====================
Add vendor directories (for example intel, marvell and so on) to
drivers/net/wireless and change the menuconfig to follow the same
categorisation. This unifies the directory layout as we already had few drivers
already using vendor directories (especially ath and ti).
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Wed, 25 Nov 2015 18:58:06 +0000 (13:58 -0500)]
Merge branch 'master' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2015-11-25

This series contains updates to fm10k, i40e and i40evf.

Alex Duyck fixes up fm10k to use napi_schedule_irqoff() instead of
napi_schedule() since the function it is called from runs from hard interrupt
context or with interrupts already disabled in netpoll.

Shannon cleans up i40e and i40evf unused cd_tunneling parameter and any code
comments that refer to it.  Then clean up a few instances of BUG_ON, based
on a Linux diatribe, especially when WARN_ON can be used.

Helin fixed a issue where using ethtool RXNFC command could let receive
flow hash could be set on disabled queues, so resolve by returning the
number of enabled queues before setting RXNFC.

Anjali fixes a MSS issue where the hardware/NVM sets a limit of no less than
256 bytes for MSS, yet the stack can send as low as 76 byte MSS.  Fixed the
issue by lowering the hardware limit to 64 bytes to avoid MDDs from firing
and causing a reset when the MSS is lower than 256.  Added a statistic to
track how many times we forced to do a write back on transmit if the
number of descriptors pending are less than a cache line.

Catherine fixes link status changes, where polling would only change link
status changes in one direction depending on what the last old data saved
off was.  This was due to the watchdog only calling link_event and not
handle_link_event.

Mitch cleans up and enhances error messages related to VF MAC/VLAN filters.

v2: Dropped patch "i40e: Properly cast type for arithmetic" patch
    based on feedback from the community
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoi40e: Bump version to 1.4.2
Catherine Sullivan [Wed, 21 Oct 2015 23:56:24 +0000 (19:56 -0400)]
i40e: Bump version to 1.4.2

Bump.

Change-ID: I2d1ce93b2ce74e4eef2394c932aef52cba99713f
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: create a generic configure rss function
Helin Zhang [Wed, 21 Oct 2015 23:56:23 +0000 (19:56 -0400)]
i40e: create a generic configure rss function

This patch renames the old pf-specific function in order to clarify
its scope. This patch also creates a more generic configure RSS
function with the old name.

This patch also creates a new more generic function to get RSS
configuration, using the appropriate method.

Change-ID: Ieddca2707b708ef19f1ebccdfd03a0a0cd63d3af
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: rework the functions to configure RSS with similar parameters
Helin Zhang [Wed, 21 Oct 2015 23:56:22 +0000 (19:56 -0400)]
i40e: rework the functions to configure RSS with similar parameters

Adjust the RSS configure functions so that there is a generic way to
hook to ethtool hooks.

Change-ID: If446e34fcfaf1bc3320d9d319829a095b5976e67
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: return the number of enabled queues for ETHTOOL_GRXRINGS
Helin Zhang [Wed, 21 Oct 2015 23:47:13 +0000 (19:47 -0400)]
i40e: return the number of enabled queues for ETHTOOL_GRXRINGS

This patch fixes a problem where using ethtool rxnfc command could
let RX flow hash be set on disabled queues. This patch fixes the
problem by returning the number of enabled queues before setting
rxnfc.

Change-ID: Idbac86b0b47ddacc8deee7cd257e41de01cbe5c0
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40evf: handle many MAC filters correctly
Mitch Williams [Wed, 21 Oct 2015 23:47:12 +0000 (19:47 -0400)]
i40evf: handle many MAC filters correctly

When a lot (many hundreds) of MAC or VLAN filters are added at one time,
we can overflow the Admin Queue buffer size with all the requests.
Unfortunately, the driver would then calculate the message size
incorrectly, causing it to be rejected by the PF. Furthermore, there was
no mechanism to trigger another request to allow for configuring the
rest of the filters that didn't fit into the first request.

To fix this, recalculate the correct buffer size when we detect the
overflow condition instead of just assuming the max buffer size. Also,
don't clear the request bit in adapter->aq_required when we have an
overflow, so that the rest of the filters can be processed later.

Change-ID: Idd7cbbc5af31315e0dcb1b10e6a02ad9817ce65c
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: clean up error messages
Mitch Williams [Wed, 21 Oct 2015 23:47:11 +0000 (19:47 -0400)]
i40e/i40evf: clean up error messages

Clean up and enhance error messages related to VF MAC/VLAN filters.
Indicate which VF is having issues, and if possible indicate the MAC
address or VLAN involved.

Also, when an error is returned from the PF driver, print useful
information about what went wrong, for the most likely cases.

Change-ID: Ib3d15eef9e3369a78fd142948671e5fa26d921b8
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: Add comment to #endif
Helin Zhang [Wed, 21 Oct 2015 23:47:10 +0000 (19:47 -0400)]
i40e/i40evf: Add comment to #endif

Add a comment to the #endif to more easily match it with its #if.

Change-ID: I47eb0a60a17dc6d2f01a930e45006d2dc82e044f
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Move the saving of old link info from handle_link_event to link_event
Catherine Sullivan [Wed, 21 Oct 2015 23:47:09 +0000 (19:47 -0400)]
i40e: Move the saving of old link info from handle_link_event to link_event

The watchdog only calls link_event not handle_link_event which means
that we need to save the old information in link_event.

Previously when polling we were comparing current data to the old data
saved the last time we actually received a link event. This means that
the polling would only fix link status changes in one direction
depending on what the last old data saved off was.

Change-ID: Ie590f30fdbcb133d0ddad4e07e3eb1aad58255b3
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: Add a stat to track how many times we have to do a force WB
Anjali Singhai Jain [Wed, 21 Oct 2015 23:47:08 +0000 (19:47 -0400)]
i40e/i40evf: Add a stat to track how many times we have to do a force WB

When in NAPI with interrupts disabled, the HW needs to be forced to do a
write back on TX if the number of descriptors pending are less than a
cache line.

This stat helps keep track of how many times we get into this situation.

Change-ID: I76c1bcc7ebccd6bffcc5aa33bfe05f2fa1c9a984
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Workaround fix for mss < 256 issue
Anjali Singhai Jain [Wed, 21 Oct 2015 23:47:07 +0000 (19:47 -0400)]
i40e: Workaround fix for mss < 256 issue

HW/NVM sets a limit of no less than 256 bytes for MSS. Stack can send as
low as 76 bytes MSS. This patch lowers the HW limit to 64 bytes to avoid
MDDs from firing and causing a reset when the MSS is lower than 256.

Change-ID: I36b500a6bb227d283c3e321a7718e0672b11fab0
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: remove BUG_ON from FCoE setup
Shannon Nelson [Wed, 21 Oct 2015 23:47:05 +0000 (19:47 -0400)]
i40e: remove BUG_ON from FCoE setup

There's no need to kill the kernel thread here. If this condition was
true, the probe() would have died long before we got here. In any case,
we'll get the same result when this code tries to use the VSI pointer
being checked.

Prompted by a recent Linus diatribe.

Change-ID: I62f531cac34d4fc28ff9657d5b2d9523ae5e33a4
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: remove BUG_ON from feature string building
Shannon Nelson [Wed, 21 Oct 2015 23:47:04 +0000 (19:47 -0400)]
i40e: remove BUG_ON from feature string building

There's really no reason to kill the kernel thread just because of a
little info string. This reworks the code to use snprintf's limiting to
assure that the string is never too long, and WARN_ON to still put out
a warning that we might want to look at the feature list length.

Prompted by a recent Linus diatribe.

Change-ID: If52ba5ca1c2344d8bf454a31bbb805eb5d2c5802
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Change BUG_ON to WARN_ON in service event complete
Shannon Nelson [Wed, 21 Oct 2015 23:47:03 +0000 (19:47 -0400)]
i40e: Change BUG_ON to WARN_ON in service event complete

There's no need to kill the thread and eventually the kernel in this
case.  In fact, the remainder of the code won't hurt anything anyway,
so just complain that we're here and move along.

Prompted by a recent Linus diatribe.

Change-ID: Iec020d8bcfedffc1cd2553cc6905fd915bb3e670
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: remove unused tunnel parameter
Shannon Nelson [Wed, 21 Oct 2015 23:47:02 +0000 (19:47 -0400)]
i40e/i40evf: remove unused tunnel parameter

Code was moved into a separate function some time ago.

Change-ID: Icabbe71ce05cf5d716d3e1152cdd9cd41d11bcb5
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agofm10k: use napi_schedule_irqoff()
Alexander Duyck [Tue, 29 Sep 2015 22:19:56 +0000 (15:19 -0700)]
fm10k: use napi_schedule_irqoff()

The fm10k_msix_clean_rings function runs from hard interrupt context or
with interrupts already disabled in netpoll.

It can use napi_schedule_irqoff() instead of napi_schedule()

Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agonet: phy: bcm7xxx: Add entry for Broadcom BCM7435
Florian Fainelli [Tue, 24 Nov 2015 23:30:21 +0000 (15:30 -0800)]
net: phy: bcm7xxx: Add entry for Broadcom BCM7435

Add a PHY entry for the Broadcom BCM7435 chips, this is a 40nm
generation Ethernet PHY which is analogous to its 7425 and 7429 counter
parts.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'sh_eth-remove-obsolete-platform_ids'
David S. Miller [Tue, 24 Nov 2015 21:48:17 +0000 (16:48 -0500)]
Merge branch 'sh_eth-remove-obsolete-platform_ids'

Geert Uytterhoeven says:

====================
sh_eth: Remove obsolete platform_device_id entries

Since commit 3d7608e4c169af03 ("ARM: shmobile: bockw: remove legacy
board file and config"), which is in v4.4-rc1, shmobile SoCs are only
supported in generic DT-only ARM multi-platform builds.  The sh_eth
driver doesn't need to match platform devices by name anymore, hence
this series removes the corresponding platform_device_id entries.

Changes since v2:
  - More Acks,
  - Platform dependency has entered mainline,

Changes since v1:
  - Protect some data and functions by #ifdef CONFIG_OF to silence
    unused compiler warnings on SH,
  - New patches 3 and 4.

Thanks for applying!
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: Remove obsolete r8a777x-ether platform_device_id entry
Geert Uytterhoeven [Tue, 24 Nov 2015 14:41:00 +0000 (15:41 +0100)]
sh_eth: Remove obsolete r8a777x-ether platform_device_id entry

Since commit 3d7608e4c169af03 ("ARM: shmobile: bockw: remove legacy
board file and config"), R-Car Gen1 SoCs are only supported in generic
DT-only ARM multi-platform builds.  The driver doesn't need to match
platform devices by name anymore, hence remove the corresponding
platform_device_id entry.

Protect sh_eth_set_rate_r8a777x() and r8a777x_data by #ifdef CONFIG_OF,
as they're now referenced on DT platforms only.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: Remove obsolete r8a7740-gether platform_device_id entry
Geert Uytterhoeven [Tue, 24 Nov 2015 14:40:59 +0000 (15:40 +0100)]
sh_eth: Remove obsolete r8a7740-gether platform_device_id entry

Since commit 1fa59bda21c7fa36 ("ARM: shmobile: Remove legacy board code
for Armadillo-800 EVA"), r8a7740 is only supported in generic DT-only
ARM multi-platform builds.  The driver doesn't need to match platform
devices by name anymore, hence remove the corresponding
platform_device_id entry.

Protect r8a7740_data by #ifdef CONFIG_OF as it's now referenced on DT
platforms only. Move it to a more logical position, in front of the
r8a777x support, so we can have a single #ifdef covering all r7s* and
r8a* support soon. This requires moving a few helper functions, too.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: Remove obsolete r8a779x-ether platform_device_id entries
Geert Uytterhoeven [Tue, 24 Nov 2015 14:40:58 +0000 (15:40 +0100)]
sh_eth: Remove obsolete r8a779x-ether platform_device_id entries

Since commit a483dcbfa21f919c ("ARM: shmobile: lager: Remove legacy
board support"), R-Car Gen2 SoCs are only supported in generic DT-only
ARM multi-platform builds.  The driver doesn't need to match platform
devices by name anymore, hence remove the corresponding
platform_device_id entry.

Protect r8a779x_data by #ifdef CONFIG_OF as it's now referenced on DT
platforms only.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: Remove obsolete r7s72100-ether platform_device_id entry
Geert Uytterhoeven [Tue, 24 Nov 2015 14:40:57 +0000 (15:40 +0100)]
sh_eth: Remove obsolete r7s72100-ether platform_device_id entry

Since commit 05104c266ae9a167 ("ARM: shmobile: r7s72100: genmai: Remove
legacy board file"), r7s72100 is only supported in generic DT-only ARM
multi-platform builds.  The driver doesn't need to match platform
devices by name anymore, hence remove the corresponding
platform_device_id entry.

Protect r7s72100_data by #ifdef CONFIG_OF as it's now referenced on DT
platforms only. Move it to a more logical position, in front of the
r8a777x support, so we can have a single #ifdef covering all r7s* and
r8a* support soon. This requires moving a helper function, too.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agovrf: remove slave queue and private slave struct
Nikolay Aleksandrov [Tue, 24 Nov 2015 13:29:16 +0000 (14:29 +0100)]
vrf: remove slave queue and private slave struct

The private slave queue and slave struct haven't been used for anything
and aren't needed, this allows to reduce memory usage and simplify
enslave/release. We can use netdev_for_each_lower_dev() to free the vrf
ports when deleting a vrf device. Also if in the future a private struct
is needed for each slave, it can be implemented via lower devices'
private member (similar to how bonding does it).

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'for-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetoot...
David S. Miller [Tue, 24 Nov 2015 21:22:40 +0000 (16:22 -0500)]
Merge branch 'for-upstream' of git://git./linux/kernel/git/bluetooth/bluetooth-next

Johan Hedberg says:

====================
pull request: bluetooth-next 2015-11-23

Here's the first bluetooth-next pull request for the 4.5 kernel.

 - Add new Get Advertising Size Information management command
 - Add support for new system note message type on monitor channel
 - Refactor LE scan changes behind separate workqueue to avoid races
 - Fix issue with privacy feature when powering on adapter
 - Various minor fixes & cleanups here and there

Please let me know if there are any issues pulling. Thanks.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: fec: no need to test for the return type of of_property_read_u32
Saurabh Sengar [Mon, 23 Nov 2015 13:51:48 +0000 (19:21 +0530)]
net: fec: no need to test for the return type of of_property_read_u32

in case of error no need to set num_tx and num_rx = 1, because in case of error
these variables will remain unchanged by of_property_read_u32 ie 1 only

Signed-off-by: Saurabh Sengar <saurabh.truth@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodrivers: net: xgene: optimizing the code
Saurabh Sengar [Mon, 23 Nov 2015 13:32:15 +0000 (19:02 +0530)]
drivers: net: xgene: optimizing the code

this patch does the following:
1 .  remove unnecessary if, else condition
2 .  reduce one variable
3 .  change the return type of 2 functions to void as there return values
turn out to be 0 always after above changes

Signed-off-by: Saurabh Sengar <saurabh.truth@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge tag 'linux-can-next-for-4.5-20151123' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Tue, 24 Nov 2015 19:22:27 +0000 (14:22 -0500)]
Merge tag 'linux-can-next-for-4.5-20151123' of git://git./linux/kernel/git/mkl/linux-can-next

Marc Kleine-Budde says:

====================
pull-request: can-next 2015-11-23

this is a pull request of a single patch for net-next/master.

The patch by Kedareswara rao Appana converts the xilinx CAN driver to
runtime_pm.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/ipv4/ipconfig: Rejoin broken lines in console output
Geert Uytterhoeven [Tue, 24 Nov 2015 13:08:23 +0000 (14:08 +0100)]
net/ipv4/ipconfig: Rejoin broken lines in console output

Commit 09605cc12c078306 ("net ipv4: use preferred log methods") replaced
a few calls of pr_cont() after a console print without a trailing
newline by pr_info(), causing lines to be split during IP
autoconfiguration, like:

    .
    ,
     OK
    IP-Config: Got DHCP answer from 192.168.97.254,
    my address is 192.168.97.44

Convert these back to using pr_cont(), so it prints again:

    ., OK
    IP-Config: Got DHCP answer from 192.168.97.254, my address is 192.168.97.44

Absorb the printing of "my address ..." into the previous call to
pr_info(), as there's no reason to use a continuation there.

Convert one more pr_info() to print nameservers while we're at it.

Fixes: 09605cc12c078306 ("net ipv4: use preferred log methods")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: include gpio consumer header file
Arnd Bergmann [Tue, 24 Nov 2015 11:34:49 +0000 (12:34 +0100)]
net: dsa: include gpio consumer header file

After the introduction of the switch gpio reset API, I'm getting
build errors in configurations that disable CONFIG_GPIOLIB:

net/dsa/dsa.c:783:16: error: implicit declaration of function 'gpio_to_desc' [-Werror=implicit-function-declaration]

The reason is that linux/gpio/consumer.h is not automatically
included without gpiolib support. This adds an explicit #include
statement to make it compile in all configurations. The reset
functionality will not work without gpiolib, which is what you
get when disabling the feature.

As far as I can tell, gpiolib is supported on all architectures
on which you can have DSA at the moment.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: cc30c16344fc ("net: dsa: Add support for a switch reset gpio")
Acked-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Tue, 24 Nov 2015 15:08:08 +0000 (10:08 -0500)]
Merge branch 'master' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2015-11-23

This series contains updates to ixgbe, ixgbevf, fm10k, i40e and i40evf.

Jacob fixes an issue where VF could attempt to read queues it does not own,
so prevent this we check queue 0 before we continue.

Matthew fixes the MTU for jumbo frames for fm10k.

Julia Lawall cleans up a unneeded NULL test in ixgbe.

Mark cleans up a redundant header inclusion.  Adds KR mode support for
CS4227 chip.  Cleaned up diagnostic code, which is no longer needed, for
the CS4227 chip.

Jean Sacren fixes kernel documentation for ixgbe.

Alex Duyck fixes an fm10k and ixgbe issue in which the polling routine would
increase the budget for receive to at least 1 per queue if multiple queues were
present.  This would result in receive packets being processed when the budget
was 0 which is meant to indicate that no receive can be handled.  Also fixes
an ixgbevf performance issue where netperf test will starve for memory in the
time form one transmit interrupt to the next, so limit lowest interrupt rate
for adaptive interrupt moderation to 12K.  Fixed up ixgbe and ixgbevf to
use napi_schedule_irqoff() where the drivers were run from hard interrupt
context or with interrupts already disabled in netpoll.

Jesse fixes a compiler warning about an unused variable for i40evf.

John Greene fixes an issue with ixgbevf, where if the VF driver is loaded
while the corresponding PH interface is down, the driver assigns a random
MAC address, can be overwritten with the value of hw->mac.perm_addr which
is 0 at that point.  So avoid this case by initializing hw->mac.perm_addr
to the randomly generated address and do not set it unless we receive an
ACK from ixgbe.

Rasmus Villemoes cleans up some confusing code in i40e debugfs code.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoVSOCK: constify vmci_transport_notify_ops structures
Julia Lawall [Sat, 21 Nov 2015 17:39:17 +0000 (18:39 +0100)]
VSOCK: constify vmci_transport_notify_ops structures

The vmci_transport_notify_ops structures are never modified, so declare
them as const.

Done with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: atm: constify in_cache_ops and eg_cache_ops structures
Julia Lawall [Sat, 21 Nov 2015 17:28:05 +0000 (18:28 +0100)]
net: atm: constify in_cache_ops and eg_cache_ops structures

The in_cache_ops and eg_cache_ops structures are never modified, so declare
them as const.

Done with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'ipmr-cleanups'
David S. Miller [Mon, 23 Nov 2015 20:06:40 +0000 (15:06 -0500)]
Merge branch 'ipmr-cleanups'

Nikolay Aleksandrov says:

====================
net: ipmr: cleanups and minor improvements

Since I'll have to work with ipmr, I decided to clean it up and do some
minor improvements. Functionally there're almost no changes except the
SLAB_PANIC removal. Most of the patches just re-design some functions to
be clearer and more concise and try to remove the ifdef web that was
inside. There's more information in each commit. This is the first set,
the end goal is to introduce complete netlink support and control over
the mfc and vif devices.
I've tried to test all of the setsockopt/getsockopt options, and also
made builds with various ipmr kconfig options turned on and off.

v2: change patch 7 to keep SLAB_PANIC and just drop the unnecessary null
check
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: factor out common vif init code
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:32 +0000 (15:57 +0100)]
net: ipmr: factor out common vif init code

Factor out common vif init code used in both tunnel and pimreg
initialization and create ipmr_init_vif_indev() function.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: rearrange and cleanup setsockopt
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:31 +0000 (15:57 +0100)]
net: ipmr: rearrange and cleanup setsockopt

Take rtnl in the beginning unconditionally as most options already need
it (one exception - MRT_DONE, see the comment inside), make the
lock/unlock places central and move out the switch() local variables.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: drop ip_mr_init() mrt_cachep null check as we'll panic if it fails
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:30 +0000 (15:57 +0100)]
net: ipmr: drop ip_mr_init() mrt_cachep null check as we'll panic if it fails

It's not necessary to check for null as SLAB_PANIC is used and we'll
panic if the alloc fails, so just drop it.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: drop an instance of CONFIG_IP_MROUTE_MULTIPLE_TABLES
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:29 +0000 (15:57 +0100)]
net: ipmr: drop an instance of CONFIG_IP_MROUTE_MULTIPLE_TABLES

Trivial replace of ifdef with IS_BUILTIN().

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: make ip_mroute_getsockopt more understandable
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:28 +0000 (15:57 +0100)]
net: ipmr: make ip_mroute_getsockopt more understandable

Use a switch to determine if optname is correct and set val accordingly.
This produces a much more straight-forward and readable code.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: fix code and comment style
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:27 +0000 (15:57 +0100)]
net: ipmr: fix code and comment style

Trivial code and comment style fixes, also removed some extra newlines,
spaces and tabs.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: remove some pimsm ifdefs and simplify
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:26 +0000 (15:57 +0100)]
net: ipmr: remove some pimsm ifdefs and simplify

Add the helper pimsm_enabled() which replaces the old CONFIG_IP_PIMSM
define and is used to check if any version of PIM-SM has been enabled.
Use a single if defined(CONFIG_IP_PIMSM_V1) || defined(CONFIG_IP_PIMSM_V2)
for the pim-sm shared code. This is okay w.r.t IGMPMSG_WHOLEPKT because
only a VIFF_REGISTER device can send such packet, and it can't be
created if pimsm_enabled() is false.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: always define mroute_reg_vif_num
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:25 +0000 (15:57 +0100)]
net: ipmr: always define mroute_reg_vif_num

Before mroute_reg_vif_num was defined only if any of the CONFIG_PIMSM_
options were set, but that's not really necessary as the size of the
struct is the same in both cases (checked with pahole, both cases size
is 3256 bytes) and we can remove some unnecessary ifdefs to simplify the
code.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipmr: move the tbl id check in ipmr_new_table
Nikolay Aleksandrov [Sat, 21 Nov 2015 14:57:24 +0000 (15:57 +0100)]
net: ipmr: move the tbl id check in ipmr_new_table

Move the table id check in ipmr_new_table and make it return error
pointer. We need this change for the upcoming netlink table manipulation
support in order to avoid code duplication and a race condition.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agointel: i40e: fix confused code
Rasmus Villemoes [Sat, 17 Oct 2015 20:58:19 +0000 (22:58 +0200)]
intel: i40e: fix confused code

This code is pretty confused. The variable name 'bytes_not_copied'
clearly indicates that the programmer knew the semantics of
copy_{to,from}_user, but then the return value is checked for being
negative and used as a -Exxx return value.

I'm not sure this is the proper fix, but at least we get rid of the
dead code which pretended to check for access faults.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbevf: fix spoofed packets with random MAC
Emil Tantilov [Mon, 12 Oct 2015 17:56:00 +0000 (10:56 -0700)]
ixgbevf: fix spoofed packets with random MAC

If ixgbevf is loaded while the corresponding PF interface is down
and the driver assigns a random MAC address, that address can be
overwritten with the value of hw->mac.perm_addr, which would be 0 at
that point.

To avoid this case we init hw->mac.perm_addr to the randomly generated
address and do not set it unless we receive ACK from ixgbe.

Reported-by: John Greene <jogreene@redhat.com>
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbevf: use ether_addr_copy instead of memcpy
Emil Tantilov [Mon, 12 Oct 2015 17:55:51 +0000 (10:55 -0700)]
ixgbevf: use ether_addr_copy instead of memcpy

replace some instances of memcpy for setting up the mac address with
ether_addr_copy()

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40evf: fix compiler warning of unused variable
Jesse Brandeburg [Sun, 4 Oct 2015 08:09:49 +0000 (01:09 -0700)]
i40evf: fix compiler warning of unused variable

Compiler complained of an unused variable, which the driver was just
using to store the result of a rd32 which is used to clear a register
unconditionally.  Just drop the unused variable and re-use one.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: Remove CS4227 diagnostic code
Mark Rustad [Fri, 2 Oct 2015 16:23:53 +0000 (09:23 -0700)]
ixgbe: Remove CS4227 diagnostic code

Testing has now shown that the diagnostic code used with the CS4227
is no longer needed, so remove it.

Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe/ixgbevf: use napi_schedule_irqoff()
Alexander Duyck [Tue, 29 Sep 2015 22:19:43 +0000 (15:19 -0700)]
ixgbe/ixgbevf: use napi_schedule_irqoff()

The ixgbe_intr and ixgbe/ixgbevf_msix_clean_rings functions run from hard
interrupt context or with interrupts already disabled in netpoll.

They can use napi_schedule_irqoff() instead of napi_schedule()

Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbevf: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
Alexander Duyck [Tue, 29 Sep 2015 20:11:15 +0000 (13:11 -0700)]
ixgbevf: Limit lowest interrupt rate for adaptive interrupt moderation to 12K

This patch is the ixgbevf version of commit 8ac34f10a5ea4 "ixgbe: Limit
lowest interrupt rate for adaptive interrupt moderation to 12K"

The same logic applies here as well as the same results since a netperf
test will starve for memory in the time from one Tx interrupt to the next.
As a result the ixgbevf driver underperformed when compared to vhost_net.

Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: Add KR mode support for CS4227 chip
Mark Rustad [Mon, 28 Sep 2015 21:37:47 +0000 (14:37 -0700)]
ixgbe: Add KR mode support for CS4227 chip

KR auto-neg mode is what we will be using going forward. The SW
interface for this mode is different that what was used for iXFI.

Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: Fix handling of NAPI budget when multiple queues are enabled per vector
Alexander Duyck [Tue, 22 Sep 2015 21:35:41 +0000 (14:35 -0700)]
ixgbe: Fix handling of NAPI budget when multiple queues are enabled per vector

This patch corrects an issue in which the polling routine would increase
the budget for Rx to at least 1 per queue if multiple queues were present.
This would result in Rx packets being processed when the budget was 0 which
is meant to indicate that no Rx can be handled.

Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agofm10k: Fix handling of NAPI budget when multiple queues are enabled per vector
Alexander Duyck [Tue, 22 Sep 2015 21:35:35 +0000 (14:35 -0700)]
fm10k: Fix handling of NAPI budget when multiple queues are enabled per vector

This patch corrects an issue in which the polling routine would increase
the budget for Rx to at least 1 per queue if multiple queues were present.
This would result in Rx packets being processed when the budget was 0 which
is meant to indicate that no Rx can be handled.

Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: fix multiple kernel-doc errors
Jean Sacren [Sat, 19 Sep 2015 11:08:44 +0000 (05:08 -0600)]
ixgbe: fix multiple kernel-doc errors

The commit dfaf891dd3e1 ("ixgbe: Refactor the RSS configuration code")
introduced a few kernel-doc errors:

1) The function name is missing;
2) The format is wrong;
3) The short description is redundant.

Fix all the above for the correct execution of the kernel doc.

Signed-off-by: Jean Sacren <sakiwit@gmail.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: Delete redundant include file
Mark Rustad [Fri, 18 Sep 2015 17:08:00 +0000 (10:08 -0700)]
ixgbe: Delete redundant include file

Delete a redundant include of net/vxlan.h.

Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: drop null test before destroy functions
Julia Lawall [Sun, 13 Sep 2015 12:15:13 +0000 (14:15 +0200)]
ixgbe: drop null test before destroy functions

Remove unneeded NULL test.

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@ expression x; @@
-if (x != NULL)
  \(kmem_cache_destroy\|mempool_destroy\|dma_pool_destroy\)(x);
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Tested-by: Darin Miller <darin.j.miller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agofm10k: Correct MTU for jumbo frames
Jacob Keller [Tue, 25 Aug 2015 20:49:11 +0000 (13:49 -0700)]
fm10k: Correct MTU for jumbo frames

Based on hardware testing, the host interface supports up to 15368 bytes
as the maximum frame size. To determine the correct MTU, we subtract 8
for the internal switch tag, 14 for the L2 header, and 4 for the
appended FCS header, resulting in 15342 bytes of payload for our maximum
MTU on jumbo frames.

Signed-off-by: Matthew Vick <matthew.vick@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agofm10k: do not assume VF always has 1 queue
Jacob Keller [Tue, 25 Aug 2015 00:27:24 +0000 (17:27 -0700)]
fm10k: do not assume VF always has 1 queue

It is possible that the PF has not yet assigned resources to the VF.
Although rare, this could result in the VF attempting to read queues it
does not own and result in FUM or THI faults in the PF. To prevent this,
check queue 0 before we continue in init_hw_vf.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <Krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoMerge branch 'rhashtable-test-enhancements'
David S. Miller [Mon, 23 Nov 2015 17:36:09 +0000 (12:36 -0500)]
Merge branch 'rhashtable-test-enhancements'

Phil Sutter says:

====================
improve fault-tolerance of rhashtable runtime-test

The following series aims to improve lib/test_rhashtable in different
situations:

Patch 1 allows the kernel to reschedule so the test does not block too
        long on slow systems.
Patch 2 fixes behaviour under pressure, retrying inserts in non-permanent
        error case (-EBUSY).
Patch 3 auto-adjusts the upper table size limit according to the number
        of threads (in concurrency test). In fact, the current default is
already too small.
Patch 4 makes it possible to retry inserts even in supposedly permanent
        error case (-ENOMEM) to expose rhashtable's remaining problem of
-ENOMEM being not as permanent as it is expected to be.

Changes since v1:
- Introduce insert_retry() which is then used in single-threaded test as
  well.
- Do not retry inserts by default if -ENOMEM was returned.
- Rename the retry counter to be a bit more verbose about what it
  contains.
- Add patch 4 as a debugging aid.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorhashtable-test: allow to retry even if -ENOMEM was returned
Phil Sutter [Fri, 20 Nov 2015 17:17:20 +0000 (18:17 +0100)]
rhashtable-test: allow to retry even if -ENOMEM was returned

This is rather a hack to expose the current issue with rhashtable to
under high pressure sometimes return -ENOMEM even though system memory
is not exhausted and a consecutive insert may succeed.

Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorhashtable-test: calculate max_entries value by default
Phil Sutter [Fri, 20 Nov 2015 17:17:19 +0000 (18:17 +0100)]
rhashtable-test: calculate max_entries value by default

A maximum table size of 64k entries is insufficient for the multiple
threads test even in default configuration (10 threads * 50000 objects =
500000 objects in total). Since we know how many objects will be
inserted, calculate the max size unless overridden by parameter.

Note that specifying the exact number of objects upon table init won't
suffice as that value is being rounded down to the next power of two -
anticipate this by rounding up to the next power of two in beforehand.

Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorhashtable-test: retry insert operations
Phil Sutter [Fri, 20 Nov 2015 17:17:18 +0000 (18:17 +0100)]
rhashtable-test: retry insert operations

After adding cond_resched() calls to threadfunc(), a surprisingly high
rate of insert failures occurred probably due to table resizes getting a
better chance to run in background. To not soften up the remaining
tests, retry inserts until they either succeed or fail permanently.

Also change the non-threaded test to retry insert operations, too.

Suggested-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorhashtable-test: add cond_resched() to thread test
Phil Sutter [Fri, 20 Nov 2015 17:17:17 +0000 (18:17 +0100)]
rhashtable-test: add cond_resched() to thread test

This should fix for soft lockup bugs triggered on slow systems.

Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>