GitHub/moto-9609/android_kernel_motorola_exynos9610.git
8 years agoi40e: Initialize pointer in client_release function
Harshitha Ramamurthy [Mon, 15 Aug 2016 21:17:19 +0000 (14:17 -0700)]
i40e: Initialize pointer in client_release function

The function i40e_client_release has a print statement that uses an
adapter pointer which is not initialized if a previous if statement
is not true. Hence, intialize it in the right place.

Change-ID: I1afdaa2c46771ac42be56edcc41bb56b455b06c8
Signed-off-by: Harshitha Ramamurthy <harshitha.ramamurthy@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Check client is open before calling client ops
Catherine Sullivan [Mon, 15 Aug 2016 21:17:18 +0000 (14:17 -0700)]
i40e: Check client is open before calling client ops

We were having a race between the completion of the client open and
calls to the client ops so don't call a client op unless we are sure the
client is open.

Testing Hints: Load IWARP driver and make sure it works as expected.

Change-Id: I741f4f2aa4fcbfdad3e40dabbbb1b005856c396b
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: use matching format identifiers
Heinrich Schuchardt [Wed, 10 Aug 2016 23:07:22 +0000 (01:07 +0200)]
i40e: use matching format identifiers

i is defined as int but output as %u several times.
Adjust the format identifiers.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Add support for HMC resource and profile for X722
Carolyn Wyborny [Thu, 4 Aug 2016 18:37:05 +0000 (11:37 -0700)]
i40e: Add support for HMC resource and profile for X722

This patch adds support for HMC resource and profile cmds for X722
firmware.

Change-ID: Icc332101f38ab15d1bfa167823100eb4f6822f7e
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Fix byte ordering in ARP NS code for X722
Carolyn Wyborny [Thu, 4 Aug 2016 18:37:04 +0000 (11:37 -0700)]
i40e: Fix byte ordering in ARP NS code for X722

This patch fixes byte ordering problems found when enabling this feature
support. Without this patch, the feature will not work correctly. This
patch fixes the definitions to have the correct byte order.

Change-ID: Ic7489fbcbe2195df7be62ff5e359201b827cefe6
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: refactor tail_bump check
Carolyn Wyborny [Thu, 4 Aug 2016 18:37:03 +0000 (11:37 -0700)]
i40e: refactor tail_bump check

This patch refactors tail bump check.

Change-ID: Ide0e19171d67d90cb2b06b8dcd4fa791ae120160
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40evf: report link speed
Mitch Williams [Thu, 4 Aug 2016 18:37:02 +0000 (11:37 -0700)]
i40evf: report link speed

The PF driver tells us the link speed, so do something with that
information. Add link speed to log messages, and report speed through
ethtool.

Change-Id: I279dc9540cc5203376406050a3e8d67e128d5882
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: use alloc_workqueue instead of create_singlethread_workqueue
Jacob Keller [Thu, 4 Aug 2016 18:37:01 +0000 (11:37 -0700)]
i40e: use alloc_workqueue instead of create_singlethread_workqueue

Replace calls to create_singlethread_workqueue instead with alloc_workqueue
as is style with other Intel drivers. This provides more control over
workqueue creation, and allows explicit setting of the desired mode of
operation. It also makes it more obvious that driver name constant is
passed to a format "%s".

Change-ID: I6192b44caf5140336cd54c5b350d51c73b541fdb
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: use configured RSS key and lookup table in i40e_vsi_config_rss
Jacob Keller [Tue, 19 Jul 2016 23:23:31 +0000 (16:23 -0700)]
i40e: use configured RSS key and lookup table in i40e_vsi_config_rss

A previous refactor added support to store user configuration for VSIs,
so that extra VSIs such as for VMDq can use this information when
configuring. Unfortunately the i40e_vsi_config_rss function was missed
in this refactor, and the values were being ignored. Fix this by
checking for the fields and using those instead of always using the
default values.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: fix broken i40e_config_rss_aq function
Jacob Keller [Tue, 19 Jul 2016 23:23:30 +0000 (16:23 -0700)]
i40e: fix broken i40e_config_rss_aq function

X722 hardware requires using the admin queue to configure RSS. This
function was previously re-written in commit e69ff813af35 ("i40e: rework
the functions to configure RSS with similar parameters").
However, the previous refactor did not work correctly for a few reasons

(a) it does not check whether seed is NULL before using it, resulting in
a NULL pointer dereference

[  402.954721] BUG: unable to handle kernel NULL pointer dereference at           (null)
[  402.955568] IP: [<ffffffffa0090ccf>] i40e_config_rss_aq.constprop.65+0x2f/0x1c0 [i40e]
[  402.956402] PGD ad610067 PUD accc0067 PMD 0
[  402.957235] Oops: 0000 [#1] SMP
[  402.958064] Modules linked in: ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 xt_conntrack ip_set nfnetlink ebtable_filter ebtable_
broute bridge stp llc ebtable_nat ebtables ip6table_mangle ip6table_raw ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv
6 ip6table_security ip6table_filter ip6_tables iptable_mangle iptable_raw iptable_nat nf_conntrack_ipv4_ nf_defrag_ipv4_ nf_nat_ip
v4_ nf_nat nf_conntrack iptable_security intel_rapl i86_kg_temp_thermal coretemp kvm_intel kvm irqbypass crct10dif_clMl crc32_
pclMl ghash_clMlni_intel iTCO_wdt iTCO_vendor_support shpchp sb_edac dcdbas pcspkr joydev ipmi_devintf wmi edac_core ipmi_ssif
 acpi_ad acpi_ower_meter ipmi_si ipmi_msghandler mei_me nfsd lpc_ich mei ioatdma tpm_tis auth_rpcgss tpm nfs_acl lockd grace s
unrpc ifs nngag200 i2c_algo_bit drm_kms_helper ttm drm iigbe bnx2x i40e dca mdio ptp pps_core libcrc32c fjes crc32c_intel
[  402.965563] CPU: 22 PID: 2461 Conm: ethtool Not tainted 4.6.0-rc7_1.2-ABNidQ+ #20
[  402.966719] Hardware name: Dell Inc. PowerEdge R720/0C4Y3R, BIOS 2.5.2 01/28/2015
[  402.967862] task: ffff880219b51dc0 ti: ffff8800b3408000 task.ti: ffff8800b3408000
[  402.969046] RIP: 0010:[<ffffffffa0090ccf>]  [<ffffffffa0090ccf>] i40e_config_rss_aq.constprop.65+0x2f/0x1c0 [i40e]
[  402.970339] RSP: 0018:ffff8800b340ba90  EFLAGS: 00010246
[  402.971616] RAX: 0000000000000000 RBX: ffff88042ec14000 RCX: 0000000000000200
[  402.972961] RDX: ffff880428eb9200 RSI: 0000000000000000 RDI: ffff88042ec14000
[  402.974312] RBP: ffff8800b340baf8 R08: ffff880237ada8f0 R09: ffff880428eb9200
[  402.975709] R10: ffff880428eb9200 R11: 0000000000000000 R12: ffff88042ec2e000
[  402.977104] R13: ffff88042ec2e000 R14: ffff88042ec14000 R15: ffff88022ea00800
[  402.978541] FS:  00007f84fd054700(0000) GS:ffff880237ac0000(0000) knlGS:0000000000000000
[  402.980003] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  402.981508] CR2: 0000000000000000 CR3: 000000003289e000 CR4: 00000000000406e0
[  402.983028] Stack:
[  402.984578]  0000000002000200 0000000000000000 ffff88023ffeda68 ffff88023ffef000
[  402.986187]  0000000000000268 ffff8800b340bbf8 ffff88023ffedd80 0000000088ce4f1d
[  402.987844]  ffff88042ec14000 ffff88022ea00800 ffff88042ec2e000 ffff88042ec14000
[  402.989509] Call Trace:
[  402.991200]  [<ffffffffa009636f>] i40e_config_rss+0x11f/0x1c0 [i40e]
[  402.992924]  [<ffffffffa00a1ae0>] i40e_set_rifh+0ic0/0x130 [i40e]
[  402.994684]  [<ffffffff816d54b7>] ethtool_set_rifh+0x1f7/0x300
[  402.996446]  [<ffffffff8136d02b>] ? cred_has_capability+0io6b/0x100
[  402.998203]  [<ffffffff8136d102>] ? selinux_capable+0x12/0x20
[  402.999968]  [<ffffffff8136277b>] ? security_capable+0x4b/0x70
[  403.001707]  [<ffffffff816d6da3>] dev_ethtool+0x1423/0x2290
[  403.003461]  [<ffffffff816eab41>] dev_ioctl+0x191/0io630
[  403.005186]  [<ffffffff811cf80a>] ? lru_cache_add+0x3a/0i80
[  403.006942]  [<ffffffff817f2a8e>] ? _raw_spin_unlock+0ie/0x20
[  403.008691]  [<ffffffff816adb95>] sock_do_ioctl+0x45/0i50
[  403.010421]  [<ffffffff816ae229>] sock_ioctl+0x209/0x2d0
[  403.012173]  [<ffffffff81262194>] do_vfs_ioctl+0u4/0io6c0
[  403.013911]  [<ffffffff81262829>] SyS_ioctl+0x79/0x90
[  403.015710]  [<ffffffff817f2e72>] entry_SYSCALL_64_fastpath+0x1a/0u4
[  403.017500] Code: 90 55 48 89 e5 41 57 41 56 41 55 41 54 53 48 89 fb 48 83 ec 40 4c 8b a7 e0 05 00 00 65 48 8b 04 25 28 00 00 00 48 89 45 d0 31 c0 <48> 8b 06 41 0f b7 bc 24 f2 0f 00 00 48 89 45 9c 48 8b 46 08 48
[  403.021454] RIP  [<ffffffffa0090ccf>] i40e_config_rss_aq.constprop.65+0x2f/0x1c0 [i40e]
[  403.023395]  RSP <ffff8800b340ba90>
[  403.025271] CR2: 0000000000000000
[  403.027169] ---[ end trace 64561b528cf61cf0 ]---

(b) it does not even bother to use the passed in *lut parameter which
defines the requested lookup table. Instead it uses its own round robin
table.

Fix these issues by re-writing it to be similar to i40e_config_rss_reg
and i40e_get_rss_aq.

Fixes: e69ff813af35 ("i40e: rework the functions to configure RSS with similar parameters", 2015-10-21)
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: move i40e_vsi_config_rss below i40e_get_rss_aq
Jacob Keller [Tue, 19 Jul 2016 23:23:29 +0000 (16:23 -0700)]
i40e: move i40e_vsi_config_rss below i40e_get_rss_aq

Move this function below the two functions related to configuring RSS
via the admin queue. This helps co-locate the two functions, and made it
easier to spot a bug in the first i40e_config_rss_aq function as
compared to the i40e_get_rss_aq function.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agonet/irda: remove pointless assignment/check
Vegard Nossum [Fri, 19 Aug 2016 16:08:57 +0000 (18:08 +0200)]
net/irda: remove pointless assignment/check

We've already set sk to sock->sk and dereferenced it, so if it's NULL
we would have crashed already. Moreover, if it was NULL we would have
crashed anyway when jumping to 'out' and trying to unlock the sock.
Furthermore, if we had assigned a different value to 'sk' we would
have been calling lock_sock() and release_sock() on different sockets.

My conclusion is that these two lines are complete nonsense and only
serve to confuse the reader.

Signed-off-by: Vegard Nossum <vegard.nossum@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed: utilize FW 8.10.10.0
Yuval Mintz [Fri, 19 Aug 2016 06:33:31 +0000 (09:33 +0300)]
qed: utilize FW 8.10.10.0

This new firmware for the qed* adpaters fixes several issues:
 - Better blocking of malicious VFs.
 - After FLR, Tx-switching [internal routing] of packets might
   be incorrect.
 - Deletion of unicast MAC filters would sometime have side-effect
   of corrupting the MAC filters configred for a device.
It also contains fixes for future qed* drivers that *hopefully* would be
sent for review in the near future.

In addition, it would allow driver some new functionality, including:
 - Allowing PF/VF driver compaitibility with old drivers [running
   pre-8.10.5.0 firmware].
 - Better debug facilities.

This would also bump the qed* driver versions to 8.10.9.20.

Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch '10GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Sat, 20 Aug 2016 00:19:20 +0000 (17:19 -0700)]
Merge branch '10GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
10GbE Intel Wired LAN Driver Updates 2016-08-18

This series contains updates to ixgbe and ixgbevf.

Emil cleans up confusing amongst the users by making an error message
into a debug message, since the TXDCTL.ENABLE (and comparable
VFTXDCTL.ENABLE for ixgbevf) bit is set only when the
transmit queue is actually enabled, which may not happen during the
configure phase eve if we waited for it.  Converts to using netdev_dbg()
macro instead of our home brewed macro for ixgbevf.  Converted the
service task to use atomic bitwise operations when setting and checking
reset requests to reduce the possibility of race conditions.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'bcm_sf2-platform-dev'
David S. Miller [Sat, 20 Aug 2016 00:15:37 +0000 (17:15 -0700)]
Merge branch 'bcm_sf2-platform-dev'

Florian Fainelli says:

====================
net: dsa: bcm_sf2: Platform conversion

This patch series converts the bcm_sf2 driver from a traditional DSA driver
into a platform_device driver and makes it use the new DSA binding that Andrew
introduced in the latest merge window.

Prior attempts used to coerce the code in net/dsa/dsa2.c to accept the old
binding, while really there is only one broken single user out there: bcm_sf2,
so instead, just assume the new DT binding is deployed and use it accordingly.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: bcm_sf2: Remove probing through old DSA binding
Florian Fainelli [Thu, 18 Aug 2016 22:30:16 +0000 (15:30 -0700)]
net: dsa: bcm_sf2: Remove probing through old DSA binding

Remove our dsa_switch_driver::drv_probe callback to prevent probing
through the old DSA binding, not that this could happen anymore now that
we have moved the matching compatible string from net/dsa/dsa.c to
drivers/net/dsa/bcm_sf2.c, so this is essentially dead code.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: bcm_sf2: Use device managed helpers
Florian Fainelli [Thu, 18 Aug 2016 22:30:15 +0000 (15:30 -0700)]
net: dsa: bcm_sf2: Use device managed helpers

Now that we have converted the drivers into a proper platform device
driver, we can use the device managed helper functions to simplify the
error paths a bit wrt. register resources and IRQs.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: bcm_sf2: Make it a real platform device driver
Florian Fainelli [Thu, 18 Aug 2016 22:30:14 +0000 (15:30 -0700)]
net: dsa: bcm_sf2: Make it a real platform device driver

The Broadcom Starfighter 2 switch driver should be a proper platform
driver, now that the DSA code has been updated to allow that, register a
switch device, feed it with the proper configuration data coming from
Device Tree and register our switch device with DSA.

The bulk of the changes consist in moving what bcm_sf2_sw_setup() did
into the platform driver probe function.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoDocumentation: dt: bindings: Update Broadcom 7445 switch document
Florian Fainelli [Thu, 18 Aug 2016 22:30:13 +0000 (15:30 -0700)]
Documentation: dt: bindings: Update Broadcom 7445 switch document

Document the updated binding which conforms to the new DSA binding in
net/dsa/dsa.txt.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: Export suspend/resume functions
Florian Fainelli [Thu, 18 Aug 2016 22:30:12 +0000 (15:30 -0700)]
net: dsa: Export suspend/resume functions

In preparation for allowing switch drivers to implement system-wide
suspend/resume functions, export dsa_switch_suspend and
dsa_switch_resume() such that these are callable from the appropriate
driver specific suspend/resume functions.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'mv88e6xxx-fix-wait'
David S. Miller [Sat, 20 Aug 2016 00:14:08 +0000 (17:14 -0700)]
Merge branch 'mv88e6xxx-fix-wait'

Andrew Lunn says:

====================
Fix mv88e6xxx wait function

The mv88e6xxx wait function can be upset of the system has nots of
other things to do and a sleep takes a lot longer than expected. Fix
this be using a fixed number of iterations, rather than a fixed
walkclock time.

Witht that change made, it is possible to consoliate another
wait function.

A wait actually timing out should not happen and when it does, it
means something serious is wrong. Make sure an error is logged,
since not all callers will log an error.
====================

Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodsa: mv88e6xxx: Make mv88e6xxx_wait() timeout verbose
Andrew Lunn [Thu, 18 Aug 2016 22:01:57 +0000 (00:01 +0200)]
dsa: mv88e6xxx: Make mv88e6xxx_wait() timeout verbose

When mv88e6xxx_wait() returns a timeout, something bad has
happened. Make sure it is noticed by logging an error.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodsa: mv88e6xxx: Use mv88e6xx_wait in mv88e6xxx_update()
Andrew Lunn [Thu, 18 Aug 2016 22:01:56 +0000 (00:01 +0200)]
dsa: mv88e6xxx: Use mv88e6xx_wait in mv88e6xxx_update()

Now that mv88e6xx_wait() iterated on a counter than a fixed time
interval, it implements the same mechanism as mv88e6xxx_update() uses.
So use it in mv88e6xx_wait().

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodsa: mv88e6xxx: Timeout based on iterations, not time
Andrew Lunn [Thu, 18 Aug 2016 22:01:55 +0000 (00:01 +0200)]
dsa: mv88e6xxx: Timeout based on iterations, not time

The mv88e6xxx driver times out operations on the switch based on
looping until an elapsed wall clock time is reached. However, if
usleep_range() sleeps much longer than expected, it could timeout with
an error without actually checking to see if the devices has completed
the operation. So replace the elapsed time with a fixed upper bound on
the number of loops.

Testing on various switches has shown that switches takes either 0 or
1 iteration, so a maximum of 16 iterations is a safe limit.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'phy-next'
David S. Miller [Sat, 20 Aug 2016 00:11:50 +0000 (17:11 -0700)]
Merge branch 'phy-next'

Andrew Lunn says:

====================
PHY Kconfig and Makefile cleanup

The Ethernet PHY directory has slowly been getting more entries.
Split the entries in the Makefile and Kconfig into MDIO bus drivers
and PHYs. Within these two groups, sort them. This should reduce merge
conflicts and aid finding what one searches for.

The Kconfig text contains redundant "Driver for" and "Support for"
which add little value, make the vendor less obvious, and defeat the
shortcut key in the menu. Remove such text.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: phy: Kconfig: Remove redundant "Support for"
Andrew Lunn [Thu, 18 Aug 2016 21:56:06 +0000 (23:56 +0200)]
net: phy: Kconfig: Remove redundant "Support for"

Remove the redundant "Support for" and "Drivers for" from the Kconfig
short description. This makes the manufacture much more prominent in
the list and makes the shortcut keys useful.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: phy: Sort Makefile and Kconfig
Andrew Lunn [Thu, 18 Aug 2016 21:56:05 +0000 (23:56 +0200)]
net: phy: Sort Makefile and Kconfig

Sort the files to reduce merge conflicts and to make it easier to find
drivers by name. Also separate the MDIO bus drivers from the PHY
drivers, again to help find what you need.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipv4: fix sparse error in fib_good_nh()
Eric Dumazet [Thu, 18 Aug 2016 17:19:34 +0000 (10:19 -0700)]
net: ipv4: fix sparse error in fib_good_nh()

Fixes following sparse errors :

net/ipv4/fib_semantics.c:1579:61: warning: incorrect type in argument 2
(different base types)
net/ipv4/fib_semantics.c:1579:61:    expected unsigned int [unsigned]
[usertype] key
net/ipv4/fib_semantics.c:1579:61:    got restricted __be32 const
[usertype] nh_gw

Fixes: a6db4494d218c ("net: ipv4: Consider failed nexthops in multipath routes")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoudp: include addrconf.h
Eric Dumazet [Thu, 18 Aug 2016 16:59:12 +0000 (09:59 -0700)]
udp: include addrconf.h

Include ipv4_rcv_saddr_equal() definition to avoid this sparse error :

net/ipv4/udp.c:362:5: warning: symbol 'ipv4_rcv_saddr_equal' was not
declared. Should it be static?

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotcp: md5: remove tcp_md5_hash_header()
Eric Dumazet [Thu, 18 Aug 2016 16:49:55 +0000 (09:49 -0700)]
tcp: md5: remove tcp_md5_hash_header()

After commit 19689e38eca5 ("tcp: md5: use kmalloc() backed scratch
areas") this function is no longer used.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoibmvnic: Handle backing device failover and reinitialization
Thomas Falcon [Thu, 18 Aug 2016 16:37:51 +0000 (11:37 -0500)]
ibmvnic: Handle backing device failover and reinitialization

An upcoming feature of IBM VNIC protocol is the ability to configure
redundant backing devices for a VNIC client. In case of a failure
on the current backing device, the driver will receive a signal
from the hypervisor indicating that a failover will occur. The driver
will then wait for a message from the backing device before
establishing a new connection.

Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: hns: Add reset function support for RoCE driver
oulijun [Thu, 18 Aug 2016 12:32:52 +0000 (20:32 +0800)]
net: hns: Add reset function support for RoCE driver

It added reset function for RoCE driver. RoCE is a feature of hns.
In hip06 SoC, in RoCE reset process, it's needed to configure dsaf
channel reset, port and sl map info. Reset function of RoCE is
located in dsaf module, we only call it in RoCE driver when needed.

This patch is used to fix the conflict, please refer to this link:
  https://www.spinics.net/lists/linux-rdma/msg39114.html

Signed-off-by: Wei Hu <xavier.huwei@huawei.com>
Signed-off-by: Nenglong Zhao <zhaonenglong@hisilicon.com>
Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Sheng Li <lisheng011@huawei.com>
Reviewed-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'rhash-raw-walkers-remove-part-1'
David S. Miller [Fri, 19 Aug 2016 21:40:25 +0000 (14:40 -0700)]
Merge branch 'rhash-raw-walkers-remove-part-1'

Herbert Xu says:

====================
rhashtable: Get rid of raw table walkers part 1

This series starts the process of getting rid of all raw rhashtable
walkers (e.g., using any of the rht_for_each helpers) from the
kernel.

We need to do this before I can fix the resize kmalloc failure issue
by using multi-layered tables.

We should do this anyway because almost all raw table walkers are
already buggy in that they don't handle multiple rhashtables during
a resize.
====================

Dave/Tomas, please keep an eye out for any new patches that try
to introduce raw table walkers and nack them.

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonetlink: Use rhashtable walk interface in diag dump
Herbert Xu [Fri, 19 Aug 2016 08:21:37 +0000 (16:21 +0800)]
netlink: Use rhashtable walk interface in diag dump

This patch converts the diag dumping code to use the rhashtable
walk code instead of going through rhashtable by hand.  The lock
nl_table_lock is now only taken while we process the multicast
list as it's not needed for the rhashtable walk.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMAINTAINERS: Add extra rhashtable maintainer
Herbert Xu [Thu, 18 Aug 2016 08:50:57 +0000 (16:50 +0800)]
MAINTAINERS: Add extra rhashtable maintainer

As I'm working actively on rhashtable it helps if people CCed me
when they work on in.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorhashtable: Remove GFP flag from rhashtable_walk_init
Herbert Xu [Thu, 18 Aug 2016 08:50:56 +0000 (16:50 +0800)]
rhashtable: Remove GFP flag from rhashtable_walk_init

The commit 8f6fd83c6c5ec66a4a70c728535ddcdfef4f3697 ("rhashtable:
accept GFP flags in rhashtable_walk_init") added a GFP flag argument
to rhashtable_walk_init because some users wish to use the walker
in an unsleepable context.

In fact we don't need to allocate memory in rhashtable_walk_init
at all.  The walker is always paired with an iterator so we could
just stash ourselves there.

This patch does that by introducing a new enter function to replace
the existing init function.  This way we don't have to churn all
the existing users again.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'cxgb-crypto'
David S. Miller [Fri, 19 Aug 2016 07:00:45 +0000 (00:00 -0700)]
Merge branch 'cxgb-crypto'

Hariprasad Shenai says:

====================
crypto/chcr: Add support for Chelsio Crypto Driver

This patch series adds support for Chelsio Crypto driver.

The patch series has been created against net-next tree and includes
patches for Chelsio Low Level Driver(cxgb4) and adds the new crypto
Upper Layer Driver(chcr) under a new directory drivers/crypto/chelsio.

Patch 1/4 ("cxgb4: Add support for dynamic allocation of resources for
ULD") adds support for dynamic allocation of resources for ULD. The
objective of this patch is to provide generic interface for upper layer
drivers to allocate and initialize hardware resources.

The present cxgb4 (network driver) apart from network functionality, also
initializes hardware and thus acts as lower layer driver for other drivers
to use hardware resources. Thus it acts as both a Low level driver for
Upper layer driver's like iw_cxgb4, cxgb4i and cxgb4it and a Network Driver.

Right now the allocation of resources for Upper layer driver's is done
statically. Patch 1/4 adds a new infrastructure for dynamic allocation of
resources. cxgb4 will read the hardware capability through firmware and
allocate/free the queues for Upper layer drivers when the respective
driver's are loaded and freed when unloaded.

Patch 2/3, 3/4 and 4/4 adds support for Chelsio Crypto Driver. The Crypto
driver will act as another ULD on top of cxgb4.

In this patch series, the ULD API framework is used only by crypto and other
ULD's will make use of it in the next series.

This patch series is only for review, if this looks ok we will test it
thoroughly and send request for merge.

We have included all the maintainers of respective drivers. Kindly
review the changes and provide feedback on the same.

V3: - Removed crypto queues from cxgb4 and added support for dynamic
      allocation of resources for Upper layer drivers
    - Dependency fix in Kconfig.

V2: - Some residual code cleanup
    - Adds pr_fmt with chcr (KBUILD_MODNAME) added
    - Changes var name to accomodate them <80 columns in the chcr_register_alg
    - Support for printing the crypto queue stats
    - Fix compile warnings reported by kbuild bot for certain architectures
    - Dependency fix in Kconfig.
    - If the request has the MAY_BACKLOG bit set and hardware queue is
      full the request is queued up else -EBUSY is returned to throttle
      the user. The queue when executed and processed returns -EINPROGRESS
      in completion.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocrypto: Added Chelsio Menu to the Kconfig file
Hariprasad Shenai [Wed, 17 Aug 2016 07:03:06 +0000 (12:33 +0530)]
crypto: Added Chelsio Menu to the Kconfig file

Adds the config entry for the Chelsio Crypto Driver, Makefile changes
for the same.

Signed-off-by: Atul Gupta <atul.gupta@chelsio.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agochcr: Support for Chelsio's Crypto Hardware
Hariprasad Shenai [Wed, 17 Aug 2016 07:03:05 +0000 (12:33 +0530)]
chcr: Support for Chelsio's Crypto Hardware

The Chelsio's Crypto Hardware can perform the following operations:
SHA1, SHA224, SHA256, SHA384 and SHA512, HMAC(SHA1), HMAC(SHA224),
HMAC(SHA256), HMAC(SHA384), HAMC(SHA512), AES-128-CBC, AES-192-CBC,
AES-256-CBC, AES-128-XTS, AES-256-XTS

This patch implements the driver for above mentioned features. This
driver is an Upper Layer Driver which is attached to Chelsio's LLD
(cxgb4) and uses the queue allocated by the LLD for sending the crypto
requests to the Hardware and receiving the responses from it.

The crypto operations can be performed by Chelsio's hardware from the
userspace applications and/or from within the kernel space using the
kernel's crypto API.

The above mentioned crypto features have been tested using kernel's
tests mentioned in testmgr.h. They also have been tested from user
space using libkcapi and Openssl.

Signed-off-by: Atul Gupta <atul.gupta@chelsio.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Register changes and fw defines for crypto
Hariprasad Shenai [Wed, 17 Aug 2016 07:03:04 +0000 (12:33 +0530)]
cxgb4: Register changes and fw defines for crypto

Signed-off-by: Atul Gupta <atul.gupta@chelsio.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Add support for dynamic allocation of resources for ULD
Hariprasad Shenai [Wed, 17 Aug 2016 07:03:03 +0000 (12:33 +0530)]
cxgb4: Add support for dynamic allocation of resources for ULD

Add a new commmon infrastructure to allocate reosurces dynamically to
Upper layer driver's(ULD) when they register with cxgb4 driver and free
them during unregistering. All the queues and the interrupts for
them will be allocated during ULD probe only and freed during remove.

Signed-off-by: Atul Gupta <atul.gupta@chelsio.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoatm: fore200e: Do not drop const qualifier
LABBE Corentin [Wed, 17 Aug 2016 13:56:45 +0000 (15:56 +0200)]
atm: fore200e: Do not drop const qualifier

The data member of structure firmware is const and this constness is
dropped by some cast.
This patch add some const for keeping the const information.

Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'bpf-next'
David S. Miller [Fri, 19 Aug 2016 06:38:17 +0000 (23:38 -0700)]
Merge branch 'bpf-next'

Daniel Borkmann says:

====================
BPF helper improvements and cleanups

This set adds various improvements to BPF helpers, a cleanup to use
skb_pkt_type_ok() helper, addition of bpf_skb_change_tail(), a follow
up for event output helper and removing ifdefs around the cgroupv2
helper bits. For details please see individual patches.

The set is based against net-next tree, but requires a merge of net
into net-next first.

Thanks a lot!
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: get rid of cgroup helper related ifdefs
Daniel Borkmann [Wed, 17 Aug 2016 23:00:41 +0000 (01:00 +0200)]
bpf: get rid of cgroup helper related ifdefs

As recently discussed during the task_under_cgroup_hierarchy() addition,
we should get rid of the ifdefs surrounding the bpf_skb_under_cgroup()
helper. If related functionality is not built-in, the helper cannot be
used anyway, which is also in line with what we do for all other helpers.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: enable event output helper also for xdp types
Daniel Borkmann [Wed, 17 Aug 2016 23:00:40 +0000 (01:00 +0200)]
bpf: enable event output helper also for xdp types

Follow-up to 555c8a8623a3 ("bpf: avoid stack copy and use skb ctx for
event output") for also adding the event output helper for XDP typed
programs. The event output helper has been very useful in particular for
debugging or event notification purposes, since it's much faster and
flexible than regular trace printk due to programmatically being able to
attach meta data. Same flags structure applies as with tc BPF programs.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: add bpf_skb_change_tail helper
Daniel Borkmann [Wed, 17 Aug 2016 23:00:39 +0000 (01:00 +0200)]
bpf: add bpf_skb_change_tail helper

This work adds a bpf_skb_change_tail() helper for tc BPF programs. The
basic idea is to expand or shrink the skb in a controlled manner. The
eBPF program can then rewrite the rest via helpers like bpf_skb_store_bytes(),
bpf_lX_csum_replace() and others rather than passing a raw buffer for
writing here.

bpf_skb_change_tail() is really a slow path helper and intended for
replies with f.e. ICMP control messages. Concept is similar to other
helpers like bpf_skb_change_proto() helper to keep the helper without
protocol specifics and let the BPF program mangle the remaining parts.
A flags field has been added and is reserved for now should we extend
the helper in future.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: use skb_pkt_type_ok helper in bpf_skb_change_type
Daniel Borkmann [Wed, 17 Aug 2016 23:00:38 +0000 (01:00 +0200)]
bpf: use skb_pkt_type_ok helper in bpf_skb_change_type

Since we have a skb_pkt_type_ok() helper for checking the type before
mangling, make use of it instead of open coding. Follow-up to commit
8b10cab64c13 ("net: simplify and make pkt_type_ok() available for other
users") that came in after d2485c4242a8 ("bpf: add bpf_skb_change_type
helper").

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotipc: add peer removal functionality
Richard Alpe [Thu, 18 Aug 2016 08:33:52 +0000 (10:33 +0200)]
tipc: add peer removal functionality

Add TIPC_NL_PEER_REMOVE netlink command. This command can remove
an offline peer node from the internal data structures.

This will be supported by the tipc user space tool in iproute2.

Signed-off-by: Richard Alpe <richard.alpe@ericsson.com>
Reviewed-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotcp: refine tcp_prune_ofo_queue() to not drop all packets
Eric Dumazet [Wed, 17 Aug 2016 21:17:09 +0000 (14:17 -0700)]
tcp: refine tcp_prune_ofo_queue() to not drop all packets

Over the years, TCP BDP has increased a lot, and is typically
in the order of ~10 Mbytes with help of clever Congestion Control
modules.

In presence of packet losses, TCP stores incoming packets into an out of
order queue, and number of skbs sitting there waiting for the missing
packets to be received can match the BDP (~10 Mbytes)

In some cases, TCP needs to make room for incoming skbs, and current
strategy can simply remove all skbs in the out of order queue as a last
resort, incurring a huge penalty, both for receiver and sender.

Unfortunately these 'last resort events' are quite frequent, forcing
sender to send all packets again, stalling the flow and wasting a lot of
resources.

This patch cleans only a part of the out of order queue in order
to meet the memory constraints.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Soheil Hassas Yeganeh <soheil@google.com>
Cc: C. Stephen Gun <csg@google.com>
Cc: Van Jacobson <vanj@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: bgmac: make it clear when setting interface type to RMII
Rafał Miłecki [Wed, 17 Aug 2016 21:11:52 +0000 (23:11 +0200)]
net: bgmac: make it clear when setting interface type to RMII

It doesn't really change anything as BGMAC_CHIPCTL_1_IF_TYPE_RMII is
equal to 0. It make code a bit clener, so far when reading it one could
think we forgot to set a proper mode. It also keeps this mode code in
sync with other ones.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: bgmac: support Ethernet core on BCM53573 SoCs
Rafał Miłecki [Wed, 17 Aug 2016 21:00:30 +0000 (23:00 +0200)]
net: bgmac: support Ethernet core on BCM53573 SoCs

BCM53573 is a new series of Broadcom's SoCs. It's based on ARM and can
be found in two packages (versions): BCM53573 and BCM47189. It shares
some code with the Northstar family, but also requires some new quirks.

First of all there can be up to 2 Ethernet cores on this SoC. If that is
the case, they are connected to two different switch ports allowing some
more complex/optimized setups. It seems the second unit doesn't come
fully configured and requires some IRQ quirk.

Other than that only the first core is connected to the PHY. For the
second one we have to register fixed PHY (similarly to the Northstar),
otherwise generic PHY driver would get some invalid info.

This has been successfully tested on Tenda AC9 (BCM47189B0).

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ethernet: nuvoton: fix spelling mistake: "aligment" -> "alignment"
Colin Ian King [Wed, 17 Aug 2016 17:31:35 +0000 (18:31 +0100)]
net: ethernet: nuvoton: fix spelling mistake: "aligment" -> "alignment"

trivial fix to spelling mistake in dev_err message

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoxen-netback: create a debugfs node for hash information
Paul Durrant [Wed, 17 Aug 2016 15:13:29 +0000 (16:13 +0100)]
xen-netback: create a debugfs node for hash information

It is useful to be able to see the hash configuration when running tests.
This patch adds a debugfs node for that purpose.

Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Cc: Wei Liu <wei.liu2@citrix.com>
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosfc: avoid division by zero
Edward Cree [Wed, 17 Aug 2016 15:00:10 +0000 (16:00 +0100)]
sfc: avoid division by zero

The division is already being done properly in efx_ef10_get_timer_config
which returns zero-on-success, unlike the old efx_ef10_get_sysclk_freq.

Fixes: d95e329a55ba ("sfc: get timer configuration from adapter")
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotcp: defer sacked assignment
Eric Dumazet [Wed, 17 Aug 2016 14:48:36 +0000 (07:48 -0700)]
tcp: defer sacked assignment

While chasing tcp_xmit_retransmit_queue() kasan issue, I found
that we could avoid reading sacked field of skb that we wont send,
possibly removing one cache line miss.

Very minor change in slow path, but why not ? ;)

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'bridge-vlan-stats-with-flags'
David S. Miller [Fri, 19 Aug 2016 06:18:42 +0000 (23:18 -0700)]
Merge branch 'bridge-vlan-stats-with-flags'

Nikolay Aleksandrov says:

====================
net: bridge: export vlan stats per-port with flags

This set adds the ability to export vlan stats per-port. Patch 01 makes
that possible by consolidating the bridge and port linkxstats calls. Then
patch 02 allows to dump the vlan entry flags in order to be able to
distinguish between bridge and port vlan entries when dumping the master
device vlan stats. That is needed because that call was implemented when
the stats API didn't have slave dumping capabilities and it dumps all vlan
stats (for both bridge and port entries). We also need it in order to print
the vlan flags when dumping the stats.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: bridge: export vlan flags with the stats
Nikolay Aleksandrov [Wed, 17 Aug 2016 10:53:10 +0000 (12:53 +0200)]
net: bridge: export vlan flags with the stats

Use one of the vlan xstats padding fields to export the vlan flags. This is
needed in order to be able to distinguish between master (bridge) and port
vlan entries in user-space when dumping the bridge vlan stats.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: bridge: consolidate bridge and port linkxstats calls
Nikolay Aleksandrov [Wed, 17 Aug 2016 10:53:09 +0000 (12:53 +0200)]
net: bridge: consolidate bridge and port linkxstats calls

In the bridge driver we usually have the same function working for both
port and bridge. In order to follow that logic and also avoid code
duplication, consolidate the bridge_ and brport_ linkxstats calls into
one since they share most of their code. As a side effect this allows us
to dump the vlan stats also via the slave call which is in preparation for
the upcoming per-port vlan stats and vlan flag dumping.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'flow-dissector-vlan-tag'
David S. Miller [Fri, 19 Aug 2016 06:13:14 +0000 (23:13 -0700)]
Merge branch 'flow-dissector-vlan-tag'

Hadar Hen Zion says:

====================
net_sched, flow_dissector, flower: Introduce vlan tag support

This patchset introduce vlan tag support to the flower classifier and the flow
dissector. In addition to adding vlan priority to act vlan.

The first 2 patches are dealing with flow-dissector:
 - The first patch is a fix, in case the vlan was already stripped from the
   skb, take it from skb->vlan_tci.
 - The second patch adds support for vlan priority.

The next 2 patches are dealing with flower:
 - The first patch is a fix, sets flow dissector 'used_keys' according to the
   mask value of each key.
 - The secound patch adds vlan tag support to the flower classifier, user space
   patches will be sent later to complete it.

The last patch adds vlan priority to act vlan since only vlan id is currently supported.

Changes from V1:
 - A new patch was added to this series "net_sched: flower: Avoid dissection of unmasked keys"
 - Adding u16 padding to struct flow_dissector_key_vlan
 - change flow_label field in struct flow_dissector_key_tags form 20 bits field to u32
 - Remove 'if (v->tcfv_push_prio)' check from tcf_vlan_dump function
 - Add support to un-stripped vlan skb and skb with multipale vlans in __skb_flow_dissect
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet_sched: act_vlan: Add priority option
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:14 +0000 (13:36 +0300)]
net_sched: act_vlan: Add priority option

The current vlan push action supports only vid and protocol options.
Add priority option.

Example script that adds vlan push action with vid and
priority:

tc filter add dev veth0 protocol ip parent ffff: \
   flower \
    indev veth0 \
   action vlan push id 100 priority 5

Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet_sched: flower: Add vlan support
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:13 +0000 (13:36 +0300)]
net_sched: flower: Add vlan support

Enhance flower to support 802.1Q vlan protocol classification.
Currently, the supported fields are vlan_id and vlan_priority.

Example:

# add a flower filter with vlan id and priority classification
tc filter add dev ens4f0 protocol 802.1Q parent ffff: \
flower \
indev ens4f0 \
vlan_ethtype ipv4 \
vlan_id 100 \
vlan_prio 3 \
action vlan pop

Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet_sched: flower: Avoid dissection of unmasked keys
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:12 +0000 (13:36 +0300)]
net_sched: flower: Avoid dissection of unmasked keys

The current flower implementation checks the mask range and set all the
keys included in that range as "used_keys", even if a specific key in
the range has a zero mask.

This behavior can cause a false positive return value of
dissector_uses_key function and unnecessary dissection in
__skb_flow_dissect.

This patch checks explicitly the mask of each key and "used_keys" will
be set accordingly.

Fixes: 77b9900ef53a ('tc: introduce Flower classifier')
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoflow_dissector: Get vlan priority in addition to vlan id
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:11 +0000 (13:36 +0300)]
flow_dissector: Get vlan priority in addition to vlan id

Add vlan priority check to the flow dissector by adding new flow
dissector struct, flow_dissector_key_vlan which includes vlan tag
fields.

vlan_id and flow_label fields were under the same struct
(flow_dissector_key_tags). It was a convenient setting since struct
flow_dissector_key_tags is used by struct flow_keys and by setting
vlan_id and flow_label under the same struct, we get precisely 24 or 48
bytes in flow_keys from flow_dissector_key_basic.

Now, when adding vlan priority support, the code will be cleaner if
flow_label and vlan tag won't be under the same struct anymore.

Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoflow_dissector: For stripped vlan, get vlan info from skb->vlan_tci
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:10 +0000 (13:36 +0300)]
flow_dissector: For stripped vlan, get vlan info from skb->vlan_tci

Early in the datapath skb_vlan_untag function is called, stripped
the vlan from the skb and set skb->vlan_tci and skb->vlan_proto fields.

The current dissection doesn't handle stripped vlan packets correctly.
In some flows, vlan doesn't exist in skb->data anymore when applying
flow dissection on the skb, fix that.

In case vlan info wasn't stripped before applying flow_dissector (RPS
flow for example), or in case of skb with multiple vlans (e.g. 802.1ad),
get the vlan info from skb->data. The flow_dissector correctly skips
any number of vlans and stores only the first level vlan.

Fixes: 0744dd00c1b1 ('net: introduce skb_flow_dissect()')
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'qed-link-fixes'
David S. Miller [Fri, 19 Aug 2016 06:04:40 +0000 (23:04 -0700)]
Merge branch 'qed-link-fixes'

Yuval Mintz says:

====================
qed*: Fix ethtool issues relating to link

This series addresses two issues that were introduced when adding
support for ethtool's link_ksettings support - the first fixes a
regression and second incorrect functionallity in the submission.

Although these are fixes, as the feature currently exists only in
'next-next' I'm aiming them for it.

Dave, please consider applying this series to 'net-next'.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqede: Fix forcing high speeds
Yuval Mintz [Fri, 19 Aug 2016 05:34:58 +0000 (08:34 +0300)]
qede: Fix forcing high speeds

While '0xdead' and '0xbeef' are "great" values, we should
use the correct SPEED_* values instead.

Fixes: 054c67d1c82a ("qed*: Add support for ethtool link_ksettings callbacks")
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed*: Fix pause setting
Yuval Mintz [Fri, 19 Aug 2016 05:34:57 +0000 (08:34 +0300)]
qed*: Fix pause setting

When moving into using ethtool's link_ksetting, qed started
supplying its own bitmask of speed/capabilities, but qede
is still checking for the SUPPORTED value to determine whether
it supports pause.

Fixes: 054c67d1c82a ("qed*: Add support for ethtool link_ksettings callbacks")
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Fri, 19 Aug 2016 06:02:50 +0000 (23:02 -0700)]
Merge branch '1GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
1GbE Intel Wired LAN Driver Updates 2016-08-18

This series contains updates to igb only.

Gangfeng Huang provides all the changes in the series to update the
igb driver to support advanced receive side filters that direct receive
packets by flows to different hardware queues. This enables a tight
control on routing a flow in the platform.  First patch allows for
receive network flow classification to insert and remove receive filters
by ethtool.  Second and third patches add the ability to insert and
remove ethertype and VLAN priority filters by ethtool.

Last patch just fixes an error message to return "Not supported" versus
"Unknown error 524".
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoixgbe: use atomic bitwise operations when handling reset requests
Emil Tantilov [Fri, 29 Jul 2016 21:46:31 +0000 (14:46 -0700)]
ixgbe: use atomic bitwise operations when handling reset requests

Use atomic bitwise operations when setting and checking reset
requests. This should help with possible races in the service task.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbevf: only check Tx queue enablement when debugging
Emil Tantilov [Fri, 29 Jul 2016 17:30:16 +0000 (10:30 -0700)]
ixgbevf: only check Tx queue enablement when debugging

Following a write the VFTXDCTL.ENABLE bit is set only when the Tx queue
is actually enabled, which may not happen during the configure phase even
if we waited for it. Make this check debug only since this is causing
confusion with users who notice the warning in dmesg.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbevf: change hw_dbg to use netdev_dbg
Emil Tantilov [Fri, 29 Jul 2016 17:30:11 +0000 (10:30 -0700)]
ixgbevf: change hw_dbg to use netdev_dbg

Instead of the home brewed macro make use of netdev_dbg same as
the ixgbe driver.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoixgbe: only check Tx queue enablement when debugging
Emil Tantilov [Fri, 29 Jul 2016 17:30:06 +0000 (10:30 -0700)]
ixgbe: only check Tx queue enablement when debugging

Following a write the TXDCTL.ENABLE bit is set only when the Tx queue
is actually enabled, which may not happen during the configure phase even
if we waited for it. Make this check debug only since this is causing
confusion with users who notice the warning in dmesg.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigb: fix error code in igb_add_ethtool_nfc_entry()
Gangfeng Huang [Wed, 6 Jul 2016 05:22:57 +0000 (13:22 +0800)]
igb: fix error code in igb_add_ethtool_nfc_entry()

Use error "rmgr: Cannot insert RX class rule: Operation not supported" is
more meaningful than "rmgr: Cannot insert RX class rule: Unknown error 524"

Signed-off-by: Gangfeng Huang <gangfeng.huang@ni.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigb: support RX flow classification by VLAN priority
Gangfeng Huang [Wed, 6 Jul 2016 05:22:56 +0000 (13:22 +0800)]
igb: support RX flow classification by VLAN priority

This patch is meant to allow for RX network flow classification to insert
and remove VLAN priority filter by ethtool

Example:
Add an VLAN priority filter:
$ ethtool -N eth0 flow-type ether vlan 0x6000 vlan-mask 0x1FFF action 2 loc 1

Show all filters:
$ ethtool -n eth0
4 RX rings available
Total 1 rules

Filter: 1
Flow Type: Raw Ethernet
Src MAC addr: 00:00:00:00:00:00 mask: FF:FF:FF:FF:FF:FF
Dest MAC addr: 00:00:00:00:00:00 mask: FF:FF:FF:FF:FF:FF
Ethertype: 0x0 mask: 0xFFFF
VLAN EtherType: 0x0 mask: 0xffff
VLAN: 0x6000 mask: 0x1fff
User-defined: 0x0 mask: 0xffffffffffffffff
Action: Direct to queue 2

Delete the filter by location:
$ ethtool -N delete 1

Signed-off-by: Ruhao Gao <ruhao.gao@ni.com>
Signed-off-by: Gangfeng Huang <gangfeng.huang@ni.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigb: support RX flow classification by ethertype
Gangfeng Huang [Wed, 6 Jul 2016 05:22:55 +0000 (13:22 +0800)]
igb: support RX flow classification by ethertype

This patch is meant to allow for RX network flow classification to insert
and remove ethertype filter by ethtool

Example:
Add an ethertype filter:
$ ethtool -N eth0 flow-type ether proto 0x88F8 action 2

Show all filters:
$ ethtool -n eth0
4 RX rings available
Total 1 rules

Filter: 15
Flow Type: Raw Ethernet
Src MAC addr: 00:00:00:00:00:00 mask: FF:FF:FF:FF:FF:FF
Dest MAC addr: 00:00:00:00:00:00 mask: FF:FF:FF:FF:FF:FF
Ethertype: 0x88F8 mask: 0x0
Action: Direct to queue 2

Delete the filter by location:
$ ethtool -N delete 15

Signed-off-by: Ruhao Gao <ruhao.gao@ni.com>
Signed-off-by: Gangfeng Huang <gangfeng.huang@ni.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigb: add support of RX network flow classification
Gangfeng Huang [Wed, 6 Jul 2016 05:22:54 +0000 (13:22 +0800)]
igb: add support of RX network flow classification

This patch is meant to allow for RX network flow classification to insert
and remove Rx filter by ethtool. Ethtool interface has it's own rules
manager

Show all filters:
$ ethtool -n eth0
4 RX rings available
Total 2 rules

Signed-off-by: Ruhao Gao <ruhao.gao@ni.com>
Signed-off-by: Gangfeng Huang <gangfeng.huang@ni.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoMerge branch 'qdisc-hash-fixes'
David S. Miller [Fri, 19 Aug 2016 04:19:09 +0000 (21:19 -0700)]
Merge branch 'qdisc-hash-fixes'

Jiri Kosina says:

====================
qdisc-hashtable fixes

The following two patches fix all the issues that have been reported
against the conversion of qdisc linked list to hashtable (currently in
net-next) so far.

First patch adjusts handling of singleton qdiscs to the new semantics, and
is rather straightforward.

The second patch, which fixes "cosmetic" issue of duplicate entries in the
qdisc dump for ingress qdiscs, is a little bit more hairy; I personally
would love to see all the already existing "if (ingress)"-like hacks go
away (by, let's say, introducing a general TCQ_F_? flag), but that's way
out of scope of this patchset (but already on my todo).

Thanks a lot to Daniel Borkmann and David Ahern for reporting the issues
and testing the patches promptly.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: sched: avoid duplicates in qdisc dump
Jiri Kosina [Tue, 16 Aug 2016 21:53:46 +0000 (23:53 +0200)]
net: sched: avoid duplicates in qdisc dump

tc_dump_qdisc() performs dumping of the per-device qdiscs in two phases;
first, the "standard" dev->qdisc is being dumped. Second, if there is/are
ingress queue(s), they are being dumped as well.

After conversion of netdevice's qdisc linked-list into hashtable, these
two sets are not in two disjunctive sets/lists any more, but are both
"reachable" directly from netdevice's hashtable. As a consequence, the
"full-depth" dump of the ingress qdiscs results in immediately hitting the
netdevice hashtable again, and duplicating the dump that has already been
performed for dev->qdisc.
What in fact needs to be dumped in case of ingress queue is "just" the
top-level ingress qdisc, as everything else has been dumped already.

Fix this by extending tc_dump_qdisc_root() in a way that it can be instructed
whether it should (while performing the "full" per-netdev qdisc dump) perform
the whole recursion, or just dump "additional" top-level (ingress) qdiscs
without performing any kind of recursion.

This fixes duplicate dumps such as

qdisc mq 0: root
qdisc pfifo_fast 0: parent :4 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: parent :3 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: parent :2 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: parent :1 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc clsact ffff: parent ffff:fff1
qdisc pfifo_fast 0: parent :4 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: parent :3 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: parent :2 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: parent :1 bands 3 priomap  1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1

Fixes: 59cc1f61f ("net: sched: convert qdisc linked list to hashtable")
Reported-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: sched: fix handling of singleton qdiscs with qdisc_hash
Jiri Kosina [Tue, 16 Aug 2016 21:52:58 +0000 (23:52 +0200)]
net: sched: fix handling of singleton qdiscs with qdisc_hash

qdisc_match_from_root() is now iterating over per-netdevice qdisc
hashtable instead of going through a linked-list of qdiscs (independently
on the actual underlying netdev), which was the case before the switch to
hashtable for qdiscs.

For singleton qdiscs, there is no underlying netdev associated though, and
therefore dumping a singleton qdisc will panic, as qdisc_dev(root) will
always be NULL.

 BUG: unable to handle kernel NULL pointer dereference at 0000000000000410
 IP: [<ffffffff8167efac>] qdisc_match_from_root+0x2c/0x70
 PGD 1aceba067 PUD 1aceb7067 PMD 0
 Oops: 0000 [#1] PREEMPT SMP
[ ... ]
 task: ffff8801ec996e00 task.stack: ffff8801ec934000
 RIP: 0010:[<ffffffff8167efac>]  [<ffffffff8167efac>] qdisc_match_from_root+0x2c/0x70
 RSP: 0018:ffff8801ec937ab0  EFLAGS: 00010203
 RAX: 0000000000000408 RBX: ffff88025e612000 RCX: ffffffffffffffd8
 RDX: 0000000000000000 RSI: 00000000ffff0000 RDI: ffffffff81cf8100
 RBP: ffff8801ec937ab0 R08: 000000000001c160 R09: ffff8802668032c0
 R10: ffffffff81cf8100 R11: 0000000000000030 R12: 00000000ffff0000
 R13: ffff88025e612000 R14: ffffffff81cf3140 R15: 0000000000000000
 FS:  00007f24b9af6740(0000) GS:ffff88026f280000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 0000000000000410 CR3: 00000001aceec000 CR4: 00000000001406e0
 Stack:
  ffff8801ec937ad0 ffffffff81681210 ffff88025dd51a00 00000000fffffff1
  ffff8801ec937b88 ffffffff81681e4e ffffffff81c42bc0 ffff880262431500
  ffffffff81cf3140 ffff88025dd51a10 ffff88025dd51a24 00000000ec937b38
 Call Trace:
  [<ffffffff81681210>] qdisc_lookup+0x40/0x50
  [<ffffffff81681e4e>] tc_modify_qdisc+0x21e/0x550
  [<ffffffff8166ae25>] rtnetlink_rcv_msg+0x95/0x220
  [<ffffffff81209602>] ? __kmalloc_track_caller+0x172/0x230
  [<ffffffff8166ad90>] ? rtnl_newlink+0x870/0x870
  [<ffffffff816897b7>] netlink_rcv_skb+0xa7/0xc0
  [<ffffffff816657c8>] rtnetlink_rcv+0x28/0x30
  [<ffffffff8168919b>] netlink_unicast+0x15b/0x210
  [<ffffffff81689569>] netlink_sendmsg+0x319/0x390
  [<ffffffff816379f8>] sock_sendmsg+0x38/0x50
  [<ffffffff81638296>] ___sys_sendmsg+0x256/0x260
  [<ffffffff811b1275>] ? __pagevec_lru_add_fn+0x135/0x280
  [<ffffffff811b1a90>] ? pagevec_lru_move_fn+0xd0/0xf0
  [<ffffffff811b1140>] ? trace_event_raw_event_mm_lru_insertion+0x180/0x180
  [<ffffffff811b1b85>] ? __lru_cache_add+0x75/0xb0
  [<ffffffff817708a6>] ? _raw_spin_unlock+0x16/0x40
  [<ffffffff811d8dff>] ? handle_mm_fault+0x39f/0x1160
  [<ffffffff81638b15>] __sys_sendmsg+0x45/0x80
  [<ffffffff81638b62>] SyS_sendmsg+0x12/0x20
  [<ffffffff810038e7>] do_syscall_64+0x57/0xb0

Fix this by special-casing singleton qdiscs (those that don't have
underlying netdevice) and introduce immediate handling of those rather
than trying to go over an underlying netdevice. We're in the same
situation in tc_dump_qdisc_root() and tc_dump_tclass_root().

Ultimately, this will have to be slightly reworked so that we are actually
able to show singleton qdiscs (noop) in the dump properly; but we're not
currently doing that anyway, so no regression there, and better do this in
a gradual manner.

Fixes: 59cc1f61f ("net: sched: convert qdisc linked list to hashtable")
Reported-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Daniel Borkmann <daniel@iogearbox.net>
Reported-by: David Ahern <dsa@cumulusnetworks.com>
Tested-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'tipc-next'
David S. Miller [Fri, 19 Aug 2016 04:14:37 +0000 (21:14 -0700)]
Merge branch 'tipc-next'

Jon Maloy says:

====================
tipc: bearer and link improvements

The first commit makes it possible to set and check the 'blocked' state
of a bearer from the generic bearer layer. The second commit is a small
improvement to the link congestion mechanism.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotipc: ensure that link congestion and wakeup use same criteria
Jon Paul Maloy [Tue, 16 Aug 2016 15:53:51 +0000 (11:53 -0400)]
tipc: ensure that link congestion and wakeup use same criteria

When a link is attempted woken up after congestion, it uses a different,
more generous criteria than when it was originally declared congested.
This has the effect that the link, and the sending process, sometimes
will be woken up unnecessarily, just to immediately return to congestion
when it turns out there is not not enough space in its send queue to
host the pending message. This is a waste of CPU cycles.

We now change the function link_prepare_wakeup() to use exactly the same
criteria as tipc_link_xmit(). However, since we are now excluding the
window limit from the wakeup calculation, and the current backlog limit
for the lowest level is too small to house even a single maximum-size
message, we have to expand this limit. We do this by evaluating an
alternative, minimum value during the setting of the importance limits.

Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotipc: make bearer packet filtering generic
Jon Paul Maloy [Tue, 16 Aug 2016 15:53:50 +0000 (11:53 -0400)]
tipc: make bearer packet filtering generic

In commit 5b7066c3dd24 ("tipc: stricter filtering of packets in bearer
layer") we introduced a method of filtering out messages while a bearer
is being reset, to avoid that links may be re-created and come back in
working state while we are still in the process of shutting them down.

This solution works well, but is limited to only work with L2 media, which
is insufficient with the increasing use of UDP as carrier media.

We now replace this solution with a more generic one, by introducing a
new flag "up" in the generic struct tipc_bearer. This field will be set
and reset at the same locations as with the previous solution, while
the packet filtering is moved to the generic code for the sending side.
On the receiving side, the filtering is still done in media specific
code, but now including the UDP bearer.

Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'qed-next'
David S. Miller [Fri, 19 Aug 2016 04:11:46 +0000 (21:11 -0700)]
Merge branch 'qed-next'

Sudarsana Reddy Kalluru says:

====================
qed*: Add support for additional statistics.

The patch series adds qed/qede support for new statistics.
Patch (1) adds couple of statistcs for "ethtool -S" display.
Patch (2) adds support for per-queue statistics to ethtool display.
Patch (3) adds qed support for NCSI statistics.

Please consider applying this to 'net-next' branch.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed: Add support for NCSI statistics.
Sudarsana Reddy Kalluru [Tue, 16 Aug 2016 14:51:03 +0000 (10:51 -0400)]
qed: Add support for NCSI statistics.

The patch adds driver support for sending the NCSI statistics to the
MFW. This is an asynchronous request from MFW. Upon receiving this, driver
populates the required data and send it to MFW.

Signed-off-by: Sudarsana Reddy 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 per-queue stats.
Sudarsana Reddy Kalluru [Tue, 16 Aug 2016 14:51:02 +0000 (10:51 -0400)]
qede: Add support for per-queue stats.

Signed-off-by: Sudarsana Reddy 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 capturing additional stats in ethtool-stats display.
Sudarsana Reddy Kalluru [Tue, 16 Aug 2016 14:51:01 +0000 (10:51 -0400)]
qede: Add support for capturing additional stats in ethtool-stats display.

The patch adds driver support for capturing stats ttl0_discard and
packet_too_big_discard in "ethtool -S" display.

Signed-off-by: Sudarsana Reddy 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 agonet: atm: remove redundant null pointer check on dev->name
Colin Ian King [Tue, 16 Aug 2016 11:08:42 +0000 (12:08 +0100)]
net: atm: remove redundant null pointer check on dev->name

dev->name is a char array of IFNAMSIZ elements, hence can never be
null, so the null pointer check is redundant. Remove it.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: phy: Update copyright info
Appana Durga Kedareswara Rao [Tue, 16 Aug 2016 06:28:29 +0000 (11:58 +0530)]
net: phy: Update copyright info

For implementing this driver most of the inputs is
provided by Andrew Lunn.

Updating the driver with Andrew Copy right.

Signed-off-by: Kedareswara rao Appana <appanad@xilinx.com>
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ethernet: macb: Add support for rx_clk
shubhrajyoti.datta@xilinx.com [Tue, 16 Aug 2016 04:44:50 +0000 (10:14 +0530)]
net: ethernet: macb: Add support for rx_clk

Some of the platforms like zynqmp ultrascale+ has a
separate clock gate for the rx clock. Add an optional
rx_clk so that the clock can be enabled.

Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Fri, 19 Aug 2016 03:45:14 +0000 (20:45 -0700)]
Merge branch '40GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
40GbE Intel Wired LAN Driver Updates 2016-08-18

This series contains updates to i40e and i40evf only.

Wei Yongjun updates i40e to use list_move() instead of list_del() &
list_add() operations.

Anjali fixes an issue where the client->open call was not protected with
the client instance mutex, which allowed client->close to be called before
the open all completed.

Catherine makes sure that the VLAN count (and stats) gets reset to 0
after reset.

Jake provides two patches, first adds the needed rtnl lock around
i40evf_set_interrupt_capability() since i40evf_init_task() does not
hold the rtnl_lock.  Second fixes an issue where users could reduce
the number of channels (queues) below the current flow director
filter rules targets.

Dave fixes a problem where a static analysis tool generates a warning
so eliminating the irrelevant check and redundant assignment for the
value of enabled_tc.

Avinash fixes an sync issue where the iWARP device open is called
before the PCI register writes are completed, so ensure the register
writes complete before exiting the setup function.

Alan fixes a bug which causes RSS to continue to work after being
disabled.

Carolyn implements a feature change which allows using ethtool to set
RDD hash options using less than four parameters if desired.

Dan Carpenter cleans up a stray unlock.

Sridhar exposes the "trust" flag to userspace via ndo_get_vf_config().
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoi40e: Expose 'trust' flag to userspace via ndo_get_vf_config.
Sridhar Samudrala [Thu, 4 Aug 2016 16:45:47 +0000 (18:45 +0200)]
i40e: Expose 'trust' flag to userspace via ndo_get_vf_config.

This enables
ip -d l
to indicate if trust is on or off for VFs.

Signed-off-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: remove a stray unlock
Dan Carpenter [Wed, 3 Aug 2016 19:01:07 +0000 (22:01 +0300)]
i40e: remove a stray unlock

We shifted the locking around a bit but forgot to delete this unlock so
now it can unlock twice.

Fixes: cd3be169a5ff ('i40e: Move the mutex lock in i40e_client_unregister')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf-Bump version from 1.6.11 to 1.6.12
Bimmy Pujari [Wed, 27 Jul 2016 19:02:41 +0000 (12:02 -0700)]
i40e/i40evf-Bump version from 1.6.11 to 1.6.12

Signed-off-by: Bimmy Pujari <Bimmy.pujari@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Allow RSS Hash set with less than four parameters
Carolyn Wyborny [Wed, 27 Jul 2016 19:02:40 +0000 (12:02 -0700)]
i40e: Allow RSS Hash set with less than four parameters

This patch implements a feature change which allows using ethtool to set
RSS hash opts using less than four parameters if desired.

Change-ID: I0fbb91255d81e997c456697c21ac39cc9754821b
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: fix memory leak
Mitch Williams [Wed, 27 Jul 2016 19:02:39 +0000 (12:02 -0700)]
i40e: fix memory leak

When we allocate memory, we must free it. It's simple courtesy.

Change-ID: Id007294096fb53344f1a8b9a0f78eddf9853c5d6
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: fix lookup table when RSS disabled/enabled
Alan Brady [Wed, 27 Jul 2016 19:02:38 +0000 (12:02 -0700)]
i40e: fix lookup table when RSS disabled/enabled

This patch fixes the bug which causes RSS to continue to work
after being disabled.  After disabling RSS, traffic would continue
to be assigned to different queues instead of falling back to a
single queue. Without this patch, attempting to disable RSS would
not work as expected. This patch fixes the bug by clearing the
lookup table used by RSS such that all traffic is assigned to a
single queue.  This patch also addresses the issue of reinstating
 the lookup table should RSS then be re-enabled.

Change-ID: Ib20c7c6a7e9f1f772bb787370f8a8c664796b141
Signed-off-by: Alan Brady <alan.brady@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Don't notify client of VF reset during VF creation
Avinash Dayanand [Wed, 27 Jul 2016 19:02:37 +0000 (12:02 -0700)]
i40e: Don't notify client of VF reset during VF creation

VF goes through reset path during VF creation which happens to also
have notification of VF reset to client. Adding conditional check to
avoid wrongly notifying VF reset during VF creation.

Also changing the call order of VF enable, calling it after VF creation
rather than before.

Change-ID: I96eabd99deae746a2f0fc465194c886f196178ce
Signed-off-by: Avinash Dayanand <avinash.dayanand@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Force register writes to mitigate sync issues with iwarp VF driver
Avinash Dayanand [Wed, 27 Jul 2016 19:02:36 +0000 (12:02 -0700)]
i40e: Force register writes to mitigate sync issues with iwarp VF driver

This patch is a fix for the bug i.e. unable to create iwarp device
in VF. This is a sync issue and the iwarp device open is called even
before the PCI register writes are done.

Forcing the PCI register writes to happen just before it exits the
function.

Change-ID: I60c6a2c709da89e845f2764cc50ce8b7373c8c44
Signed-off-by: Avinash Dayanand <avinash.dayanand@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: don't allow reduction of channels below active FD rules
Jacob Keller [Wed, 27 Jul 2016 19:02:35 +0000 (12:02 -0700)]
i40e: don't allow reduction of channels below active FD rules

If a driver is unable to maintain all current user supplied settings
from ethtool (or other sources), it is not ok for a user request to
succeed and silently trample over previous configuration.

To that end, if you change the number of channels, it must not be
allowed to reduce the number of channels (queues) below the current
flow director filter rules targets. In this case, return -EINVAL when
a request to reduce the number of channels would do so. In addition
log a warning to the kernel buffer explaining why we failed, and report
the rules which prevent us from lowering the number of channels.

Change-ID: If41464d63d7aab11cedf09e4f3aa1a69e21ffd88
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Fix static analysis tool warning
Dave Ertman [Wed, 27 Jul 2016 19:02:34 +0000 (12:02 -0700)]
i40e: Fix static analysis tool warning

This patch fixes a problem where a static analysis tool generates
a warning for "INVARIANT_CONDITION: Expression 'enabled_tc' used
in the condition always yields the same result."

Without this patch, the driver will not pass the static analysis
tool checks without generating warnings.

This patch fixes the problem by eliminating the irrelevant check
and redundant assignment for the value of enabled_tc.

Change-ID: Ia7d44cb050f507df7de333e96369d322e08bf408
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>