platform/kernel/linux-rpi3.git
8 years agonet: snmp: kill various STATS_USER() helpers
Eric Dumazet [Wed, 27 Apr 2016 23:44:27 +0000 (16:44 -0700)]
net: snmp: kill various STATS_USER() helpers

In the old days (before linux-3.0), SNMP counters were duplicated,
one for user context, and one for BH context.

After commit 8f0ea0fe3a03 ("snmp: reduce percpu needs by 50%")
we have a single copy, and what really matters is preemption being
enabled or disabled, since we use this_cpu_inc() or __this_cpu_inc()
respectively.

We therefore kill SNMP_INC_STATS_USER(), SNMP_ADD_STATS_USER(),
NET_INC_STATS_USER(), NET_ADD_STATS_USER(), SCTP_INC_STATS_USER(),
SNMP_INC_STATS64_USER(), SNMP_ADD_STATS64_USER(), TCP_ADD_STATS_USER(),
UDP_INC_STATS_USER(), UDP6_INC_STATS_USER(), and XFRM_INC_STATS_USER()

Following patches will rename __BH helpers to make clear their
usage is not tied to BH being disabled.

Signed-off-by: Eric Dumazet <edumazet@google.com>
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 [Thu, 28 Apr 2016 01:59:08 +0000 (21:59 -0400)]
Merge branch '40GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
40GbE Intel Wired LAN Driver Updates 2016-04-27

This series contains updates to i40e and i40evf.

Alex Duyck cleans up the feature flags since they are becoming pretty
"massive", the primary change being that we now build our features list
around hw_encap_features.  Added support for IPIP and SIT offloads,
which should improvement in throughput for IPIP and SIT tunnels with
the offload enabled.

Mitch adds support for configuring RSS on behalf of the VFs, which removes
the burden of dealing with different hardware interfaces from the VF
drivers and improves future compatibility.  Fix to ensure that we do not
panic by checking that the vsi_res pointer is valid before dereferencing
it, after which we can drink beer and eat peanuts.

Shannon does come housekeeping in i40e_add_fdir_ethtool() in preparation
for more cloud filter work.  Added flexibility to the nvmupdate
facility by adding the ability to specify an AQ event opcode to wait on
after Exec_AQ request.

Michal adds device capability which defines if an update is available and
if a security check is needed during the update process.

Kamil just adds a device id to support X722 QSFP+ device.

Greg fixes an issue where a mirror rule ID may be zero, so do not return
invalid parameter when the user passes in a zero for a rule ID.  Adds
support to steer packets to VSIs by VLAN tag alone while being in
promiscuous mode for multicast and unicast MAC addresses.

Jesse fixes the driver from offloading the VLAN tag into the skb any
time there was a VLAN tag and the hardware stripping was enabled, to
making sure it is enabled before put_tag.

v2: Dropped patch 8 ("i40e: Allow user to change input set mask for flow
    director") while Kiran reworks a more generalized solution based
    on feedback from David Miller.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet-rfs: fix false sharing accessing sd->input_queue_head
Eric Dumazet [Tue, 26 Apr 2016 22:30:07 +0000 (15:30 -0700)]
net-rfs: fix false sharing accessing sd->input_queue_head

sd->input_queue_head is incremented for each processed packet
in process_backlog(), and read from other cpus performing
Out Of Order avoidance in get_rps_cpu()

Moving this field in a separate cache line keeps it mostly
hot for the cpu in process_backlog(), as other cpus will
only read it.

In a stress test, process_backlog() was consuming 6.80 % of cpu cycles,
and the patch reduced the cost to 0.65 %

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: w5100: support W5500
Akinobu Mita [Tue, 26 Apr 2016 20:43:48 +0000 (05:43 +0900)]
net: w5100: support W5500

This adds support for W5500 chip.

W5500 has similar register and memory organization with W5100 and W5200.
There are a few important differences listed below but it is still
possible to share common code with W5100 and W5200.

* W5500 register and memory are organized by multiple blocks.  Each one
is selected by 16bits offset address and 5bits block select bits.

But the existing register access operations take u16 address.  This change
extends the addess by u32 address and put offset address to lower 16bits
and block select bits to upper 16bits.

This change also adds the offset addresses for socket register and TX/RX
memory blocks to the driver private data structure in order to reduce
conditional switches for each chip.

* W5500 has the different register offset for socket interrupt mask
register.  Newly added internal functions w5100_enable_intr() and
w5100_disable_intr() take care of the diffrence.

* W5500 has the different register offset for retry time-value register.
But this register is only used to verify that the reset value is correctly
read at initialization.  So move the verification to w5100_hw_reset()
which already does different things for different chips.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Mike Sinkovsky <msink@permonline.ru>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoi40evf: Add driver support for promiscuous mode
Anjali Singhai Jain [Tue, 12 Apr 2016 15:30:52 +0000 (08:30 -0700)]
i40evf: Add driver support for promiscuous mode

Add necessary Linux Ethernet driver support for promiscuous mode
operation. Add a flag so the VF knows it is in promiscuous mode
and two state flags to discreetly track multicast and unicast
promiscuous states.

Change-Id: Ib2f2dc7a7582304fec90fc917ebb7ded21ba1de4
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
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: Add VF promiscuous mode driver support
Anjali Singhai Jain [Tue, 12 Apr 2016 15:30:51 +0000 (08:30 -0700)]
i40e: Add VF promiscuous mode driver support

Add infrastructure for Network Function Virtualization VLAN tagged
packet steering feature.

Change-Id: I9b873d8fcc253858e6baba65ac68ec5b9363944e
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Add promiscuous on VLAN support
Greg Rose [Tue, 12 Apr 2016 15:30:50 +0000 (08:30 -0700)]
i40e: Add promiscuous on VLAN support

NFV use cases require the ability to steer packets to VSIs by VLAN tag
alone while being in promiscuous mode for multicast and unicast MAC
addresses.  These two new functions support that ability.

Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e/i40evf: Only offload VLAN tag if enabled
Jesse Brandeburg [Tue, 12 Apr 2016 15:30:49 +0000 (08:30 -0700)]
i40e/i40evf: Only offload VLAN tag if enabled

The driver was offloading the VLAN tag into the skb
any time there was a VLAN tag and the hardware stripping was
enabled.  Just check to make sure it's enabled before put_tag.

Change-Id: Ife95290c06edd9a616393b38679923938b382241
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: Remove zero check
Greg Rose [Tue, 12 Apr 2016 15:30:48 +0000 (08:30 -0700)]
i40e: Remove zero check

A mirror rule ID may be zero so do not return invalid parameter when the
user passes in a zero value for a rule ID.

Change-ID: I261b8c24725ce2c6ed32f859da81093dfcbe2970
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Add DeviceID for X722 QSFP+
Kamil Krawczyk [Tue, 12 Apr 2016 15:30:47 +0000 (08:30 -0700)]
i40e: Add DeviceID for X722 QSFP+

Change-ID: I1370fbc7774e815ac1ad56561e97488e829592fc
Signed-off-by: Kamil Krawczyk <kamil.krawczyk@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoi40e: Add device capability which defines if update is available
Michal Kosiarz [Tue, 12 Apr 2016 15:30:46 +0000 (08:30 -0700)]
i40e: Add device capability which defines if update is available

Add device capability which defines if update is available and security
check is needed during update process.

Change-ID: I380787c878275e1df18b39198df3ee3666342282
Signed-off-by: Michal Kosiarz <michal.kosiarz@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Wed, 27 Apr 2016 19:43:10 +0000 (15:43 -0400)]
Merge git://git./linux/kernel/git/davem/net

Minor overlapping changes in the conflicts.

In the macsec case, the change of the default ID macro
name overlapped with the 64-bit netlink attribute alignment
fixes in net-next.

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipv6: Use passed in table for nexthop lookups
David Ahern [Mon, 25 Apr 2016 04:26:04 +0000 (21:26 -0700)]
net: ipv6: Use passed in table for nexthop lookups

Similar to 3bfd847203c6 ("net: Use passed in table for nexthop lookups")
for IPv4, if the route spec contains a table id use that to lookup the
next hop first and fall back to a full lookup if it fails (per the fix
4c9bcd117918b ("net: Fix nexthop lookups")).

Example:

    root@kenny:~# ip -6 ro ls table red
    local 2100:1::1 dev lo  proto none  metric 0  pref medium
    2100:1::/120 dev eth1  proto kernel  metric 256  pref medium
    local 2100:2::1 dev lo  proto none  metric 0  pref medium
    2100:2::/120 dev eth2  proto kernel  metric 256  pref medium
    local fe80::e0:f9ff:fe09:3cac dev lo  proto none  metric 0  pref medium
    local fe80::e0:f9ff:fe1c:b974 dev lo  proto none  metric 0  pref medium
    fe80::/64 dev eth1  proto kernel  metric 256  pref medium
    fe80::/64 dev eth2  proto kernel  metric 256  pref medium
    ff00::/8 dev red  metric 256  pref medium
    ff00::/8 dev eth1  metric 256  pref medium
    ff00::/8 dev eth2  metric 256  pref medium
    unreachable default dev lo  metric 240  error -113 pref medium

    root@kenny:~# ip -6 ro add table red 2100:3::/64 via 2100:1::64
    RTNETLINK answers: No route to host

Route add fails even though 2100:1::64 is a reachable next hop:
    root@kenny:~# ping6 -I red  2100:1::64
    ping6: Warning: source address might be selected on device other than red.
    PING 2100:1::64(2100:1::64) from 2100:1::1 red: 56 data bytes
    64 bytes from 2100:1::64: icmp_seq=1 ttl=64 time=1.33 ms

With this patch:
    root@kenny:~# ip -6 ro add table red 2100:3::/64 via 2100:1::64
    root@kenny:~# ip -6 ro ls table red
    local 2100:1::1 dev lo  proto none  metric 0  pref medium
    2100:1::/120 dev eth1  proto kernel  metric 256  pref medium
    local 2100:2::1 dev lo  proto none  metric 0  pref medium
    2100:2::/120 dev eth2  proto kernel  metric 256  pref medium
    2100:3::/64 via 2100:1::64 dev eth1  metric 1024  pref medium
    local fe80::e0:f9ff:fe09:3cac dev lo  proto none  metric 0  pref medium
    local fe80::e0:f9ff:fe1c:b974 dev lo  proto none  metric 0  pref medium
    fe80::/64 dev eth1  proto kernel  metric 256  pref medium
    fe80::/64 dev eth2  proto kernel  metric 256  pref medium
    ff00::/8 dev red  metric 256  pref medium
    ff00::/8 dev eth1  metric 256  pref medium
    ff00::/8 dev eth2  metric 256  pref medium
    unreachable default dev lo  metric 240  error -113 pref medium

Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agotaskstats: fix nl parsing in accounting/getdelays.c
Nicolas Dichtel [Wed, 27 Apr 2016 15:53:08 +0000 (17:53 +0200)]
taskstats: fix nl parsing in accounting/getdelays.c

The type TASKSTATS_TYPE_NULL should always be ignored.

When jumping to the next attribute, only the length of the current
attribute should be added, not the length of all nested attributes.
This last bug was not visible before commit 80df554275c2, because the
kernel didn't put more than two nested attributes.

Fixes: a3baf649ca9c ("[PATCH] per-task-delay-accounting: documentation")
Fixes: 80df554275c2 ("taskstats: use the libnl API to align nlattr on 64-bit")
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Tue, 26 Apr 2016 23:25:51 +0000 (16:25 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Handle v4/v6 mixed sockets properly in soreuseport, from Craig
    Gallak.

 2) Bug fixes for the new macsec facility (missing kmalloc NULL checks,
    missing locking around netdev list traversal, etc.) from Sabrina
    Dubroca.

 3) Fix handling of host routes on ifdown in ipv6, from David Ahern.

 4) Fix double-fdput in bpf verifier.  From Jann Horn.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (31 commits)
  bpf: fix double-fdput in replace_map_fd_with_map_ptr()
  net: ipv6: Delete host routes on an ifdown
  Revert "ipv6: Revert optional address flusing on ifdown."
  net/mlx4_en: fix spurious timestamping callbacks
  net: dummy: remove note about being Y by default
  cxgbi: fix uninitialized flowi6
  ipv6: Revert optional address flusing on ifdown.
  ipv4/fib: don't warn when primary address is missing if in_dev is dead
  net/mlx5: Add pci shutdown callback
  net/mlx5_core: Remove static from local variable
  net/mlx5e: Use vport MTU rather than physical port MTU
  net/mlx5e: Fix minimum MTU
  net/mlx5e: Device's mtu field is u16 and not int
  net/mlx5_core: Add ConnectX-5 to list of supported devices
  net/mlx5e: Fix MLX5E_100BASE_T define
  net/mlx5_core: Fix soft lockup in steering error flow
  qlcnic: Update version to 5.3.64
  net: stmmac: socfpga: Remove re-registration of reset controller
  macsec: fix netlink attribute validation
  macsec: add missing macsec prefix in uapi
  ...

8 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Tue, 26 Apr 2016 23:17:01 +0000 (16:17 -0700)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Arnd Bergmann:
 "Here are the latest bug fixes for ARM SoCs, mostly addressing recent
  regressions.  Changes are across several platforms, so I'm listing
  every change separately here.

  Regressions since 4.5:

   - A correction of the psci firmware DT binding, to prevent users from
     relying on unintended semantics

   - Actually getting the newly merged clock driver for some OMAP
     platforms to work

   - A revert of patches for the Qualcomm BAM, these need to be reworked
     for 4.7 to avoid breaking boards other than the one they were
     intended for

   - A correction for the I2C device nodes on the Socionext Uniphier
     platform

   - i.MX SDHCI was broken for non-DT platforms due to a change with the
     setting of the DMA mask

   - A revert of a patch that accidentally added a nonexisting clock on
     the Rensas "Porter" board

   - A couple of OMAP fixes that are all related to suspend after the
     power domain changes for dra7

   - On Mediatek, revert part of the power domain initialization changes
     that broke mt8173-evb

  Fixes for older bugs:

   - Workaround for an "external abort" in the omap34xx suspend/resume
     code.

   - The USB1/eSATA should not be listed as an excon device on
     am57xx-beagle-x15 (broken since v4.0)

   - A v4.5 regression in the TI AM33xx and AM43XX DT specifying
     incorrect DMA request lines for the GPMC

   - The jiffies calibration on Renesas platforms was incorrect for some
     modern CPU cores.

   - A hardware errata woraround for clockdomains on TI DRA7"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  drivers: firmware: psci: unify enable-method binding on ARM {64,32}-bit systems
  arm64: dts: uniphier: fix I2C nodes of PH1-LD20
  ARM: shmobile: timer: Fix preset_lpj leading to too short delays
  Revert "ARM: dts: porter: Enable SCIF_CLK frequency and pins"
  ARM: dts: r8a7791: Don't disable referenced optional clocks
  Revert "ARM: OMAP: Catch callers of revision information prior to it being populated"
  ARM: OMAP3: Fix external abort on 36xx waking from off mode idle
  ARM: dts: am57xx-beagle-x15: remove extcon_usb1
  ARM: dts: am437x: Fix GPMC dma properties
  ARM: dts: am33xx: Fix GPMC dma properties
  Revert "soc: mediatek: SCPSYS: Fix double enabling of regulators"
  ARM: mach-imx: sdhci-esdhc-imx: initialize DMA mask
  ARM: DRA7: clockdomain: Implement timer workaround for errata i874
  ARM: OMAP: Catch callers of revision information prior to it being populated
  ARM: dts: dra7: Correct clock tree for sys_32k_ck
  ARM: OMAP: DRA7: Provide proper class to omap2_set_globals_tap
  ARM: OMAP: DRA7: wakeupgen: Skip SAR save for wakeupgen
  Revert "dts: msm8974: Add dma channels for blsp2_i2c1 node"
  Revert "dts: msm8974: Add blsp2_bam dma node"
  ARM: dts: Add clocks for dm814x ADPLL

8 years agodevpts: more pty driver interface cleanups
Linus Torvalds [Tue, 26 Apr 2016 03:04:08 +0000 (20:04 -0700)]
devpts: more pty driver interface cleanups

This is more prep-work for the upcoming pty changes.  Still just code
cleanup with no actual semantic changes.

This removes a bunch pointless complexity by just having the slave pty
side remember the dentry associated with the devpts slave rather than
the inode.  That allows us to remove all the "look up the dentry" code
for when we want to remove it again.

Together with moving the tty pointer from "inode->i_private" to
"dentry->d_fsdata" and getting rid of pointless inode locking, this
removes about 30 lines of code.  Not only is the end result smaller,
it's simpler and easier to understand.

The old code, for example, depended on the d_find_alias() to not just
find the dentry, but also to check that it is still hashed, which in
turn validated the tty pointer in the inode.

That is a _very_ roundabout way to say "invalidate the cached tty
pointer when the dentry is removed".

The new code just does

dentry->d_fsdata = NULL;

in devpts_pty_kill() instead, invalidating the tty pointer rather more
directly and obviously.  Don't do something complex and subtle when the
obvious straightforward approach will do.

The rest of the patch (ie apart from code deletion and the above tty
pointer clearing) is just switching the calling convention to pass the
dentry or file pointer around instead of the inode.

Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Peter Anvin <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Peter Hurley <peter@hurleysoftware.com>
Cc: Serge Hallyn <serge.hallyn@ubuntu.com>
Cc: Willy Tarreau <w@1wt.eu>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: Alan Cox <gnomes@lxorguk.ukuu.org.uk>
Cc: Jann Horn <jann@thejh.net>
Cc: Greg KH <greg@kroah.com>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
8 years agobpf: fix double-fdput in replace_map_fd_with_map_ptr()
Jann Horn [Tue, 26 Apr 2016 20:26:26 +0000 (22:26 +0200)]
bpf: fix double-fdput in replace_map_fd_with_map_ptr()

When bpf(BPF_PROG_LOAD, ...) was invoked with a BPF program whose bytecode
references a non-map file descriptor as a map file descriptor, the error
handling code called fdput() twice instead of once (in __bpf_map_get() and
in replace_map_fd_with_map_ptr()). If the file descriptor table of the
current task is shared, this causes f_count to be decremented too much,
allowing the struct file to be freed while it is still in use
(use-after-free). This can be exploited to gain root privileges by an
unprivileged user.

This bug was introduced in
commit 0246e64d9a5f ("bpf: handle pseudo BPF_LD_IMM64 insn"), but is only
exploitable since
commit 1be7f75d1668 ("bpf: enable non-root eBPF programs") because
previously, CAP_SYS_ADMIN was required to reach the vulnerable code.

(posted publicly according to request by maintainer)

Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agopch_gbe: fix bogus trylock conversion
Florian Westphal [Tue, 26 Apr 2016 21:14:30 +0000 (23:14 +0200)]
pch_gbe: fix bogus trylock conversion

Should have converted 'if (trylock)' to 'lock'.

Fixes: a6086a893718db ("drivers: net: remove NETDEV_TX_LOCKED")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'sh_eth-next'
David S. Miller [Tue, 26 Apr 2016 20:07:21 +0000 (16:07 -0400)]
Merge branch 'sh_eth-next'

Sergei Shtylyov says:

====================
sh_eth: couple of software reset bit cleanups

   Here's a set of 2 patches against DaveM's 'net-next.git' repo. We clean up
the use of the software reset bits...
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: rename ARSTR register bit
Sergei Shtylyov [Sun, 24 Apr 2016 20:46:15 +0000 (23:46 +0300)]
sh_eth: rename ARSTR register bit

The Renesas RZ/A1H manual names the software reset bit in the software reset
register (ARSTR) ARST which makes a bit more sense than the ARSTR_ARSTR name
used now  by the driver -- rename the latter to ARSTR_ARST.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: use EDMR_SRST_GETHER in sh_eth_check_reset()
Sergei Shtylyov [Sun, 24 Apr 2016 20:45:23 +0000 (23:45 +0300)]
sh_eth: use EDMR_SRST_GETHER in sh_eth_check_reset()

sh_eth_check_reset() uses a bare number where EDMR_SRST_GETHER would fit,
i.e. the receive/trasmit software reset bits that comprise EDMR_SRST_GETHER
read as 1  while the corresponding reset is in progress and thus, when both
are 0, the reset is complete.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'mlx5-next'
David S. Miller [Tue, 26 Apr 2016 19:58:03 +0000 (15:58 -0400)]
Merge branch 'mlx5-next'

Saeed Mahameed says:

====================
Mellanox 100G extending mlx5 ethtool support

Changes from V0:
- Dropped: net/mlx5e: Disable link up on INIT HCA command
  Due to Ido's and Or's requests we will submit this patch to net and will need it for -stable.
- Rebased to: 11afbff86168 ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next")

This series is centralized around extending and improving mlx5 ethernet driver ethtool
support. We've done some code refactoring for ethtool statistics reporting, making it
more scalable and robust, now each reported ethtool counter belongs to a group and has
its own descriptor within that group, the descriptor holds the counter name and offset
in memory in that group memory block.

Added new counters:
- Reporting more error and drop counter in ifconig/ip tool.
- Per priority pause and traffic counter in ethtool.
- link down events counter in ethtool.

Set features handling was also refactored a little bit to be more resilient and generic,
now setting more than one feature will not stop on the first failed one, but instead
it will try to continue setting others. We made it generic to make it simpler for adding
more features support, it is now done easily by only introducing a handler function of
the new supported netdev feature, and let the generic handler do the job.

New netdev features and ethtool support:
- Netdev feature RXALL, set on/off FCS check offload.
- Netdev feature HW_VLAN_CTAG_RX, set on/off rx-vlan stripping offload.
- Ethtool interface identify.
- Ethtool dump module EEPROM.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Fix checksum handling for non-stripped vlan packets
Saeed Mahameed [Sun, 24 Apr 2016 19:51:56 +0000 (22:51 +0300)]
net/mlx5e: Fix checksum handling for non-stripped vlan packets

Now as rx-vlan offload can be disabled, packets can be received
with vlan tag not stripped, which means is_first_ethertype_ip will
return false, for that we need to check if the hardware reported
csum OK so we will report CHECKSUM_UNNECESSARY for those packets.

Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Add ethtool support for rxvlan-offload (vlan stripping)
Gal Pressman [Sun, 24 Apr 2016 19:51:55 +0000 (22:51 +0300)]
net/mlx5e: Add ethtool support for rxvlan-offload (vlan stripping)

Use ethtool -K <interface> rxvlan <on/off> to enable/disable
C-TAG vlan stripping by hardware.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Add ethtool support for dump module EEPROM
Gal Pressman [Sun, 24 Apr 2016 19:51:54 +0000 (22:51 +0300)]
net/mlx5e: Add ethtool support for dump module EEPROM

Add query MCIA, PMLP registers infrastructure and commands.
Add ethtool support for get_module_info() and get_module_eeprom()
callbacks.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Add ethtool support for interface identify (LED blinking)
Gal Pressman [Sun, 24 Apr 2016 19:51:53 +0000 (22:51 +0300)]
net/mlx5e: Add ethtool support for interface identify (LED blinking)

Add the needed hardware command and mlx5_ifc structs for managing LED
control.
Add set_phys_id ethtool callback to support ethtool -p flag.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Add support for RXALL netdev feature
Eran Ben Elisha [Sun, 24 Apr 2016 19:51:52 +0000 (22:51 +0300)]
net/mlx5e: Add support for RXALL netdev feature

Introduce new access register named Ports Check Mask Register (PCMR) to
control all HW checks on port. With this register, the driver can
enable/disable Hardware FCS validation.

When RXALL is enabled/disabled using ndo_set_features, enable/disable
fcs check at HW.
User can change HW configuration using rx-all flag at ethtool.

Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Improve set features ndo resiliency
Gal Pressman [Sun, 24 Apr 2016 19:51:51 +0000 (22:51 +0300)]
net/mlx5e: Improve set features ndo resiliency

In current mlx5e ndo_set_features implementation, setting some features
can success while others can fail. Today, we return one error code which
doesn't reflect the current features status of the netdev at the end of
the ndo callback.

Set netdev->features with features which were successfully set in order
to keep the current status in case of failure. For this purpose, define
new Macro to set/unset specific feature in netdev->features.

This patch introduces a mechanism that uses feature handlers for each
feature.
Set features will call a generic handler, which will then call a specific
handler in his turn and update netdev->features according to it's return
value. Each specific handler is responsible to perform driver specific
actions, and updating params if needed.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Add link down events counter
Gal Pressman [Sun, 24 Apr 2016 19:51:50 +0000 (22:51 +0300)]
net/mlx5e: Add link down events counter

Expose link_down_events counter through ethtool -S.
This counter is read from PPort statistics, then proccessed and stored as
a special handling software counter.
This counter is stored along software counters since it is the only PPort
counter that it's size is not 64 bits.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Add per priority group to PPort counters
Gal Pressman [Sun, 24 Apr 2016 19:51:49 +0000 (22:51 +0300)]
net/mlx5e: Add per priority group to PPort counters

Expose counters providing information for each priority level (PCP) through
ethtool -S option and DCBNL.
This includes rx/tx bytes, frames, and pause counters.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Rename VPort counters
Gal Pressman [Sun, 24 Apr 2016 19:51:48 +0000 (22:51 +0300)]
net/mlx5e: Rename VPort counters

VPort and software counters names are confusing and may be unclear, all
VPort counters now have a prefix of rx/tx_vport_*.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Statistics handling refactoring
Gal Pressman [Sun, 24 Apr 2016 19:51:47 +0000 (22:51 +0300)]
net/mlx5e: Statistics handling refactoring

Redesign ethtool statistics handling and reporting in the driver:
1. Move counters to a separate file (en_stats.h).
2. Remove unnecessary dependencies between stats and strings.
3. Use counter descriptors which hold a name and offset for each counter,
   and will be used to decide which counters will be exposed.

For example when adding a new software counter to ethtool, instead of:
1. Add to stats struct.
2. Add to strings struct in the same order.
3. Change macro defining number of software counters.
The only thing needed is to link the new counter to a counter descriptor.

VPort counters are a set of hardware traffic counters created automatically
for each virtual port opened.
PPort counters are a set of counters describing per physical port
performance statistics.
These counters are gathered from hardware register and divided to groups
according to different protocols.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx5e: Report additional error statistics in get stats ndo
Gal Pressman [Sun, 24 Apr 2016 19:51:46 +0000 (22:51 +0300)]
net/mlx5e: Report additional error statistics in get stats ndo

Provide rtnl_link_stats64 with information regarding physical errors to be
seen in ifconfig and ip tool.

Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoinfiniband: nes: Kill unused variable in nes_netdev_start_xmit()
David S. Miller [Tue, 26 Apr 2016 19:57:19 +0000 (15:57 -0400)]
infiniband: nes: Kill unused variable in nes_netdev_start_xmit()

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'netdev_tx_locked-removal'
David S. Miller [Tue, 26 Apr 2016 19:53:06 +0000 (15:53 -0400)]
Merge branch 'netdev_tx_locked-removal'

Florian Westphal says:

====================
net: core: remove TX_LOCKED support

Not that many users left, lets kill it.

TX_LOCKED was meant to be used by LLTX drivers when spin_trylock()
failed.  Stack then re-queued if collisions happened on different
cpus or free'd the skb to prevent deadlocks.

Most of the driver removal patches fall into one of three categories:
1. remove the driver-private tx lock (and LLTX flag), or...
2. convert spin_trylock to plain spin_lock, or...
3. convert TX_LOCKED to free+TX_OK

Patches are grouped by these categories, last patch is the actual removal.
All driver changes were compile tested only with exception of atl1e.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: remove NETDEV_TX_LOCKED support
Florian Westphal [Sun, 24 Apr 2016 19:38:14 +0000 (21:38 +0200)]
net: remove NETDEV_TX_LOCKED support

No more users in the tree, remove NETDEV_TX_LOCKED support.
Adds another hole in softnet_stats struct, but better than keeping
the unused collision counter around.

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodrivers: net: remove NETDEV_TX_LOCKED
Florian Westphal [Sun, 24 Apr 2016 19:38:13 +0000 (21:38 +0200)]
drivers: net: remove NETDEV_TX_LOCKED

replace the trylock by a full spin_lock and remove TX_LOCKED return value.
Followup patch will remove TX_LOCKED from the kernel.

Cc: Jon Mason <jdmason@kudzu.us>
Cc: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodrivers: net: use NETDEV_TX_OK instead of NETDEV_TX_LOCKED
Florian Westphal [Sun, 24 Apr 2016 19:38:12 +0000 (21:38 +0200)]
drivers: net: use NETDEV_TX_OK instead of NETDEV_TX_LOCKED

These drivers already call netif_stop_queue() so we should not be called
unless tx space is available.  Just free the skb and return TX_OK.

Followup patch will remove NETDEV_TX_LOCKED from the kernel.

Cc: linux-parisc@vger.kernel.org
Cc: linux-hams@vger.kernel.org
Cc: Thomas Sailer <t.sailer@alumni.ethz.ch>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoatle1: remove LLTX support and TX_UNLOCKED
Florian Westphal [Sun, 24 Apr 2016 19:38:11 +0000 (21:38 +0200)]
atle1: remove LLTX support and TX_UNLOCKED

similar to atl1c: lock is only used in ndo_start_xmit, but we also
advertised LLTX, so remove that as well and let core stack handle
tx locking.

Allows to remove the TX_LOCKED return value from the driver.

Cc: Jay Cliburn <jcliburn@gmail.com>
Cc: Chris Snook <chris.snook@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoatl1c: remove private tx lock
Florian Westphal [Sun, 24 Apr 2016 19:38:10 +0000 (21:38 +0200)]
atl1c: remove private tx lock

AFAICS this is safe: the lock is only used in the .ndo_start_xmit
function and this driver does not set LLTX.

Gets rid of TX_LOCKED return value, followup patches will remove it.

Cc: Jay Cliburn <jcliburn@gmail.com>
Cc: Chris Snook <chris.snook@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoRDMA/nes: remove use of NETDEV_TX_LOCKED
Florian Westphal [Sun, 24 Apr 2016 19:38:09 +0000 (21:38 +0200)]
RDMA/nes: remove use of NETDEV_TX_LOCKED

ndo_start_xmit never returns it to stack, but nes_nic_send helper used it if
skb could not be queued to hardware.  Switch to bool instead.

Cc: <linux-rdma@vger.kernel.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'get_phy_device-retval'
David S. Miller [Tue, 26 Apr 2016 19:41:21 +0000 (15:41 -0400)]
Merge branch 'get_phy_device-retval'

Sergei Shtylyov says:

====================
Don't return NULL from get_phy_device() anymore

   Here's the set of 5 patches against DaveM's 'net-next.git' repo. The first
patch makes get_phy_device() return only error values on error, the rest of
the patches clean up the callers of that function...
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoof_mdio: get_phy_device() doesn't return NULL anymore
Sergei Shtylyov [Sun, 24 Apr 2016 17:31:42 +0000 (20:31 +0300)]
of_mdio: get_phy_device() doesn't return NULL anymore

Now that get_phy_device() no longer returns NULL on error, we don't need
to check for it...

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomdio_bus: get_phy_device() doesn't return NULL anymore
Sergei Shtylyov [Sun, 24 Apr 2016 17:30:53 +0000 (20:30 +0300)]
mdio_bus: get_phy_device() doesn't return NULL anymore

Now that get_phy_device() no longer returns NULL on error, we don't need
to check for it...

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agofixed_phy: get_phy_device() doesn't return NULL anymore
Sergei Shtylyov [Sun, 24 Apr 2016 17:29:23 +0000 (20:29 +0300)]
fixed_phy: get_phy_device() doesn't return NULL anymore

Now that get_phy_device() no longer returns NULL on error, we don't need
to check for it...

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoxgene: get_phy_device() doesn't return NULL anymore
Sergei Shtylyov [Sun, 24 Apr 2016 17:27:49 +0000 (20:27 +0300)]
xgene: get_phy_device() doesn't return NULL anymore

Now that get_phy_device() no longer returns NULL on error, we don't need
to check for it...

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agophylib: don't return NULL from get_phy_device()
Sergei Shtylyov [Sun, 24 Apr 2016 17:25:23 +0000 (20:25 +0300)]
phylib: don't return NULL from get_phy_device()

Arnd Bergmann asked that get_phy_device() returns either NULL or the error
value,  not both on error.  Do as he said, return ERR_PTR(-ENODEV) instead
of NULL when the PHY ID registers read as  all ones.

Suggested-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: get rid of the 2nd parameter to sh_eth_dev_init()
Sergei Shtylyov [Sun, 24 Apr 2016 16:11:07 +0000 (19:11 +0300)]
sh_eth: get rid of the 2nd parameter to sh_eth_dev_init()

sh_eth_dev_init()  is now always called with  'true' as the  2nd argument,
so that there's no more sense in having 2 parameters to this function...

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosctp: sctp_diag should fill RMEM_ALLOC with asoc->rmem_alloc when rcvbuf_policy is set
Xin Long [Sun, 24 Apr 2016 15:21:22 +0000 (23:21 +0800)]
sctp: sctp_diag should fill RMEM_ALLOC with asoc->rmem_alloc when rcvbuf_policy is set

For sctp assoc, when rcvbuf_policy is set, it will has it's own
rmem_alloc, when we dump asoc info in sctp_diag, we should use that
value on RMEM_ALLOC as well, just like WMEM_ALLOC.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomacvtap: check minor when unregistering
Francesco Ruggeri [Sat, 23 Apr 2016 22:04:31 +0000 (15:04 -0700)]
macvtap: check minor when unregistering

macvtap_device_event(NETDEV_UNREGISTER) should check vlan->minor to
determine if it is being invoked in the context of a macvtap_newlink
that failed, for example in this code sequence:

macvtap_newlink
  macvlan_common_newlink
    register_netdevice
      call_netdevice_notifiers(NETDEV_REGISTER, dev)
        macvtap_device_event(NETDEV_REGISTER)
          <fail here, vlan->minor = 0>
      rollback_registered(dev);
        rollback_registered_many
          call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
            macvtap_device_event(NETDEV_UNREGISTER)
              <nothing to clean up here>

Signed-off-by: Francesco Ruggeri <fruggeri@arista.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomacvlan: fix failure during registration v3
Francesco Ruggeri [Sat, 23 Apr 2016 22:03:32 +0000 (15:03 -0700)]
macvlan: fix failure during registration v3

If macvlan_common_newlink fails in register_netdevice after macvlan_init
then it decrements port->count twice, first in macvlan_uninit (from
register_netdevice or rollback_registered) and then again in
macvlan_common_newlink.
A similar problem may exist in the ipvlan driver.
This patch consolidates modifications to port->count into macvlan_init
and macvlan_uninit (thanks to Eric Biederman for suggesting this approach).

v3: remove macvtap specific bits.

Signed-off-by: Francesco Ruggeri <fruggeri@arista.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'qed-bandwidth-config'
David S. Miller [Tue, 26 Apr 2016 17:56:27 +0000 (13:56 -0400)]
Merge branch 'qed-bandwidth-config'

Manish Chopra says:

====================
qed: Bandwidth configuration support

This series adds support in driver for min/max bandwidth configuration
for the PF on the link change notification or on explicit request of
bandwidth update from the MFW [management firmware].

The same infrastructure would later be used by user based flows
[for example, rate shaping for the VFs]

Please consider applying this series to "net-next"
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed: Add PF min bandwidth configuration support
Manish Chopra [Tue, 26 Apr 2016 14:56:10 +0000 (10:56 -0400)]
qed: Add PF min bandwidth configuration support

This patch adds support for PF minimum bandwidth update
or configuration notified by management firmware.

Signed-off-by: Manish Chopra <manish.chopra@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed: Add PF max bandwidth configuration support
Manish Chopra [Tue, 26 Apr 2016 14:56:09 +0000 (10:56 -0400)]
qed: Add PF max bandwidth configuration support

This patch adds support for PF maximum bandwidth update
or configuration notified by management firmware.

Signed-off-by: Manish Chopra <manish.chopra@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoqed: Add vport WFQ configuration APIs
Manish Chopra [Tue, 26 Apr 2016 14:56:08 +0000 (10:56 -0400)]
qed: Add vport WFQ configuration APIs

This patch adds relevant APIs needed to configure WFQ
(Weighted fair queueing) values for the vports. WFQ configuration
is used per vport basis when minimum bandwidth update/configuration
is notified to the PF by the management firmware.

Signed-off-by: Manish Chopra <manish.chopra@qlogic.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'cxgb4-next'
David S. Miller [Tue, 26 Apr 2016 17:23:58 +0000 (13:23 -0400)]
Merge branch 'cxgb4-next'

Hariprasad Shenai says:

====================
cxgb4: minor fixes, decode msgs and code refactor for few functions

This patch series adds new routine to get adapter information and removes
some redundant messages logged in dmesg. Fixes race while freeing tx
descriptors. Decodes module type and link down reason codes obtained from
firmware. Refactor port initialization and FW_PORT_CMD handling. Pass
correct port id in DCB message handler to obtain netdev associated.

This patch series has been created against net-next tree and includes
patches on cxgb4 driver.

We have included all the maintainers of respective drivers. Kindly review
the change and let us know in case of any review comments.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Decode link down reason code obtained from firmware
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:29 +0000 (20:10 +0530)]
cxgb4: Decode link down reason code obtained from firmware

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Don't assume FW_PORT_CMD reply is always port info msg
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:28 +0000 (20:10 +0530)]
cxgb4: Don't assume FW_PORT_CMD reply is always port info msg

The firmware can send a set of asynchronous replies through FW_PORT_CMD
with DCBX information when that's negotiated with the Link Peer. The old
code always assumed that a FW_PORT_CMD reply was always a Get Port
Information message. This change conditionalizes the code to only handle
the Get Port Information messages and throws a warning if we don't
understand what we've been given.

Also refactor t4_handle_fw_rpl() so that core functionality performed by
t4_handle_get_port_info() for a specified port.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: DCB message handler needs to use correct portid to netdev mapping
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:27 +0000 (20:10 +0530)]
cxgb4: DCB message handler needs to use correct portid to netdev mapping

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Refactor t4_port_init function
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:26 +0000 (20:10 +0530)]
cxgb4: Refactor t4_port_init function

Refactor t4_port_init() so that the core functionality is done by
t4_init_portinfo() for a particular port. Also rename variables to
sensible ones.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Properly decode port module type
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:25 +0000 (20:10 +0530)]
cxgb4: Properly decode port module type

Decode and log port module error, unknown modules and unsupported
modules.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Avoids race and deadlock while freeing tx descriptor
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:24 +0000 (20:10 +0530)]
cxgb4: Avoids race and deadlock while freeing tx descriptor

There could be race between t4_eth_xmit() and t4_free_sge_resources() while
freeing tx descriptors, take txq lock in t4_free_sge_resources(). We need
to stop the xmit frame path which runs in bottom half context while
unloading the driver using _bh variant of the lock. This is to prevent
the deadlock between xmit and driver unload.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: Add llseek operation for flash debugfs entry
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:23 +0000 (20:10 +0530)]
cxgb4: Add llseek operation for flash debugfs entry

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agocxgb4: add new routine to get adapter info
Hariprasad Shenai [Tue, 26 Apr 2016 14:40:22 +0000 (20:10 +0530)]
cxgb4: add new routine to get adapter info

Add new routine to print out general adapter information (various version
numbers, adapter name, part number, serial number, etc.) and remove
redundant information dumped in the Port Information.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.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, 26 Apr 2016 17:15:56 +0000 (13:15 -0400)]
Merge branch 'for-upstream' of git://git./linux/kernel/git/bluetooth/bluetooth-next

Johan Hedberg says:

====================
pull request: bluetooth-next 2016-04-26

Here's another set of Bluetooth & 802.15.4 patches for the 4.7 kernel:

 - Cleanups & refactoring of ieee802154 & 6lowpan code
 - Security related additions to ieee802154 and mrf24j40 driver
 - Memory corruption fix to Bluetooth 6lowpan code
 - Race condition fix in vhci driver
 - Enhancements to the atusb 802.15.4 driver

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

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoi40evf: Allow PF driver to configure RSS
Mitch Williams [Tue, 12 Apr 2016 15:30:44 +0000 (08:30 -0700)]
i40evf: Allow PF driver to configure RSS

If the PF driver reports proper support, allow the PF driver to
configure RSS on the behalf of the VF driver. This will allow for RSS
support on future hardware without changes to the VF driver.

Unfortunately, the old RSS code still needs to stay as the driver needs
to be compatible with PF drivers that don't support this interface. But
this change still simplifies the data structures a bunch and makes this
code simpler to read and maintain.

Change-ID: I0375aad40788ecdc0cb24d5cfeccf07804e69771
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: Specify AQ event opcode to wait for
Shannon Nelson [Tue, 12 Apr 2016 15:30:43 +0000 (08:30 -0700)]
i40e: Specify AQ event opcode to wait for

To add a little flexibility to the nvmupdate facility, this code adds the
ability to specify an AQ event opcode to wait on after the Exec_AQ request.

Change-ID: Iddbfd63c3de8df3edb9d3e90678b08989bc4946e
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoMerge branch 'nla_align-set-4'
David S. Miller [Tue, 26 Apr 2016 16:00:50 +0000 (12:00 -0400)]
Merge branch 'nla_align-set-4'

Nicolas Dichtel says:

====================
netlink: align attributes when needed (patchset #3)

This is the continuation (series #3) of the work done to align netlink
attributes when these attributes contain some 64-bit fields.

It's the last patchset from what I've seen.

The last user of nla_put_u64() is block/drbd. This module does not use
standard netlink API (see all the stuff in include/linux/genl_magic_struct.h
and include/linux/genl_magic_func.h). I didn't modify it because it's seems
hard to do it whithout testing and fully understanding the context (for
example, why include/linux/drbd_genl.h is not part of uapi?).
Any thoughts?
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosched: align nlattr properly when needed
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:18 +0000 (10:06 +0200)]
sched: align nlattr properly when needed

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoneigh: align nlattr properly when needed
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:17 +0000 (10:06 +0200)]
neigh: align nlattr properly when needed

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agortnl: align nlattr properly when needed
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:16 +0000 (10:06 +0200)]
rtnl: align nlattr properly when needed

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoovs: align nlattr properly when needed
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:15 +0000 (10:06 +0200)]
ovs: align nlattr properly when needed

I also fix commit 8b32ab9e6ef1: use nla_total_size_64bit() for
OVS_FLOW_ATTR_USED in ovs_flow_cmd_msg_size().

Fixes: 8b32ab9e6ef1 ("ovs: use nla_put_u64_64bit()")
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosock_diag: align nlattr properly when needed
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:14 +0000 (10:06 +0200)]
sock_diag: align nlattr properly when needed

I also fix the value of INET_DIAG_MAX. It's wrong since commit 8f840e47f190
which is only in net-next right now, thus I didn't make a separate patch.

Fixes: 8f840e47f190 ("sctp: add the sctp_diag.c file")
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agofs/quota: use nla_put_u64_64bit()
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:13 +0000 (10:06 +0200)]
fs/quota: use nla_put_u64_64bit()

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodrivers/wireless: use nla_put_u64_64bit()
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:12 +0000 (10:06 +0200)]
drivers/wireless: use nla_put_u64_64bit()

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomacsec: use nla_put_u64_64bit()
Nicolas Dichtel [Tue, 26 Apr 2016 08:06:11 +0000 (10:06 +0200)]
macsec: use nla_put_u64_64bit()

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ipv6: Delete host routes on an ifdown
David Ahern [Fri, 22 Apr 2016 03:56:12 +0000 (20:56 -0700)]
net: ipv6: Delete host routes on an ifdown

It was a simple idea -- save IPv6 configured addresses on a link down
so that IPv6 behaves similar to IPv4. As always the devil is in the
details and the IPv6 stack as too many behavioral differences from IPv4
making the simple idea more complicated than it needs to be.

The current implementation for keeping IPv6 addresses can panic or spit
out a warning in one of many paths:

1. IPv6 route gets an IPv4 route as its 'next' which causes a panic in
   rt6_fill_node while handling a route dump request.

2. rt->dst.obsolete is set to DST_OBSOLETE_DEAD hitting the WARN_ON in
   fib6_del

3. Panic in fib6_purge_rt because rt6i_ref count is not 1.

The root cause of all these is references related to the host route for
an address that is retained.

So, this patch deletes the host route every time the ifdown loop runs.
Since the host route is deleted and will be re-generated an up there is
no longer a need for the l3mdev fix up. On the 'admin up' side move
addrconf_permanent_addr into the NETDEV_UP event handling so that it
runs only once versus on UP and CHANGE events.

All of the current panics and warnings appear to be related to
addresses on the loopback device, but given the catastrophic nature when
a bug is triggered this patch takes the conservative approach and evicts
all host routes rather than trying to determine when it can be re-used
and when it can not. That can be a later optimizaton if desired.

Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoRevert "ipv6: Revert optional address flusing on ifdown."
David S. Miller [Tue, 26 Apr 2016 15:47:41 +0000 (11:47 -0400)]
Revert "ipv6: Revert optional address flusing on ifdown."

This reverts commit 841645b5f2dfceac69b78fcd0c9050868d41ea61.

Ok, this puts the feature back.  I've decided to apply David A.'s
bug fix and run with that rather than make everyone wait another
whole release for this feature.

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoi40e: Code cleanup in i40e_add_fdir_ethtool
Shannon Nelson [Tue, 12 Apr 2016 15:30:42 +0000 (08:30 -0700)]
i40e: Code cleanup in i40e_add_fdir_ethtool

A little bit of code cleanup in prep for more cloud filter work.

Change-ID: I0dc33ce0d4c207944336a07437640fef920c100c
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 agodrivers: firmware: psci: unify enable-method binding on ARM {64,32}-bit systems
Sudeep Holla [Fri, 22 Apr 2016 15:18:02 +0000 (16:18 +0100)]
drivers: firmware: psci: unify enable-method binding on ARM {64,32}-bit systems

Currently ARM CPUs DT bindings allows different enable-method value for
PSCI based systems. On ARM 64-bit this property is required and must be
"psci" while on ARM 32-bit systems this property is optional and must
be "arm,psci" if present.

However, "arm,psci" has always been the compatible string for the PSCI
node, and was never intended to be the enable-method. So this is a bug
in the binding and not a deliberate attempt at specifying 32-bit
differently.

This is problematic if 32-bit OS is run on 64-bit system which has
"psci" as enable-method rather than the expected "arm,psci".

So let's unify the value into "psci" and remove support for "arm,psci"
before it finds any users.

Reported-by: Soby Mathew <Soby.Mathew@arm.com>
Cc: Rob Herring <robh+dt@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
8 years agoi40evf: Don't Panic
Mitch Williams [Tue, 12 Apr 2016 15:30:41 +0000 (08:30 -0700)]
i40evf: Don't Panic

Under some circumstances the driver remove function may be called before
the driver is fully initialized. So we can't assume that we know where
our towel is at, or that all of the data structures are initialized.

To ensure that we don't panic, check that the vsi_res pointer is valid
before dereferencing it. Then drink beer and eat peanuts.

Change-ID: If697b4db57348e39f9538793e16aa755e3e1af03
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: Add support for configuring VF RSS
Mitch Williams [Tue, 12 Apr 2016 15:30:40 +0000 (08:30 -0700)]
i40e: Add support for configuring VF RSS

Add support for configuring RSS on behalf of the VFs. This removes the
burden of dealing with different hardware interfaces from the VF
drivers, allowing for better future compatibility.

Change-ID: Icea75d3f37241ee8e447be5779e5abb53ddf04c0
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 support for IPIP and SIT offloads
Alexander Duyck [Sat, 2 Apr 2016 07:06:56 +0000 (00:06 -0700)]
i40e/i40evf: Add support for IPIP and SIT offloads

Looking over the documentation it turns out enabling IPIP and SIT offloads
for i40e is pretty straightforward.  As such I decided to enable them with
this patch.  In my testing I am seeing an improvement of 8 to 10 Gb/s
for IPIP and SIT tunnels with this offload enabled.

Signed-off-by: Alexander Duyck <aduyck@mirantis.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 feature flags
Alexander Duyck [Sat, 2 Apr 2016 07:05:14 +0000 (00:05 -0700)]
i40e/i40evf: Clean up feature flags

The feature flags list for i40e and i40evf is beginning to become pretty
massive.  I plan to add another 4 or so features to these drivers and
duplicating the flags for each and every flags list is becoming a bit
repetitive.

The primary change here is that we now build our features list around
hw_encap_features.  After that we assign that to vlan_features,
hw_features, and finally map that onto features.  In addition we end up
throwing features onto hw_encap_features that end up having no effect such
as the Rx offloads and SCTP_CRC.  However that should have no impact and
makes things a bit easier for us as hw_encap_features is one of the less
updated features maps available.

For i40evf I went through and sanity checked a few features as well.
Specifically RXCSUM was being set as a read-only feature which didn't make
much sense.  I have updated things so we can clear the NETIF_F_RXCSUM flag
since that is really a software feature and not a hardware one anyway so
disabling it is just a matter of ignoring the result from the hardware.

Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoMerge branch 'ila-csum-neutral'
David S. Miller [Tue, 26 Apr 2016 05:32:01 +0000 (01:32 -0400)]
Merge branch 'ila-csum-neutral'

Tom Herbert says:

====================
ila: Support for checksum neutral translations

This patch set updates ILA to support draft-herbert-nvo3-ila-02. The
primary addition is support checksum neutral ILA translation.
This allows address to be performed and still keep any transport
layer checksums that include the addresses in their pseudo header to
still be correct without the translator needing to parse L4.

Other items are:
   - Structures for ILA addresses, identifiers, locators
   - Disallow translation on non-ILA addresses (check by
     type in identifier).
   - Change xlat (nf_input) to translates solely based
     on matching locators not identifiers (since identifiers
     are not obfuscated by checksum neutral).
   - Side effect if above is that multiple ILA domains are
     supported. Each local locator can map to a different
     SIR address (ILA domain), and each domain defines its
     own identifier space.

Tested: Ran TCP_RR with 200 cnxs. ILA performance is slightly better
than previously since we are not longer parsing L4 for checksum
handling. I amd seeing about 1% performance overhead. Also ran
TCP_STREAM and tested non-ILA address (type=0) are not translated.

v2: Fix compilation errors
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoila: add checksum neutral ILA translations
Tom Herbert [Sat, 23 Apr 2016 18:46:57 +0000 (11:46 -0700)]
ila: add checksum neutral ILA translations

Support checksum neutral ILA as described in the ILA draft. The low
order 16 bits of the identifier are used to contain the checksum
adjustment value.

The csum-mode parameter is added to described checksum processing. There
are three values:
 - adjust transport checksum (previous behavior)
 - do checksum neutral mapping
 - do nothing

On output the csum-mode in the ila_params is checked and acted on. If
mode is checksum neutral mapping then to mapping and set C-bit.

On input, C-bit is checked. If it is set checksum-netural mapping is
done (regardless of csum-mode in ila params) and C-bit will be cleared.
If it is not set then action in csum-mode is taken.

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoila: xlat changes
Tom Herbert [Sat, 23 Apr 2016 18:46:56 +0000 (11:46 -0700)]
ila: xlat changes

Change model of xlat to be used only for input where lookup is done on
the locator part of an address (comparing to locator_match as key
in rhashtable). This is needed for checksum neutral translation
which obfuscates the low order 16 bits of the identifier. It also
permits hosts to be in muliple ILA domains (each locator can map
to a different SIR address). A check is also added to disallow
translating non-ILA addresses (check of type in identifier).

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoila: Add struct definitions and helpers
Tom Herbert [Sat, 23 Apr 2016 18:46:55 +0000 (11:46 -0700)]
ila: Add struct definitions and helpers

Add structures for identifiers, locators, and an ila address which
is composed of a locator and identifier and in6_addr can be cast to
it. This includes a three bit type field and enums for the types defined
in ILA I-D.

In ILA lwt don't allow user to set a translation for a non-ILA
address (type of identifier is zero meaning it is an IID). This also
requires that the destination prefix is at least 65 bytes (64
bit locator and first byte of identifier).

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/mlx4_en: fix spurious timestamping callbacks
Eric Dumazet [Sat, 23 Apr 2016 18:35:46 +0000 (11:35 -0700)]
net/mlx4_en: fix spurious timestamping callbacks

When multiple skb are TX-completed in a row, we might incorrectly keep
a timestamp of a prior skb and cause extra work.

Fixes: ec693d47010e8 ("net/mlx4_en: Add HW timestamping (TS) support")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dummy: remove note about being Y by default
Ivan Babrou [Sat, 23 Apr 2016 12:58:03 +0000 (12:58 +0000)]
net: dummy: remove note about being Y by default

Signed-off-by: Ivan Babrou <ivan@cloudflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: tsi108: use NULL for pointer-typed argument
Julia Lawall [Sat, 23 Apr 2016 09:07:02 +0000 (11:07 +0200)]
net: tsi108: use NULL for pointer-typed argument

The first argument of pci_free_consistent has type struct pci_dev *, so use
NULL instead of 0.

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

// <smpl>
@@
@@
pci_free_consistent(
- 0
+ NULL
  , ...)
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'hns-debug-dsaf'
David S. Miller [Tue, 26 Apr 2016 05:09:18 +0000 (01:09 -0400)]
Merge branch 'hns-debug-dsaf'

Yisen Zhuang says:

====================
net: hns: add support of debug dsaf device

There are two kinds of dsaf device in hns, one is for service ports,
contains crossbar in it, can work under different mode. Another is for
debug port, only can work under single port mode. The current code only
declares a dsaf device for both service ports and debug ports.It is not so
readability. This patch separates it to three platform devices to make the
code more simple and readability.

The diagram of all port in one platform device(old):
                                        CPU
                                         |
                                         |        DSAF(one platform device)
    --------------------------------------------------------------      /
    |                           |                  |      |      |     /
    |                          PPE                PPE    PPE     |    /
    |                           |                  |      |      |   /
    |                           |                  |      |      |  /
    |                        crossbar              |      |      | /
    |                           |                  |      |      |/
    |    -----------------------------------       |      |      |
    |    |      |      |      |      |      |      |      |      |
    |    |      |      |      |      |      |      |      |      |
    |   MAC    MAC    MAC    MAC    MAC    MAC    MAC    MAC     |
    |    |      |      |      |      |      |      |      |      |
    --------------------------------------------------------------
         |      |      |      |      |      |      |      |
        PHY    PHY    PHY    PHY    PHY    PHY    PHY    PHY

The diagram of separating ports to three platform(new):
                                                         CPU
                                                          |
                                    -----------------------------------
                                    |                     |           |
    ----------------------------------------------    ---------   ---------
    |                         |                  |    |       |   |   |   |
    |                        PPE                 |    |  PPE  |   |  PPE  |
    |                         |                  |    |   |   |   |   |   |
    |                         |                  |    |   |   |   |   |   |
    |                      crossbar              |    |   |   |   |   |   |
    |                         |                  |    |   |   |   |   |   |
    |   ----------------------------------       |    |   |   |   |   |   |
    |   |     |     |      |      |      |       |    |   |   |   |   |   |
    |   |     |     |      |      |      |       |    |   |   |   |   |   |
    |  MAC   MAC   MAC    MAC    MAC    MAC      |    |  MAC  |   |  MAC  |
    |   |     |     |      |      |      |       |    |   |   |   |   |   |
    ----------------------------------------------    ---------   ---------
        |     |     |      |      |      |    \        /  |        /  |
       PHY   PHY   PHY    PHY    PHY    PHY    \      /  PHY      /  PHY
                                                \    /           /
                                                 \  /           /
                                             DSAF(three platform device)

We take the compatibility into consideration, and it works well by using the
old dts file(tested on d02 board).

For more details, please see individual patches.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodts: hisi: update hns dst for separating dsaf dev support
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:17 +0000 (17:05 +0800)]
dts: hisi: update hns dst for separating dsaf dev support

Because debug dsaf port was separated from service dsaf port, this patch
updates the related configurations of hns dts, changes it to match with
the new binding files. This also removes enet nodes which don't exist in
d02 board.

Signed-off-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoDocumentation: Bindings: add port-idx-in-ae for enet node
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:16 +0000 (17:05 +0800)]
Documentation: Bindings: add port-idx-in-ae for enet node

This patch adds description for port-idx-in-ae attribute.

Signed-off-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoDocumentation: Bindings: Update DT binding for separating dsaf dev support
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:15 +0000 (17:05 +0800)]
Documentation: Bindings: Update DT binding for separating dsaf dev support

Because debug dsaf port was separated from service dsaf port, this patch
updates the related information of DT binding.

Signed-off-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: hns: add attribute port-mode-offset for dsaf port node
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:14 +0000 (17:05 +0800)]
net: hns: add attribute port-mode-offset for dsaf port node

Port mode offset for each dsaf port is different. The current code is not
so readability. This patch adds configuration named port-mode-offset to
make the code simple and more readability. If port-mode-offset isn't
exists, default value 0 will be used.

Signed-off-by: Daode Huang <huangdaode@hisilicon.com>
Signed-off-by: Yisen Zhuang <Yisen.Zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: hns: add attribute port-rst-offset for dsaf port node
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:13 +0000 (17:05 +0800)]
net: hns: add attribute port-rst-offset for dsaf port node

The reset offset for each port in a dsaf is different. The current code is
not so readability. This patch adds configuration named port-rst-offset to
make the code simple and more readability. If this attribute doesn't exist,
default value of this attribute is equal to its port index.

Signed-off-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: hns: add attribute cpld_ctrl for dsaf port node
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:12 +0000 (17:05 +0800)]
net: hns: add attribute cpld_ctrl for dsaf port node

This patch adds attribute cpld_ctrl for dsaf port node, parses the syscon
for mac_cb from dts, and changes the method of access the cpld related
registers through syscon.

Signed-off-by: Daode Huang <huangdaode@hisilicon.com>
Signed-off-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: hns: separate debug dsaf device from service dsaf device
Yisen.Zhuang\(Zhuangyuzeng\) [Sat, 23 Apr 2016 09:05:11 +0000 (17:05 +0800)]
net: hns: separate debug dsaf device from service dsaf device

There are two kinds of dsaf device in hns, one is for service ports,
contains crossbar in it, can work under different mode. Another is for
debug port, only can work under "single-port" mode. The current code only
declared a dsaf device for both service ports and debug ports. This patch
separate it to three platform devices.

Here is the diagram of all port in one platform device(old):
                                        CPU
                                         |
                                         |        DSAF(one platform device)
    --------------------------------------------------------------      /
    |                           |                  |      |      |     /
    |                          PPE                PPE    PPE     |    /
    |                           |                  |      |      |   /
    |                           |                  |      |      |  /
    |                        crossbar              |      |      | /
    |                           |                  |      |      |/
    |    -----------------------------------       |      |      |
    |    |      |      |      |      |      |      |      |      |
    |    |      |      |      |      |      |      |      |      |
    |   MAC    MAC    MAC    MAC    MAC    MAC    MAC    MAC     |
    |    |      |      |      |      |      |      |      |      |
    --------------------------------------------------------------
         |      |      |      |      |      |      |      |
        PHY    PHY    PHY    PHY    PHY    PHY    PHY    PHY

Here is the diagram of separate all ports to three platform(new):
                                                         CPU
                                                          |
                                    -----------------------------------
                                    |                     |           |
    ----------------------------------------------    ---------   ---------
    |                         |                  |    |       |   |   |   |
    |                        PPE                 |    |  PPE  |   |  PPE  |
    |                         |                  |    |   |   |   |   |   |
    |                         |                  |    |   |   |   |   |   |
    |                      crossbar              |    |   |   |   |   |   |
    |                         |                  |    |   |   |   |   |   |
    |   ----------------------------------       |    |   |   |   |   |   |
    |   |     |     |      |      |      |       |    |   |   |   |   |   |
    |   |     |     |      |      |      |       |    |   |   |   |   |   |
    |  MAC   MAC   MAC    MAC    MAC    MAC      |    |  MAC  |   |  MAC  |
    |   |     |     |      |      |      |       |    |   |   |   |   |   |
    ----------------------------------------------    ---------   ---------
        |     |     |      |      |      |    \        /  |        /  |
       PHY   PHY   PHY    PHY    PHY    PHY    \      /  PHY      /  PHY
                                                \    /           /
                                                 \  /           /
                                             DSAF(three platform device)

Signed-off-by: Daode Huang <huangdaode@hisilicon.com>
Signed-off-by: Yisen Zhuang <yisen.zhuang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>