openwrt/staging/blogic.git
9 years agocxgb4: Add ethtool support to get adapter stats
Hariprasad Shenai [Wed, 3 Jun 2015 15:34:39 +0000 (21:04 +0530)]
cxgb4: Add ethtool support to get adapter stats

Add ethtool support to get adapter specific hardware statistics

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge tag 'batman-adv-for-davem' of git://git.open-mesh.org/linux-merge
David S. Miller [Thu, 4 Jun 2015 03:22:46 +0000 (20:22 -0700)]
Merge tag 'batman-adv-for-davem' of git://git.open-mesh.org/linux-merge

Antonio Quartulli says:

====================
pull request: batman-adv 20150603

here you have our second batch of patches intended for net-next.

In this patchset you won't find any new features, but quite some code
cleanup work, a bunch of code style fixes and also comments corrections
by Markus Pargmann.

Moreover you have a patch from Sven Eckelmann removing an unnecessary
NULL check in batadv_iv_ogm_update_seqnos().

Please pull or let me know of any problem!
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: tulip: rearrange order of searching for substrings
Rasmus Villemoes [Wed, 3 Jun 2015 11:44:03 +0000 (13:44 +0200)]
net: tulip: rearrange order of searching for substrings

Currently, two of the branches are dead code, since an earlier smaller
substring would have been found ("TP" in the "TP_NW" case and either
of "BNC" and "AUI" in the "BNC_AUI" case). Rearrange the strstr()
calls so that the longer strings are searched for first.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Thu, 4 Jun 2015 03:19:16 +0000 (20:19 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2015-06-03

This series contains updates to e1000e only.

Yanir provides 8 fixes and 1 version bump for e1000e.  First fix resolves
a possible unit hang if multiple descriptors are available in the rings
during reset or runtime suspend.  Fixed a warning on SPARC compile based
on a suggested solution from Alex Duyck.  Fixed a logical error, where
a "if" condition under which a flush should occur, was revered.  Fixed
a hardware issue that prevented i219 from working in legacy interrupts
mode.  Fixed the hardware clock configuration and suprious non-linear
increment.  Lastly, fixed a system hang which occurred during execution
of "ethtool -t <NIC>", by disabling MULR for the loopback test to avoid
the hand state.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobpf: introduce bpf_clone_redirect() helper
Alexei Starovoitov [Tue, 2 Jun 2015 23:03:14 +0000 (16:03 -0700)]
bpf: introduce bpf_clone_redirect() helper

Allow eBPF programs attached to classifier/actions to call
bpf_clone_redirect(skb, ifindex, flags) helper which will
mirror or redirect the packet by dynamic ifindex selection
from within the program to a target device either at ingress
or at egress. Can be used for various scenarios, for example,
to load balance skbs into veths, split parts of the traffic
to local taps, etc.

Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet/mlx4_core: fix typo in mlx4_set_vf_mac
Carol Soto [Tue, 2 Jun 2015 21:07:25 +0000 (16:07 -0500)]
net/mlx4_core: fix typo in mlx4_set_vf_mac

fix typo in mlx4_set_vf_mac

Acked-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Carol L Soto <clsoto@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet/mlx4_core: need to call close fw if alloc icm is called twice
Carol Soto [Tue, 2 Jun 2015 21:07:24 +0000 (16:07 -0500)]
net/mlx4_core: need to call close fw if alloc icm is called twice

If mlx4_enable_sriov is called by adapter without this
feature MLX4_DEV_CAP_FLAG2_SYS_EQS then during this path the function alloc
icm is called twice without freeing the structures from the first time.

Acked-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Carol L Soto <clsoto@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet/mlx4_core: double free of dev_vfs
Carol L Soto [Tue, 2 Jun 2015 21:07:23 +0000 (16:07 -0500)]
net/mlx4_core: double free of dev_vfs

If user loads mlx4_core with num_vfs greater than
supported then variable dev->dev_vfs is freed 2 times after unloading the
driver.

Acked-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Carol L Soto <clsoto@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agodrivers/net/ethernet/dec/tulip/uli526x.c: fix misleading indentation in uli526x_timer
David Malcolm [Tue, 2 Jun 2015 19:31:17 +0000 (15:31 -0400)]
drivers/net/ethernet/dec/tulip/uli526x.c: fix misleading indentation in uli526x_timer

This code in drivers/net/ethernet/dec/tulip/uli526x.c
function "uli526x_timer":

  1086          } else
  1087                  if ((tmp_cr12 & 0x3) && db->link_failed) {
  [...snip...]
  1109                  }
  1110                  else if(!(tmp_cr12 & 0x3) && db->link_failed)
  1111                  {
  [...snip...]
  1117                  }
  1118                  db->init=0;

is misleadingly indented: the
  db->init=0
is indented as if part of the else clause at line 1086, but it is
independent of it (no braces before the "if" at line 1087).

This patch fixes the indentation to reflect the actual meaning of the code,
though is it actually meant to be part of the "else" clause?  (I'm a
compiler developer, not a kernel person).  It also adds spaces around
the assignment, to placate checkpatch.pl.

Seen via an experimental new gcc warning I'm working on for gcc 6,
-Wmisleading-indentation, using gcc r223098 adding
-Werror=misleading-indentation to KBUILD_CFLAGS in Makefile.
The experimental GCC emits this warning (as an error), rightly IMHO:

drivers/net/ethernet/dec/tulip/uli526x.c: In function ‘uli526x_timer’:
drivers/net/ethernet/dec/tulip/uli526x.c:1118:3: error: statement is
indented as if it were guarded by... [-Werror=misleading-indentation]
   db->init=0;
    ^
drivers/net/ethernet/dec/tulip/uli526x.c:1086:4: note: ...this ‘else’
clause, but it is not
  } else
     ^

Hope this is helpful
Dave

Signed-off-by: David Malcolm <dmalcolm@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: phy: dp83867: Add TI dp83867 phy
Dan Murphy [Tue, 2 Jun 2015 14:34:37 +0000 (09:34 -0500)]
net: phy: dp83867: Add TI dp83867 phy

Add support for the TI dp83867 Gigabit ethernet phy
device.

The DP83867 is a robust, low power, fully featured
Physical Layer transceiver with integrated PMD
sublayers to support 10BASE-T, 100BASE-TX and
1000BASE-T Ethernet protocols.

Signed-off-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoethernet: micrel: use time_after_eq
Antonio Murdaca [Tue, 2 Jun 2015 12:35:52 +0000 (14:35 +0200)]
ethernet: micrel: use time_after_eq

use the time_after_eq macro for jiffies comparison operation

Signed-off-by: Antonio Murdaca <antonio.murdaca@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet/mlx4_core: Fix build failure introduced by the EQ pool changes
Or Gerlitz [Tue, 2 Jun 2015 07:29:48 +0000 (10:29 +0300)]
net/mlx4_core: Fix build failure introduced by the EQ pool changes

When CONFIG_RFS_ACCEL or SMP aren't set, we fail to build, fix it.

Also, avoid build warning as of unused function on that setup.

Fixes: c66fa19c405a ('net/mlx4: Add EQ pool')
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobatman-adv: Remove unnecessary ret variable in algo_register
Markus Pargmann [Fri, 26 Dec 2014 11:41:41 +0000 (12:41 +0100)]
batman-adv: Remove unnecessary ret variable in algo_register

Remove ret variable and all jumps.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: Remove unnecessary ret variable
Markus Pargmann [Fri, 26 Dec 2014 11:41:40 +0000 (12:41 +0100)]
batman-adv: Remove unnecessary ret variable

We can avoid this indirect return variable by directly returning the
error values.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: main, batadv_compare_eth return bool
Markus Pargmann [Fri, 26 Dec 2014 11:41:39 +0000 (12:41 +0100)]
batman-adv: main, batadv_compare_eth return bool

Declare the returntype of batadv_compare_eth as bool.
The function called inside this helper function
(ether_addr_equal_unaligned) also uses bool as return value, so there is
no need to return int.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: main, Convert is_my_mac() to bool
Markus Pargmann [Fri, 26 Dec 2014 11:41:38 +0000 (12:41 +0100)]
batman-adv: main, Convert is_my_mac() to bool

It is much clearer to see a bool type as return value than 'int' for
functions that are supposed to return true or false.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: Remove unnecessary check for orig_ifinfo not NULL
Sven Eckelmann [Sun, 1 Mar 2015 15:56:26 +0000 (16:56 +0100)]
batman-adv: Remove unnecessary check for orig_ifinfo not NULL

orig_ifinfo is dereferenced multiple times in batadv_iv_ogm_update_seqnos
before the check for NULL is done. The function also exists at the
beginning when orig_ifinfo would have been NULL. This makes the check at
the end unnecessary and only confuses the reader/code analyzers.

Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: types, Fix comment on bcast_own
Markus Pargmann [Fri, 26 Dec 2014 11:41:37 +0000 (12:41 +0100)]
batman-adv: types, Fix comment on bcast_own

batadv_orig_bat_iv->bcast_own is actually not a bitfield, it is an
array. Adjust the comment accordingly.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Antonio Quartulli <antonio@meshcoding.com>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agoe1000e: Bump the version to 3.2.5
Yanir Lubetkin [Tue, 2 Jun 2015 14:05:50 +0000 (17:05 +0300)]
e1000e: Bump the version to 3.2.5

Bump the version to reflect the driver changes and bug fixes for i219.
Also update the copyright, while we are at it.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: fix unit hang during loopback test
Yanir Lubetkin [Tue, 2 Jun 2015 14:05:47 +0000 (17:05 +0300)]
e1000e: fix unit hang during loopback test

System would hang during execution of "ethtool -t <NIC>" for the same
reason that required flushing the descriptor rings. This fix disables
MULR for the loopback test to avoid the hang state.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: fix systim issues
Yanir Lubetkin [Tue, 2 Jun 2015 14:05:45 +0000 (17:05 +0300)]
e1000e: fix systim issues

Two issues involving systim were reported.
1. Clock is not running in the correct frequency
2. In some situations, systim values were not incremented linearly
This patch fixes the hardware clock configuration and the spurious
non-linear increment.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: fix legacy interrupt handling in i219
Yanir Lubetkin [Tue, 2 Jun 2015 14:05:42 +0000 (17:05 +0300)]
e1000e: fix legacy interrupt handling in i219

This fix handles a hardware issue that prevented i219 from
working in legacy interrupts mode (IntMode=0)

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: fix flush_desc_ring implementation
Yanir Lubetkin [Tue, 2 Jun 2015 14:05:38 +0000 (17:05 +0300)]
e1000e: fix flush_desc_ring implementation

The indication that a descriptor ring flush is required was read from
FEXTNVM7 by mistake. It should be read from the PCI config space.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: fix logical error in flush_desc_rings
Yanir Lubetkin [Wed, 22 Apr 2015 16:25:17 +0000 (19:25 +0300)]
e1000e: fix logical error in flush_desc_rings

The condition under which the flush should occur was reversed.  The fix
should be applied before any HW reset (unless followed by bus reset)
and before any power state transition from D0.

If E1000_FEXTNVM7_NEED_DESCRING_FLUSH bit is set in FEXTNVM7 and TDLEN > 0
the Tx ring should be flushed. (fixes ~95% of the hang states).
If the E1000_FEXTNVM7_NEED_DESCRING_FLUSH did not clear, we should also
flush the RX ring. Bug was caught by Alexander Duyck during a code review
when examining this fix.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: remove call to do_div and sign mismatch warning
Yanir Lubetkin [Wed, 22 Apr 2015 02:55:43 +0000 (05:55 +0300)]
e1000e: remove call to do_div and sign mismatch warning

Fixes a warning that was reported by Yanjiang Jin
<yanjiang.jin@windriver.com> by implementing the solution suggested by
Alexander Duyck <alexander.h.duyck@redhat.com>.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: i219 execute unit hang fix on every reset or power state transition
Yanir Lubetkin [Wed, 22 Apr 2015 01:15:01 +0000 (04:15 +0300)]
e1000e: i219 execute unit hang fix on every reset or power state transition

After testing various cases, the conclusion is that the fix MUST be
executed BEFORE any event that the HW is reset or transition to D3.
To fix that I moved the execution to the relevant places but per
Alexander Duyck's review, ensure now that the DMA is valid and was not
freed before manipulating the ring.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agoe1000e: i219 fix unit hang on reset and runtime D3
Yanir Lubetkin [Mon, 13 Apr 2015 23:20:21 +0000 (02:20 +0300)]
e1000e: i219 fix unit hang on reset and runtime D3

Unit hang may occur if multiple descriptors are available in the rings
during reset or runtime suspend. This state can be detected by testing
bit 8 in the FEXTNVM7 register. If this bit is set and there are pending
descriptors in one of the rings, we must flush them prior to reset. Same
applies entering runtime suspend.

Signed-off-by: Yanir Lubetkin <yanirx.lubetkin@intel.com>
Reviewed-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agobatman-adv: iv_ogm, fix comment function name
Markus Pargmann [Fri, 26 Dec 2014 11:41:36 +0000 (12:41 +0100)]
batman-adv: iv_ogm, fix comment function name

This is a small copy paste fix for batadv_ing_buffer_avg.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: iv_ogm, fix coding style
Markus Pargmann [Fri, 26 Dec 2014 11:41:35 +0000 (12:41 +0100)]
batman-adv: iv_ogm, fix coding style

The kernel coding style says, that there should not be multiple
assignments in one row.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: iv_ogm, Fix dup_status comment
Markus Pargmann [Fri, 26 Dec 2014 11:41:34 +0000 (12:41 +0100)]
batman-adv: iv_ogm, Fix dup_status comment

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: iv_ogm_orig_update, style, add missing brackets
Markus Pargmann [Fri, 26 Dec 2014 11:41:33 +0000 (12:41 +0100)]
batman-adv: iv_ogm_orig_update, style, add missing brackets

CodingStyle describes that either none or both branches of a conditional
have to have brackets.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: iv_ogm_queue_add, Simplify expressions
Markus Pargmann [Fri, 26 Dec 2014 11:41:32 +0000 (12:41 +0100)]
batman-adv: iv_ogm_queue_add, Simplify expressions

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agobatman-adv: iv_ogm_aggregate_new, simplify error handling
Markus Pargmann [Fri, 26 Dec 2014 11:41:31 +0000 (12:41 +0100)]
batman-adv: iv_ogm_aggregate_new, simplify error handling

It is just a bit easier to put the error handling at one place and let
multiple error paths use the same calls.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch>
9 years agoMerge branch 'sfc-next'
David S. Miller [Tue, 2 Jun 2015 19:57:39 +0000 (12:57 -0700)]
Merge branch 'sfc-next'

Shradha Shah says:

====================
sfc: ndo_get_phys_port_id, vadaptor stats and PF unload when Vf's assigned to guest

This is the third and last instalment of SRIOV for EF10 patches.

This patch set includes implementation of ndo_get_phys_port_id
and changes to the MAC statistics code in order to support
vadaptor statistics.

It also includes code to deal with PF unload when Vf's are still
assigned to the guest.

The first couple of patches create sysfs files for physical port
and link control flags which are particularly useful when we have
enabled a large number of VF's.

These patches have been tested with and without CONFIG_SFC_SRIOV.
The creation and content of the sysfs files has been tested.
The statistics are tested using ethtool for monitoring.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: leak vports if a VF is assigned during PF unload
Daniel Pieczko [Tue, 2 Jun 2015 10:41:00 +0000 (11:41 +0100)]
sfc: leak vports if a VF is assigned during PF unload

If any VF is assigned as the PF is unloaded, do not attempt to
remove its vport or the vswitch.  These will be removed if the
driver binds to the PF again, as an entity reset occurs during
probe.

A 'force' flag is added to efx_ef10_pci_sriov_disable() to
distinguish between disabling SR-IOV and driver unload.
SR-IOV cannot be disabled if VFs are assigned to guests.

If the PF driver is unloaded while VFs are assigned, the driver
may try to bind to the VF again at a later point if the driver
has been reloaded and the VF returns to the same domain as the PF.
In this case, the PF will not have a VF data structure, so the VF
can check this and drop out of probe early.

In this case, efx->vf_count will be zero but VFs will be present.
The user is advised to remove the VF and re-create it. The check
at the beginning of efx_ef10_pci_sriov_disable() that
efx->vf_count is non-zero is removed to allow SR-IOV to be
disabled in this case. Also, if the PF driver is unloaded, it
will disable SR-IOV to remove these unknown VFs.

By not disabling bus-mastering if VFs are still assigned, the VF
will continue to pass traffic after the PF has been removed.

When using the max_vfs module parameter, if VFs are already
present do not try to initialise any more.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: force removal of VF and vport on driver removal
Daniel Pieczko [Tue, 2 Jun 2015 10:40:46 +0000 (11:40 +0100)]
sfc: force removal of VF and vport on driver removal

When the driver unloads, force the unbind and removal of any
VFs in the host with the PF.  The PF cannot remove vports and
vswitches if they are still being used by a VF driver, and when
unloading the sfc driver the removal order is not guaranteed,
so the instruction from the PF to the VF to unbind enforces a
suitable ordering so that vswitches and vports can be removed.

As a result of this, manually unbinding the driver from a single
PF will result in all of its VFs in the host also being removed.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: do not allow VFs to be destroyed if assigned to guests
Daniel Pieczko [Tue, 2 Jun 2015 10:40:31 +0000 (11:40 +0100)]
sfc: do not allow VFs to be destroyed if assigned to guests

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: don't update stats on VF when called in atomic context
Daniel Pieczko [Tue, 2 Jun 2015 10:40:18 +0000 (11:40 +0100)]
sfc: don't update stats on VF when called in atomic context

The ifenslave command to set up a bond runs in an atomic
context, and it queries the stats on the devices that are
being enslaved. A VF needs to make an MCDI call to update
its stats, which is not allowed in atomic context.

The releasing of the stats_lock is moved to the beginning of
the VF stats update function so that in_interrupt() can be
used; it must be taken again before returning from this
function.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: suppress vadaptor stats when EVB is not present
Daniel Pieczko [Tue, 2 Jun 2015 10:40:05 +0000 (11:40 +0100)]
sfc: suppress vadaptor stats when EVB is not present

The raw_mask array is not initialised, so it needs to be
explicitly set to zero in the 'else' branch.

If the EVB capability is not present, a port cannot have multiple
functions so the per-port MAC stats are correct and should match
the corresponding vadaptor stats, so this redundancy can be
removed from the ethtool stats output.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: suppress ENOENT error messages from MC_CMD_MAC_STATS
Daniel Pieczko [Tue, 2 Jun 2015 10:39:49 +0000 (11:39 +0100)]
sfc: suppress ENOENT error messages from MC_CMD_MAC_STATS

MC_CMD_MAC_STATS can be called on a function before a
vadaptor has been created, as the kernel can call into this
through ndo_get_stats/ndo_get_stats64.

If MC_CMD_MAC_STATS is called before the DMA queues have been
setup, so that a vadaptor has not been created yet, firmware
will return ENOENT. This is expected, so suppress the MCDI
error message in this case.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: update netdevice statistics to use vadaptor stats
Daniel Pieczko [Tue, 2 Jun 2015 10:39:33 +0000 (11:39 +0100)]
sfc: update netdevice statistics to use vadaptor stats

The netdevice statistics (in /proc/net/dev) are per-function
stats so they must use the vadaptor stats. Change the use of
MAC stats to vadaptor stats, and remove any statistics that
can only be measured per-port.  All stats that are removed
will be shown as zeroes when these statistics are displayed.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: DMA the VF stats only when requested
Daniel Pieczko [Tue, 2 Jun 2015 10:39:20 +0000 (11:39 +0100)]
sfc: DMA the VF stats only when requested

Firmware does not support a periodic DMA of vadaptor-stats
on VFs, so only update the stats buffer when stats are
requested (when running "ethtool -S" or an ip/ifconfig
command that reports stats).

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: display vadaptor statistics for all interfaces
Daniel Pieczko [Tue, 2 Jun 2015 10:39:06 +0000 (11:39 +0100)]
sfc: display vadaptor statistics for all interfaces

All interfaces will display vadaptor statistics, so set all the
relevant bits in the stats bitmask. Only functions with the
LINKCTRL flag will see other stats, including (per-port) MAC stats.

The vadaptor stats are from rx_unicast to tx_overflow.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: set the port-id when calling MC_CMD_MAC_STATS
Daniel Pieczko [Tue, 2 Jun 2015 10:38:49 +0000 (11:38 +0100)]
sfc: set the port-id when calling MC_CMD_MAC_STATS

The port-id must be known so that the RMON level can be
set for the collection of vadapter stats.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: add "port_" prefix to MAC stats
Daniel Pieczko [Tue, 2 Jun 2015 10:38:34 +0000 (11:38 +0100)]
sfc: add "port_" prefix to MAC stats

The MAC stats are per-port and will only be displayed on the PF
with control of the link (one per physical port). Vadapter stats
will also be displayed for this PF, so distinguish the MAC stats
by adding a prefix of "port_".

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: Implement ndo_gets_phys_port_id() for EF10 VFs
Shradha Shah [Tue, 2 Jun 2015 10:38:16 +0000 (11:38 +0100)]
sfc: Implement ndo_gets_phys_port_id() for EF10 VFs

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: Add sysfs entry for flags (link control and primary)
Shradha Shah [Tue, 2 Jun 2015 10:37:58 +0000 (11:37 +0100)]
sfc: Add sysfs entry for flags (link control and primary)

On  every adapter there will be one primary PF per adaptor and
one link control PF per port.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: Add paranthesis correctly on all branches of the if statement
Shradha Shah [Tue, 2 Jun 2015 10:37:41 +0000 (11:37 +0100)]
sfc: Add paranthesis correctly on all branches of the if statement

This change is a stylistic change and does not affect
functionality.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosfc: Add code to export port_num in netdev->dev_port
Shradha Shah [Tue, 2 Jun 2015 10:37:25 +0000 (11:37 +0100)]
sfc: Add code to export port_num in netdev->dev_port

In the case where we have multiple functions (PFs and VFs), this
sysfs entry is useful to identify the physical port corresponding
to the function we are interested in.

Signed-off-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'thunderx-next'
David S. Miller [Tue, 2 Jun 2015 19:49:39 +0000 (12:49 -0700)]
Merge branch 'thunderx-next'

Aleksey Makarov says:

====================
net: thunderx: fix problems reported by static check tools

These are fixes for the problems that were reported by static check tools.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: use GFP_KERNEL in thread context
Aleksey Makarov [Tue, 2 Jun 2015 18:00:27 +0000 (11:00 -0700)]
net: thunderx: use GFP_KERNEL in thread context

GFP_KERNEL should be used in the thread context

Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: check if memory allocation was successful
Aleksey Makarov [Tue, 2 Jun 2015 18:00:26 +0000 (11:00 -0700)]
net: thunderx: check if memory allocation was successful

This fixes a coccinelle warning:

coccinelle warnings: (new ones prefixed by >>)

>> drivers/net/ethernet/cavium/thunder/nicvf_queues.c:360:1-11: alloc
>> with no test, possible model on line 367

vim +360 drivers/net/ethernet/cavium/thunder/nicvf_queues.c

   354 err = nicvf_alloc_q_desc_mem(nic, &sq->dmem, q_len,
SND_QUEUE_DESC_SIZE,
   355      NICVF_SQ_BASE_ALIGN_BYTES);
   356 if (err)
   357 return err;
   358
   359 sq->desc = sq->dmem.base;
 > 360 sq->skbuff = kcalloc(q_len, sizeof(u64), GFP_ATOMIC);
   361 sq->head = 0;
   362 sq->tail = 0;
   363 atomic_set(&sq->free_cnt, q_len - 1);
   364 sq->thresh = SND_QUEUE_THRESH;
   365
   366 /* Preallocate memory for TSO segment's header */
 > 367 sq->tso_hdrs = dma_alloc_coherent(&nic->pdev->dev,
   368   q_len *
TSO_HEADER_SIZE,
   369   &sq->tso_hdrs_phys,
GFP_KERNEL);
   370 if (!sq->tso_hdrs)

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: remove unneeded type conversions
Aleksey Makarov [Tue, 2 Jun 2015 18:00:25 +0000 (11:00 -0700)]
net: thunderx: remove unneeded type conversions

No need to cast void* to u8*: pointer arithmetics
works same way for both.

Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: fix nicvf_set_rxfh()
Aleksey Makarov [Tue, 2 Jun 2015 18:00:24 +0000 (11:00 -0700)]
net: thunderx: fix nicvf_set_rxfh()

This fixes a copypaste bug that was discovered by a static analysis
tool:

The patch 4863dea3fab0: "net: Adding support for Cavium ThunderX
network controller" from May 26, 2015, leads to the following static
checker warning:

        drivers/net/ethernet/cavium/thunder/nicvf_ethtool.c:517
nicvf_set_rxfh()
        warn: we tested 'hkey' before and it was 'false'

drivers/net/ethernet/cavium/thunder/nicvf_ethtool.c
   506          /* We do not allow change in unsupported parameters */
   507          if (hkey ||
                    ^^^^
We return here.

   508              (hfunc != ETH_RSS_HASH_NO_CHANGE && hfunc !=
ETH_RSS_HASH_TOP))
   509                  return -EOPNOTSUPP;
   510
   511          rss->enable = true;
   512          if (indir) {
   513                  for (idx = 0; idx < rss->rss_size; idx++)
   514                          rss->ind_tbl[idx] = indir[idx];
   515          }
   516
   517          if (hkey) {
                    ^^^^
So this is dead code.

   518                  memcpy(rss->key, hkey, RSS_HASH_KEY_SIZE *
sizeof(u64));
   519                  nicvf_set_rss_key(nic);
   520          }
   521
   522          nicvf_config_rss(nic);
   523          return 0;
   524  }

regards,
dan carpenter

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: add static
Aleksey Makarov [Tue, 2 Jun 2015 18:00:23 +0000 (11:00 -0700)]
net: thunderx: add static

This fixes sparse messages like this:

drivers/net/ethernet/cavium/thunder/nicvf_main.c:1141:26: sparse: symbol
'nicvf_get_stats64' was not declared. Should it be static?

Also remove unused declarations

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: delete unused variables
Aleksey Makarov [Tue, 2 Jun 2015 18:00:22 +0000 (11:00 -0700)]
net: thunderx: delete unused variables

They were left from development stage

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: rework mac address handling
Aleksey Makarov [Tue, 2 Jun 2015 18:00:21 +0000 (11:00 -0700)]
net: thunderx: rework mac address handling

This fixes sparse message:

drivers/net/ethernet/cavium/thunder/nicvf_main.c:385:40: sparse: cast to
restricted __le64

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: introduce a function for mailbox access
Aleksey Makarov [Tue, 2 Jun 2015 18:00:20 +0000 (11:00 -0700)]
net: thunderx: introduce a function for mailbox access

This fixes sparse message:

drivers/net/ethernet/cavium/thunder/nicvf_main.c:153:25: sparse: cast to
restricted __le64

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: fix constants
Aleksey Makarov [Tue, 2 Jun 2015 18:00:19 +0000 (11:00 -0700)]
net: thunderx: fix constants

This fixes sparse messages like this:

drivers/net/ethernet/cavium/thunder/thunder_bgx.c:897:24: sparse:
constant 0x300000000000 is so big it is long

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Aleksey Makarov <aleksey.makarov@caviumnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: thunderx: Cleanup duplicate NODE_ID macros, add nic_get_node_id()
Robert Richter [Tue, 2 Jun 2015 18:00:18 +0000 (11:00 -0700)]
net: thunderx: Cleanup duplicate NODE_ID macros, add nic_get_node_id()

There are duplicate NODE_ID macro definitions. Move all of them to
nic.h for usage in nic and bgx driver and introduce nic_get_node_id()
helper function.

This patch also fixes 64bit mask which should have been ULL by
reworking the node calculation.

Signed-off-by: Robert Richter <rrichter@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Tue, 2 Jun 2015 05:33:25 +0000 (22:33 -0700)]
Merge git://git./linux/kernel/git/davem/net

Conflicts:
drivers/net/phy/amd-xgbe-phy.c
drivers/net/wireless/iwlwifi/Kconfig
include/net/mac80211.h

iwlwifi/Kconfig and mac80211.h were both trivial overlapping
changes.

The drivers/net/phy/amd-xgbe-phy.c file got removed in 'net-next' and
the bug fix that happened on the 'net' side is already integrated
into the rest of the amd-xgbe driver.

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'cxgb4-next'
David S. Miller [Tue, 2 Jun 2015 05:14:35 +0000 (22:14 -0700)]
Merge branch 'cxgb4-next'

Hariprasad Shenai says:

====================
cxgb4/cxgb4vf: Adds support for Chelsio T6 adapter

This patch series adds the following:
Adds NIC driver support for T6 adapter
Adds vNIC driver support for T6 adapter

This patch series has been created against net-next tree and includes
patches on cxgb4 and cxgb4vf driver.

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

Thanks

V2:
 Fixed compilation issue, when CHELSIO_T4_FCOE is set
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4vf: Adds SRIOV driver changes for T6 adapter
Hariprasad Shenai [Tue, 2 Jun 2015 08:29:40 +0000 (13:59 +0530)]
cxgb4vf: Adds SRIOV driver changes for T6 adapter

Adds vnic driver register related changes for T6 adapter

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4: Adds support for T6 adapter
Hariprasad Shenai [Tue, 2 Jun 2015 08:29:39 +0000 (13:59 +0530)]
cxgb4: Adds support for T6 adapter

Adds NIC driver related changes for T6 adapter. Register related
changes, MC related changes, VF related changes, doorbell related
changes, debugfs changes, etc

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4: Add is_t6 macro and T6 register ranges
Hariprasad Shenai [Tue, 2 Jun 2015 08:29:38 +0000 (13:59 +0530)]
cxgb4: Add is_t6 macro and T6 register ranges

Adds new macro is_t6 and adds the register address range for T6 adapter

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Tue, 2 Jun 2015 03:51:18 +0000 (20:51 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Various VTI tunnel (mark handling, PMTU) bug fixes from Alexander
    Duyck and Steffen Klassert.

 2) Revert ethtool PHY query change, it wasn't correct.  The PHY address
    selected by the driver running the PHY to MAC connection decides
    what PHY address GET ethtool operations return information from.

 3) Fix handling of sequence number bits for encryption IV generation in
    ESP driver, from Herbert Xu.

 4) UDP can return -EAGAIN when we hit a bad checksum on receive, even
    when there are other packets in the receive queue which is wrong.
    Just respect the error returned from the generic socket recv
    datagram helper.  From Eric Dumazet.

 5) Fix BNA driver firmware loading on big-endian systems, from Ivan
    Vecera.

 6) Fix regression in that we were inheriting the congestion control of
    the listening socket for new connections, the intended behavior
    always was to use the default in this case.  From Neal Cardwell.

 7) Fix NULL deref in brcmfmac driver, from Arend van Spriel.

 8) OTP parsing fix in iwlwifi from Liad Kaufman.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (26 commits)
  vti6: Add pmtu handling to vti6_xmit.
  Revert "net: core: 'ethtool' issue with querying phy settings"
  bnx2x: Move statistics implementation into semaphores
  xen: netback: read hotplug script once at start of day.
  xen: netback: fix printf format string warning
  Revert "netfilter: ensure number of counters is >0 in do_replace()"
  net: dsa: Properly propagate errors from dsa_switch_setup_one
  tcp: fix child sockets to use system default congestion control if not set
  udp: fix behavior of wrong checksums
  sfc: free multiple Rx buffers when required
  bna: fix soft lock-up during firmware initialization failure
  bna: remove unreasonable iocpf timer start
  bna: fix firmware loading on big-endian machines
  bridge: fix br_multicast_query_expired() bug
  via-rhine: Resigning as maintainer
  brcmfmac: avoid null pointer access when brcmf_msgbuf_get_pktid() fails
  mac80211: Fix mac80211.h docbook comments
  iwlwifi: nvm: fix otp parsing in 8000 hw family
  iwlwifi: pcie: fix tracking of cmd_in_flight
  ip_vti/ip6_vti: Preserve skb->mark after rcv_cb call
  ...

9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc
Linus Torvalds [Tue, 2 Jun 2015 03:44:51 +0000 (20:44 -0700)]
Merge git://git./linux/kernel/git/davem/sparc

Pull Sparc fixes from David Miller:

 1) Setup the core/threads/sockets bitmaps correctly so that 'lscpus'
    and friends operate properly.  Frtom Chris Hyser.

 2) The bit that normally means "Cached Virtually" on sun4v systems,
    actually changes meaning in M7 and later chips.  Fix from Khalid
    Aziz.

 3) One some PCI-E systems we need to probe different OF properties to
    fill in the PCI slot information properly, from Eric Snowberg.

 4) Kill an extraneous memset after kzalloc(), from Christophe Jaillet.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc: Resolve conflict between sparc v9 and M7 on usage of bit 9 of TTE
  sparc64: pci slots information is not populated in sysfs
  sparc: kernel: GRPCI2: Remove a useless memset
  sparc64: Setup sysfs to mark LDOM sockets, cores and threads correctly

9 years agoMerge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Linus Torvalds [Tue, 2 Jun 2015 01:49:45 +0000 (18:49 -0700)]
Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull virtio fix from Michael Tsirkin:
 "Last-minute virtio fix for 4.1

  This tweaks an exported user-space header to fix build breakage for
  userspace using it"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  include/uapi/linux/virtio_balloon.h: include linux/virtio_types.h

9 years agogeneve: allow user to specify TOS info for tunnel frames
John W. Linville [Mon, 1 Jun 2015 19:51:35 +0000 (15:51 -0400)]
geneve: allow user to specify TOS info for tunnel frames

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agogeneve: allow user to specify TTL for tunnel frames
John W. Linville [Mon, 1 Jun 2015 19:51:34 +0000 (15:51 -0400)]
geneve: allow user to specify TTL for tunnel frames

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'rocker-next'
David S. Miller [Tue, 2 Jun 2015 00:00:09 +0000 (17:00 -0700)]
Merge branch 'rocker-next'

Scott Feldman says:

====================
rocker: enable by default untagged VLAN support

This patch set is a followup to Simon Horman's RFC patch:

   [PATCH/RFC net-next] rocker: by default accept untagged packets

Now, on port probe, we install untagged VLAN (vid=0) support for each port
as the default.  This is equivalent to the command:

   bridge vlan add vid 0 dev DEV self

Accepting untagged VLAN pkts is a reasonable default, but the user could
override this with:

   bridge vlan del vid 0 dev DEV self

With this, we no longer need 8021q module to install vid=0 when port interface
opens.  In fact, we don't need support for legacy VLAN ndo ops at all since
they're superseded by bridge_setlink/dellink.  So remove legacy VLAN ndo ops
support in driver.  (The legacy VLAN ndo ops are supported by bonding/team
drivers, but don't fit into the transaction model offered by switchdev, so
switching all VLAN functions to bridge_setlink/dellink switchdev support gets
us stacked driver + transaction model support).
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: remove support for legacy VLAN ndo ops
Scott Feldman [Mon, 1 Jun 2015 18:39:06 +0000 (11:39 -0700)]
rocker: remove support for legacy VLAN ndo ops

Remove support for legacy ndo ops
.ndo_vlan_rx_add_vid/.ndo_vlan_rx_kill_vid.  Rocker will use
bridge_setlink/dellink exclusively for VLAN add/del operations.

The legacy ops are needed if using 8021q driver module to setup VLANs on
the port.  But an alternative exists in using bridge_setlink/delink to
setup VLANs, which doesn't depend on 8021q module.  So rocker will switch
to the newer setlink/dellink ops.  VLANs can added/delete from the port,
regardless if port is bridged or not, using the bridge commands:

bridge vlan [add|del] vid VID dev DEV self

(Yes, I agree it's confusing to use the "bridge" command to set a VLAN on a
non-bridged port).

Using setlink/dellink over legacy ops let's us handle the stacked driver
case automatically.  It's built-in.  setlink also pass additional flags
(PVID, egress untagged) that aren't available with the legacy ops.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: install/remove router MAC for untagged VLAN when joining/leaving bridge
Scott Feldman [Mon, 1 Jun 2015 18:39:05 +0000 (11:39 -0700)]
rocker: install/remove router MAC for untagged VLAN when joining/leaving bridge

When the port joins a bridge, the port's internal VLAN ID needs to change
to the bridge's internal VLAN ID.  Likewise, when leaving the bridge, the
internal VLAN ID reverts back the port's original internal VLAN ID.  (The
internal VLAN ID is used by device to internally mark untagged pkts with
some VLAN, which will eventually be removed on egress...think PVID).  When
the internal VLAN ID changes, we need to update the VLAN table entries and
the router MAC entries for IP/IPv6 to reflect the new internal VLAN ID.

This patch makes use of the common rocker_port_vlan_add/del functions to
make sure the tables are updated for the current internal VLAN ID.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: install untagged VLAN (vid=0) support for each port
Scott Feldman [Mon, 1 Jun 2015 18:39:04 +0000 (11:39 -0700)]
rocker: install untagged VLAN (vid=0) support for each port

On port probe, install by default untagged VLAN support.  This is
equivalent to running the command:

bridge vlan add vid 0 dev DEV self

A user could, if they wanted, manaully removing untagged support from the
port by running the command:

bridge vlan del vid 0 dev DEV self

But installing it by default on port initialization gives the normal
expected behavior.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: cleanup vlan table on error adding vlan
Scott Feldman [Mon, 1 Jun 2015 18:39:03 +0000 (11:39 -0700)]
rocker: cleanup vlan table on error adding vlan

Basic house keeping: If there is an error adding the router MAC for this
vlan, removing the just installed VLAN table entry to leave device in same
state as before failure.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: zero allocate ports array
Scott Feldman [Mon, 1 Jun 2015 18:39:02 +0000 (11:39 -0700)]
rocker: zero allocate ports array

When allocating the array of rocker port pointers, zero the array values so
we can test for !NULL to see if port is allocated/registered.  We'll need
this later when installing untagged VLAN support for each port, during port
probe.  It's a long story, but to install a VLAN (vid=0 for untagged, in
this case) on a port, we'll need to scan other ports to see if the VLAN
group for that VLAN has been setup.  To scan the other ports, we need to
walk the port array.

Signed-off-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf
David S. Miller [Mon, 1 Jun 2015 23:56:43 +0000 (16:56 -0700)]
Merge git://git./pub/scm/linux/kernel/git/pablo/nf

Pablo Neira Ayuso says:

====================
Netfilter fix for net

The following patch reverts the ebtables chunk that enforces counters that was
introduced in the recently applied d26e2c9ffa38 ('Revert "netfilter: ensure
number of counters is >0 in do_replace()"') since this breaks ebtables.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agovlan: Add GRO support for non hardware accelerated vlan
Toshiaki Makita [Mon, 1 Jun 2015 12:55:06 +0000 (21:55 +0900)]
vlan: Add GRO support for non hardware accelerated vlan

Currently packets with non-hardware-accelerated vlan cannot be handled
by GRO. This causes low performance for 802.1ad and stacked vlan, as their
vlan tags are currently not stripped by hardware.

This patch adds GRO support for non-hardware-accelerated vlan and
improves receive performance of them.

Test Environment:
 vlan device (.1Q) on vlan device (.1ad) on ixgbe (82599)

Result:

- Before

$ netperf -t TCP_STREAM -H 192.168.20.2 -l 60
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 87380  16384  16384    60.00    5233.17

Rx side CPU usage:
  %usr      %sys      %irq     %soft     %idle
  0.27     58.03      0.00     41.70      0.00

- After

$ netperf -t TCP_STREAM -H 192.168.20.2 -l 60
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 87380  16384  16384    60.00    7586.85

Rx side CPU usage:
  %usr      %sys      %irq     %soft     %idle
  0.50     25.83      0.00     59.53     14.14

[ Register VLAN offloads with priority 10 -DaveM ]

Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4: remove unused fn to enable/disable db coalescing
Hariprasad Shenai [Mon, 1 Jun 2015 14:37:41 +0000 (20:07 +0530)]
cxgb4: remove unused fn to enable/disable db coalescing

Remove unused function cxgb4_enable_db_coalescing() and
cxgb4_disable_db_coalescing()

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge tag 'wireless-drivers-for-davem-2015-06-01' of git://git.kernel.org/pub/scm...
David S. Miller [Mon, 1 Jun 2015 23:06:29 +0000 (16:06 -0700)]
Merge tag 'wireless-drivers-for-davem-2015-06-01' of git://git./linux/kernel/git/kvalo/wireless-drivers

Kalle Valo says:

====================
iwlwifi:

* fix OTP parsing 8260
* fix powersave handling for 8260

brcmfmac:

* fix null pointer crash
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: remove rocker parameter from functions that have rocker_port parameter
Simon Horman [Mon, 1 Jun 2015 04:25:04 +0000 (13:25 +0900)]
rocker: remove rocker parameter from functions that have rocker_port parameter

The rocker (switch) of a rocker_port may be trivially obtained from
the latter it seems cleaner not to pass the former to a function when
the latter is being passed anyway.

rocker_port_rx_proc() is omitted from this change as it is a hot path case.

Signed-off-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Scott Feldman <sfeldma@gmail.com>
Acked-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agovti6: Add pmtu handling to vti6_xmit.
Steffen Klassert [Fri, 29 May 2015 18:28:26 +0000 (11:28 -0700)]
vti6: Add pmtu handling to vti6_xmit.

We currently rely on the PMTU discovery of xfrm.
However if a packet is localy sent, the PMTU mechanism
of xfrm tries to to local socket notification what
might not work for applications like ping that don't
check for this. So add pmtu handling to vti6_xmit to
report MTU changes immediately.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobnx2x: Alloc 4k fragment for each rx ring buffer element
Gabriel Krisman Bertazi [Wed, 27 May 2015 16:51:43 +0000 (13:51 -0300)]
bnx2x: Alloc 4k fragment for each rx ring buffer element

The driver allocates one page for each buffer on the rx ring, which is
too much on architectures like ppc64 and can cause unexpected allocation
failures when the system is under stress.  Now, we keep a memory pool
per queue, and if the architecture's PAGE_SIZE is greater than 4k, we
fragment pages and assign each 4k segment to a ring element, which
reduces the overall memory consumption on such architectures.  This
helps avoiding errors like the example below:

[bnx2x_alloc_rx_sge:435(eth1)]Can't alloc sge
[c00000037ffeb900] [d000000075eddeb4] .bnx2x_alloc_rx_sge+0x44/0x200 [bnx2x]
[c00000037ffeb9b0] [d000000075ee0b34] .bnx2x_fill_frag_skb+0x1ac/0x460 [bnx2x]
[c00000037ffebac0] [d000000075ee11f0] .bnx2x_tpa_stop+0x160/0x2e8 [bnx2x]
[c00000037ffebb90] [d000000075ee1560] .bnx2x_rx_int+0x1e8/0xc30 [bnx2x]
[c00000037ffebcd0] [d000000075ee2084] .bnx2x_poll+0xdc/0x3d8 [bnx2x] (unreliable)

Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Reviewed-by: Lino Sanfilippo <LinoSanfilippo@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoopenvswitch: include datapath actions with sampled-packet upcall to userspace
Neil McKee [Wed, 27 May 2015 03:59:43 +0000 (20:59 -0700)]
openvswitch: include datapath actions with sampled-packet upcall to userspace

If new optional attribute OVS_USERSPACE_ATTR_ACTIONS is added to an
OVS_ACTION_ATTR_USERSPACE action, then include the datapath actions
in the upcall.

This Directly associates the sampled packet with the path it takes
through the virtual switch. Path information currently includes mangling,
encapsulation and decapsulation actions for tunneling protocols GRE,
VXLAN, Geneve, MPLS and QinQ, but this extension requires no further
changes to accommodate datapath actions that may be added in the
future.

Adding path information enhances visibility into complex virtual
networks.

Signed-off-by: Neil McKee <neil.mckee@inmon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: Add priority to packet_offload objects.
David S. Miller [Mon, 1 Jun 2015 21:56:09 +0000 (14:56 -0700)]
net: Add priority to packet_offload objects.

When we scan a packet for GRO processing, we want to see the most
common packet types in the front of the offload_base list.

So add a priority field so we can handle this properly.

IPv4/IPv6 get the highest priority with the implicit zero priority
field.

Next comes ethernet with a priority of 10, and then we have the MPLS
types with a priority of 15.

Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Suggested-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoRevert "net: core: 'ethtool' issue with querying phy settings"
David S. Miller [Mon, 1 Jun 2015 21:43:50 +0000 (14:43 -0700)]
Revert "net: core: 'ethtool' issue with querying phy settings"

This reverts commit f96dee13b8e10f00840124255bed1d8b4c6afd6f.

It isn't right, ethtool is meant to manage one PHY instance
per netdevice at a time, and this is selected by the SET
command.  Therefore by definition the GET command must only
return the settings for the configured and selected PHY.

Reported-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobnx2x: Move statistics implementation into semaphores
Yuval Mintz [Mon, 1 Jun 2015 12:08:18 +0000 (15:08 +0300)]
bnx2x: Move statistics implementation into semaphores

Commit dff173de84958 ("bnx2x: Fix statistics locking scheme") changed the
bnx2x locking around statistics state into using a mutex - but the lock
is being accessed via a timer which is forbidden.

[If compiled with CONFIG_DEBUG_MUTEXES, logs show a warning about
accessing the mutex in interrupt context]

This moves the implementation into using a semaphore [with size '1']
instead.

Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoxen: netback: read hotplug script once at start of day.
Ian Campbell [Mon, 1 Jun 2015 10:30:24 +0000 (11:30 +0100)]
xen: netback: read hotplug script once at start of day.

When we come to tear things down in netback_remove() and generate the
uevent it is possible that the xenstore directory has already been
removed (details below).

In such cases netback_uevent() won't be able to read the hotplug
script and will write a xenstore error node.

A recent change to the hypervisor exposed this race such that we now
sometimes lose it (where apparently we didn't ever before).

Instead read the hotplug script configuration during setup and use it
for the lifetime of the backend device.

The apparently more obvious fix of moving the transition to
state=Closed in netback_remove() to after the uevent does not work
because it is possible that we are already in state=Closed (in
reaction to the guest having disconnected as it shutdown). Being
already in Closed means the toolstack is at liberty to start tearing
down the xenstore directories. In principal it might be possible to
arrange to unregister the device sooner (e.g on transition to Closing)
such that xenstore would still be there but this state machine is
fragile and prone to anger...

A modern Xen system only relies on the hotplug uevent for driver
domains, when the backend is in the same domain as the toolstack it
will run the necessary setup/teardown directly in the correct sequence
wrt xenstore changes.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoxen: netback: fix printf format string warning
Ian Campbell [Mon, 1 Jun 2015 10:30:04 +0000 (11:30 +0100)]
xen: netback: fix printf format string warning

drivers/net/xen-netback/netback.c: In function ‘xenvif_tx_build_gops’:
drivers/net/xen-netback/netback.c:1253:8: warning: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 5 has type ‘int’ [-Wformat=]
        (txreq.offset&~PAGE_MASK) + txreq.size);
        ^

PAGE_MASK's type can vary by arch, so a cast is needed.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
----
v2: Cast to unsigned long, since PAGE_MASK can vary by arch.
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoRevert "netfilter: ensure number of counters is >0 in do_replace()"
Bernhard Thaler [Thu, 28 May 2015 08:26:18 +0000 (10:26 +0200)]
Revert "netfilter: ensure number of counters is >0 in do_replace()"

This partially reverts commit 1086bbe97a07 ("netfilter: ensure number of
counters is >0 in do_replace()") in net/bridge/netfilter/ebtables.c.

Setting rules with ebtables does not work any more with 1086bbe97a07 place.

There is an error message and no rules set in the end.

e.g.

~# ebtables -t nat -A POSTROUTING --src 12:34:56:78:9a:bc -j DROP
Unable to update the kernel. Two possible causes:
1. Multiple ebtables programs were executing simultaneously. The ebtables
   userspace tool doesn't by default support multiple ebtables programs
running

Reverting the ebtables part of 1086bbe97a07 makes this work again.

Signed-off-by: Bernhard Thaler <bernhard.thaler@wvnet.at>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
9 years agoinclude/uapi/linux/virtio_balloon.h: include linux/virtio_types.h
Mikko Rapeli [Sat, 30 May 2015 15:39:25 +0000 (17:39 +0200)]
include/uapi/linux/virtio_balloon.h: include linux/virtio_types.h

Fixes userspace compilation error:

error: unknown type name ‘__virtio16’
  __virtio16 tag;

Signed-off-by: Mikko Rapeli <mikko.rapeli@iki.fi>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoxen-netfront: Use setup_timer
Vaishali Thakkar [Mon, 1 Jun 2015 04:58:37 +0000 (10:28 +0530)]
xen-netfront: Use setup_timer

Use the timer API function setup_timer instead of structure field
assignments to initialize a timer.

A simplified version of the Coccinelle semantic patch that performs
this transformation is as follows:

@change@
expression e, func, da;
@@

-init_timer (&e);
+setup_timer (&e, func, da);
-e.data = da;
-e.function = func;

Signed-off-by: Vaishali Thakkar <vthakkar1994@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosparc: Resolve conflict between sparc v9 and M7 on usage of bit 9 of TTE
Khalid Aziz [Wed, 27 May 2015 16:00:46 +0000 (10:00 -0600)]
sparc: Resolve conflict between sparc v9 and M7 on usage of bit 9 of TTE

sparc: Resolve conflict between sparc v9 and M7 on usage of bit 9 of TTE

Bit 9 of TTE is CV (Cacheable in V-cache) on sparc v9 processor while
the same bit 9 is MCDE (Memory Corruption Detection Enable) on M7
processor. This creates a conflicting usage of the same bit. Kernel
sets TTE.cv bit on all pages for sun4v architecture which works well
for sparc v9 but enables memory corruption detection on M7 processor
which is not the intent. This patch adds code to determine if kernel
is running on M7 processor and takes steps to not enable memory
corruption detection in TTE erroneously.

Signed-off-by: Khalid Aziz <khalid.aziz@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosparc64: pci slots information is not populated in sysfs
Eric Snowberg [Wed, 27 May 2015 15:59:19 +0000 (11:59 -0400)]
sparc64: pci slots information is not populated in sysfs

Add PCI slot numbers within sysfs for PCIe hardware.  Larger
PCIe systems with nested PCI bridges and slots further
down on these bridges were not being populated within sysfs.
This will add ACPI style PCI slot numbers for these systems
since the OF 'slot-names' information is not available on
all PCIe platforms.

Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agosparc: kernel: GRPCI2: Remove a useless memset
Christophe Jaillet [Fri, 1 May 2015 12:05:39 +0000 (14:05 +0200)]
sparc: kernel: GRPCI2: Remove a useless memset

grpci2priv is allocated using kzalloc, so there is no need to memset it.

Signed-off-by: Christophe Jaillet <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: dsa: Properly propagate errors from dsa_switch_setup_one
Florian Fainelli [Fri, 29 May 2015 17:29:46 +0000 (10:29 -0700)]
net: dsa: Properly propagate errors from dsa_switch_setup_one

While shuffling some code around, dsa_switch_setup_one() was introduced,
and it was modified to return either an error code using ERR_PTR() or a
NULL pointer when running out of memory or failing to setup a switch.

This is a problem for its caler: dsa_switch_setup() which uses IS_ERR()
and expects to find an error code, not a NULL pointer, so we still try
to proceed with dsa_switch_setup() and operate on invalid memory
addresses. This can be easily reproduced by having e.g: the bcm_sf2
driver built-in, but having no such switch, such that drv->setup will
fail.

Fix this by using PTR_ERR() consistently which is both more informative
and avoids for the caller to use IS_ERR_OR_NULL().

Fixes: df197195a5248 ("net: dsa: split dsa_switch_setup into two functions")
Reported-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agotcp: fix child sockets to use system default congestion control if not set
Neal Cardwell [Fri, 29 May 2015 17:47:07 +0000 (13:47 -0400)]
tcp: fix child sockets to use system default congestion control if not set

Linux 3.17 and earlier are explicitly engineered so that if the app
doesn't specifically request a CC module on a listener before the SYN
arrives, then the child gets the system default CC when the connection
is established. See tcp_init_congestion_control() in 3.17 or earlier,
which says "if no choice made yet assign the current value set as
default". The change ("net: tcp: assign tcp cong_ops when tcp sk is
created") altered these semantics, so that children got their parent
listener's congestion control even if the system default had changed
after the listener was created.

This commit returns to those original semantics from 3.17 and earlier,
since they are the original semantics from 2007 in 4d4d3d1e8 ("[TCP]:
Congestion control initialization."), and some Linux congestion
control workflows depend on that.

In summary, if a listener socket specifically sets TCP_CONGESTION to
"x", or the route locks the CC module to "x", then the child gets
"x". Otherwise the child gets current system default from
net.ipv4.tcp_congestion_control. That's the behavior in 3.17 and
earlier, and this commit restores that.

Fixes: 55d8694fa82c ("net: tcp: assign tcp cong_ops when tcp sk is created")
Cc: Florian Westphal <fw@strlen.de>
Cc: Daniel Borkmann <dborkman@redhat.com>
Cc: Glenn Judd <glenn.judd@morganstanley.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'rds-next'
David S. Miller [Mon, 1 Jun 2015 04:47:23 +0000 (21:47 -0700)]
Merge branch 'rds-next'

Sowmini Varadhan says:

====================
net/rds: SOL_RDS socket option to explicitly select transport

Today the underlying transport (TCP or IB) for a PF_RDS socket is
implicitly selected based on the local address used to bind(2) the
PF_RDS socket. This results in some non-deterministic behavior when
there are un-numbered and IPoIB interfaces sharing the same IP address.
It also places the constraint that the IB interface must have an IP
address (and thus, IPoIB) configured on it.

The non-determinism may be avoided by providing the user-space application
a socket option that allows it to explicitly select the transport
prior to bind(2).

Patch 1 of this series provides the constant definitions needed by
the application via <linux/rds.h>.

Patch 2 provides the setsockopt support, and Patch 3 provides the
getsockopt support.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet/rds Add getsockopt support for SO_RDS_TRANSPORT
Sowmini Varadhan [Fri, 29 May 2015 21:28:09 +0000 (17:28 -0400)]
net/rds Add getsockopt support for SO_RDS_TRANSPORT

The currently attached transport for a PF_RDS socket may be obtained
from user space by invoking getsockopt(2) using the SO_RDS_TRANSPORT
option at the SOL_RDS level. The integer optval returned will be one
of the RDS_TRANS_* constants defined in linux/rds.h.

Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet/rds: Add setsockopt support for SO_RDS_TRANSPORT
Sowmini Varadhan [Fri, 29 May 2015 21:28:08 +0000 (17:28 -0400)]
net/rds: Add setsockopt support for SO_RDS_TRANSPORT

An application may deterministically attach the underlying transport for
a PF_RDS socket by invoking setsockopt(2) with the SO_RDS_TRANSPORT
option at the SOL_RDS level. The integer argument to setsockopt must be
one of the RDS_TRANS_* transport types, e.g., RDS_TRANS_TCP. The option
must be specified before invoking bind(2) on the socket, and may only
be used once on the socket. An attempt to set the option on a bound
socket, or to invoke the option after a successful SO_RDS_TRANSPORT
attachment, will return EOPNOTSUPP.

Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>