Net_device is a vast and important structure, but it has no kernel-doc
compliant documentation. This patch extracts the comments from the structure
to clean it up, and let the scripts extract documentation from it. I know that
the patch is big, but it's just reordering of comments into the appropriate
form, and adding a few more, for the missing members.
Signed-off-by: Karoly Kemeny <karoly.kemeny@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Vince Bridgers says:

====================
net: stmmac: Improve mcast/ucast filter for snps

This patch series adds Synopsys specific bindings for the Synopsys EMAC
filter characteristics since those are implementation dependent. The
multicast and unicast filtering code was improved to handle different
configuration variations based on device tree settings.

I verified the operation of the multicast and unicast filters through
Synopsys support as requested during the V1 review, and tested the GMAC
configuration on an Altera Cyclone 5 SOC (which supports 256 multicast
bins and 128 Unicast addresses). The 10/100 variant of this driver
modification was not tested, although it was compile tested. I shared
the email thread results of the investigation through Synopsys with the
stmmac maintainer.

V4: Remove patch from series that addressed a sparse issue from a
    down rev'd version of sparse that does not show up in the
    latest version of sparse.
V3: Break up the patch into interface and functional change patches
    per review comments
V2: Confirm with Synopsys methods to determine number of Multicast bins
    and Unicast address filter entries per first round review comments.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds and modifies code to support multiple Multicast and Unicast
Synopsys MAC filter configurations. The default configuration is defined to
support legacy driver behavior, which is 64 Multicast bins. The Unicast
filter code previously assumed all controllers support 32 or 16 Unicast
addresses based on controller version number, but this has been corrected
to support a default of 1 Unicast address. The filter configuration may
be specified through the devicetree using a Synopsys specific device tree
entry. This information was verified with Synopsys through
Synopsys Support Case #8000684337 and shared with the maintainer.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds socfpga Ethernet filter attributes for multicast
and unicast filters per Synopsys Ethernet IP configuration chosen
by Altera for the Cyclone 5 and Arria SOC FPGAs.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This change adds bindings for the number of multicast hash bins and perfect
filter entries supported by the Synopsys EMAC. The Synopsys EMAC core is
configurable at device creation time, and can be configured for a different
number of multicast hash bins and a different number of perfect filter
entries. The device does not provide a way to query these parameters,
therefore parameters are required. The Altera Cyclone V SOC has support for
256 multicast hash bins and 128 perfect filter entries, and is different
than what's currently provided in the stmmac driver.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch removes the check for the number of mulitcast addresses
when using hash based filtering since it's not necessary. If the number
of multicast addresses in the list exceeds the number of multicast hash
bins, the bins will "fold" over into one of the bins configured and
enabled for the particular component instance.

The default number of maximum unicast addresses was changed from 32 to 1
since this number is not dependent on the component revision. The maximum
number of multicast and unicast addresses is dependent on the configuration
of the Synopsys EMAC configured by the SOC architect at the time the
features were selected and configured for a particular component. Sadly,
Synopsys does not provide a way to query the precise number supported
by a particular component, so we must fall back on a devicetree entry.
This configuration could vary from vendor to vendor (such as STMicro,
Altera, etc).

The multicast bins are set for every possible filtering case (including
no entries) - previously the bits were set only if multicast filter entries
were present.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The synopsys EMAC can be configured for different numbers of multicast hash
bins and perfect filter entries at device creation time and there's no way
to query this configuration information at runtime. As a result, a devicetree
parameter is required in order for the driver to program these filters
correctly for a particular device instance. This patch modifies the
10/100/1000 MAC software interface such that these configuration parameters
can be set at initialization time.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

====================
Netfilter updates for net-next

The following patchset contains netfilter updates for net-next, they are:

1) Add the reject expression for the nf_tables bridge family, this
   allows us to send explicit reject (TCP RST / ICMP dest unrech) to
   the packets matching a rule.

2) Simplify and consolidate the nf_tables set dumping logic. This uses
   netlink control->data to filter out depending on the request.

3) Perform garbage collection in xt_hashlimit using a workqueue instead
   of a timer, which is problematic when many entries are in place in
   the tables, from Eric Dumazet.

4) Remove leftover code from the removed ulog target support, from
   Paul Bolle.

5) Dump unmodified flags in the netfilter packet accounting when resetting
   counters, so userspace knows that a counter was in overquota situation,
   from Alexey Perevalov.

6) Fix wrong usage of the bitwise functions in nfnetlink_acct, also from
   Alexey.

7) Fix a crash when adding new set element with an empty NFTA_SET_ELEM_LIST
   attribute.

This patchset also includes a couple of cleanups for xt_LED from
Duan Jiong and for nf_conntrack_ipv4 (using coccinelle) from
Himangi Saraogi.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

commit d23ff701 (tcp: add generic netlink support for tcp_metrics) introduced
netlink support for the new tcp_metrics, however it restricted getting of
tcp_metrics to root user only. This is a change from how these values could
have been fetched when in the old route cache. Unless there's a legitimate
reason to restrict the reading of these values it would be better if normal
users could fetch them.

Cc: Julian Anastasov <ja@ssi.bg>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Debabrata Banerjee <dbanerje@akamai.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 34c5bd66 introduced the possibility that an
uninitialized pointer on the stack (orig_fp) can call into
sk_unattached_filter_destroy() when its value is non NULL.

Before that commit orig_fp was only destroyed in the same
block where it was assigned a valid BPF prog before. Fix it
up by initializing it to NULL.

Fixes: 34c5bd66 ("net: filter: don't release unattached filter through call_rcu()")
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Pablo Neira <pablo@netfilter.org>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>

Otherwise, the kernel oopses in nla_for_each_nested when iterating over
the unset attribute NFTA_SET_ELEM_LIST_ELEMENTS in the
nf_tables_{new,del}setelem() path.

netlink: 65524 bytes leftover after parsing attributes in process `nft'.
[...]
Oops: 0000 [#1] SMP
[...]
CPU: 2 PID: 6287 Comm: nft Not tainted 3.16.0-rc2+ #169
RIP: 0010:[<ffffffffa0526e61>]  [<ffffffffa0526e61>] nf_tables_newsetelem+0x82/0xec [nf_tables]
[...]
Call Trace:
 [<ffffffffa05178c4>] nfnetlink_rcv+0x2e7/0x3d7 [nfnetlink]
 [<ffffffffa0517939>] ? nfnetlink_rcv+0x35c/0x3d7 [nfnetlink]
 [<ffffffff8137d300>] netlink_unicast+0xf8/0x17a
 [<ffffffff8137d6a5>] netlink_sendmsg+0x323/0x351
[...]

Fix this by returning -EINVAL if this attribute is not set, which
doesn't make sense at all since those commands are there to add and to
delete elements from the set.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

Bit helper functions were used for manipulation with NFACCT_F_OVERQUOTA,
but they are accepting pit position, but not a bit mask. As a result
not a third bit for NFACCT_F_OVERQUOTA was set, but forth. Such
behaviour was dangarous and could lead to unexpected overquota report
result.
Signed-off-by: Alexey Perevalov <a.perevalov@samsung.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

Steffen Klassert says:

====================
pull request (net-next): ipsec-next 2014-07-30

This is the last pull request for ipsec-next before I'll be
off for two weeks starting on friday. David, can you please
take urgent ipsec patches directly into net/net-next during
this time?

1) Error handling simplifications for vti and vti6.
   From Mathias Krause.

2) Remove a duplicate semicolon after a return statement.
   From Christoph Paasch.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Vince Bridgers says:

====================
net: libphy: Add phy specific functions to access mmd regs

This set of patches addresses a problem found with the Micrel ksz9021 phy and
libphy, where the ksz9021 phy does not support mmd extended register access
per the IEEE specification as assumed by libphy. The first patch adds a
framework for phy specific support to specify their own function to access
extended phy registers, return a failure code if not supported, or to default
to libphy's IEEE defined method for accessing the mmd extended phy registers.

This issue was found by using the Synopsys EMAC and a Micrel ksz9021 phy on the
Altera Cyclone 5 SOC development kit. This patch was tested on the same system
in both positive and negative test cases.

V5: Revert name of mmd register access functions, check for phy specific
    driver override functions in mmd register access functions per
    Florian's comments to minimize source code changes
V4: Correct error when formatting V3 patch - erroneous text cut from code
V3: Correct formatting of function arguments, remove return statement from
    NULL functions, and add patch for PHY driver documentation per review
    comments.
V2: Split the original patch submission into seperate patches for the libphy
    framework required for the modification and for the Micrel Phy.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Update the PHY library documentation to describe how a specific PHY
driver can use the PAL MMD register access routines or override those
routines with it's own in the event the PHY does not support the IEEE
standard for reading and writing MMD phy registers.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The Micrel ksz9021 PHY does not support standard IEEE standard MMD
extended register access, therefore requires stubs to fail the read
register method and do nothing for the write register method when
libphy attempts to read and/or configure Energy Efficient Ethernet
features in PHYS that do support those features. This problem
was observed on an Altera Cyclone V SOC development kit that
uses the Synopsys EMAC and the Micrel ksz9021 PHY. This patch
was tested on the same board, and Energy Efficient Ethernet is
now disabled as expected since the Micrel PHY does not support that
feature.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

libphy was originally written assuming all phy devices support clause 45
access extensions to the mmd registers through the indirection registers
located within the first 16 phy registers. This assumption is not true
in all cases, and one specific example is the Micrel ksz9021 10/100/1000
Mbps phy. Using the stmmac driver, accessing the mmd registers to query
and configure energy efficient Ethernet (EEE) features yielded unexpected
behavior.

This patch adds mmd access functions to the phy driver that can be
overriden by the phy specific driver if the phy does not support this
mechanism or uses it's own non-standard access mechanism. By default,
the IEEE Compatible clause 45 access mechanism described in clause 22
is used. With this patch, EEE query/configure functions as expected
using the stmmac and the Micrel ksz9021 phy.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Shruti Kanetkar <Shruti@Freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix one misspelled word reported by codespell.
Signed-off-by: Madalin Bucur <madalin.bucur@freescale.com>
Signed-off-by: Shruti Kanetkar <Shruti@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sk_unattached_filter_destroy() does not always need to release the
filter object via rcu. Since this filter is never attached to the
socket, the caller should be responsible for releasing the filter
in a safe way, which may not necessarily imply rcu.

This is a short summary of clients of this function:

1) xt_bpf.c and cls_bpf.c use the bpf matchers from rules, these rules
   are removed from the packet path before the filter is released. Thus,
   the framework makes sure the filter is safely removed.

2) In the ppp driver, the ppp_lock ensures serialization between the
   xmit and filter attachment/detachment path. This doesn't use rcu
   so deferred release via rcu makes no sense.

3) In the isdn/ppp driver, it is called from isdn_ppp_release()
   the isdn_ppp_ioctl(). This driver uses mutex and spinlocks, no rcu.
   Thus, deferred rcu makes no sense to me either, the deferred releases
   may be just masking the effects of wrong locking strategy, which
   should be fixed in the driver itself.

4) In the team driver, this is the only place where the rcu
   synchronization with unattached filter is used. Therefore, this
   patch introduces synchronize_rcu() which is called from the
   genetlink path to make sure the filter doesn't go away while packets
   are still walking over it. I think we can revisit this once struct
   bpf_prog (that only wraps specific bpf code bits) is in place, then
   add some specific struct rcu_head in the scope of the team driver if
   Jiri thinks this is needed.

Deferred rcu release for unattached filters was originally introduced
in 302d6637 ("filter: Allow to create sk-unattached filters").
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

This structure is not exposed to userspace, so fix this by defining
struct sk_filter; so we skip the casting in kernelspace. This is safe
since userspace has no way to lurk with that internal pointer.

Fixes: e6f30c73 ("netfilter: x_tables: add xt_bpf match")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Correct typo in the name of the type given to sizeof.  Because it is the
size of a pointer that is wanted, the typo has no impact on compilation or
execution.

This problem was found using Coccinelle (http://coccinelle.lip6.fr/).  The
semantic patch used can be found in message 0 of this patch series.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

Tom Lendacky says:

====================
amd-xgbe: AMD XGBE driver update 2014-07-25

This patch series is dependent on the following patch that was
applied to the net tree and needs to be applied to the net-next
tree:
  332cfc82 - amd-xgbe: Fix error return code in xgbe_probe()

The following series of patches includes fixes and new support in the
driver.

- Device bindings documentation update
- Hardware timestamp support
- 2.5GbE support changes
- Fifo sizes based on active queues/rings
- Phylib driver updates for:
  - Rate change completion check
  - KR training initiation
  - Auto-negotiation results
- Traffic class support, including DCB support

This patch series is based on net-next.

Changes in V2:
  - Remove DBGPR(...., __func__) calls
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds support for traffic classes as well as support
for Data Center Bridging interfaces related to traffic classes
and priority flow control.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a netdev_info statement detailing whether auto-negotiation was
completed through parallel detection or through the auto-negotiation
protocol.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

As part of changing rates to KR mode, KR training is initiated. If
the KR training is restarted it is possible to enter an invalid logic
state. This can be avoided by asserting a training reset bit before
initiating the KR training and then clearing the training reset bit.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, the logic will loop endlessly waiting for a rate change
to complete.  Add a counter so that if the rate change signals
never indicate complete the loop will eventually exit.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When setting the fifo sizes for the queues and enabling the queues
use the number of active Tx and Rx queues that have been enabled
not the maximum number available.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Update the amd-xgbe driver and phylib driver to better support
the 2.5GbE mode for the hardware. In order to be able establish
2.5GbE using clause 73 auto negotiation the device will support
speed sets of 1GbE/10GbE and 2.5GbE/10GbE.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds support for Tx and Rx hardware timestamping.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

An earlier patch added support for the "dma-coherent" device property.
This patch adds this optional property to the amd-xgbe device bindings
documentation.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

No need for the unlikely(), WARN_ON() and BUG_ON() internally use
unlikely() on the condition.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Current explanation of dcb_app->priority is wrong. It says priority is
expected to be a 3-bit unsigned integer which is only true when working with
DCBx-IEEE. Use of dcb_app->priority by DCBx-CEE expects it to be 802.1p user
priority bitmap. Updated accordingly

This affects the cxgb4 driver, but I will post those changes as part of a
larger changeset shortly.

Fixes: 3e29027a ("dcbnl: add support for ieee8021Qaz attributes")
Signed-off-by: Anish Bhatt <anish@chelsio.com>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds platform init/exit functions and modifications to support
suspend/resume for the Altera Cyclone 5 SOC Ethernet controller. The platform
exit function puts the controller into reset using the socfpga reset
controller driver. The platform init function sets up the Synopsys mac by
first making sure the Ethernet controller is held in reset, programming the
phy mode through external support logic, then deasserts reset through
the socfpga reset manager driver.
Signed-off-by: Vince Bridgers <vbridgers2013@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Eli Cohen says:

====================
mlx5 driver changes related to PCI handling ***

The first of these patches is changing the pci device driver from mlx5_ib to
mlx5_core in a similar manner it is done in mlx4. This set the grounds for us
to introduce Ethernet driver for HW which uses mlx5.

The other two patches contain minor fixes.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

In the event flow, we currently pass only a port number in the
void *data argument.  Rather than pass a pointer to the event handlers,
we should use an "unsigned long" parameter, and pass the port number
value directly.

In the future, if necessary for some events, we can use the unsigned long
parameter to pass a pointer.

Based on a patch by Eli Cohen <eli@mellanox.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There were many places where parameters which should be u8/u16 were
integer type.

Additionally, in 2 places, a check for a non-null pointer was added
before dereferencing the pointer (this is actually a bug fix).
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In preparation for a new mlx5 device which is VPI (i.e., ports can be
either IB or ETH), move the pci device functionality from mlx5_ib
to mlx5_core.

This involves the following changes:
1. Move mlx5_core_dev struct out of mlx5_ib_dev. mlx5_core_dev
   is now an independent structure maintained by mlx5_core.
   mlx5_ib_dev now has a pointer to that struct.
   This requires changing a lot of places where the core_dev
   struct was accessed via mlx5_ib_dev (now, this needs to
   be a pointer dereference).
2. All PCI initializations are now done in mlx5_core. Thus,
   it is now mlx5_core which does pci_register_device (and not
   mlx5_ib, as was previously).
3. mlx5_ib now registers itself with mlx5_core as an "interface"
   driver. This is very similar to the mechanism employed for
   the mlx4 (ConnectX) driver. Once the HCA is initialized
   (by mlx5_core), it invokes the interface drivers to do
   their initializations.
4. There is a new event handler which the core registers:
   mlx5_core_event(). This event handler invokes the
   event handlers registered by the interfaces.

Based on a patch by Eli Cohen <eli@mellanox.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, we have a 3-stage seeding process in prandom():

Phase 1 is from the early actual initialization of prandom()
subsystem which happens during core_initcall() and remains
most likely until the beginning of late_initcall() phase.
Here, the system might not have enough entropy available
for seeding with strong randomness from the random driver.
That means, we currently have a 32bit weak LCG() seeding
the PRNG status register 1 and mixing that successively
into the other 3 registers just to get it up and running.

Phase 2 starts with late_initcall() phase resp. when the
random driver has initialized its non-blocking pool with
enough entropy. At that time, we throw away *all* inner
state from its 4 registers and do a full reseed with strong
randomness.

Phase 3 starts right after that and does a periodic reseed
with random slack of status register 1 by a strong random
source again.

A problem in phase 1 is that during bootup data structures
can be initialized, e.g. on module load time, and thus access
a weakly seeded prandom and are never changed for the rest
of their live-time, thus carrying along the results from a
week seed. Lets make sure that current but also future users
access a possibly better early seeded prandom.

This patch therefore improves phase 1 by trying to make it
more 'unpredictable' through mixing in seed from a possible
hardware source. Now, the mix-in xors inner state with the
outcome of either of the two functions arch_get_random_{,seed}_int(),
preferably arch_get_random_seed_int() as it likely represents
a non-deterministic random bit generator in hw rather than
a cryptographically secure PRNG in hw. However, not all might
have the first one, so we use the PRNG as a fallback if
available. As we xor the seed into the current state, the
worst case would be that a hardware source could be unverifiable
compromised or backdoored. In that case nevertheless it
would be as good as our original early seeding function
prandom_seed_very_weak() since we mix through xor which is
entropy preserving.

Joint work with Daniel Borkmann.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>