When the VLAN tag is present in the packet buffer (i.e VLAN stripping disabled, QinQ)
the driver will currently report CHECKSUM_UNNECESSARY.
Instead of using CHECKSUM_COMPLETE offload for packets with first
ethertype of IPv4/6, use it for packets with last ethertype of IPv4/6 to
cover the former cases as well.

The checksum field present in the CQE is calculated from the IP header
until the end of the packet. When the first ethertype is different than
IPv4/6 (for ex. 802.1Q VLAN) a checksum of the VLAN header/s should be
added. The small header/s checksum calculation will allow us to use
CHECKSUM_COMPLETE instead of CHECKSUM_UNNECESSARY.

Testing bandwidth of one and 8 TCP streams to a single RQ,
LRO and VLAN stripping offloads disabled:
CPU: Intel(R) Xeon(R) CPU E5-2660 v2 @ 2.20GHz
NIC: Mellanox Technologies MT27710 Family [ConnectX-4 Lx]

Before:
+--------------+--------------------+---------------------+----------------------+
| Traffic type | 1 Stream BW [Mbps] | 8 Streams BW [Mbps] |   Checksum offload   |
+--------------+--------------------+---------------------+----------------------+
| Untagged     |          28,247.35 |           24,716.88 | CHECKSUM_COMPLETE    |
| VLAN         |          27,516.69 |           23,752.26 | CHECKSUM_UNNECESSARY |
| QinQ         |           6,961.30 |           20,667.04 | CHECKSUM_UNNECESSARY |
+--------------+--------------------+---------------------+----------------------+

Now:
+--------------+--------------------+---------------------+-------------------+
| Traffic type | 1 Stream BW [Mbps] | 8 Streams BW [Mbps] | Checksum offload  |
+--------------+--------------------+---------------------+-------------------+
| Untagged     |          28,521.28 |           24,926.32 | CHECKSUM_COMPLETE |
| VLAN         |          27,389.37 |           23,715.34 | CHECKSUM_COMPLETE |
| QinQ         |           6,901.77 |           20,845.73 | CHECKSUM_COMPLETE |
+--------------+--------------------+---------------------+-------------------+

No performance degradation observed.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

The following counters are now exposed through ethtool -S:
rx[i]_removed_vlan_packets (per channel)
rx_removed_vlan_packets
tx[i]_added_vlan_packets (per channel)
tx_added_vlan_packets

rx_removed_vlan_packets: The number of packets that had their
outer VLAN header stripped to the CQE by the hardware.
tx_added_vlan_packets: The number of packets that had their
outer VLAN header inserted by the hardware.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Report VLAN insertion support for S-tagged packets and add support by
choosing the correct VLAN type in the WQE.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

When a user chooses to use 802.1ad VLAN the proper steering rules will
be added to the VLAN flow table (matching the specific S-tag VID).
Due to current hardware limitation, when using 802.1ad, we must disable
C-tag VLAN stripping on the RQs.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Replace explicit declaration of bitmap with DECLARE_BITMAP kernel macro.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Extend the net device error logging with netdev_*_once macros.
netdev_*_once are the equivalents of the dev_*_once macros which are
useful for messages that should only be logged once.

Also add netdev_WARN_ONCE, which is the "once" extension for the already
existing netdev_WARN macro.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

When add VLAN rule fails the active vlan bit should be cleared.

Fixes: afb736e9 ("net/mlx5: Ethernet resource handling files")
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Rename VLAN related symbols to better reflect the fact that they
are associated to C-tag VLAN.
Signed-off-by: Gal Pressman <galp@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Vivien Didelot says:

====================
net: dsa: setup stage

When probing a DSA switch, there is basically two stages.

The first stage is the parsing of the switch device, from either device
tree or platform data. It fetches the DSA tree to which it belongs, and
validates its ports. The switch device is then added to the tree, and
the second stage is called if this was the last switch of the tree.

The second stage is the setup of the tree, which validates that the tree
is complete, sets up the routing tables, the default CPU port for user
ports, sets up the switch drivers and finally the master interfaces,
which makes the whole switch fabric functional.

This patch series covers the second setup stage. The setup and teardown
of a switch tree have been separated into logical steps, and the probing
of a switch now simply parses and adds a switch to a tree.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This commit brings no functional changes. It gets rid of the underscore
prefixed _dsa_register_switch and _dsa_unregister_switch functions in
favor of dsa_switch_probe() which parses and adds a switch to a tree and
dsa_switch_remove() which removes a switch from a tree.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Now that the tree setup is centralized, we can simplify the code a bit
more by setting up or tearing down the tree directly when adding or
removing a switch to/from it.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The *_complete() functions take too much arguments to do only one thing:
they try to fetch the dsa_port structures corresponding to device nodes
under the "link" list property of DSA ports, and use them to setup the
routing table of switches.

This patch simplifies them by providing instead simpler
dsa_{port,switch,tree}_setup_routing_table functions which return a
boolean value, true if the tree is complete.

dsa_tree_setup_routing_table is called inside dsa_tree_setup which
simplifies the switch registering function as well.

A switch's routing table is now initialized before its setup.

This also makes dsa_port_is_valid obsolete, remove it.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The OF code provides a of_for_each_phandle() helper to iterate over
phandles. Use it instead of arbitrary iterating ourselves over the list
of phandles hanging to the "link" property of the port's device node.

The of_phandle_iterator_next() helper calls of_node_put() itself on
it.node. Thus We must only do it ourselves if we break the loop.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Instead of having two dsa_ds_find_port_dn (which returns a bool) and
dsa_dst_find_port_dn (which returns a switch) functions, provide a more
explicit dsa_tree_find_port_by_node function which returns a matching
port.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The dsa_dsa_port_apply and dsa_cpu_port_apply functions do exactly the
same. The dsa_user_port_apply function does not try to register a fixed
link but try to create a slave.

This commit factorizes and scopes all that in two convenient
dsa_port_setup and dsa_port_teardown functions.

It won't hurt to register a devlink_port for unused port as well.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patches brings no functional changes. It removes the unused dst
argument from the dsa_ds_apply and dsa_ds_unapply functions, rename them
to dsa_switch_setup and dsa_switch_teardown for a more explicit scope.

This clarifies the steps of the setup or teardown of a switch fabric.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This commit provides better scope for the DSA tree setup and teardown
functions. It renames the "applied" bool to "setup" and print a message
when the tree is setup, as it is done during teardown.

At the same time, check dst->setup in dsa_tree_setup, where it is set to
true.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add DSA helpers to setup and teardown a master net device wired to its
CPU port. This centralizes the dsa_ptr assignment.

This also makes the master ethtool helpers static at the same time.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The dsa_dst_parse function called just before dsa_dst_apply does not
parse the tree but does only one thing: it assigns the default CPU port
to dst->cpu_dp and to each user ports.

This patch simplifies this by calling a dsa_tree_setup_default_cpu
function at the beginning of dsa_dst_apply directly.

A dsa_port_is_user helper is added for convenience.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

A DSA port has a dedicated CPU port assigned to it, stored in the cpu_dp
member. It is not meant to be modified by a port, thus make it const.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

v16->17
 - Fixed disputed check code: keep them in nsh_push and nsh_pop
   but also add them in __ovs_nla_copy_actions

v15->v16
 - Add csum recalculation for nsh_push, nsh_pop and set_nsh
   pointed out by Pravin
 - Move nsh key into the union with ipv4 and ipv6 and add
   check for nsh key in match_validate pointed out by Pravin
 - Add nsh check in validate_set and __ovs_nla_copy_actions

v14->v15
 - Check size in nsh_hdr_from_nlattr
 - Fixed four small issues pointed out By Jiri and Eric

v13->v14
 - Rename skb_push_nsh to nsh_push per Dave's comment
 - Rename skb_pop_nsh to nsh_pop per Dave's comment

v12->v13
 - Fix NSH header length check in set_nsh

v11->v12
 - Fix missing changes old comments pointed out
 - Fix new comments for v11

v10->v11
 - Fix the left three disputable comments for v9
   but not fixed in v10.

v9->v10
 - Change struct ovs_key_nsh to
       struct ovs_nsh_key_base base;
       __be32 context[NSH_MD1_CONTEXT_SIZE];
 - Fix new comments for v9

v8->v9
 - Fix build error reported by daily intel build
   because nsh module isn't selected by openvswitch

v7->v8
 - Rework nested value and mask for OVS_KEY_ATTR_NSH
 - Change pop_nsh to adapt to nsh kernel module
 - Fix many issues per comments from Jiri Benc

v6->v7
 - Remove NSH GSO patches in v6 because Jiri Benc
   reworked it as another patch series and they have
   been merged.
 - Change it to adapt to nsh kernel module added by NSH
   GSO patch series

v5->v6
 - Fix the rest comments for v4.
 - Add NSH GSO support for VxLAN-gpe + NSH and
   Eth + NSH.

v4->v5
 - Fix many comments by Jiri Benc and Eric Garver
   for v4.

v3->v4
 - Add new NSH match field ttl
 - Update NSH header to the latest format
   which will be final format and won't change
   per its author's confirmation.
 - Fix comments for v3.

v2->v3
 - Change OVS_KEY_ATTR_NSH to nested key to handle
   length-fixed attributes and length-variable
   attriubte more flexibly.
 - Remove struct ovs_action_push_nsh completely
 - Add code to handle nested attribute for SET_MASKED
 - Change PUSH_NSH to use the nested OVS_KEY_ATTR_NSH
   to transfer NSH header data.
 - Fix comments and coding style issues by Jiri and Eric

v1->v2
 - Change encap_nsh and decap_nsh to push_nsh and pop_nsh
 - Dynamically allocate struct ovs_action_push_nsh for
   length-variable metadata.

OVS master and 2.8 branch has merged NSH userspace
patch series, this patch is to enable NSH support
in kernel data path in order that OVS can support
NSH in compat mode by porting this.
Signed-off-by: Yi Yang <yi.y.yang@intel.com>
Acked-by: Jiri Benc <jbenc@redhat.com>
Acked-by: Eric Garver <e@erig.me>
Acked-by: Pravin Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Timestamps in pktgen are currently retrieved using the deprecated
do_gettimeofday() function that wraps its signed 32-bit seconds in 2038
(on 32-bit architectures) and requires a division operation to calculate
microseconds.

The pktgen header is also defined with the same limitations, hardcoding
to a 32-bit seconds field that can be interpreted as unsigned to produce
times that only wrap in 2106. Whatever code reads the timestamps should
be aware of that problem in general, but probably doesn't care too
much as we are mostly interested in the time passing between packets,
and that is correctly represented.

Using 64-bit nanoseconds would be cheaper and good for 584 years. Using
monotonic times would also make this unambiguous by avoiding the overflow,
but would make it harder to correlate to the times with those on remote
machines. Either approach would require adding a new runtime flag and
implementing the same thing on the remote side, which we probably don't
want to do unless someone sees it as a real problem. Also, this should
be coordinated with other pktgen implementations and might need a new
magic number.

For the moment, I'm documenting the overflow in the source code, and
changing the implementation over to an open-coded ktime_get_real_ts64()
plus division, so we don't have to look at it again while scanning for
deprecated time interfaces.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

====================
Netfilter/IPVS updates for net-next

The following patchset contains Netfilter/IPVS updates for your net-next
tree, they are:

1) Speed up table replacement on busy systems with large tables
   (and many cores) in x_tables. Now xt_replace_table() synchronizes by
   itself by waiting until all cpus had an even seqcount and we use no
   use seqlock when fetching old counters, from Florian Westphal.

2) Add nf_l4proto_log_invalid() and nf_ct_l4proto_log_invalid() to speed
   up packet processing in the fast path when logging is not enabled, from
   Florian Westphal.

3) Precompute masked address from configuration plane in xt_connlimit,
   from Florian.

4) Don't use explicit size for set selection if performance set policy
   is selected.

5) Allow to get elements from an existing set in nf_tables.

6) Fix incorrect check in nft_hash_deactivate(), from Florian.

7) Cache netlink attribute size result in l4proto->nla_size, from
   Florian.

8) Handle NFPROTO_INET in nf_ct_netns_get() from conntrack core.

9) Use power efficient workqueue in conntrack garbage collector, from
   Vincent Guittot.

10) Remove unnecessary parameter, in conntrack l4proto functions, also
    from Florian.

11) Constify struct nf_conntrack_l3proto definitions, from Florian.

12) Remove all typedefs in nf_conntrack_h323 via coccinelle semantic
    patch, from Harsha Sharma.

13) Don't store address in the rbtree nodes in xt_connlimit, they are
    never used, from Florian.

14) Fix out of bound access in the conntrack h323 helper, patch from
    Eric Sesterhenn.

15) Print symbols for the address returned with %pS in IPVS, from
    Helge Deller.

16) Proc output should only display its own netns in IPVS, from
    KUWAZAWA Takuya.

17) Small clean up in size_entry_mwt(), from Colin Ian King.

18) Use test_and_clear_bit from nf_nat_proto_clean() instead of separated
    non-atomic test and then clear bit, from Florian Westphal.

19) Consolidate prefix length maps in ipset, from Aaron Conole.

20) Fix sparse warnings in ipset, from Jozsef Kadlecsik.

21) Simplify list_set_memsize(), from simran singhal.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Add ethtool statistics support by reading the GOP statistics from the
hardware counters. Also implement a workqueue to gather the statistics
every second or some 32-bit counters could overflow.
Suggested-by: Stefan Chulski <stefanc@marvell.com>
Signed-off-by: Miquel Raynal <miquel.raynal@free-electrons.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Christophe JAILLET says:

====================
fsl/fman: Fix some error handling code in mac_probe

Commit c6e26ea8 ("dpaa_eth: change device used") generated some
conflicts in my patches waiting for submission. So I took a closer look at
it.

So here is a serie of 4 patches.

The 1st one is just about a spurious call to 'dev_set_drvdata()', which is
done in only 1 error handling path in the function.

The 2nd one removes some devm_iounmap/release/kfree functions which look
useless to me.

The 3rd one fixes a missing of_node_put.

The 4th one is just cosmetic and removes a useless message.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Memory allocation functions already display some informaton in case of
memory allocation failure. There is no need to add an extra 'dev_err' here.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

If 'of_phy_find_device()' fails, we must undo the previous 'of_node_get()'
call, as done the the following error handling code.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

There is no need to release explicitly some devm_ allocated resources.
If the 'mac_probe()' probe function fails, they will be released
automatically, as already done in the other error handling paths of
this function.

Also goto '_return_of_get_parent' as in the other error handling paths.
This is useless (priv->fixed_link is NULL at this point), but at least
it is consistent.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit c6e26ea8 ("dpaa_eth: change device used") has removed usage of
'dev_set_drvdata()' in the 'mac_probe() function.

This call should also be axed.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

The size for IFLA_IF_NETNSID is missing from the size calculation
because the proceeding semicolon was not removed. Fix this by removing
the semicolon.

Detected by CoverityScan, CID#1461135 ("Structurally dead code")

Fixes: 79e1ad14 ("rtnetlink: use netnsid to query interface")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Now that ds->num_ports is 3, there is no need to check range of "port"
parameter.
Signed-off-by: Egil Hjelmeland <privat@egil-hjelmeland.no>
Signed-off-by: David S. Miller <davem@davemloft.net>

On 32-bit architectures, rtc_time_to_tm() returns incorrect results
in 2038 or later, and do_gettimeofday() is broken for the same reason.

This changes the code to use ktime_get_real_seconds() and time64_to_tm()
instead, both of them are 2038-safe, and we can also get rid of the
CONFIG_RTC_LIB dependency that way.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit a67e9472 ("of: Add array read functions with min/max size
limits") added a new interface for reading variable-length arrays from
DT properties. One user was added in dsa recently and this causes a
build error because that code can be built with CONFIG_OF disabled:

net/dsa/dsa2.c: In function 'dsa_switch_parse_member_of':
net/dsa/dsa2.c:678:7: error: implicit declaration of function 'of_property_read_variable_u32_array'; did you mean 'of_property_read_u32_array'? [-Werror=implicit-function-declaration]

This adds a dummy functions for of_property_read_variable_u32_array()
and a few others that had been missing here. I decided to move
of_property_read_string() and of_property_read_string_helper() in the
process to make it easier to compare the two sets of function prototypes
to make sure they match.

Fixes: 975e6e32 ("net: dsa: rework switch parsing")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We return on the previous line so this "return 0;" statement should just
be deleted.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Egil Hjelmeland says:

====================
net: dsa: lan9303: Linting

This series is non-functional.
 - Correct some errors in comments and documentation.
Remove scripts/checkpatch.pl WARNINGs and most CHECKs:
 - Replace msleep(1) with usleep_range()
 - Adjust indenting

Changes v1 -> v2:
 - Removed patch 4 "Remove unnecessary parentheses", to be addressed later
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove scripts/checkpatch.pl CHECKs by adjusting indenting.
Signed-off-by: Egil Hjelmeland <privat@egil-hjelmeland.no>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove scripts/checkpatch.pl WARNING by replacing msleep(1) with usleep_range()
Signed-off-by: Egil Hjelmeland <privat@egil-hjelmeland.no>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Egil Hjelmeland <privat@egil-hjelmeland.no>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Two comments refer to registers, but lack the LAN9303_ prefix.
Fix that.
Signed-off-by: Egil Hjelmeland <privat@egil-hjelmeland.no>
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix to return a negative error code from the VID  create error handling
case instead of 0, as done elsewhere in this function.

Fixes: c57529e1 ("mlxsw: spectrum: Replace vPorts with Port-VLAN")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>