- Adding new notifier block (struct mlx5_nb) monitor_counters_nb
  for handeling MONITOR_COUNTER new event type.
- Adding work queue element: monitor_counters_work for re-arm and
  update stats.
- We re-queue the update stat work, only when working over firmware
  that doesn't support the monitored counters.
Signed-off-by: Eyal Davidovich <eyald@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

new file monitor_stats.c for the new API.
add arm_monitor_counter new command support.
add set_monitor_counter new command support.

The device can monitor specific counters and provide an event to notify
when these counters are changed.
The monitoring is done in best effort manner where the minimum
notification period is 200 ms, however when the device is loaded, the
notification might be delayed.
To configure the required counters to be monitored, the
SET_MONITOR_COUNTER command shall be used with a list of counters to be
monitored.
The device firmware can monitor up to HCA_CAP.max_num_of_monitor_counters.
The configuration is done based on counter type (such as ppcnt, q counter,
etc) and additional param according to the type of counter selected.
Upon monitor counter change, the device will generate
Monitor_Counter_Change event.
The device will not generate new events unless the driver re-arms the
monitoring functionality, using the ARM_MONITOR_COUNTER command.
Signed-off-by: Eyal Davidovich <eyald@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

PPCNT is not supported if PCAM access reg is supported and ppcnt bit is clear.
Signed-off-by: Eyal Davidovich <eyald@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Writing 64B CQEs to 128B cache lines results in a RMW operation. Padding
the CQEs to 128B if possible improves performance on 128B cache line
systems like PPC.

Testing on PPC showed up to a 24% improvement in small packet throughput
vs the default behavior, depending on the workload and system topology.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently a flow is associated with a single encap structure. The FW
extended destination features enables the driver to associate a flow
with multiple encap instances.

Change the encap id field from a flow scope to a per destination value
in the flow attributes struct. Use the encaps array to associate a flow
table entry with multiple encap entries.

Update the neigh logic to offload only if all encapsulations used in a
flow are connected, and un-offload upon the first one disconnected.

Note that the driver can now support up to two encap destinations.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently a flow can be associated with a single encap entry. The
extended destination feature enables the driver to configure multiple
encap entries per flow.

Change the encap flow association field to array as a pre-step towards
supporting multiple encap destinations. Use only the first array
element, with no functional change.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently the driver can support only a single TC tunnel_set action.
Change the tunnel info fields to arrays, as a pre-step to support
multiple encapsulations for a single flow, with no functional change.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

A rule with the following actions is split to a two level FDB:
1. Forward to local mirror vport
2. Header rewrite
3. Forward to local vport
In the first level flow table, forward the packet to the local port and
forward the packet to the second level flow table for header rewrite and
local port forwarding. This configuration fails when mirroring to a
remote encapsulated destination because currently an FTE cannot support
encap and table destinations.

Use the extended destination capabilities to configure the first level
flow table with a multi-destination FTE to the uplink and second level
table and the second level flow table for the header rewrite and local
port forwarding.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently the FTE encap flag applies to all destinations.
To support mirroring encapsulated traffic to a local port the driver
split the two destinations to two flow table entries:
Table#0: - FWD to the local vport
         - Goto table#1
Table#1: - Encap and FWD to wire
The firmware extended destination capabilities enable the driver to set
an encapsulation flag per destination.

Remove the split logic and use the extended destination mechanism
instead.

Note that split technique is still required for pedit and VLAN push
scenarios.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently a FW syndrome is emitted if the driver configures a
multi-destination FTE where the first destination is a tunneled uplink
port and the second destination is a local vPort.

Support this scenario by creating a multi-destination FTE using the
firmware's extended destination capabilities.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Flow attr reformat action bit is moved from the global action bits to a
per destination flags field, as a pre-step for adding additional flags
to support encapsulation properties per destination, with no
functionality change.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently the eswitch flow attr structure stores each destination
specific property in its own specific array.
Group them in an array of destination structures as a pre-step towards
adding additional destination specific field properties.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

The mirror count esw attributes field is used to determine if splitting
the rule to two FTEs is required while programming e-switch mirroring.
Rename it to split count, making it clearer with no functional change.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently the driver identifies identical vport destinations by
comparing the vport ID. The FW extended destination feature enables the
driver to forward the packet to the same vport with multiple
encapsulation properties.

Change the vport destination comparison logic to compare
the encapsulation properties in addition to the vport ID.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Acked-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Minor markup fixup from bpf-next into net-next merge in the BPF helper
description of bpf_sk_lookup_tcp() and bpf_sk_lookup_udp(). Also sync
up the copy of bpf.h from tooling infrastructure.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Daniel Borkmann says:

====================
pull-request: bpf-next 2018-12-11

The following pull-request contains BPF updates for your *net-next* tree.

It has three minor merge conflicts, resolutions:

1) tools/testing/selftests/bpf/test_verifier.c

 Take first chunk with alignment_prevented_execution.

2) net/core/filter.c

  [...]
  case bpf_ctx_range_ptr(struct __sk_buff, flow_keys):
  case bpf_ctx_range(struct __sk_buff, wire_len):
        return false;
  [...]

3) include/uapi/linux/bpf.h

  Take the second chunk for the two cases each.

The main changes are:

1) Add support for BPF line info via BTF and extend libbpf as well
   as bpftool's program dump to annotate output with BPF C code to
   facilitate debugging and introspection, from Martin.

2) Add support for BPF_ALU | BPF_ARSH | BPF_{K,X} in interpreter
   and all JIT backends, from Jiong.

3) Improve BPF test coverage on archs with no efficient unaligned
   access by adding an "any alignment" flag to the BPF program load
   to forcefully disable verifier alignment checks, from David.

4) Add a new bpf_prog_test_run_xattr() API to libbpf which allows for
   proper use of BPF_PROG_TEST_RUN with data_out, from Lorenz.

5) Extend tc BPF programs to use a new __sk_buff field called wire_len
   for more accurate accounting of packets going to wire, from Petar.

6) Improve bpftool to allow dumping the trace pipe from it and add
   several improvements in bash completion and map/prog dump,
   from Quentin.

7) Optimize arm64 BPF JIT to always emit movn/movk/movk sequence for
   kernel addresses and add a dedicated BPF JIT backend allocator,
   from Ard.

8) Add a BPF helper function for IR remotes to report mouse movements,
   from Sean.

9) Various cleanups in BPF prog dump e.g. to make UAPI bpf_prog_info
   member naming consistent with existing conventions, from Yonghong
   and Song.

10) Misc cleanups and improvements in allowing to pass interface name
    via cmdline for xdp1 BPF example, from Matteo.

11) Fix a potential segfault in BPF sample loader's kprobes handling,
    from Daniel T.

12) Fix SPDX license in libbpf's README.rst, from Andrey.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Adding and removing neighbor entries to / from the gc_list need to be
done while holding the table lock; a couple of places were missed in the
original patch.

Move the list_add_tail in neigh_alloc to ___neigh_create where the lock
is already obtained. Since neighbor entries should rarely be moved
to/from PERMANENT state, add lock/unlock around the gc_list changes in
neigh_change_state rather than extending the lock hold around all
neighbor updates.

Fixes: 58956317 ("neighbor: Improve garbage collection")
Reported-by: Andrei Vagin <avagin@gmail.com>
Reported-by: syzbot+6cc2fd1d3bdd2e007363@syzkaller.appspotmail.com
Reported-by: syzbot+35e87b87c00f386b041f@syzkaller.appspotmail.com
Reported-by: syzbot+b354d1fb59091ea73c37@syzkaller.appspotmail.com
Reported-by: syzbot+3ddead5619658537909b@syzkaller.appspotmail.com
Reported-by: syzbot+424d47d5c456ce8b2bbe@syzkaller.appspotmail.com
Reported-by: syzbot+e4d42eb35f6a27b0a628@syzkaller.appspotmail.com
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Saeed Mahameed:

====================
mlx5e-updates-2018-12-10 (gre)

This patch set adds GRE offloading support to Mellanox ethernet driver.

Patches 1-5 replace the existing egdev mechanism with the new TC indirect
block binds mechanism that was introduced by Netronome:
7f76fa36 ("net: sched: register callbacks for indirect tc block binds")

Patches 6-9 add GRE offloading support along with some required
refactoring work.

Patch 10, Add netif_is_gretap()/netif_is_ip6gretap()
 - Changed the is_gretap_dev and is_ip6gretap_dev logic from structure
   comparison to string comparison of the rtnl_link_ops kind field.

Patch 11, add GRE offloading support to mlx5.

Patch 12 removes the egdev mechanism from TC as it is no longer used by
any of the drivers.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The egdev mechanism was replaced by the TC indirect block notifications
platform.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Cc: John Hurley <john.hurley@netronome.com>
Cc: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Add HW offloading support for TC flower filters configured on
gretap/ip6gretap net devices.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Changed the is_gretap_dev and is_ip6gretap_dev logic from structure
comparison to string comparison of the rtnl_link_ops kind field.

This approach aligns with the current identification methods and function
names of vxlan and geneve network devices.

Convert mlxsw to use these helpers and use them in downstream mlx5 patch.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Move tunnel offloading related code to a separate source file for better
code maintainability.

Code refactoring with no functional change.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently the tunnel offloading encap/decap methods assumes that VXLAN
is the sole tunneling protocol. Lay the infrastructure for supporting
multiple tunneling protocols by branching according to the tunnel
net device kind.

Encap filters tunnel type is determined according to the egress/mirred
net device. Decap filters classify the tunnel type according to the
filter's ingress net device kind.

Distinguish between the tunnel type as defined by the SW model and
the FW reformat type that specifies the HW operation being made.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Separates the vxlan header match handling from the matching on the
general fields of ipv4/6 tunnels, thus allowing the common IP tunnel
match code to branch in down stream patch, to multiple IP tunnels.

This patch doesn't add any functionality.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Separates the vxlan header encap logic from the general ipv4/6
encapsulation methods, thus allowing the common IP encap/decap code to
branch in downstream patch to multiple IP tunnels.

Code refactoring with no functional change.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Use TC indirect block notifications to offload filters that
are configured on higher level device interfaces (e.g. tunnel
devices). This mechanism replaces the current egdev implementation.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Propagate the filter's net_device parameter to the tc flower parsed
attributes structure so that it can later be used in tunnel decap
offloading sequences.

Pre-step for replacing egdev logic with the indirect block
notification mechanism.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently the driver controls flower filters that are installed on its
devices. However, with the introduction of the indirect block
notifications platform the driver may receive control events for filters
that are installed on higher level net devices (e.g. tunnel devices).
Therefore, the driver filter control API will not be able to implicitly
assume the filter's net device.

Explicitly specify the filter's net device, no functional change
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Towards using this mechanism as the means to offload tunnel decap rules
set on SW tunnel devices instead of egdev, add the supporting structures
and functions.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Currently only a single field in the representor private structure
is relevant for uplink representors.  As a pre-step to allow adding
additional uplink representor fields, introduce uplink representor
private structure.

This is prepration step towards replacing egdev logic with the
indirect block notification mechanism. This patch doesn't change
any functionality.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Acked-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

mlx5-next shared branch with rdma subtree to avoid mlx5 rdma v.s. netdev
conflicts.

Highlights:

1) RDMA ODP  (On Demand Paging) improvements and moving ODP logic to
mlx5 RDMA driver
2) Improved mlx5 core driver and device events handling and provided API
for upper layers to subscribe to device events.
3) RDMA only code cleanup from mlx5 core
4) Add helper to get CQE opcode
5) Rework handling of port module events
6) shared mlx5_ifc.h updates to avoid conflicts
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Yonghong Song says:

====================
Before func_info and line_info are added to the kernel, there are several
fields in structure bpf_prog_info specifying the "count" of a user buffer, e.g.,
        __u32 nr_jited_ksyms;
        __u32 nr_jited_func_lens;
The naming convention has the prefix "nr_".

The func_info and line_info support added several fields
        __u32 func_info_cnt;
        __u32 line_info_cnt;
        __u32 jited_line_info_cnt;
to indicate the "count" of buffers func_info, line_info and jited_line_info.
The original intention is to keep the field names the same as those in
structure bpf_attr, so it will be clear that the "count" returned to user
space will be the same as the one passed to the kernel during prog load.

Unfortunately, the field names *_info_cnt are not consistent with
other existing fields in bpf_prog_info.
This patch set renamed the fields *_info_cnt to nr_*_info
to keep naming convention consistent.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Rename all occurances of *_info_cnt field access
to nr_*_info in tools directory.

The local variables finfo_cnt, linfo_cnt and jited_linfo_cnt
in function do_dump() of tools/bpf/bpftool/prog.c are also
changed to nr_finfo, nr_linfo and nr_jited_linfo to
keep naming convention consistent.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Sync kernel uapi bpf.h "*_info_cnt => nr_*_info"
changes to tools directory.
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

In uapi bpf.h, currently we have the following fields in
the struct bpf_prog_info:
	__u32 func_info_cnt;
	__u32 line_info_cnt;
	__u32 jited_line_info_cnt;
The above field names "func_info_cnt" and "line_info_cnt"
also appear in union bpf_attr for program loading.

The original intention is to keep the names the same
between bpf_prog_info and bpf_attr
so it will imply what we returned to user space will be
the same as what the user space passed to the kernel.

Such a naming convention in bpf_prog_info is not consistent
with other fields like:
        __u32 nr_jited_ksyms;
        __u32 nr_jited_func_lens;

This patch made this adjustment so in bpf_prog_info
newly introduced *_info_cnt becomes nr_*_info.
Acked-by: Song Liu <songliubraving@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

info.nr_jited_ksyms and info.nr_jited_func_lens cannot be 0 in these two
statements, so we don't need to check them.
Signed-off-by: Song Liu <songliubraving@fb.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

This isn't used anywhere across the mlx5 driver stack,
remove it.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Update the flow steering command formatting according to the extended
destination API.
Note that the FW dictates that multi destination FTEs that involve at
least one encap must use the extended destination format, while single
destination ones must use the legacy format.
Using extended destination format requires FW support. Check for its
capabilities and return error if not supported.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Change the driver flow destination struct to use bit flags with the vhca
id valid being the 1st one. The flags field is more extendable and will
be used in downstream patch.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Extended destinations provide the ability to configure different
encapsulation properties per destination on a single FTE. This is
needed for use-cases such as remote mirroring over tunneled networks.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>