Register devlink ports upon NIC init. TX and RX health reporters handle
errors which may occur early on at driver initialization. And because
these reporters are to be moved to port context, they require devlink
ports to be already registered.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In order to use new devlink port health reporters infrastructure, add
corresponding constructor and destructor functions.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add devlink-health reporter support on per-port basis.
The main difference existing devlink-health is that port reporters are
stored in per-devlink_port lists. Upon creation of such health reporter the
reference to a port it belongs to is stored in reporter struct.

Fill the port index attribute in devlink-health response to
allow devlink userspace utility to distinguish between device and port
reporters.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a generic __devlink_health_reporter_find_by_name() that can be used
with arbitrary devlink health reporter list.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Devlink keeps its own reference to every reporter in a list and inits
refcount to 1 upon reporter's creation. Existing destructor waits to
free the memory indefinitely using msleep() until all references except
devlink's own are put.

Rework this mechanism by moving memory free routine to a separate
function, which is called when the last reporter reference is put.

Besides, it allows to call __devlink_health_reporter_destroy() while
locked on a reporters list mutex in symmetry to
__devlink_health_reporter_create(), which is required in follow-up
patch.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Prepare a common routine in devlink_health_reporter_create() for usage
in similar functions for devlink port health reporters.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Saeed Mahameed says:

====================
mlx5-updates-2020-07-09

This series provides updates to mlx5 CT (connection tracking) offloads
For more information please see tag log below.

Please pull and let me know if there is any problem.

The following conflict is expected when net is merged into net-next:
to resolve just use the hunks from net-next.

<<<<<<< HEAD (net-next)
	mlx5_tc_ct_del_ft_entry(ct_priv, entry);
	kfree(entry);
======= (net)
	mlx5_tc_ct_entry_del_rules(ct_priv, entry);
	kfree(entry);
>>>>>>> b1a7d5bdfe54c98eca46e2c997d4e3b1484a49af

mlx5 connection tracking offloads updates:

1)  Restore CT state from lookup in zone instead of tupleid

    On a miss, Use this zone + 5 tuple taken from the skb, to lookup the CT
    entry and restore it, instead of the driver allocated tuple id.

    This improves flow insertion rate by avoiding the allocation of a header
    rewrite context to maintain the tupleid.

2) Re-use modify header HW objects for identical modify actions.

3) Expand tunnel register mappings
   Reg_c1 is 32 bits wide. Before this patchset, 24 bit were allocated
   for the tuple_id,  6 bits for tunnel mapping and 2 bits for tunnel
   options mappings.

   Restoring the ct state from zone lookup instead of tuple id requires
   reg_c1 to store 8 bits mapping the ct zone, leaving 24 bits for tunnel
   mappings.

   Expand tunnel and tunnel options register mappings to 12 bit each.

4) Trivial cleanup and fixes.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Jakub Kicinski says:

====================
udp_tunnel: add NIC RX port offload infrastructure

Kernel has a facility to notify drivers about the UDP tunnel ports
so that devices can recognize tunneled packets. This is important
mostly for RX - devices which don't support CHECKSUM_COMPLETE can
report checksums of inner packets, and compute RSS over inner headers.
Some drivers also match the UDP tunnel ports also for TX, although
doing so may lead to false positives and negatives.

Unfortunately the user experience when trying to take adavantage
of these facilities is suboptimal. First of all there is no way
for users to check which ports are offloaded. Many drivers resort
to printing messages to aid debugging, other use debugfs. Even worse
the availability of the RX features (NETIF_F_RX_UDP_TUNNEL_PORT)
is established purely on the basis of the driver having the ndos
installed. For most drivers, however, the ability to perform offloads
is contingent on device capabilities (driver support multiple device
and firmware versions). Unless driver resorts to hackish clearing
of features set incorrectly by the core - users are left guessing
whether their device really supports UDP tunnel port offload or not.

There is currently no way to indicate or configure whether RX
features include just the checksum offload or checksum and using
inner headers for RSS. Many drivers default to not using inner
headers for RSS because most implementations populate the source
port with entropy from the inner headers. This, however, is not
always the case, for example certain switches are only able to
use a fixed source port during encapsulation.

We have also seen many driver authors get the intricacies of UDP
tunnel port offloads wrong. Most commonly the drivers forget to
perform reference counting, or take sleeping locks in the callbacks.

This work tries to improve the situation by pulling the UDP tunnel
port table maintenance out of the drivers. It turns out that almost
all drivers maintain a fixed size table of ports (in most cases one
per tunnel type), so we can take care of all the refcounting in the
core, and let the driver specify if they need to sleep in the
callbacks or not. The new common implementation will also support
replacing ports - when a port is removed from a full table it will
try to find a previously missing port to take its place.

This patch only implements the core functionality along with a few
drivers I was hoping to test manually [1] along with a test based
on a netdevsim implementation. Following patches will convert all
the drivers. Once that's complete we can remove the ndos, and rely
directly on the new infrastrucutre.

Then after RSS (RXFH) is converted to netlink we can add the ability
to configure the use of inner RSS headers for UDP tunnels.

[1] Unfortunately I wasn't able to, turns out 2 of the devices
I had access to were older generation or had old FW, and they
did not actually support UDP tunnel port notifications (see
the second paragraph). The thrid device appears to program
the UDP ports correctly but it generates bad UDP checksums with
or without these patches. Long story short - I'd appreciate
reviews and testing here..

v4:
 - better build fix (hopefully this one does it..)
v3:
 - fix build issue;
 - improve bnxt changes.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Convert to new infra, make use of the ability to sleep in the callback.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Convert to new infra, taking advantage of sleeping in callbacks.

v2:
 - use bp->*_fw_dst_port_id != INVALID_HW_RING_ID as indication
   that the offload is active.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make use of new common udp_tunnel_nic infra. ixgbe supports
IPv4 only, and only single VxLAN and Geneve ports (one each).

v2:
 - split out the RXCSUM feature handling to separate change;
 - declare structs separately;
 - use ti.type instead of assuming table 0 is VxLAN;
 - move setting netdev->udp_tunnel_nic_info to its own switch.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

It appears the clearing of UDP tunnel ports when RXCSUM
is disabled is unnecessary. Driver will not pay attention
to checksum bits if RXCSUM is not set, so we can let
the hardware parse the packets.

Note that the UDP tunnel port NDO handlers don't pay attention
to the state of RXCSUM, so the ports could had been re-programmed,
anyway.

This cleanup simplifies later conversion patch.

v2:
 - break this out of the following patch.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add validating the UDP tunnel infra works.

$ ./udp_tunnel_nic.sh
PASSED all 383 checks
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add UDP tunnel port handlers to our fake driver so we can test
the core infra.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add an interface to report offloaded UDP ports via ethtool netlink.

Now that core takes care of tracking which UDP tunnel ports the NICs
are aware of we can quite easily export this information out to
user space.

The responsibility of writing the netlink dumps is split between
ethtool code and udp_tunnel_nic.c - since udp_tunnel module may
not always be loaded, yet we should always report the capabilities
of the NIC.

$ ethtool --show-tunnels eth0
Tunnel information for eth0:
  UDP port table 0:
    Size: 4
    Types: vxlan
    No entries
  UDP port table 1:
    Size: 4
    Types: geneve, vxlan-gpe
    Entries (1):
        port 1230, vxlan-gpe

v4:
 - back to v2, build fix is now directly in udp_tunnel.h
v3:
 - don't compile ETHTOOL_MSG_TUNNEL_INFO_GET in if CONFIG_INET
   not set.
v2:
 - fix string set count,
 - reorder enums in the uAPI,
 - fix type of ETHTOOL_A_TUNNEL_UDP_TABLE_TYPES to bitset
   in docs and comments.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Cater to devices which:
 (a) may want to sleep in the callbacks;
 (b) only have IPv4 support;
 (c) need all the programming to happen while the netdev is up.

Drivers attach UDP tunnel offload info struct to their netdevs,
where they declare how many UDP ports of various tunnel types
they support. Core takes care of tracking which ports to offload.

Use a fixed-size array since this matches what almost all drivers
do, and avoids a complexity and uncertainty around memory allocations
in an atomic context.

Make sure that tunnel drivers don't try to replay the ports when
new NIC netdev is registered. Automatic replays would mess up
reference counting, and will be removed completely once all drivers
are converted.

v4:
 - use a #define NULL to avoid build issues with CONFIG_INET=n.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make it possible to use tunnel types as flags more easily.
There doesn't appear to be any user using the type as an
array index, so this should make no difference.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

debugfs_create_u32_array() allocates a small structure to wrap
the data and size information about the array. If users ever
try to remove the file this leads to a leak since nothing ever
frees this wrapper.

That said there are no upstream users of debugfs_create_u32_array()
that'd remove a u32 array file (we only have one u32 array user in
CMA), so there is no real bug here.

Make callers pass a wrapper they allocated. This way the lifetime
management of the wrapper is on the caller, and we can avoid the
potential leak in debugfs.

CC: Chucheng Luo <luochucheng@vivo.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Before this commit, on ft flush, ft entries were not removed
from the ct_tuple hashtables. Fix it.

Fixes: ac991b48 ("net/mlx5e: CT: Offload established flows")
Signed-off-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

"flow" parameter is not used in __mlx5_tc_ct_flow_offload_clear(),
remove it.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>

Instead of having to deal with converting between int and ERR_PTR for
return values in mlx5_tc_ct_flow_offload(), make the internal helper
functions return a ptr to mlx5_flow_handle instead of passing it as
output param, this will also avoid gcc confusion and false alarms,
thus we remove the redundant ERR_PTR rule initialization.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Suggested-by: Jason Gunthorpe <jgg@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>

Reg_c1 is 32 bits wide. Originally, 24 bit were allocated for the tuple_id,
6 bits for tunnel mapping and 2 bits for tunnel options mappings.

Restoring the ct state from zone lookup instead of tuple id requires
reg_c1 to store 8 bits mapping the ct zone, leaving 24 bits for tunnel
mappings.

Expand tunnel and tunnel options register mappings to 12 bit each.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Use a single byte mapping for zone restore register (zone matching
remains 16 bit).

This makes room for using the freed 8 bits on register C1 for
mapping more tunnels and tunnel options.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

After removing the tupleid register which changed per tuple,
tuple modify headers set the ct_state, zone, mark, and label registers.
For non-natted tuples going through the same tc rules path, their values
will be the same, and all their modify headers will be the same.

Re-use tuple modify header when possible, by adding each new modify
header to an hahstable, and looking up identical ones before creating
a new one.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Refactor sharing of mod headers to new file and while there,
remove spin lock and flows list, as this is only used for warn on.

Use the generic API in the next patch to re-use tuple modify headers
for identical modify actions,
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Remove tupleid, and replace it with zone_restore, which is the zone an
established tuple sets after match. On miss, Use this zone + tuple
taken from the skb, to lookup the ct entry and restore it.

This improves flow insertion rate by avoiding the allocation of a header
rewrite context.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Next patches will remove the tupleid registers that is used
to restore the ct state on miss, and instead use the tuple on
the missed packet to lookup which state to restore.
Disable tuple rewrites after connection tracking.

For tuple rewrites, inject a ct_state=-trk match so it won't
change the tuple for established flows (+trk) that passed connection
tracking, and instead miss to software.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

The next patch will pass the mlx5e_priv struct to the
modify_header_match_supported method. Use this opportunity to refactor
the existing pr_info call to a netdev_info call.
Signed-off-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

With ct clear we don't jump to the ct tables, so header rewrite
of 5-tuple can be done in place (and not moved to after the CT action).

Check for ct clear action, and if so, allow 5-tuple header
rewrite.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Save original tuple and natted tuple in two new hashtables.

This is a pre-step for restoring ct state after hw miss by performing a
5-tuple lookup on the hash tables.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

When eswitch is unsupported, currently -EPERM error code is returned
instead of -EOPNOTSUPP.

Due to this VF device's devlink virtual port is not enumerated because
port_function_get() callback returned -EPERM instead of -EOPNOTSUPP.

Hence, return the error code -EOPNOTSUPP when eswitch is unsupported.

Fixes: bd939753 ("net/mlx5: E-switch, Introduce and use eswitch support check helper")
Signed-off-by: Parav Pandit <parav@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

Ido Schimmel says:

====================
Expose port split attributes

Danielle says:

Currently, user space has no way of knowing if a port can be split and
into how many ports. Among other things, this makes it impossible to
write generic tests for port split functionality.

Therefore, this set exposes two new devlink port attributes to user
space: Number of lanes and whether the port can be split or not.

Patch set overview:

Patches #1-#4 cleanup 'struct devlink_port_attrs' and reduce the number
of parameters passed between drivers and devlink via
devlink_port_attrs_set()

Patch #5 adds devlink port lanes attributes

Patches #6-#7 add devlink port splittable attribute

Patch #8 exploits the fact that devlink is now aware of port's number of
lanes and whether the port can be split or not and moves some checks
from drivers to devlink

Patch #9 adds a port split test

Changes since v2:
* Remove some local variables from patch #3
* Reword function description in patch #5
* Fix a bug in patch #8
* Add a test for the splittable attribute in patch #9

Changes since v1:
* Rename 'width' attribute to 'lanes'
* Add 'splittable' attribute
* Move checks from drivers to devlink
====================
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Test port split configuration using previously added number of port lanes
attribute.

Check that all the splittable ports are successfully split to their maximum
number of lanes and below, and that those which are not splittable fail to
be split.

Test output example:

TEST: swp4 is unsplittable                                         [ OK ]
TEST: split port swp53 into 4                                      [ OK ]
TEST: Unsplit port pci/0000:03:00.0/25                             [ OK ]
TEST: split port swp53 into 2                                      [ OK ]
TEST: Unsplit port pci/0000:03:00.0/25                             [ OK ]
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, all the input checks are done in driver.

After adding the split capability to devlink port, move the checks to
devlink.
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a new attribute that indicates the split ability of devlink port.

Drivers are expected to set it via devlink_port_attrs_set(), before
registering the port.
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, port attributes like flavour, port number and whether the port
was split are set when initializing a port.

Set the split ability of the port as well, based on port_mapping->width
field and split attribute of devlink port in spectrum, so that it could be
easily passed to devlink in the next patch.
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a new devlink port attribute that indicates the port's number of lanes.

Drivers are expected to set it via devlink_port_attrs_set(), before
registering the port.

The attribute is not passed to user space in case the number of lanes is
invalid (0).
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, port attributes like flavour, port number and whether the
port was split are set when initializing a port.

Set the number of lanes of the port as well so that it could be easily
passed to devlink in the next patch.
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, devlink_port_attrs_set accepts a long list of parameters,
that most of them are devlink port's attributes.

Use the devlink_port_attrs struct to replace the relevant parameters.
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The struct devlink_port_attrs holds the attributes of devlink_port.

Similarly to the previous patch, 'switch_port' attribute is another
exception.

Move 'switch_port' to be devlink_port's field.
Signed-off-by: Danielle Ratson <danieller@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>