net/mlx5e: Support cross-vhca RSS

Implement driver support for the HW feature that allows RX steering of
one device to target other device's RQs.

In SD multi-pf netdev mode, we set the secondaries into silent mode,
disconnecting them from the network. This feature is then used to steer
traffic from the primary to the secondaries.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

Revert "mlx5 updates 2023-12-20"

Revert "net/mlx5: Implement management PF Ethernet profile"
This reverts commit 22c4640698a1d47606b5a4264a584e8046641784.
Revert "net/mlx5: Enable SD feature"
This reverts commit c88c49ac9c18fb7c3fa431126de1d8f8f555e912.
Revert "net/mlx5e: Block TLS device offload on combined SD netdev"
This reverts commit 83a59ce0057b7753d7fbece194b89622c663b2a6.
Revert "net/mlx5e: Support per-mdev queue counter"
This reverts commit d72baceb92539a178d2610b0e9ceb75706a75b55.
Revert "net/mlx5e: Support cross-vhca RSS"
This reverts commit c73a3ab8fa6e93a783bd563938d7cf00d62d5d34.
Revert "net/mlx5e: Let channels be SD-aware"
This reverts commit e4f9686bdee7b4dd89e0ed63cd03606e4bda4ced.
Revert "net/mlx5e: Create EN core HW resources for all secondary devices"
This reverts commit c4fb94aa822d6c9d05fc3c5aee35c7e339061dc1.
Revert "net/mlx5e: Create single netdev per SD group"
This reverts commit e2578b4f983cfcd47837bbe3bcdbf5920e50b2ad.
Revert "net/mlx5: SD, Add informative prints in kernel log"
This reverts commit c82d360325112ccc512fc11a3b68cdcdf04a1478.
Revert "net/mlx5: SD, Implement steering for primary and secondaries"
This reverts commit 605fcce33b2d1beb0139b6e5913fa0b2062116b2.
Revert "net/mlx5: SD, Implement devcom communication and primary election"
This reverts commit a45af9a96740873db9a4b5bb493ce2ad81ccb4d5.
Revert "net/mlx5: SD, Implement basic query and instantiation"
This reverts commit 63b9ce944c0e26c44c42cdd5095c2e9851c1a8ff.
Revert "net/mlx5: SD, Introduce SD lib"
This reverts commit 4a04a31f49320d078b8078e1da4b0e2faca5dfa3.
Revert "net/mlx5: Fix query of sd_group field"
This reverts commit e04984a37398b3f4f5a79c993b94c6b1224184cc.
Revert "net/mlx5e: Use the correct lag ports number when creating TISes"
This reverts commit a7e7b40c4bc115dbf2a2bb453d7bbb2e0ea99703.

There are some unanswered questions on the list, and we don't
have any docs. Given the lack of replies so far and the fact
that v6.8 merge window has started - let's revert this and
revisit for v6.9.

Link: https://lore.kernel.org/all/20231221005721.186607-1-saeed@kernel.org/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/mlx5e: Support cross-vhca RSS

Implement driver support for the HW feature that allows RX steering of
one device to target other device's RQs.

In SD multi-mdev netdev mode, we set the secondaries into silent mode,
disconnecting them from the network. This feature is then used to steer
traffic from the primary to the secondaries.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Preparations for supporting larger number of channels

Data center server CPUs number keeps getting larger with time.
Currently, our driver limits the number of channels to 128.

Maximum channels number is enforced and bounded by hardcoded
defines (en.h/MLX5E_MAX_NUM_CHANNELS) even though the device and machine
(CPUs num) can allow more.

Refactor current implementation in order to handle further channels.

The maximum supported channels number will be increased in the followup
patch.

Introduce RQT size calculation/allocation scheme below:
1) Preserve current RQT size of 256 for channels number up to 128 (the
old limit).
2) For greater channels number, RQT size is calculated by multiplying
the channels number by 2 and rounding up the result to the nearest
power of 2. If the calculated RQT size exceeds the maximum supported
size by the NIC, fallback to this maximum RQT size
(1 << log_max_rqt_size).

Since RQT size is no more static, allocate and free the indirection
table SW shadow dynamically.

Signed-off-by: Adham Faris <afaris@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Refactor mlx5e_rss_init() and mlx5e_rss_free() API's

Introduce code refactoring below:
1) Introduce single API for creating and destroying rss object,
mlx5e_rss_create() and mlx5e_rss_destroy() respectively.
2) mlx5e_rss_create() constructs and initializes RSS object depends
on a function new param enum mlx5e_rss_create_type. Callers (like
rx_res.c) will no longer need to allocate RSS object via
mlx5e_rss_alloc() and initialize it immediately via
mlx5e_rss_init_no_tirs() or mlx5e_rss_init(), this will be done by
a single call to mlx5e_rss_create(). Hence, mlx5e_rss_alloc() and
mlx5e_rss_init_no_tirs() have been removed from rss.h file and became
static functions.

Signed-off-by: Adham Faris <afaris@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Refactor rx_res_init() and rx_res_free() APIs

Refactor mlx5e_rx_res_init() and mlx5e_rx_res_free() by wrapping
mlx5e_rx_res_alloc() and mlx5e_rx_res_destroy() API's respectively.

Signed-off-by: Adham Faris <afaris@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5: Fix flowhash key set/get for custom RSS

mlx5 flow hash field retrieval and set only worked on the default
RSS context as of commit f01cc58c18d6 ("net/mlx5e: Support multiple RSS
contexts"), not custom RSS contexts.

For example, before this patch attempting to retrieve the flow hash fields
for RSS context 1 fails:

$ sudo ethtool -u eth1 rx-flow-hash tcp4 context 1
Cannot get RX network flow hashing options: Invalid argument

This patch fixes getting and setting the flow hash fields for contexts
other than the default context.

Getting the flow hash fields for RSS context 1:

$ sudo ethtool -u eth1 rx-flow-hash tcp4 context 1
For RSS context 1:
TCP over IPV4 flows use these fields for computing Hash flow key:
IP DA

Now, setting the flash hash fields to a custom value:

$ sudo ethtool -U eth1 rx-flow-hash tcp4 sdfn context 1

And retrieving them again:

$ sudo ethtool -u eth1 rx-flow-hash tcp4 context 1
For RSS context 1:
TCP over IPV4 flows use these fields for computing Hash flow key:
IP SA
IP DA
L4 bytes 0 & 1 [TCP/UDP src port]
L4 bytes 2 & 3 [TCP/UDP dst port]

Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx5e: xsk: Use queue indices starting from 0 for XSK queues

In the initial implementation of XSK in mlx5e, XSK RQs coexisted with
regular RQs in the same channel. The main idea was to allow RSS work the
same for regular traffic, without need to reconfigure RSS to exclude XSK
queues.

However, this scheme didn't prove to be beneficial, mainly because of
incompatibility with other vendors. Some tools don't properly support
using higher indices for XSK queues, some tools get confused with the
double amount of RQs exposed in sysfs. Some use cases are purely XSK,
and allocating the same amount of unused regular RQs is a waste of
resources.

This commit changes the queuing scheme to the standard one, where XSK
RQs replace regular RQs on the channels where XSK sockets are open. Two
RQs still exist in the channel to allow failsafe disable of XSK, but
only one is exposed at a time. The next commit will achieve the desired
memory save by flushing the buffers when the regular RQ is unused.

As the result of this transition:

1. It's possible to use RSS contexts over XSK RQs.

2. It's possible to dedicate all queues to XSK.

3. When XSK RQs coexist with regular RQs, the admin should make sure no
unwanted traffic goes into XSK RQs by either excluding them from RSS or
settings up the XDP program to return XDP_PASS for non-XSK traffic.

4. When using a mixed fleet of mlx5e devices and other netdevs, the same
configuration can be applied. If the application supports the fallback
to copy mode on unsupported drivers, it will work too.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/mlx5e: Split out channel (de)activation in rx_res

To decrease the nesting level and reduce duplication of code, create
functions to redirect direct RQTs to the actual RQs or drop_rq, which
are used in the activation and deactivation flows of channels.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net/mlx5e: Take packet_merge params directly from the RX res struct

As packet_merge params structure is saved on the RX resources structure, there
is no need to pass it separately.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Sync TIR params updates against concurrent create/modify

Transport Interface Receive (TIR) objects perform the packet processing and
reassembly and is also responsible for demultiplexing the packets into the
different RQs.

There are certain TIR context attributes that propagate to the pointed RQs
and applied to them (like packet_merge offloads (LRO/SHAMPO) and
tunneled_offload_en). When TIRs do not agree on attributes values, a "last
one wins" policy is applied. Hence, if not synced properly, a race between
TIR params update and a concurrent TIR create/modify operation might yield
to a mismatch between the shadow parameters in SW and the actual applied
state of the RQs in HW.

tunneled_offload_en is a fixed attribute per profile, while packet merge
offload state might be toggled and get out-of-sync. When this happens,
packet_merge offload might be working although not requested, or the
opposite.

All updates to packet_merge state and all create/modify operations of
regular redirection/steering TIRs are done under the same priv->state_lock,
so they do not run in parallel, and no race is possible.

However, there are other kind of TIRs (acceleration offloads TIRs, like TLS
TIRs) which are created on demand for each new connection without holding
the coarse priv->state_lock, hence might race.

Fix this by synchronizing all packet_merge state reads and writes against
all TIR create/modify operations. Include the modify operations of the
regular redirection steering TIRs under the new lock, for better code
layering and division of responsibilities.

Fixes: 1182f3659357 ("net/mlx5e: kTLS, Add kTLS RX HW offload support")
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Rename TIR lro functions to TIR packet merge functions

This series introduces new packet merge type, therefore rename lro
functions to packet merge to support the new merge type:
- Generalize + rename mlx5e_build_tir_ctx_lro to
mlx5e_build_tir_ctx_packet_merge.
- Rename mlx5e_modify_tirs_lro to mlx5e_modify_tirs_packet_merge.
- Rename lro bit in mlx5_ifc_modify_tir_bitmask_bits to packet_merge.
- Rename lro_en in mlx5e_params to packet_merge_type type and combine
packet_merge params into one struct mlx5e_packet_merge_param.

Signed-off-by: Khalid Manaa <khalidm@nvidia.com>
Signed-off-by: Ben Ben-Ishay <benishay@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Fix condition when retrieving PTP-rqn

When activating the PTP-RQ, redirect the RQT from drop-RQ to PTP-RQ.
Use mlx5e_channels_get_ptp_rqn to retrieve the rqn. This helper returns
a boolean (not status), hence caller should consider return value 0 as a
fail. Change the caller interpretation of the return value.

Fixes: 43ec0f41fa73 ("net/mlx5e: Hide all implementation details of mlx5e_rx_res")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Support flow classification into RSS contexts

Extend the existing flow classification support, to steer
flows not only directly to a receive ring, but also into
the new RSS contexts.

Create needed TIR objects on demand, and hold reference
on the RSS context.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Support multiple RSS contexts

Add support to multiple RSS contexts. Resources of the non-default
RSS contexts are allocated and created on demand. Each RSS context
can be controlled and configured separately, via the implemented
ethtool ops. Here we limit the num of total contexts to 16.

We do not enforce any kind of new limitation over the indirection table
content. More specifically, two separate contexts can be configured to
fully or partially point to the same set of receive rings.

The default RSS context (index 0) is created with its full set of TIRs.
All other contexts are created with an empty set, then TIRs are added
upon first usage when steering rules are added.
We use a reference counting mechanism to make sure an RSS context is
not removed before the rules pointing to it.

Block ethtool set_channels operations when multiple RSS contexts exist,
as currently the kernel doesn't protect against inconsistent channels
configs that break non-default RSS contexts.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Convert RSS to a dedicated object

Code related to RSS is now encapsulated into a dedicated object and put
into new files en/rss.{c,h}. All usages are converted.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Introduce abstraction of RSS context

Bring all fields that define and maintain RSS behavior together
into a new structure.
Align all usages with this new structure. Keep it hidden within
rx_res.c.
This helps supporting multiple RSS contexts in downstream patch.

Use dynamic allocations for the RSS context.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Introduce TIR create/destroy API in rx_res

Take TIR control operations in rx_res into functions.
This is in preparation to supporting on-demand TIR operations in
downstream patches.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Do not try enable RSS when resetting indir table

All calls to mlx5e_rx_res_rss_set_indir_uniform() occur while the RSS
state is inactive, i.e. the RQT is pointing to the drop RQ, not to the
channels' RQs.
It means that the "apply" part of the function is not called.
Remove this part from the function, and document the change. It will be
useful for next patches in the series, allows code simplifications when
multiple RSS contexts are introduced.

Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Rename traffic type enums

Rename traffic type enums as part of the preparation for moving
the traffic type logic to a separate file.

Signed-off-by: Maor Gottlieb <maorg@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Allocate the array of channels according to the real max_nch

The channels array in struct mlx5e_rx_res is converted to a dynamic one,
which will use the dynamic value of max_nch instead of
implementation-defined maximum of MLX5E_MAX_NUM_CHANNELS.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Hide all implementation details of mlx5e_rx_res

This commit moves all implementation details of struct mlx5e_rx_res
under en/rx_res.c. All access to RX resources is now done using methods.
Encapsulating RX resources into an object allows for better
manageability, because all the implementation details are now in a
single place, and external code can use only a limited set of API
methods to init/teardown the whole thing, reconfigure RSS and LRO
parameters, connect TIRs to flow steering and activate/deactivate TIRs.

mlx5e_rx_res is self-contained and doesn't depend on struct mlx5e_priv
or include en.h.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>

net/mlx5e: Move management of indir traffic types to rx_res

This commit moves the responsibility of keeping the RSS configuration
for different traffic types to en/rx_res.{c,h}, hiding the
implementation details behind the new getters, and abandons all usage of
struct mlx5e_tirc_config, which is no longer useful and superseded by
struct mlx5e_rss_params_traffic_type.

Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>