The internal key APIs, when called with group keys where mac80211
doesn't pass the (AP) station pointer, are still sometimes called
with the AP station pointer on internal calls. This is confusing.
Clean it up and always call them with the AP STA when it exists,
even when coming in from mac80211, by looking it up immediately.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://patch.msgid.link/20260511203428.0efcbac3e96e.I5b78af51df283d452fc4f1366b04b4769f60f24e@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
We should stick to mac80211's flow to start / stop beaconing. This
allows to stop beaconing before we remove the BIGTK.
Note that the start and stop beaconing flows are not exactly symmetric.
When we start beaconing, we just update the beacon template. We assume
that mac80211 won't update the beacons, if we're not supposed to be
sending it.
Also note that we now send the beacon template after the broadcast
station was added to the firmware: the broadcast station is added in
the start_ap() flow, while the beacon template is now added in the
link_changed() flow which happens later. This is not what we did
before this patch, but this sequence is supported by the firmware as
well.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Link: https://patch.msgid.link/20260511203428.cf10e5754171.I8022517c6c5aedb4b56fba30a5545de8f62dddbe@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
The max_channel_switch_time in wiphy_nan_capa is in microseconds, but
the value was set to 4, which is only 4 microseconds instead of the
intended 4 milliseconds.
Fix by using 4 * USEC_PER_MSEC.
Signed-off-by: Israel Kozitz <israel.kozitz@intel.com>
Reviewed-by: Ilan Peer <ilan.peer@intel.com>
Link: https://patch.msgid.link/20260510234534.cb3c05e5d334.I89d9f336aaf388c6e48769de5fe1e5db19295057@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Frames that are sent to an NMI station are always NAN management frames.
Therefore there is no need to configure TLC for such a station.
Always use host rate for the frames going to that station.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://patch.msgid.link/20260510234534.fb7272ddbb45.I843dba36f52363bc7e9c81f937547ead12147539@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
The firmware requires links for NAN which mac80211 doesn't use,
so introduce a new NAN link data structure that the driver has
for itself only, and handle the link command sending code for
NAN using this data structure, most of the bss_conf data isn't
used for NAN anyway, so those structures aren't useful.
With that, add, activate, deactivate or remove links depending
on the local NAN schedule updates.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Co-developed-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20260510234534.39ee3533ef30.I81ff6dc02e912396040f922e21888fd94b3c26e6@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Until now we maintained the NAN vif in the driver only. The fw used the
AUX MAC for sync and discovery operations.
But when we want to configure a local schedule, we need to add the MAC
first.
NAN_DATA interfaces are not added to the FW. Instead, the local
address of these interfaces are configured to the FW via the NAN MAC.
Add the add/remove/update operations for the NAN interface, and fill the
NAN special parameters in it.
Note that this doesn't fully implement the schedule change, but only the
addition/removal of the NAN MAC. The full schedule management
implementation will come in a later patch.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Co-developed-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20260510234534.dd12944c140b.I3578198660a533faf9f6a94432ef2114f4a9dfae@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
While NAN stations have the deflink link STA and that even
carries their information, having link STAs mostly implies
having real links, and NAN muddies that by having stations
with deflink carrying their capabilities and links at the
NAN level, but no link stations corresponding to NAN links.
Separate out the data needed to build TLC commands into a
new struct iwl_mld_tlc_sta_capa data structure so that the
whole data usage in the TLC code is clarified and we won't
make assumptions, say about being able to look up the link
of an interface from the (NAN) link sta correctly, which
would result in a link but not with a chanctx.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://patch.msgid.link/20260510234534.2c6589aab9db.I01729ec8be4687884831fc8761bf3fd8bba83056@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Copy the HT, VHT and HE capabilities from the sbands:
- The HT capabilities from the 2.4 GHz sband (there is no difference
between the bands anyway).
- The VHT capabilities from the 5 GHz sband, obviously.
- The HE capabilities from the 2.4 GHz and for NL80211_IFTYPE_STATION.
Fix it up to include also the needed 5 GHz bits.
For HE, there are bits that are band-dependent and iftype-dependent. For
those set to what makes most sense, and leave a comment to re-visit.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://patch.msgid.link/20260510234534.3a5e5251dbaf.I566e1749cd17baa672bdd43712114e434f810a69@changeid
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Three BA session handlers use ffs(ba_data->sta_mask) - 1 to derive a
station ID without checking that sta_mask is non-zero. When sta_mask is
zero, ffs() returns 0 and the subtraction wraps to 0xFFFFFFFF, causing
an out-of-bounds access on fw_id_to_link_sta[].
Add WARN_ON_ONCE(!ba_data->sta_mask) guards before each ffs() call,
consistent with the existing check in iwl_mld_ampdu_rx_start().
Reported-by: Yuhao Jiang <danisjiang@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Junrui Luo <moonafterrain@outlook.com>
Link: https://patch.msgid.link/SYBPR01MB788115C6CE873271A9A15A25AF51A@SYBPR01MB7881.ausprd01.prod.outlook.com
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Remove header files that are included but provide no symbols,
types, or macros used by the including translation unit.
In iwl-trans.c, fw/api/tx.h defines TX command structures
(iwl_tx_cmd, iwl_tx_resp, TX_CMD_* flags) used by the PCIe TX
path, not by the transport core itself. Similarly, iwl-fh.h
defines Flow Handler register addresses and DMA-related constants
(FH_*, RFH_*, TFD_*) that are consumed by PCIe-specific code,
none of which are referenced in iwl-trans.c.
In iwl-nvm-parse.c, fw/acpi.h defines ACPI/SAR/GEO/PPAG
interfaces (iwl_acpi_*, iwl_sar_*, iwl_geo_*). No references to
any of these interfaces exist in this file.
Signed-off-by: Jay Ng <jayng9663@gmail.com>
Link: https://patch.msgid.link/20260408034413.1875915-1-jayng9663@gmail.com
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
iwl_mvm_ptp_remove() calls cancel_delayed_work_sync() only after
ptp_clock_unregister() and clearing ptp_data state (ptp_clock,
ptp_clock_info, last_gp2).
This creates a race where the delayed work iwl_mvm_ptp_work() can
execute between ptp_clock_unregister() and cancel_delayed_work_sync(),
observing partially cleared PTP state.
Move cancel_delayed_work_sync() before ptp_clock_unregister() to
ensure the delayed work is fully stopped before any PTP cleanup
begins.
Cc: stable@vger.kernel.org
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Signed-off-by: Junjie Cao <junjie.cao@intel.com>
Link: https://patch.msgid.link/20260212125035.1345718-1-junjie.cao@intel.com
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
iwl_mld_ptp_remove() calls cancel_delayed_work_sync() only after
ptp_clock_unregister() and clearing ptp_data state (ptp_clock,
last_gp2, wrap_counter).
This creates a race where the delayed work iwl_mld_ptp_work() can
execute between ptp_clock_unregister() and cancel_delayed_work_sync(),
observing partially cleared PTP state.
Move cancel_delayed_work_sync() before ptp_clock_unregister() to
ensure the delayed work is fully stopped before any PTP cleanup
begins.
Cc: stable@vger.kernel.org
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Signed-off-by: Junjie Cao <junjie.cao@intel.com>
Link: https://patch.msgid.link/20260212125035.1345718-2-junjie.cao@intel.com
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
The reason for pick_next_task_fair() is the put/set optimization that
avoids touching the common ancestors. However, it is possible to
implement this in the put_prev_task() and set_next_task() calls as
used in put_prev_set_next_task().
Notably, put_prev_set_next_task() is the only site that:
- calls put_prev_task() with a .next argument;
- calls set_next_task() with .first = true.
This means that put_prev_task() can determine the common hierarchy and
stop there, and then set_next_task() can terminate where put_prev_task
stopped.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://patch.msgid.link/20260511120628.057634261@infradead.org
With commit 50653216e4 ("sched: Add support to pick functions to
take rf") removing the balance callback, the pick_task() callback is
in charge of newidle balancing.
This means pick_task_fair() should do so too. This hasn't been a
problem in practise because pick_next_task_fair() is used. However,
since we'll be removing that one shortly, make sure pick_next_task()
is up to scratch.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://patch.msgid.link/20260511120627.944705718@infradead.org
Robots figured out you can read and write this concurrently and got
'upset'. Gemini even noted sched_dynamic_show() can generate
'confusing' output if it observed different values during the
printing.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260511120627.176946327@infradead.org
Commit c9d93a73ce ("sched/fair: Drop redundant RCU read lock in NOHZ
kick path") removed the rcu_read_lock()/unlock() pair from
set_cpu_sd_state_busy() and set_cpu_sd_state_idle() on the assumption
that all callers run in a safe context for rcu_dereference_all(): IRQs
disabled or cpus_write_lock() held.
That assumption is wrong for the CPU hotplug teardown path. When CPUs
are taken offline, set_cpu_sd_state_busy() is invoked via:
cpuhp/N kthread
cpuhp_thread_fun()
cpuhp_invoke_callback()
sched_cpu_deactivate()
nohz_balance_exit_idle()
set_cpu_sd_state_busy()
rcu_dereference_all(per_cpu(sd_llc, cpu))
The cpuhp kthread holds cpu_hotplug_lock (percpu-rwsem) but runs with
preemption and IRQs enabled. As a result, lockdep correctly reports a
suspicious RCU usage on CPU offline, e.g.:
# echo 0 > /sys/devices/system/cpu/cpu1/online
=============================
WARNING: suspicious RCU usage
-----------------------------
kernel/sched/fair.c:12793 suspicious rcu_dereference_check() usage!
...
2 locks held by cpuhp/1/20:
#0: (cpu_hotplug_lock){++++}-{0:0}, at: cpuhp_thread_fun+0x42/0x1ae
#1: (cpuhp_state-down){+.+.}-{0:0}, at: cpuhp_thread_fun+0x72/0x1ae
Call Trace:
lockdep_rcu_suspicious
nohz_balance_exit_idle
sched_cpu_deactivate
cpuhp_invoke_callback
cpuhp_thread_fun
smpboot_thread_fn
Fix this by adding RCU read lock coverage to the one caller that lacks
it: nohz_balance_exit_idle() in the CPU hotplug teardown.
The other callers (nohz_balancer_kick() and nohz_balance_enter_idle())
genuinely run with IRQs disabled, so they remain unchanged.
Fixes: c9d93a73ce ("sched/fair: Drop redundant RCU read lock in NOHZ kick path")
Closes: https://lore.kernel.org/all/38fe0a1d-1a48-435a-910a-c278024d9ac9@samsung.com/
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Andrea Righi <arighi@nvidia.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260522092523.2046095-1-arighi@nvidia.com
Some DM-Firmware are not able to restore the clock rates and the
clock parents after a suspend-resume. The CLK_CONTEXT_LOST firmware
capability has been introduced to identify this characteristic.
In this case the responsibility is therefore delegated to the
ti_sci driver, which uses clk_restore_context() to trigger the
context_restore() operation for all registered clocks, including
those managed by the sci-clk. The sci-clk driver implements the
context_restore() operation to ensure rates and clock parents are
correctly restored.
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Kendall Willis <k-willis@ti.com>
Signed-off-by: Thomas Richard (TI) <thomas.richard@bootlin.com>
Link: https://patch.msgid.link/20260519-ti-sci-jacinto-s2r-restore-irq-v9-4-c550a8ae0f31@bootlin.com
Signed-off-by: Nishanth Menon <nm@ti.com>
Implement the restore_context() operation to restore the clock rate
and the clock parent state. The clock rate is saved in sci_clk struct
during set_rate() and recalc_rate() operations. The parent index
is saved in sci_clk struct during set_parent() operation. During
clock registration, the core retrieves each clock’s parent using
get_parent() operation to ensure the internal clock tree reflects
the actual hardware state, including any configurations made by the
bootloader. So we also save the parent index in get_parent().
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Kendall Willis <k-willis@ti.com>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Reviewed-by: Brian Masney <bmasney@redhat.com>
Signed-off-by: Thomas Richard (TI) <thomas.richard@bootlin.com>
Link: https://patch.msgid.link/20260519-ti-sci-jacinto-s2r-restore-irq-v9-3-c550a8ae0f31@bootlin.com
Signed-off-by: Nishanth Menon <nm@ti.com>
Some DM-Firmware are not able to restore the IRQ context after
a suspend-resume. The IRQ_CONTEXT_LOST firmware capability has
been introduced to identify this characteristic. In this case the
responsibility is delegated to the ti_sci driver, which maintains
an internal list of all requested IRQs. This list is updated on
each set()/free() operation, and all IRQs are restored during the
resume_noirq() phase.
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Kendall Willis <k-willis@ti.com>
Signed-off-by: Thomas Richard (TI) <thomas.richard@bootlin.com>
Link: https://patch.msgid.link/20260519-ti-sci-jacinto-s2r-restore-irq-v9-2-c550a8ae0f31@bootlin.com
Signed-off-by: Nishanth Menon <nm@ti.com>
In BOARDCFG_MANAGED mode, the low power mode configuration is done
statically for the DM via the boardcfg. Constraints are not supported,
and prepare_sleep() is not needed.
Reviewed-by: Dhruva Gole <d-gole@ti.com>
Reviewed-by: Kendall Willis <k-willis@ti.com>
Signed-off-by: Thomas Richard (TI) <thomas.richard@bootlin.com>
Link: https://patch.msgid.link/20260519-ti-sci-jacinto-s2r-restore-irq-v9-1-c550a8ae0f31@bootlin.com
Signed-off-by: Nishanth Menon <nm@ti.com>
Without this, CRYPTO_LZ4 had to be manually enabled in the config
to use LZ4 for hibernation compression. Add the select so it gets
pulled in automatically when hibernation is enabled, just like
CRYPTO_LZO already does.
Tested-by: l1rox3 <l1rox3.developer@gmail.com>
Signed-off-by: l1rox3 <l1rox3.developer@gmail.com>
Link: https://patch.msgid.link/20260520081254.13493-1-l1rox3.developer@gmail.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Replace the custom implementation of dt_node_to_map with
pinctrl_generic_dt_node_to_map() to demonstrate its use.
spacemit_pin_mux_config didn't provide much value in the first place,
because the group contains the information required to look up the
spacemit_pin struct corresponding to a pin, so there's no loss in
functionality as a result of the generic function carrying only the mux
data in the group's data pointer rather than having an array of
spacemit_pin_mux_config structs.
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Troy Mitchell <troy.mitchell@linux.spacemit.com>
Signed-off-by: Linus Walleij <linusw@kernel.org>
As far as I can tell spacemit_pctrl_check_power(), called during the
custom implementation of dt_node_to_map, is redundant because
the driver's implementation generate_config performs the check too.
Removing this would allow the driver to use the newly added common
function pinctrl_generic_pinmux_dt_node_to_map().
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Linus Walleij <linusw@kernel.org>
Akin to my recently added pinctrl_generic_pins_functions_dt_node_to_map(),
create an analogue that performs the same role of dynamically creating
groups at runtime for controllers using the pinmux property.
The pinmux property is freeform, so this function mandates that the
upper 16 bits contain the pin and the lower 16 bits contains the mux
setting. The group's data pointer is populated with an array of the mux
settings for each pin it contains.
Since the node parsing and subsequent pinctrl core function calls are
practically identical to the pins + functions case, other than which
properties are examined, it makes sense to extract the common code from
pinctrl_generic_pins_function_dt_node_to_map() into a generic function
that takes the case-specific devicetree parsing function as an argument.
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Linus Walleij <linusw@kernel.org>
In order to make pinctrl_generic_to_map() usable for controllers that
use pinmux, change the functions char array pointer that it passes to
pinctrl_generic_add_group() to a void pointer. In the pinmux case this
property will contain the mux setting as a number rather than as strings
in the pins + functions case.
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Linus Walleij <linusw@kernel.org>
Add SM_THERMAL_CALIB_READ at SMC ID 0x82000047 in the command
table and implement meson_sm_get_thermal_calib(), which forwards the
tsensor_id argument to the secure monitor and returns the calibration data.
Also realign the CMD() column to improve readability.
Signed-off-by: Ronald Claveau <linux-kernel-dev@aliel.fr>
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
[ dlezcano: Fixed kernel-doc format warning ]
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20260424-add-thermal-t7-vim4-v5-3-9040ca36afe2@aliel.fr
Add SM_THERMAL_CALIB_READ to the secure monitor command enum and
introduce meson_sm_get_thermal_calib() to allow drivers to retrieve
thermal sensor calibration data through the firmware interface.
Signed-off-by: Ronald Claveau <linux-kernel-dev@aliel.fr>
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://patch.msgid.link/20260424-add-thermal-t7-vim4-v5-2-9040ca36afe2@aliel.fr
Add the amlogic,t7-thermal compatible for the Amlogic T7 thermal sensor.
Unlike existing variants which use a phandle to the ao-secure syscon,
the T7 relies on a secure monitor interface described by a phandle and
a sensor index argument.
The T7 integrates multiple thermal sensors, all accessed through the
same SMC call. The sensor index argument is required to identify which
sensor's calibration data the secure monitor should return, as a single
SM_THERMAL_CALIB_READ command serves all of them.
Introduce the amlogic,secure-monitor property as a phandle-array and
make amlogic,ao-secure or amlogic,secure-monitor conditionally required
depending on the compatible.
Signed-off-by: Ronald Claveau <linux-kernel-dev@aliel.fr>
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://patch.msgid.link/20260424-add-thermal-t7-vim4-v5-1-9040ca36afe2@aliel.fr
Use devm_add_action_or_reset() which does the replaced code. It
results in a simpler and more concise code.
Signed-off-by: Daniel Lezcano <daniel.lezcano@oss.qualcomm.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Reviewed-by: Rafael J. Wysocki (Intel) <rafael@kernel.org>
Link: https://patch.msgid.link/20260429161430.3802970-2-daniel.lezcano@oss.qualcomm.com
Use devm_thermal_of_cooling_device_register() to simplify resource
management and avoid manual cleanup in error paths.
As a side effect this change has the benefit of solving an existing
issue. Before, the function tegra_soctherm_remove() only called
debugfs_remove_recursive() and never called thermal_cooling_device_unregister()
for any of the cooling devices registered here.
After the driver removal, the thermal framework's cdev list would
still hold references to thermal_cooling_device objects whose devdata
pointer (ts) pointed to memory already freed by the platform device's
devm cleanup.
With this change, the cooling device is unregistered when the driver
is removed, thus fixing the issue above.
Signed-off-by: Daniel Lezcano <daniel.lezcano@oss.qualcomm.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://patch.msgid.link/20260424160019.41710-2-daniel.lezcano@oss.qualcomm.com
Replace the manual error handling paths disabling the clocks with
devm_add_action_or_reset(). This ensures the clocks are properly
disabled on probe failure and driver removal, while simplifying the
code by removing the explicit error paths.
Signed-off-by: Daniel Lezcano <daniel.lezcano@oss.qualcomm.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@kernel.org>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://patch.msgid.link/20260424160019.41710-1-daniel.lezcano@oss.qualcomm.com
Wei Fang says:
====================
net: enetc: Prepare for ENETC v4 VF support
This patch series refactors and extends the ENETC driver infrastructure
to prepare for upcoming ENETC v4 Virtual Function (VF) support. The main
focus is on code commonization, improved VF-PF communication, and dynamic
resource allocation.
The ENETC IP has evolved across different revisions, and the existing
driver architecture was primarily designed around v1 hardware. To support
v4 VFs efficiently, we need to share common code between PF drivers of
different IP versions while maintaining compatibility.
Key changes in this series:
1. VF-PF Messaging Infrastructure:
- Convert mailbox messages to new formats
- Use read_poll_timeout() for simplifying VF mailbox polling
- Add support for IP minor revision query via messaging
2. Code Commonization:
- Relocate SR-IOV configuration helpers to common PF code
- Move VF message handlers to dedicated enetc_msg.c
- Integrate enetc_msg.c into enetc-pf-common driver
3. CBDR (Control Buffer Descriptor Ring) Improvements:
- Align v1 CBDR API with v4 for VF driver sharing
- Add CBDR setup/teardown hooks to enetc_si_ops
4. Dynamic Resource Management:
- Dynamically allocate rxmsg based on actual VF count
- Use MADDR_TYPE constant for MAC filter array sizing
5. Generic SR-IOV Initialization:
- Add generic helper to initialize SR-IOV resources
This refactoring lays the groundwork for cleanly integrating ENETC v4 VF
support in subsequent patch series, allowing code reuse between v1 and v4
PF drivers while maintaining a clean separation of version-specific
logic.
v1: https://lore.kernel.org/imx/20260511080805.2052495-1-wei.fang@nxp.com/
====================
Link: https://patch.msgid.link/20260522092438.1264020-1-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
