Add device tree binding support for the Gigabit Ethernet (GBETH) IP on
Renesas RZ/G3L SoC. This SoC uses different Synopsys DesignWare MAC
version 5.30 compared to RZ/G3E.
RZ/G3L requires an extra clock compared to RZ/G3E and has pps interrupts.
Add a new compatible string "renesas,r9a08g046-gbeth" for RZ/G3L SoC and
update the schema to handle hardware differences between SoC variants.
Extend the base snps,dwmac.yaml schema to accommodate the PPS interrupts.
Acked-by: Conor Dooley <conor.dooley@microchip.com> Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com> Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com> Link: https://patch.msgid.link/20260131161250.5047-2-biju.das.jz@bp.renesas.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
mptcp: implement .read_sock and .splice_read
This series is a preparation work for future in-kernel MPTCP sockets
usage. Here, two interfaces are implemented: read_sock and splice_read.
As a result of this series, splice() with MPTCP sockets -- which was
already supported -- is now improved.
- Patches 1-2: .read_sock implementation
- Patches 3-4: .splice_read implementation
- Patches 5-6: validate splice() support with MPTCP sockets.
====================
Geliang Tang [Fri, 30 Jan 2026 19:24:28 +0000 (20:24 +0100)]
selftests: mptcp: add splice io mode
This patch adds a new 'splice' io mode for mptcp_connect to test
the newly added read_sock() and splice_read() functions of MPTCP.
do_splice() efficiently transfers data directly between two file
descriptors (infd and outfd) without copying to userspace, using
Linux's splice() system call.
Geliang Tang [Fri, 30 Jan 2026 19:24:25 +0000 (20:24 +0100)]
mptcp: implement .read_sock
Current in-kernel TCP sockets -- i.e. from nvme_tcp_try_recv() -- need
to call .read_sock interface of struct proto_ops, but it's not
implemented in MPTCP.
This patch implements it with reference to __tcp_read_sock() and
__mptcp_recvmsg_mskq().
Corresponding to tcp_recv_skb(), a new helper for MPTCP named
mptcp_recv_skb() is added to peek a skb from sk->sk_receive_queue.
Compared with __mptcp_recvmsg_mskq(), mptcp_read_sock() uses
sk->sk_rcvbuf as the max read length. The LISTEN status is checked
before the while loop, and mptcp_recv_skb() and mptcp_cleanup_rbuf()
are invoked after the loop. In the loop, all flags checks for
__mptcp_recvmsg_mskq() are removed.
Reviewed-by: Hannes Reinecke <hare@kernel.org> Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn> Reviewed-by: Mat Martineau <martineau@kernel.org> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> Link: https://patch.msgid.link/20260130-net-next-mptcp-splice-v2-2-31332ba70d7f@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
ptp: vmclock: Add VM generation counter and ACPI notification
Similarly to live migration, starting a VM from some serialized state
(aka snapshot) is an event which calls for adjusting guest clocks, hence
a hypervisor should increase the disruption_marker before resuming the
VM vCPUs, letting the guest know.
However, loading a snapshot, is slightly different than live migration,
especially since we can start multiple VMs from the same serialized
state. Apart from adjusting clocks, the guest needs to take additional
action during such events, e.g. recreate UUIDs, reset network
adapters/connections, reseed entropy pools, etc. These actions are not
necessary during live migration. This calls for a differentiation
between the two triggering events.
We differentiate between the two events via an extra field in the
vmclock_abi, called vm_generation_counter. Whereas hypervisors should
increase the disruption marker in both cases, they should only increase
vm_generation_counter when a snapshot is loaded in a VM (not during live
migration).
Additionally, we attach an ACPI notification to VMClock. Implementing
the notification is optional for the device. VMClock device will declare
that it implements the notification by setting
VMCLOCK_FLAG_NOTIFICATION_PRESENT bit in vmclock_abi flags. Hypervisors
that implement the notification must send an ACPI notification every
time seq_count changes to an even number. The driver will propagate
these notifications to userspace via the poll() interface.
====================
David Woodhouse [Fri, 30 Jan 2026 17:36:06 +0000 (17:36 +0000)]
ptp: ptp_vmclock: return TAI not UTC
To output UTC would involve complex calculations about whether the time
elapsed since the reference time has crossed the end of the month when
a leap second takes effect. I've prototyped that, but it made me sad.
Much better to report TAI, which is what PHCs should do anyway.
And much much simpler.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk> Signed-off-by: Babis Chalios <bchalios@amazon.es> Tested-by: Takahiro Itazuri <itazur@amazon.com> Link: https://patch.msgid.link/20260130173704.12575-8-itazur@amazon.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
David Woodhouse [Fri, 30 Jan 2026 17:36:04 +0000 (17:36 +0000)]
ptp: ptp_vmclock: add 'VMCLOCK' to ACPI device match
As we finalised the spec, we spotted that vmgenid actually says that the
_HID is supposed to be hypervisor-specific. Although in the 13 years
since the original vmgenid doc was published, nobody seems to have cared
about using _HID to distinguish between implementations on different
hypervisors, and we only ever use the _CID.
For consistency, match the _CID of "VMCLOCK" too.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk> Signed-off-by: Babis Chalios <bchalios@amazon.es> Tested-by: Takahiro Itazuri <itazur@amazon.com> Link: https://patch.msgid.link/20260130173704.12575-6-itazur@amazon.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
David Woodhouse [Fri, 30 Jan 2026 17:36:03 +0000 (17:36 +0000)]
ptp: ptp_vmclock: Add device tree support
Add device tree support to the ptp_vmclock driver, allowing it to probe
via device tree in addition to ACPI.
Handle optional interrupt for clock disruption notifications, mirroring
the ACPI notification behaviour.
Although the interrupt is marked as 'optional' in the DT bindings, if
the device *advertises* the VMCLOCK_FLAG_NOTIFICATION_ABSENT then it
*should* have an interrupt. The driver will refuse to initialize if not.
Babis Chalios [Fri, 30 Jan 2026 17:36:01 +0000 (17:36 +0000)]
ptp: vmclock: support device notifications
Add optional support for device notifications in VMClock. When
supported, the hypervisor will send a device notification every time it
updates the seq_count to a new even value.
Moreover, add support for poll() in VMClock as a means to propagate this
notification to user space. poll() will return a POLLIN event to
listeners every time seq_count changes to a value different than the one
last seen (since open() or last read()/pread()). This means that when
poll() returns a POLLIN event, listeners need to use read() to observe
what has changed and update the reader's view of seq_count. In other
words, after a poll() returned, all subsequent calls to poll() will
immediately return with a POLLIN event until the listener calls read().
The device advertises support for the notification mechanism by setting
flag VMCLOCK_FLAG_NOTIFICATION_PRESENT in vmclock_abi flags field. If
the flag is not present the driver won't setup the ACPI notification
handler and poll() will always immediately return POLLHUP.
Signed-off-by: Babis Chalios <bchalios@amazon.es> Signed-off-by: David Woodhouse <dwmw@amazon.co.uk> Signed-off-by: Takahiro Itazuri <itazur@amazon.com> Reviewed-by: David Woodhouse <dwmw@amazon.co.uk> Tested-by: Takahiro Itazuri <itazur@amazon.com> Link: https://patch.msgid.link/20260130173704.12575-3-itazur@amazon.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Babis Chalios [Fri, 30 Jan 2026 17:36:00 +0000 (17:36 +0000)]
ptp: vmclock: add vm generation counter
Similar to live migration, loading a VM from some saved state (aka
snapshot) is also an event that calls for clock adjustments in the
guest. However, guests might want to take more actions as a response to
such events, e.g. as discarding UUIDs, resetting network connections,
reseeding entropy pools, etc. These are actions that guests don't
typically take during live migration, so add a new field in the
vmclock_abi called vm_generation_counter which informs the guest about
such events.
Hypervisor advertises support for vm_generation_counter through the
VMCLOCK_FLAG_VM_GEN_COUNTER_PRESENT flag. Users need to check the
presence of this bit in vmclock_abi flags field before using this flag.
Signed-off-by: Babis Chalios <bchalios@amazon.es> Signed-off-by: David Woodhouse <dwmw@amazon.co.uk> Reviewed-by: David Woodhouse <dwmw@amazon.co.uk> Tested-by: Takahiro Itazur <itazur@amazon.com> Link: https://patch.msgid.link/20260130173704.12575-2-itazur@amazon.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Many network drivers have unnecessary empty module_init and module_exit
functions. Remove them (including some that just print a message). Note
that if a module_init function exists, a module_exit function must also
exist; otherwise, the module cannot be unloaded.
Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com> Acked-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Acked-by: Ping-Ke Shih <pkshih@realtek.com> Acked-by: Toke Høiland-Jørgensen <toke@toke.dk> Link: https://patch.msgid.link/20260131004327.18112-1-enelsonmoore@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
net: ethernet: use module_pci_driver; remove useless driver versions
The module version is useless, and the only thing these drivers' init
routines did besides pci_register_driver was to print the driver name
and/or version.
Acked-by: Francois Romieu <romieu@fr.zoreil.com> (epic100) Reviewed-by: Simon Horman <horms@kernel.org> (epic100, sis900) Reviewed-by: Sai Krishna <saikrishnag@marvell.com> (epic100) Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com> Link: https://patch.msgid.link/20260131022441.56274-1-enelsonmoore@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
net: phy: dp83867: Always program R/SGMII enable bits
The hardware designers at my company neglected to read the datasheet for
this PHY and did not add appropriate resistors to configure it for
SGMII. Add support for configuring the it based on phy-mode instead of
relying on the resistors for a suitable default.
====================
Sean Anderson [Thu, 29 Jan 2026 17:12:05 +0000 (12:12 -0500)]
net: phy: dp83867: Always program R/SGMII enable bits
If the board designers have neglected to populate the appropriate
resistors on the strapping pins then the phy may default to the wrong
interface mode. Enable/disable the RGMII/SGMII enable bits as necessary
to select the correct interface.
The dp83867 strapping pins have four levels and typically configure two
features at once. LED_0 controls both port mirroring and whether SGMII
is enabled. If it is pulled to VDDIO, both port mirroring and SGMII
will be enabled. For variants of the dp83867 that do not support SGMII,
this will prevent data from being transferred. As we now explicitly set
the SGMII and RGMII enable bits, we do not need to detect whether SGMII
has been inadvertently enabled.
Sean Anderson [Thu, 29 Jan 2026 17:12:04 +0000 (12:12 -0500)]
net: phy: dp83867: Program TX FIFO for all interfaces
All supported interfaces use the TX FIFO register at least some of the
time, so there's no point in checking the interface. Retain the check
for the RX FIFO level since it is only used by SGMII.
The issue is that the existing code uses ethtool_get_flow_spec_ring_vf,
which will return a non-zero value if the ring_cookie is set to
RX_CLS_FLOW_DISC, which then causes bnxt_add_ntuple_cls_rule to return
-EOPNOTSUPP because it thinks the user is trying to set an ntuple filter
for a vf.
Fix this by first checking that the ring_cookie is not RX_CLS_FLOW_DISC.
After this patch, ntuple filters for drops can be added:
% sudo ethtool -U eth0 flow-type udp4 src-ip 1.1.1.1 action -1
Added rule with ID 0
% ethtool -n eth0
44 RX rings available
Total 1 rules
Filter: 0
Rule Type: UDP over IPv4
Src IP addr: 1.1.1.1 mask: 0.0.0.0
Dest IP addr: 0.0.0.0 mask: 255.255.255.255
TOS: 0x0 mask: 0xff
Src port: 0 mask: 0xffff
Dest port: 0 mask: 0xffff
Action: Drop
Jakub Kicinski [Sat, 31 Jan 2026 20:30:29 +0000 (12:30 -0800)]
tools: ynl: cli: make the output compact
Make the default (non-JSON) output more compact. Looking at RSS
context dumps is pretty much impossible without this, because
default print shows the indirection table with line per entry:
Jakub Kicinski [Sat, 31 Jan 2026 22:54:54 +0000 (14:54 -0800)]
docs: networking: mention that RSS table should be 4x the queue count
Spell out the recommendation that the RSS table should be
4x the queue count to avoid traffic imbalance. Include minor
rephrasing and removal of the explicit 128 entry example
since a 128 entry table is inadequate on modern machines.
Jakub Kicinski [Sat, 31 Jan 2026 22:54:53 +0000 (14:54 -0800)]
selftests: drv-net: rss: validate min RSS table size
Add a test which checks that the RSS table is at least 4x the max
queue count supported by the device. The original RSS spec from
Microsoft stated that the RSS indirection table should be 2 to 8
times the CPU count, presumably assuming queue per CPU. If the
CPU count is not a power of two, however, a power-of-2 table
2x larger than queue count results in a 33% traffic imbalance.
Validate that the indirection table is at least 4x the queue
count. This lowers the imbalance to 16% which empirically
appears to be more acceptable to memcache-like workloads.
This first 2 patches are by Biju Das, target the rcar_canfd driver and
add support for FD-only mode.
Lad Prabhakar's patches, also for the rcar_canfd driver add support
for the RZ/T2H SoC.
The last 2 patches are by Michael Tretter and me, target the sja1000
driver and clean up the CAN state handling.
* tag 'linux-can-next-for-6.20-20260131' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next:
can: sja1000: sja1000_err(): use error counter for error state
can: sja1000: sja1000_err(): make use of sja1000_get_berr_counter() to read error counters
can: rcar_canfd: Add RZ/T2H support
dt-bindings: can: renesas,rcar-canfd: Document RZ/T2H and RZ/N2H SoCs
dt-bindings: can: renesas,rcar-canfd: Document RZ/V2H(P) and RZ/V2N SoCs
dt-bindings: can: renesas,rcar-canfd: Specify reset-names
can: rcar_canfd: Add support for FD-Only mode
dt-bindings: can: renesas,rcar-canfd: Document renesas,fd-only property
====================
As reported by Al Viro, the TCP ULP support for SMC is fundamentally
broken. The implementation attempts to convert an active TCP socket
into an SMC socket by modifying the underlying `struct file`, dentry,
and inode in-place, which violates core VFS invariants that assume
these structures are immutable for an open file, creating a risk of
use after free errors and general system instability.
Given the severity of this design flaw and the fact that cleaner
alternatives (e.g., LD_PRELOAD, BPF) exist for legacy application
transparency, the correct course of action is to remove this feature
entirely.
net: ax25: remove plumbing for never-implemented DAMA Master support
The AX25_DAMA_MASTER option has been unimplemented and marked broken
ever since it was introduced in 2007 in commit 954b2e7f4c37 ("[NET]
AX.25 Kconfig and docs updates and fixes"). At this point, it is very
unlikely it will be implemented. Remove it.
====================
net: wwan: add NMEA port type support
The series introduces a long discussed NMEA port type support for the
WWAN subsystem. There are two goals. From the WWAN driver perspective,
NMEA exported as any other port type (e.g. AT, MBIM, QMI, etc.). From
user space software perspective, the exported chardev belongs to the
GNSS class what makes it easy to distinguish desired port and the WWAN
device common to both NMEA and control (AT, MBIM, etc.) ports makes it
easy to locate a control port for the GNSS receiver activation.
Done by exporting the NMEA port via the GNSS subsystem with the WWAN
core acting as proxy between the WWAN modem driver and the GNSS
subsystem.
The series starts from a cleanup patch. Then three patches prepares the
WWAN core for the proxy style operation. Followed by a patch introding a
new WWNA port type, integration with the GNSS subsystem and demux. The
series ends with a couple of patches that introduce emulated EMEA port
to the WWAN HW simulator.
The series is the product of the discussion with Loic about the pros and
cons of possible models and implementation. Also Muhammad and Slark did
a great job defining the problem, sharing the code and pushing me to
finish the implementation. Daniele has caught an issue on driver
unloading and suggested an investigation direction. What was concluded
by Loic. Many thanks.
====================
Sergey Ryazanov [Mon, 26 Jan 2026 06:21:57 +0000 (14:21 +0800)]
net: wwan: hwsim: support NMEA port emulation
Support NMEA port emulation for the WWAN core GNSS port testing purpose.
Emulator produces pair of GGA + RMC sentences every second what should
be enough to fool gpsd into believing it is working with a NMEA GNSS
receiver.
If the GNSS system is enabled then one NMEA port will be created
automatically for the simulated WWAN device. Manual NMEA port creation
is not supported at the moment.
Sergey Ryazanov [Mon, 26 Jan 2026 06:21:56 +0000 (14:21 +0800)]
net: wwan: hwsim: refactor to support more port types
Just introduced WWAN NMEA port type needs a testing option. The WWAN HW
simulator was developed with the AT port type in mind and cannot be
easily extended. Refactor it now to make it capable to support more port
types.
No big functional changes, mostly renaming with a little code
rearrangement.
Sergey Ryazanov [Mon, 26 Jan 2026 06:21:55 +0000 (14:21 +0800)]
net: wwan: add NMEA port support
Many WWAN modems come with embedded GNSS receiver inside and have a
dedicated port to output geopositioning data. On the one hand, the
GNSS receiver has little in common with WWAN modem and just shares a
host interface and should be exported using the GNSS subsystem. On the
other hand, GNSS receiver is not automatically activated and needs a
generic WWAN control port (AT, MBIM, etc.) to be turned on. And a user
space software needs extra information to find the control port.
Introduce the new type of WWAN port - NMEA. When driver asks to register
a NMEA port, the core allocates common parent WWAN device as usual, but
exports the NMEA port via the GNSS subsystem and acts as a proxy between
the device driver and the GNSS subsystem.
From the WWAN device driver perspective, a NMEA port is registered as a
regular WWAN port without any difference. And the driver interacts only
with the WWAN core. From the user space perspective, the NMEA port is a
GNSS device which parent can be used to enumerate and select the proper
control port for the GNSS receiver management.
Sergey Ryazanov [Mon, 26 Jan 2026 06:21:54 +0000 (14:21 +0800)]
net: wwan: core: split port unregister and stop
Upcoming GNSS (NMEA) port type support requires exporting it via the
GNSS subsystem. On another hand, we still need to do basic WWAN core
work: call the port stop operation, purge queues, release the parent
WWAN device, etc. To reuse as much code as possible, split the port
unregistering function into the deregistration of a regular WWAN port
device, and the common port tearing down code.
In order to keep more code generic, break the device_unregister() call
into device_del() and put_device(), which release the port memory
uniformly.
Sergey Ryazanov [Mon, 26 Jan 2026 06:21:53 +0000 (14:21 +0800)]
net: wwan: core: split port creation and registration
Upcoming GNSS (NMEA) port type support requires exporting it via the
GNSS subsystem. On another hand, we still need to do basic WWAN core
work: find or allocate the WWAN device, make it the port parent, etc. To
reuse as much code as possible, split the port creation function into
the registration of a regular WWAN port device, and basic port struct
initialization.
To be able to use put_device() uniformly, break the device_register()
call into device_initialize() and device_add() and call device
initialization earlier.
While at it, fix a minor number leak upon WWAN port registration
failure.
We need information about existing WWAN device children since we remove
the device after removing the last child. Previously, we tracked users
implicitly by checking whether ops was registered and existence of a
child device of the wwan_class class. Upcoming GNSS (NMEA) port type
support breaks this approach by introducing a child device of the
gnss_class class.
And a modem driver can easily trigger a kernel Oops by removing regular
(e.g., MBIM, AT) ports first and then removing a GNSS port. The WWAN
device will be unregistered on removal of a last regular WWAN port. And
subsequent GNSS port removal will cause NULL pointer dereference in
simple_recursive_removal().
In order to support ports of classes other than wwan_class, switch to
explicit references counting. Introduce a dedicated counter to the WWAN
device struct, increment it on every wwan_create_dev() call, decrement
on wwan_remove_dev(), and actually unregister the WWAN device when there
are no more references.
Run tested with wwan_hwsim with NMEA support patches applied and
different port removing sequences.
Sergey Ryazanov [Mon, 26 Jan 2026 06:21:51 +0000 (14:21 +0800)]
net: wwan: core: remove unused port_id field
It was used initially for a port id allocation, then removed, and then
accidently introduced again, but it is still unused. Drop it again to
keep code clean.
Add debugfs hooks to display tx/rx rings for each napi
vector.
Note that the cloning mechanism in fbnic_ethtool.c for configuration
changes protects against concurrency issues with simultaneous config
changes along with debugs ring accesses.
The configuration switch builds up the new configuration offline,
takes the current config down, which removes the debugfs nv files, and
switches to the new configuration. The new configuration is brought
up which brings the debugfs files back on top of the new configuration
rings.
The interaction with fbnic_queue_stop() and fbnic_queue_start() will
similarly delete and add the files for the indicated vector.
net: usb: remove unnecessary get_drvinfo code and driver versions
Many USB network drivers define get_drvinfo functions which add no
value over usbnet_get_drvinfo, only setting the driver name and
version. usbnet_get_drvinfo automatically sets the driver name, and
separate driver versions are now frowned upon in the kernel. Remove all
driver versions and replace these get_drvinfo functions with references
to usbnet_get_drvinfo where possible. Where that is not possible,
remove unnecessary code to set the driver name. Also remove two
unnecessary initializations from aqc111_get_drvinfo, an inaccurate
comment in pegasus.c, and an unused macro in catc.c.
Reviewed-by: Andrew Lunn <andrew@lunn.ch> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Peter Korsgaard <peter@korsgaard.com> (for dm9601.c) Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com> Link: https://patch.msgid.link/20260129042435.13395-2-enelsonmoore@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This commit adds SNMP drop count increment for the packets in
per NUMA queues which were introduced in commit b650bf0977d3
("udp: remove busylock and add per NUMA queues"). note that SNMP
counters are incremented currently by the caller for skb. And
that these skbs on the intermediate queue cannot be counted
there so need similar logic in their error path.
Signed-off-by: Mahdi Faramarzpour <mahdifrmx@gmail.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20260129083806.204752-1-mahdifrmx@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Dumazet [Thu, 29 Jan 2026 15:34:58 +0000 (15:34 +0000)]
tcp: reduce tcp sockets size by one cache line
By default, when a kmem_cache is created with SLAB_TYPESAFE_BY_RCU,
slub has to use extra storage for the freelist pointer after each
object, because slub assumes that any bit in the object
can be used by RCU readers.
Because proto_register() is also using SLAB_HWCACHE_ALIGN,
this forces slub to use one extra cache line per object.
We can instead put the slub freelist anywhere in the object,
granted the concurrent RCU readers are not supposed to
use the pointer value.
Add a new (struct sock)sk_freeptr field, in an union
with sk_rcu: No RCU readers would need to look at sk_rcu,
which is only used at free phase.
Add functions to support xmit along with TSO/GSO.
Also, add functions to handle TX completion events in the NAPI context.
This commit introduces the fundamental transmit data path
Add support to receive packet using NAPI, build and deliver the skb
to stack. With help of meta data available in completions, fill the
appropriate information in skb.
Geetha sowjanya [Wed, 28 Jan 2026 02:24:48 +0000 (07:54 +0530)]
octeontx2-pf: cn10k/cn20k: Update count_eot in NPA_LF_AURA_BATCH_FREE0
This patch updates the count_eot calculation for CN20K devices.
Where the count_eot feild extended to 2 bits, while maintaining
CN10K compatibility where only bit 0 is used.
Jakub Kicinski [Fri, 30 Jan 2026 03:17:42 +0000 (19:17 -0800)]
Merge tag 'wireless-next-2026-01-29' of https://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next
Johannes Berg says:
====================
Another fairly large set of changes, notably:
- cfg80211/mac80211
- most of EPPKE/802.1X over auth frames support
- additional FTM capabilities
- split up drop reasons better, removing generic RX_DROP
- NAN cleanups/fixes
- ath11k:
- support for Channel Frequency Response measurement
- ath12k:
- support for the QCC2072 chipset
- iwlwifi:
- partial NAN support
- UNII-9 support
- some UHR/802.11bn FW APIs
- remove most of MLO/EHT from iwlmvm
(such devices use iwlmld)
- rtw89:
- preparations for RTL8922DE support
* tag 'wireless-next-2026-01-29' of https://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (184 commits)
wifi: iwlegacy: add missing mutex protection in il4965_store_tx_power()
wifi: iwlegacy: add missing mutex protection in il3945_store_measurement()
wifi: mac80211: use u64_stats_t with u64_stats_sync properly
wifi: p54: Fix memory leak in p54_beacon_update()
wifi: cfg80211: treat deprecated INDOOR_SP_AP_OLD control value as LPI mode
wifi: rtw88: sdio: Migrate to use sdio specific shutdown function
wifi: rsi: sdio: Migrate to use sdio specific shutdown function
sdio: Provide a bustype shutdown function
wifi: nl80211/cfg80211: support operating as RSTA in PMSR FTM request
wifi: nl80211/cfg80211: add negotiated burst period to FTM result
wifi: nl80211/cfg80211: clarify periodic FTM parameters for non-EDCA based ranging
wifi: nl80211/cfg80211: add new FTM capabilities
wifi: iwlwifi: rename struct iwl_mcc_allowed_ap_type_cmd::offset_map
wifi: iwlwifi: mvm: Remove link_id from time_events
wifi: iwlwifi: mld: change cluster_id type to u8 array
wifi: iwlwifi: support V13 of iwl_lari_config_change_cmd
wifi: iwlwifi: split bios_value_u32 to separate the header
wifi: iwlwifi: uefi: cache the DSM functions
wifi: iwlwifi: acpi: cache the DSM functions
wifi: iwlwifi: mvm: Cleanup MLO code
...
====================
Jakub Kicinski [Wed, 28 Jan 2026 19:32:58 +0000 (11:32 -0800)]
eth: bnxt: make sure we populate the qcfg defaults on old FW/HW
The driver now depends on the core to tell it what the rx page size
should be for the agg ring. We must populate the ndo_default_qcfg
callback even if we don't support any queue ops.
Jakub Kicinski [Fri, 30 Jan 2026 02:50:11 +0000 (18:50 -0800)]
Merge branch 'net-hinic3-fix-code-styles'
Fan Gong says:
====================
net: hinic3: Fix code styles
This patchset provides 4 code styles fixes:
Remove empty lines between error handling.
Remove defensive txq_num check.
Use array_size instead of multiplying.
SQ use SQ_CTXT_PREF_CI_HI to improve readability.
====================
Fan Gong [Thu, 29 Jan 2026 04:01:10 +0000 (12:01 +0800)]
hinic3: Use array_size instead of multiplying
Since commit 17fcb3dc12bb ("hinic3: module initialization and tx/rx logic")
use "size * sizeof(u64)" in hinic3_feature_nego.
Use array_size instead of multiplying can make it clearer.
Fan Gong [Thu, 29 Jan 2026 04:01:08 +0000 (12:01 +0800)]
hinic3: Fix code Style(remove empty lines between error handling)
According to previous review comment, fix code style of removing empty
lines between the actions on the error handling path to make it more
idiomatic in the merged code.
Geetha sowjanya [Tue, 27 Jan 2026 12:51:47 +0000 (18:21 +0530)]
octeontx2-af: Workaround SQM/PSE stalls by disabling sticky
NIX SQ manager sticky mode is known to cause stalls when multiple SQs
share an SMQ and transmit concurrently. Additionally, PSE may deadlock
on transitions between sticky and non-sticky transmissions. There is
also a credit drop issue observed when certain condition clocks are
gated.
work around these hardware errata by:
- Disabling SQM sticky operation:
- Clear TM6 (bit 15)
- Clear TM11 (bit 14)
- Disabling sticky → non-sticky transition path that can deadlock PSE:
- Clear TM5 (bit 23)
- Preventing credit drops by keeping the control-flow clock enabled:
- Set TM9 (bit 21)
These changes are applied via NIX_AF_SQM_DBG_CTL_STATUS. With this
configuration the SQM/PSE maintain forward progress under load without
credit loss, at the cost of disabling sticky optimizations.
The test_rss_flow_label_6only test case fails on devices that do not
support IPv6 flow label hashing. Make it skip neatly, consistent with
the behavior of the test_rss_flow_label case.
====================
net: stmmac: rk: second chunk of cleanups
This series is a cut-down installment of the dwmac-rk cleanups, covering
up to the point that the AI review had its first issue with the patches.
Unfortunately, this means that we introduce ->init but do not add any
users for it yet. That will be in the next round once this has got
through AI review and merged.
====================
All the definitions of the RGMII related xxx_GMAC_CLK_xxx definitions
use the same field values to select the clock rate. Provide common
definitions for these field values, passing them in to a single macro
for each variant that generates the appropriate values for the speed
register.
rk_gmac_setup() delves into the PHY's DT node to retrieve its reset
control using of_reset_control_get(). However, it never releases it
when the driver is removed. Add reset_control_put() to rk_gmac_exit()
to clean this up.
net: phylink: fix NULL pointer deref in phylink_major_config()
When a MAC driver returns a PCS for an interface mode, and then we
attempt to switch to a different mode that doesn't require a PCS,
this causes phylink to oops:
This is caused by "pcs" being NULL when we attempt to execute:
pcs->phylink = pl;
Make this conditional on pcs being non-null.
Fixes: 486dc391ef43 ("net: phylink: allow mac_select_pcs() to remove a PCS") Reported-by: Mohd Ayaan Anwar <mohd.anwar@oss.qualcomm.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1vl39Q-00000006utm-229h@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Report the active PHY interface from the point of view of the dwmac
hardware to the kernel log, where the core supports reading this.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Tested-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Link: https://patch.msgid.link/E1vl36F-00000006url-1fWA@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In commit 9b443e58a896 ("net: stmmac: qcom-ethqos: remove MAC_CTRL_REG
modification"), ethqos->mac_base is only written, never read. Let's
remove it.
Reviewed-by: Andrew Lunn <andrew@lunn.ch> Tested-by: Mohd Ayaan Anwar <mohd.anwar@oss.qualcomm.com> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://patch.msgid.link/E1vl39x-00000006uvc-0A3S@rmk-PC.armlinux.org.uk Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ivan Vecera [Mon, 26 Jan 2026 16:22:51 +0000 (17:22 +0100)]
dpll: expose fractional frequency offset in ppt
Currently, the dpll subsystem exports the fractional frequency offset
(FFO) in parts per million (ppm). This granularity is insufficient for
high-precision synchronization scenarios which often require parts per
trillion (ppt) resolution.
Add a new netlink attribute DPLL_A_PIN_FRACTIONAL_FREQUENCY_OFFSET_PPT
to expose the FFO in ppt.
Update the dpll netlink core to expect the driver-provided FFO value
to be in ppt. To maintain backward compatibility with existing userspace
tools, populate the legacy DPLL_A_PIN_FRACTIONAL_FREQUENCY_OFFSET
attribute by dividing the new ppt value by 1,000,000.
Update the zl3073x and mlx5 drivers to provide the FFO value in ppt:
- zl3073x: adjust the fixed-point calculation to produce ppt (10^12)
instead of ppm (10^6).
- mlx5: scale the existing ppm value by 1,000,000.
Signed-off-by: Ivan Vecera <ivecera@redhat.com> Reviewed-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Link: https://patch.msgid.link/20260126162253.27890-1-ivecera@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Andre Carvalho [Tue, 27 Jan 2026 19:39:20 +0000 (19:39 +0000)]
netconsole: selftests: Move netconsole selftests to separate target
This patch moves netconsole selftests from drivers/net to its own target
in drivers/net/netconsole.
This change helps saving some resources from CI since tests in
drivers/net automatically run against real hardware which are not used
by netconsole tests as they rely solely on netdevsim.
lib_netcons.sh is kept under drivers/net/lib since it is also used by
bonding selftests. Finally, drivers/net config remains unchanged as
netpoll_basic.py requires netconsole (and does leverage real HW testing).
Jakub Kicinski [Fri, 30 Jan 2026 01:22:35 +0000 (17:22 -0800)]
Merge tag 'for-net-next-2026-01-29' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next
Luiz Augusto von Dentz says:
====================
bluetooth-next pull request for net-next:
core:
- L2CAP: Add support for setting BT_PHY
- HCI: Add LE Channel Sounding HCI Command/event structures
- hci_conn: Set link_policy on incoming ACL connections
- MGMT: Add idle_timeout to configurable system parameters
drivers:
- btusb: Add support for MediaTek7920 0489:e158
- btusb: Add device ID for Realtek RTL8761BU
- btusb: Reject autosuspend if discovery is active
- btusb: Add new VID/PID for RTL8852CE
- btusb: Add USB ID 0489:e112 for Realtek 8851BE
- dt-bindings: qcom: Split to separate schema
- btqca: Add WCN6855 firmware priority selection feature
* tag 'for-net-next-2026-01-29' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next: (40 commits)
Bluetooth: hci_qca: Enable HFP hardware offload for WCN6855 and WCN7850
Bluetooth: hci_qca: Refactor HFP hardware offload capability handling
Bluetooth: btusb: Add USB ID 7392:e611 for Edimax EW-7611UXB
Bluetooth: L2CAP: Fix not tracking outstanding TX ident
Bluetooth: MGMT: Fix memory leak in set_ssp_complete
Bluetooth: hci_uart: fix null-ptr-deref in hci_uart_write_work
Bluetooth: btusb: Remove duplicate entry for 0x13d3/0x3618
Bluetooth: btusb: Add device ID for Realtek RTL8761BU
Bluetooth: btintel: Remove unneeded CONFIG_PM* #ifdef's
Bluetooth: btnxpuart: Remove unneeded CONFIG_PM ifdef
Bluetooth: btusb: Add new VID/PID for RTL8852CE
Bluetooth: btintel_pcie: Remove unnecessary check before kfree_skb()
Bluetooth: btusb: Reject autosuspend if discovery is active
Bluetooth: hci_core: Export hci_discovery_active
Bluetooth: btusb: Use pm_ptr instead of #ifdef CONFIG_PM
dt-bindings: bluetooth: qcom,wcn7850-bt: Deprecate old supplies
dt-bindings: bluetooth: qcom,wcn7850-bt: Split to separate schema
dt-bindings: bluetooth: qcom,wcn6855-bt: Deprecate old supplies
dt-bindings: bluetooth: qcom,wcn6855-bt: Split to separate schema
dt-bindings: bluetooth: qcom,wcn6750-bt: Deprecate old supplies
...
====================
Mengshi Wu [Tue, 27 Jan 2026 02:54:23 +0000 (10:54 +0800)]
Bluetooth: hci_qca: Enable HFP hardware offload for WCN6855 and WCN7850
Add QCA_CAP_HFP_HW_OFFLOAD capability flag to WCN6855 and WCN7850
device data structures to enable Hands-Free Profile (HFP) hardware
offload support on these Qualcomm Bluetooth chipsets.
Signed-off-by: Mengshi Wu <mengshi.wu@oss.qualcomm.com> Acked-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Replace SoC-specific check with capability-based approach for HFP
hardware offload configuration. Add QCA_CAP_HFP_HW_OFFLOAD capability
flag and support_hfp_hw_offload field to qca_serdev structure. Add
QCA_CAP_HFP_HW_OFFLOAD capability flag to QCA2066 device data
structures.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com> Signed-off-by: Mengshi Wu <mengshi.wu@oss.qualcomm.com> Acked-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Bluetooth: L2CAP: Fix not tracking outstanding TX ident
This attempts to proper track outstanding request by using struct ida
and allocating from it in l2cap_get_ident using ida_alloc_range which
would reuse ids as they are free, then upon completion release
the id using ida_free.
This fixes the qualification test case L2CAP/COS/CED/BI-29-C which
attempts to check if the host stack is able to work after 256 attempts
to connect which requires Ident field to use the full range of possible
values in order to pass the test.
projects / thirdparty / kernel / linux.git / log
