wireguard-linux/drivers/net/ethernet/intel, branch stable

xsk: honor XDP_TX_METADATA in zero-copy path

2026-08-24T18:12:00Z

The zero-copy path reads TX metadata whenever the UMEM has metadata space, even if the descriptor does not set XDP_TX_METADATA. Pass descriptor options through the metadata helpers and ignore metadata unless the option is set. This does not fix the existing per-WQE metadata handling for mlx5 MPWQEs. Only the descriptor that starts a session passes through xsk_tx_metadata_request() and configures offload state shared by the batch. Metadata on descriptors joining an open session is therefore not validated and does not configure its requested offloads. In addition, a non-NULL metadata pointer from such a descriptor is treated as a timestamp completion request even when XDP_TXMD_FLAGS_TIMESTAMP is not set, so its metadata union can be overwritten with an unrequested timestamp. Fixing mixed metadata states within one MPWQE requires a separate change. Fixes: 48eb03dd2630 ("xsk: Add TX timestamp and TX checksum offload support") Reviewed-by: Alexander Lobakin Signed-off-by: Stanislav Fomichev Reviewed-by: Maciej Fijalkowski Reviewed-by: Aleksandr Loktionov Link: https://patch.msgid.link/20260819160535.1472459-3-sdf@fomichev.me Signed-off-by: Jakub Kicinski

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2026-08-18T17:42:41Z

Merge in late fixes in preparation for the net-next PR. Conflicts: drivers/dpll/dpll_core.c drivers/dpll/dpll_netlink.c 33f016b23a219 ("dpll: fix NULL deref in dpll_device_ops() during teardown race") b1d0c412088e3 ("dpll: add STATE_CONNECTED_OVERRIDE pin capability") https://lore.kernel.org/aoR9YYY2P5--3x0N@sirena.org.uk https://lore.kernel.org/aoR9VmKllVGwmQn_@sirena.org.uk No adjacent changes. Signed-off-by: Jakub Kicinski

Merge branch '200GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue

2026-08-18T01:01:55Z

Tony Nguyen says: ==================== Introduce iXD driver Larysa Zaremba says: This patch series adds the iXD driver, which supports the Intel(R) Control Plane PCI Function on Intel E2100 and later IPUs and FNICs. It facilitates a centralized control over multiple IDPF PFs/VFs/SFs exposed by the same card. The reason for the separation is to be able to offload the control plane to the host different from where the data plane is running. This is the first phase in the release of this driver where we implement the initialization of the core PCI driver. Subsequent phases will implement advanced features like usage of idpf ethernet aux device, link management, NVM update via devlink, switchdev port representors, data and exception path, flow rule programming, etc. The first phase entails the following aspects: 1. Additional libie functionalities: Patches 1-5 introduce additional common library API for drivers to communicate with the control plane through mailbox communication. A control queue is a hardware interface which is used by the driver to interact with other subsystems (like firmware). The library APIs allow the driver to setup and configure the control queues to send and receive virtchnl messages. The library has an internal bookkeeping (XN API) mechanism to keep track of the send messages. It supports both synchronous as well as asynchronous way of handling the messages. The library also handles the timeout internally for synchronous messages using events. This reduces the driver's overhead in handling the timeout error cases. The current patch series supports only APIs that are needed for device initialization. These include APIs in the libie_pci module: * Allocating/freeing the DMA memory and mapping the MMIO regions for BAR0, read/write APIs for drivers to access the MMIO memory and libie_cp module: * Control queue initialization and configuration * Transport initialization for bookkeeping * Blocking and asynchronous mailbox transactions Once the mailbox is initialized, the drivers can send and receive virtchnl messages to/from the control plane. The modules above are not supposed to be linked with the main libie library, but do share the folder with it. 2. idpf: Patches 6-11 refactor the idpf driver to use the libie APIs for control queue configuration, virtchnl transaction, device initialization and reset and adjust related code accordingly. 3. ixd: Patches 12-15 add the ixd driver and implement multiple pieces of the initialization flow as follows: * Add the ability to load * A reset is issued to ensure a clean device state, followed by initialization of the mailbox * Device capabilities: As part of initialization, the driver has to determine what the device is capable of (ex. max queues, vports, etc). This information is obtained from the firmware and stored by the driver. * Enable initial support for the devlink interface * '200GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue: ixd: add devlink support ixd: add the core initialization ixd: add reset checks and initialize the mailbox ixd: add basic driver framework for Intel(R) Control Plane Function idpf: print a debug message and bail in case of non-event ctlq message idpf: make mbx_task queueing and cancelling more consistent idpf: refactor idpf to use libie control queues idpf: refactor idpf to use libie_pci APIs idpf: remove unused code for getting RSS info from device idpf: remove 'vport_params_reqd' field libie: add bookkeeping support for control queue messages libie: add control queue support libeth: allow to create fill queues without NAPI libie: add PCI device initialization helpers to libie virtchnl: move virtchnl and virtchnl2 headers to 'include/linux/net/intel' ==================== Link: https://patch.msgid.link/20260812212532.905873-1-anthony.l.nguyen@intel.com Signed-off-by: Jakub Kicinski

net: add READ_ONCE()/WRITE_ONCE() annotations for dev->num_tc

2026-08-17T17:27:48Z

Several fast-path and control-path lockless readers access dev->num_tc (e.g., skb_tx_hash(), netdev_txq_to_tc(), netdev_get_num_tc(), and qdisc/driver lookups) while concurrent writers update dev->num_tc during TC setup, device reset, or channel configuration. Add READ_ONCE() and WRITE_ONCE() annotations to prevent compiler reordering and load/store tearing when accessing dev->num_tc. Update inline helpers in netdevice.h (netdev_get_num_tc(), netdev_set_prio_tc_map(), and netdev_get_sb_channel()) as well as writers and lockless readers in core networking code and drivers. Signed-off-by: Eric Dumazet Link: https://patch.msgid.link/20260812085440.3917924-3-edumazet@google.com Signed-off-by: Jakub Kicinski

net: prevent torn reads in netdev_tc_txq

2026-08-17T17:27:48Z

netdev_set_tc_queue() (and related helpers/drivers such as netdev_bind_sb_channel_queue(), netdev_reset_tc(), and netdev_unbind_sb_channel()) perform separate 16-bit writes to dev->tc_to_txq[tc].count and dev->tc_to_txq[tc].offset. Furthermore, memset() in netdev_reset_tc() and netdev_unbind_sb_channel() provides no guarantee of performing full 32-bit word stores. Concurrent lockless readers (e.g. skb_tx_hash(), netdev_txq_to_tc(), ixgbe_select_queue(), taprio, mqprio, FPE drivers) can observe torn values where offset and count belong to inconsistent configurations. Redefine struct netdev_tc_txq to embed count and offset inside a union with a u32 combined field, allowing atomic manipulation via READ_ONCE() and WRITE_ONCE(). Update all lockless readers and writers across the kernel to use READ_ONCE() and WRITE_ONCE() on the combined field. Signed-off-by: Eric Dumazet Link: https://patch.msgid.link/20260812085440.3917924-2-edumazet@google.com Signed-off-by: Jakub Kicinski

ixd: add devlink support

2026-08-12T20:56:51Z

Enable initial support for the devlink interface with the ixd driver. The ixd hardware is a single function PCIe device. So, the PCIe adapter gets its own devlink instance to manage device-wide resources or configuration. $ devlink dev show pci/0000:83:00.6 $ devlink dev info pci/0000:83:00.6 pci/0000:83:00.6: driver ixd serial_number 00-a0-c9-ff-ff-23-45-67 versions: fixed: device.type MEV running: fw.mgmt.api 2.0 Signed-off-by: Amritha Nambiar Reviewed-by: Michal Swiatkowski Reviewed-by: Maciej Fijalkowski Reviewed-by: Przemek Kitszel Tested-by: Bharath R Signed-off-by: Larysa Zaremba Signed-off-by: Tony Nguyen

ixd: add the core initialization

2026-08-12T20:56:51Z

As the mailbox is setup, initialize the core. This makes use of the send and receive mailbox message framework for virtchnl communication between the driver and device Control Plane (CP). To start with, driver confirms the virtchnl version with the CP. Once that is done, it requests and gets the required capabilities and resources needed such as max vectors, queues, vports etc. Use a unified way of handling the virtchnl messages, where a single function handles all related memory management and the caller only provides the callbacks to fill the send buffer and to handle the response. Place generic control queue message handling separately to facilitate the addition of protocols other than virtchannel in the future. Co-developed-by: Amritha Nambiar Signed-off-by: Amritha Nambiar Reviewed-by: Maciej Fijalkowski Tested-by: Bharath R Signed-off-by: Larysa Zaremba Signed-off-by: Tony Nguyen

ixd: add reset checks and initialize the mailbox

2026-08-12T20:56:51Z

At the end of the probe, trigger hard reset, initialize and schedule the after-reset task. If the reset is complete in a pre-determined time, initialize the default mailbox, through which other resources will be negotiated. Co-developed-by: Amritha Nambiar Signed-off-by: Amritha Nambiar Reviewed-by: Maciej Fijalkowski Reviewed-by: Aleksandr Loktionov Tested-by: Bharath R Signed-off-by: Larysa Zaremba Signed-off-by: Tony Nguyen

ixd: add basic driver framework for Intel(R) Control Plane Function

2026-08-12T20:56:51Z

Add module register and probe functionality. Add the required support to register IXD PCI driver, as well as probe, remove and shutdown callbacks. Enable the PCI device and request to reserve the memory resources that will be used by the driver. Finally map the BAR0 address space. For now, use devm_kzalloc() to allocate adapter, as it requires the least amount of code. In a later commit, it will be replaced with a devlink alternative. Co-developed-by: Amritha Nambiar Signed-off-by: Amritha Nambiar Reviewed-by: Maciej Fijalkowski Tested-by: Bharath R Signed-off-by: Larysa Zaremba Signed-off-by: Tony Nguyen

idpf: print a debug message and bail in case of non-event ctlq message

2026-08-12T20:56:51Z

Unlike previous internal idpf ctlq implementation, libie_cp calls the default message handler for all received messages that do not have a matching xn transaction, not only for VIRTCHNL2_OP_EVENT. This leads to many error messages printing garbage, because the parsing expected a valid event message, but got e.g. a delayed response for a timed-out transaction. The information about timed-out transactions and otherwise unhandleable messages can still be valuable for developers, so print the information with dynamic debug and exit the function, so the following functions can parse valid events in peace. Reviewed-by: Aleksandr Loktionov Reviewed-by: Michal Kubiak Tested-by: Samuel Salin Signed-off-by: Larysa Zaremba Signed-off-by: Tony Nguyen