Extend the MAC_TCR_SS (Speed Select) register field width from 2 bits
to 3 bits to properly support all speed settings.
The MAC_TCR register's SS field encoding requires 3 bits to represent
all supported speeds:
- 0x00: 10Gbps (XGMII)
- 0x02: 2.5Gbps (GMII) / 100Mbps
- 0x03: 1Gbps / 10Mbps
- 0x06: 2.5Gbps (XGMII) - P100a only
With only 2 bits, values 0x04-0x07 cannot be represented, which breaks
2.5G XGMII mode on newer platforms and causes incorrect speed select
values to be programmed.
Fixes: 07445f3c7c ("amd-xgbe: Add support for 10 Mbps speed")
Co-developed-by: Guruvendra Punugupati <Guruvendra.Punugupati@amd.com>
Signed-off-by: Guruvendra Punugupati <Guruvendra.Punugupati@amd.com>
Signed-off-by: Raju Rangoju <Raju.Rangoju@amd.com>
Link: https://patch.msgid.link/20260226170753.250312-1-Raju.Rangoju@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When both eth interfaces with links up are added to a bridge or hsr
interface, ping fails if the link speed is not 1Gbps (e.g., 100Mbps).
The issue is seen because when switching to offload (bridge/hsr) mode,
prueth_emac_restart() restarts the firmware and clears DRAM with
memset_io(), setting all memory to 0. This includes PORT_LINK_SPEED_OFFSET
which firmware reads for link speed. The value 0 corresponds to
FW_LINK_SPEED_1G (0x00), so for 1Gbps links the default value is correct
and ping works. For 100Mbps links, the firmware needs FW_LINK_SPEED_100M
(0x01) but gets 0 instead, causing ping to fail. The function
emac_adjust_link() is called to reconfigure, but it detects no state change
(emac->link is still 1, speed/duplex match PHY) so new_state remains false
and icssg_config_set_speed() is never called to correct the firmware speed
value.
The fix resets emac->link to 0 before calling emac_adjust_link() in
prueth_emac_common_start(). This forces new_state=true, ensuring
icssg_config_set_speed() is called to write the correct speed value to
firmware memory.
Fixes: 06feac1540 ("net: ti: icssg-prueth: Fix emac link speed handling")
Signed-off-by: MD Danish Anwar <danishanwar@ti.com>
Link: https://patch.msgid.link/20260226102356.2141871-1-danishanwar@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit 31a7a0bbeb ("dpaa2-switch: add bounds check for if_id in IRQ
handler") introduces a range check for if_id to avoid an out-of-bounds
access. If an out-of-bounds if_id is detected, the interrupt status is
not cleared. This may result in an interrupt storm.
Clear the interrupt status after detecting an out-of-bounds if_id to avoid
the problem.
Found by an experimental AI code review agent at Google.
Fixes: 31a7a0bbeb ("dpaa2-switch: add bounds check for if_id in IRQ handler")
Cc: Junrui Luo <moonafterrain@outlook.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Link: https://patch.msgid.link/20260227055812.1777915-1-linux@roeck-us.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2026-02-19 (idpf, ice, i40e, ixgbevf, e1000e)
For idpf:
Li Li moves the check for software marker to occur after incrementing
next to clean to avoid re-encountering the same packet. He also adds a
couple of checks to prevent NULL pointer dereferences and NULLs rss_key,
after free, in error path so that later checks are properly evaluated.
Brian Vazquez adjusts IRQ naming to have correlation with netdev naming.
Sreedevi removes validation of action type as part of ntuple rule
deletion.
For ice:
Aaron Ma breaks RDMA initialization into two steps and adjusts calls so
that VSIs are entirely configured before plugging.
Michal Schmidt fixes initialization of loopback VSI to have proper
resources allocated to allow for loopback testing to occur.
For i40e:
Thomas Gleixner fixes a leak of preempt count by replacing get_cpu()
with smp_processor_id().
For ixgbevf:
Jedrzej adds a check for mailbox version before attempting to call an
associated link state call that is supported in that mailbox version.
For e1000e:
Vitaly clears power gating feature for Panther Lake systems to avoid
packet issues.
* '200GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
e1000e: clear DPG_EN after reset to avoid autonomous power-gating
e1000e: introduce new board type for Panther Lake PCH
ixgbevf: fix link setup issue
i40e: Fix preempt count leak in napi poll tracepoint
ice: fix crash in ethtool offline loopback test
ice: recap the VSI and QoS info after rebuild
idpf: Fix flow rule delete failure due to invalid validation
idpf: change IRQ naming to match netdev and ethtool queue numbering
idpf: nullify pointers after they are freed
idpf: skip deallocating txq group's txqs if it is NULL
idpf: skip deallocating bufq_sets from rx_qgrp if it is NULL
idpf: increment completion queue next_to_clean in sw marker wait routine
====================
Link: https://patch.msgid.link/20260225211546.1949260-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MANA hardware requires at least one doorbell ring every 8 wraparounds
of the CQ. The driver rings the doorbell as a form of flow control to
inform hardware that CQEs have been consumed.
The NAPI poll functions mana_poll_tx_cq() and mana_poll_rx_cq() can
poll up to CQE_POLLING_BUFFER (512) completions per call. If the CQ
has fewer than 512 entries, a single poll call can process more than
4 wraparounds without ringing the doorbell. The doorbell threshold
check also uses ">" instead of ">=", delaying the ring by one extra
CQE beyond 4 wraparounds. Combined, these issues can cause the driver
to exceed the 8-wraparound hardware limit, leading to missed
completions and stalled queues.
Fix this by capping the number of CQEs polled per call to 4 wraparounds
of the CQ in both TX and RX paths. Also change the doorbell threshold
from ">" to ">=" so the doorbell is rung as soon as 4 wraparounds are
reached.
Cc: stable@vger.kernel.org
Fixes: 58a63729c9 ("net: mana: Fix doorbell out of order violation and avoid unnecessary doorbell rings")
Signed-off-by: Long Li <longli@microsoft.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Link: https://patch.msgid.link/20260226192833.1050807-1-longli@microsoft.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In the current implementation, flushing multicast entries in MAC mode
incorrectly deletes entries for all ports instead of only the target port,
disrupting multicast traffic on other ports. The cause is adding multicast
entries by setting only host port bit, and not setting the MAC port bits.
Fix this by setting the MAC port's bit in the port mask while adding the
multicast entry. Also fix the flush logic to preserve the host port bit
during removal of MAC port and free ALE entries when mask contains only
host port.
Fixes: 5c50a856d5 ("drivers: net: ethernet: cpsw: add multicast address to ALE table")
Signed-off-by: Chintan Vankar <c-vankar@ti.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20260224181359.2055322-1-c-vankar@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Paolo Abeni:
"Including fixes from IPsec, Bluetooth and netfilter
Current release - regressions:
- wifi: fix dev_alloc_name() return value check
- rds: fix recursive lock in rds_tcp_conn_slots_available
Current release - new code bugs:
- vsock: lock down child_ns_mode as write-once
Previous releases - regressions:
- core:
- do not pass flow_id to set_rps_cpu()
- consume xmit errors of GSO frames
- netconsole: avoid OOB reads, msg is not nul-terminated
- netfilter: h323: fix OOB read in decode_choice()
- tcp: re-enable acceptance of FIN packets when RWIN is 0
- udplite: fix null-ptr-deref in __udp_enqueue_schedule_skb().
- wifi: brcmfmac: fix potential kernel oops when probe fails
- phy: register phy led_triggers during probe to avoid AB-BA deadlock
- eth:
- bnxt_en: fix deleting of Ntuple filters
- wan: farsync: fix use-after-free bugs caused by unfinished tasklets
- xscale: check for PTP support properly
Previous releases - always broken:
- tcp: fix potential race in tcp_v6_syn_recv_sock()
- kcm: fix zero-frag skb in frag_list on partial sendmsg error
- xfrm:
- fix race condition in espintcp_close()
- always flush state and policy upon NETDEV_UNREGISTER event
- bluetooth:
- purge error queues in socket destructors
- fix response to L2CAP_ECRED_CONN_REQ
- eth:
- mlx5:
- fix circular locking dependency in dump
- fix "scheduling while atomic" in IPsec MAC address query
- gve: fix incorrect buffer cleanup for QPL
- team: avoid NETDEV_CHANGEMTU event when unregistering slave
- usb: validate USB endpoints"
* tag 'net-7.0-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (72 commits)
netfilter: nf_conntrack_h323: fix OOB read in decode_choice()
dpaa2-switch: validate num_ifs to prevent out-of-bounds write
net: consume xmit errors of GSO frames
vsock: document write-once behavior of the child_ns_mode sysctl
vsock: lock down child_ns_mode as write-once
selftests/vsock: change tests to respect write-once child ns mode
net/mlx5e: Fix "scheduling while atomic" in IPsec MAC address query
net/mlx5: Fix missing devlink lock in SRIOV enable error path
net/mlx5: E-switch, Clear legacy flag when moving to switchdev
net/mlx5: LAG, disable MPESW in lag_disable_change()
net/mlx5: DR, Fix circular locking dependency in dump
selftests: team: Add a reference count leak test
team: avoid NETDEV_CHANGEMTU event when unregistering slave
net: mana: Fix double destroy_workqueue on service rescan PCI path
MAINTAINERS: Update maintainer entry for QUALCOMM ETHQOS ETHERNET DRIVER
dpll: zl3073x: Remove redundant cleanup in devm_dpll_init()
selftests/net: packetdrill: Verify acceptance of FIN packets when RWIN is 0
tcp: re-enable acceptance of FIN packets when RWIN is 0
vsock: Use container_of() to get net namespace in sysctl handlers
net: usb: kaweth: validate USB endpoints
...
The driver obtains sw_attr.num_ifs from firmware via dpsw_get_attributes()
but never validates it against DPSW_MAX_IF (64). This value controls
iteration in dpaa2_switch_fdb_get_flood_cfg(), which writes port indices
into the fixed-size cfg->if_id[DPSW_MAX_IF] array. When firmware reports
num_ifs >= 64, the loop can write past the array bounds.
Add a bound check for num_ifs in dpaa2_switch_init().
dpaa2_switch_fdb_get_flood_cfg() appends the control interface (port
num_ifs) after all matched ports. When num_ifs == DPSW_MAX_IF and all
ports match the flood filter, the loop fills all 64 slots and the control
interface write overflows by one entry.
The check uses >= because num_ifs == DPSW_MAX_IF is also functionally
broken.
build_if_id_bitmap() silently drops any ID >= 64:
if (id[i] < DPSW_MAX_IF)
bmap[id[i] / 64] |= ...
Fixes: 539dda3c5d ("staging: dpaa2-switch: properly setup switching domains")
Signed-off-by: Junrui Luo <moonafterrain@outlook.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Link: https://patch.msgid.link/SYBPR01MB78812B47B7F0470B617C408AAF74A@SYBPR01MB7881.ausprd01.prod.outlook.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Fix a "scheduling while atomic" bug in mlx5e_ipsec_init_macs() by
replacing mlx5_query_mac_address() with ether_addr_copy() to get the
local MAC address directly from netdev->dev_addr.
The issue occurs because mlx5_query_mac_address() queries the hardware
which involves mlx5_cmd_exec() that can sleep, but it is called from
the mlx5e_ipsec_handle_event workqueue which runs in atomic context.
The MAC address is already available in netdev->dev_addr, so no need
to query hardware. This avoids the sleeping call and resolves the bug.
Call trace:
BUG: scheduling while atomic: kworker/u112:2/69344/0x00000200
__schedule+0x7ab/0xa20
schedule+0x1c/0xb0
schedule_timeout+0x6e/0xf0
__wait_for_common+0x91/0x1b0
cmd_exec+0xa85/0xff0 [mlx5_core]
mlx5_cmd_exec+0x1f/0x50 [mlx5_core]
mlx5_query_nic_vport_mac_address+0x7b/0xd0 [mlx5_core]
mlx5_query_mac_address+0x19/0x30 [mlx5_core]
mlx5e_ipsec_init_macs+0xc1/0x720 [mlx5_core]
mlx5e_ipsec_build_accel_xfrm_attrs+0x422/0x670 [mlx5_core]
mlx5e_ipsec_handle_event+0x2b9/0x460 [mlx5_core]
process_one_work+0x178/0x2e0
worker_thread+0x2ea/0x430
Fixes: cee137a634 ("net/mlx5e: Handle ESN update events")
Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20260224114652.1787431-6-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The cited commit miss to add locking in the error path of
mlx5_sriov_enable(). When pci_enable_sriov() fails,
mlx5_device_disable_sriov() is called to clean up. This cleanup function
now expects to be called with the devlink instance lock held.
Add the missing devl_lock(devlink) and devl_unlock(devlink)
Fixes: 84a433a40d ("net/mlx5: Lock mlx5 devlink reload callbacks")
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20260224114652.1787431-5-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The cited commit introduced MLX5_PRIV_FLAGS_SWITCH_LEGACY to identify
when a transition to legacy mode is requested via devlink. However, the
logic failed to clear this flag if the mode was subsequently changed
back to MLX5_ESWITCH_OFFLOADS (switchdev). Consequently, if a user
toggled from legacy to switchdev, the flag remained set, leaving the
driver with wrong state indicating
Fix this by explicitly clearing the MLX5_PRIV_FLAGS_SWITCH_LEGACY bit
when the requested mode is MLX5_ESWITCH_OFFLOADS.
Fixes: 2a4f56fbcc ("net/mlx5e: Keep netdev when leave switchdev for devlink set legacy only")
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20260224114652.1787431-4-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
While testing corner cases in the driver, a use-after-free crash
was found on the service rescan PCI path.
When mana_serv_reset() calls mana_gd_suspend(), mana_gd_cleanup()
destroys gc->service_wq. If the subsequent mana_gd_resume() fails
with -ETIMEDOUT or -EPROTO, the code falls through to
mana_serv_rescan() which triggers pci_stop_and_remove_bus_device().
This invokes the PCI .remove callback (mana_gd_remove), which calls
mana_gd_cleanup() a second time, attempting to destroy the already-
freed workqueue. Fix this by NULL-checking gc->service_wq in
mana_gd_cleanup() and setting it to NULL after destruction.
Call stack of issue for reference:
[Sat Feb 21 18:53:48 2026] Call Trace:
[Sat Feb 21 18:53:48 2026] <TASK>
[Sat Feb 21 18:53:48 2026] mana_gd_cleanup+0x33/0x70 [mana]
[Sat Feb 21 18:53:48 2026] mana_gd_remove+0x3a/0xc0 [mana]
[Sat Feb 21 18:53:48 2026] pci_device_remove+0x41/0xb0
[Sat Feb 21 18:53:48 2026] device_remove+0x46/0x70
[Sat Feb 21 18:53:48 2026] device_release_driver_internal+0x1e3/0x250
[Sat Feb 21 18:53:48 2026] device_release_driver+0x12/0x20
[Sat Feb 21 18:53:48 2026] pci_stop_bus_device+0x6a/0x90
[Sat Feb 21 18:53:48 2026] pci_stop_and_remove_bus_device+0x13/0x30
[Sat Feb 21 18:53:48 2026] mana_do_service+0x180/0x290 [mana]
[Sat Feb 21 18:53:48 2026] mana_serv_func+0x24/0x50 [mana]
[Sat Feb 21 18:53:48 2026] process_one_work+0x190/0x3d0
[Sat Feb 21 18:53:48 2026] worker_thread+0x16e/0x2e0
[Sat Feb 21 18:53:48 2026] kthread+0xf7/0x130
[Sat Feb 21 18:53:48 2026] ? __pfx_worker_thread+0x10/0x10
[Sat Feb 21 18:53:48 2026] ? __pfx_kthread+0x10/0x10
[Sat Feb 21 18:53:48 2026] ret_from_fork+0x269/0x350
[Sat Feb 21 18:53:48 2026] ? __pfx_kthread+0x10/0x10
[Sat Feb 21 18:53:48 2026] ret_from_fork_asm+0x1a/0x30
[Sat Feb 21 18:53:48 2026] </TASK>
Fixes: 505cc26bca ("net: mana: Add support for auxiliary device servicing events")
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Dipayaan Roy <dipayanroy@linux.microsoft.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/aZ2bzL64NagfyHpg@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Panther Lake systems introduced an autonomous power gating feature for
the integrated Gigabit Ethernet in shutdown state (S5) state. As part of
it, the reset value of DPG_EN bit was changed to 1. Clear this bit after
performing hardware reset to avoid errors such as Tx/Rx hangs, or packet
loss/corruption.
Fixes: 0c9183ce61 ("e1000e: Add support for the next LOM generation")
Signed-off-by: Vitaly Lifshits <vitaly.lifshits@intel.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
It may happen that VF spawned for E610 adapter has problem with setting
link up. This happens when ixgbevf supporting mailbox API 1.6 cooperates
with PF driver which doesn't support this version of API, and hence
doesn't support new approach for getting PF link data.
In that case VF asks PF to provide link data but as PF doesn't support
it, returns -EOPNOTSUPP what leads to early bail from link configuration
sequence.
Avoid such situation by using legacy VFLINKS approach whenever negotiated
API version is less than 1.6.
To reproduce the issue just create VF and set its link up - adapter must
be any from the E610 family, ixgbevf must support API 1.6 or higher while
ixgbevf must not.
Fixes: 53f0eb62b4 ("ixgbevf: fix getting link speed data for E610 devices")
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Piotr Kwapulinski <piotr.kwapulinski@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Cc: stable@vger.kernel.org
Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Fix IRDMA hardware initialization timeout (-110) after resume by
separating VSI-dependent configuration from RDMA resource allocation,
ensuring VSI is rebuilt before IRDMA accesses it.
After resume from suspend, IRDMA hardware initialization fails:
ice: IRDMA hardware initialization FAILED init_state=4 status=-110
Separate RDMA initialization into two phases:
1. ice_init_rdma() - Allocate resources only (no VSI/QoS access, no plug)
2. ice_rdma_finalize_setup() - Assign VSI/QoS info and plug device
This allows:
- ice_init_rdma() to stay in ice_resume() (mirrors ice_deinit_rdma()
in ice_suspend())
- VSI assignment deferred until after ice_vsi_rebuild() completes
- QoS info updated after ice_dcb_rebuild() completes
- Device plugged only when control queues, VSI, and DCB are all ready
Fixes: bc69ad7486 ("ice: avoid IRQ collision to fix init failure on ACPI S3 resume")
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Aaron Ma <aaron.ma@canonical.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
When deleting a flow rule using "ethtool -N <dev> delete <location>",
idpf_sideband_action_ena() incorrectly validates fsp->ring_cookie even
though ethtool doesn't populate this field for delete operations. The
uninitialized ring_cookie may randomly match RX_CLS_FLOW_DISC or
RX_CLS_FLOW_WAKE, causing validation to fail and preventing legitimate
rule deletions. Remove the unnecessary sideband action enable check and
ring_cookie validation during delete operations since action validation
is not required when removing existing rules.
Fixes: ada3e24b84 ("idpf: add flow steering support")
Signed-off-by: Sreedevi Joshi <sreedevi.joshi@intel.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
The code uses the vidx for the IRQ name but that doesn't match ethtool
reporting nor netdev naming, this makes it hard to tune the device and
associate queues with IRQs. Sequentially requesting irqs starting from
'0' makes the output consistent.
This commit changes the interrupt numbering but preserves the name
format, maintaining ABI compatibility. Existing tools relying on the old
numbering are already non-functional, as they lack a useful correlation
to the interrupts.
Before:
ethtool -L eth1 tx 1 combined 3
grep . /proc/irq/*/*idpf*/../smp_affinity_list
/proc/irq/67/idpf-Mailbox-0/../smp_affinity_list:0-55,112-167
/proc/irq/68/idpf-eth1-TxRx-1/../smp_affinity_list:0
/proc/irq/70/idpf-eth1-TxRx-3/../smp_affinity_list:1
/proc/irq/71/idpf-eth1-TxRx-4/../smp_affinity_list:2
/proc/irq/72/idpf-eth1-Tx-5/../smp_affinity_list:3
ethtool -S eth1 | grep -v ': 0'
NIC statistics:
tx_q-0_pkts: 1002
tx_q-1_pkts: 2679
tx_q-2_pkts: 1113
tx_q-3_pkts: 1192 <----- tx_q-3 vs idpf-eth1-Tx-5
rx_q-0_pkts: 1143
rx_q-1_pkts: 3172
rx_q-2_pkts: 1074
After:
ethtool -L eth1 tx 1 combined 3
grep . /proc/irq/*/*idpf*/../smp_affinity_list
/proc/irq/67/idpf-Mailbox-0/../smp_affinity_list:0-55,112-167
/proc/irq/68/idpf-eth1-TxRx-0/../smp_affinity_list:0
/proc/irq/70/idpf-eth1-TxRx-1/../smp_affinity_list:1
/proc/irq/71/idpf-eth1-TxRx-2/../smp_affinity_list:2
/proc/irq/72/idpf-eth1-Tx-3/../smp_affinity_list:3
ethtool -S eth1 | grep -v ': 0'
NIC statistics:
tx_q-0_pkts: 118
tx_q-1_pkts: 134
tx_q-2_pkts: 228
tx_q-3_pkts: 138 <--- tx_q-3 matches idpf-eth1-Tx-3
rx_q-0_pkts: 111
rx_q-1_pkts: 366
rx_q-2_pkts: 120
Fixes: d4d5587182 ("idpf: initialize interrupts and enable vport")
Signed-off-by: Brian Vazquez <brianvv@google.com>
Reviewed-by: Brett Creeley <brett.creeley@amd.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Tested-by: Samuel Salin <Samuel.salin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
rss_data->rss_key needs to be nullified after it is freed.
Checks like "if (!rss_data->rss_key)" in the code could fail
if it is not nullified.
Tested: built and booted the kernel.
Fixes: 83f38f210b ("idpf: Fix RSS LUT NULL pointer crash on early ethtool operations")
Signed-off-by: Li Li <boolli@google.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Tested-by: Samuel Salin <Samuel.salin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Currently, in idpf_wait_for_sw_marker_completion(), when an
IDPF_TXD_COMPLT_SW_MARKER packet is found, the routine breaks out of
the for loop and does not increment the next_to_clean counter. This
causes the subsequent NAPI polls to run into the same
IDPF_TXD_COMPLT_SW_MARKER packet again and print out the following:
[ 23.261341] idpf 0000:05:00.0 eth1: Unknown TX completion type: 5
Instead, we should increment next_to_clean regardless when an
IDPF_TXD_COMPLT_SW_MARKER packet is found.
Tested: with the patch applied, we do not see the errors above from NAPI
polls anymore.
Fixes: 9d39447051 ("idpf: remove SW marker handling from NAPI")
Signed-off-by: Li Li <boolli@google.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
When stmmac_init_timestamping() is called, it clears the receive and
transmit path booleans that allow timestamps to be read. These are
never re-initialised until after userspace requests timestamping
features to be enabled.
However, our copy of the timestamp configuration is not cleared, which
means we return the old configuration to userspace when requested.
This is inconsistent. Fix this by clearing the timestamp configuration.
Fixes: d6228b7cdd ("net: stmmac: implement the SIOCGHWTSTAMP ioctl")
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/E1vuUu4-0000000Afea-0j9B@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In DQ-QPL mode, gve_tx_clean_pending_packets() incorrectly uses the RDA
buffer cleanup path. It iterates num_bufs times and attempts to unmap
entries in the dma array.
This leads to two issues:
1. The dma array shares storage with tx_qpl_buf_ids (union).
Interpreting buffer IDs as DMA addresses results in attempting to
unmap incorrect memory locations.
2. num_bufs in QPL mode (counting 2K chunks) can significantly exceed
the size of the dma array, causing out-of-bounds access warnings
(trace below is how we noticed this issue).
UBSAN: array-index-out-of-bounds in
drivers/net/ethernet/drivers/net/ethernet/google/gve/gve_tx_dqo.c:178:5 index 18 is out of
range for type 'dma_addr_t[18]' (aka 'unsigned long long[18]')
Workqueue: gve gve_service_task [gve]
Call Trace:
<TASK>
dump_stack_lvl+0x33/0xa0
__ubsan_handle_out_of_bounds+0xdc/0x110
gve_tx_stop_ring_dqo+0x182/0x200 [gve]
gve_close+0x1be/0x450 [gve]
gve_reset+0x99/0x120 [gve]
gve_service_task+0x61/0x100 [gve]
process_scheduled_works+0x1e9/0x380
Fix this by properly checking for QPL mode and delegating to
gve_free_tx_qpl_bufs() to reclaim the buffers.
Cc: stable@vger.kernel.org
Fixes: a6fb8d5a8b ("gve: Tx path for DQO-QPL")
Signed-off-by: Ankit Garg <nktgrg@google.com>
Reviewed-by: Jordan Rhee <jordanrhee@google.com>
Reviewed-by: Harshitha Ramamurthy <hramamurthy@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20260220215324.1631350-1-joshwash@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This converts some of the visually simpler cases that have been split
over multiple lines. I only did the ones that are easy to verify the
resulting diff by having just that final GFP_KERNEL argument on the next
line.
Somebody should probably do a proper coccinelle script for this, but for
me the trivial script actually resulted in an assertion failure in the
middle of the script. I probably had made it a bit _too_ trivial.
So after fighting that far a while I decided to just do some of the
syntactically simpler cases with variations of the previous 'sed'
scripts.
The more syntactically complex multi-line cases would mostly really want
whitespace cleanup anyway.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the exact same thing as the 'alloc_obj()' version, only much
smaller because there are a lot fewer users of the *alloc_flex()
interface.
As with alloc_obj() version, this was done entirely with mindless brute
force, using the same script, except using 'flex' in the pattern rather
than 'objs*'.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
Ntuple filters can be deleted when the interface
is down. The current code blindly sends the filter
delete command to FW. When the interface is down, all
the VNICs are deleted in the FW. When the VNIC is
freed in the FW, all the associated filters are also
freed. We need not send the free command explicitly.
Sending such command will generate FW error in the
dmesg.
In order to fix this, we can safely return from
bnxt_hwrm_cfa_ntuple_filter_free() when BNXT_STATE_OPEN
is not true which confirms the VNICs have been deleted.
Fixes: 8336a974f3 ("bnxt_en: Save user configured filters in a lookup list")
Suggested-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20260219185313.2682148-3-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We need to free the corresponding RSS context VNIC
in FW everytime an RSS context is deleted in driver.
Commit 667ac333db added a check to delete the VNIC
in FW only when netif_running() is true to help delete
RSS contexts with interface down.
Having that condition will make the driver leak VNICs
in FW whenever close() happens with active RSS contexts.
On the subsequent open(), as part of RSS context restoration,
we will end up trying to create extra VNICs for which we
did not make any reservation. FW can fail this request,
thereby making us lose active RSS contexts.
Suppose an RSS context is deleted already and we try to
process a delete request again, then the HWRM functions
will check for validity of the request and they simply
return if the resource is already freed. So, even for
delete-when-down cases, netif_running() check is not
necessary.
Remove the netif_running() condition check when deleting
an RSS context.
Reported-by: Jakub Kicinski <kicinski@meta.com>
Fixes: 667ac333db ("eth: bnxt: allow deleting RSS contexts when the device is down")
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20260219185313.2682148-2-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In ixp4xx_get_ts_info() ixp46x_ptp_find() is called
unconditionally despite this feature only existing on
ixp46x, leading to the following splat from tcpdump:
root@OpenWrt:~# tcpdump -vv -X -i eth0
(...)
Unable to handle kernel NULL pointer dereference at virtual address
00000238 when read
(...)
Call trace:
ptp_clock_index from ixp46x_ptp_find+0x1c/0x38
ixp46x_ptp_find from ixp4xx_get_ts_info+0x4c/0x64
ixp4xx_get_ts_info from __ethtool_get_ts_info+0x90/0x108
__ethtool_get_ts_info from __dev_ethtool+0xa00/0x2648
__dev_ethtool from dev_ethtool+0x160/0x234
dev_ethtool from dev_ioctl+0x2cc/0x460
dev_ioctl from sock_ioctl+0x1ec/0x524
sock_ioctl from sys_ioctl+0x51c/0xa94
sys_ioctl from ret_fast_syscall+0x0/0x44
(...)
Segmentation fault
Check for ixp46x in ixp46x_ptp_find() before trying to set up
PTP to avoid this.
To avoid altering the returned error code from ixp4xx_hwtstamp_set()
which before this patch was -EOPNOTSUPP, we return -EOPNOTSUPP
from ixp4xx_hwtstamp_set() if ixp46x_ptp_find() fails no matter
the error code. The helper function ixp46x_ptp_find() helper
returns -ENODEV.
Fixes: 9055a2f591 ("ixp4xx_eth: make ptp support a platform driver")
Signed-off-by: Linus Walleij <linusw@kernel.org>
Link: https://patch.msgid.link/20260219-ixp4xx-fix-ethernet-v3-1-f235ccc3cd46@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Jakub Kicinski:
"Including fixes from Netfilter.
Current release - new code bugs:
- net: fix backlog_unlock_irq_restore() vs CONFIG_PREEMPT_RT
- eth: mlx5e: XSK, Fix unintended ICOSQ change
- phy_port: correctly recompute the port's linkmodes
- vsock: prevent child netns mode switch from local to global
- couple of kconfig fixes for new symbols
Previous releases - regressions:
- nfc: nci: fix false-positive parameter validation for packet data
- net: do not delay zero-copy skbs in skb_attempt_defer_free()
Previous releases - always broken:
- mctp: ensure our nlmsg responses to user space are zero-initialised
- ipv6: ioam: fix heap buffer overflow in __ioam6_fill_trace_data()
- fixes for ICMP rate limiting
Misc:
- intel: fix PCI device ID conflict between i40e and ipw2200"
* tag 'net-7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (85 commits)
net: nfc: nci: Fix parameter validation for packet data
net/mlx5e: Use unsigned for mlx5e_get_max_num_channels
net/mlx5e: Fix deadlocks between devlink and netdev instance locks
net/mlx5e: MACsec, add ASO poll loop in macsec_aso_set_arm_event
net/mlx5: Fix misidentification of write combining CQE during poll loop
net/mlx5e: Fix misidentification of ASO CQE during poll loop
net/mlx5: Fix multiport device check over light SFs
bonding: alb: fix UAF in rlb_arp_recv during bond up/down
bnge: fix reserving resources from FW
eth: fbnic: Advertise supported XDP features.
rds: tcp: fix uninit-value in __inet_bind
net/rds: Fix NULL pointer dereference in rds_tcp_accept_one
octeontx2-af: Fix default entries mcam entry action
net/mlx5e: XSK, Fix unintended ICOSQ change
ipv6: icmp: icmpv6_xrlim_allow() optimization if net.ipv6.icmp.ratelimit is zero
ipv4: icmp: icmpv4_xrlim_allow() optimization if net.ipv4.icmp_ratelimit is zero
ipv6: icmp: remove obsolete code in icmpv6_xrlim_allow()
inet: move icmp_global_{credit,stamp} to a separate cache line
icmp: prevent possible overflow in icmp_global_allow()
selftests/net: packetdrill: add ipv4-mapped-ipv6 tests
...
The max number of channels is always an unsigned int, use the correct
type to fix compilation errors done with strict type checking, e.g.:
error: call to ‘__compiletime_assert_1110’ declared with attribute
error: min(mlx5e_get_devlink_param_num_doorbells(mdev),
mlx5e_get_max_num_channels(mdev)) signedness error
Fixes: 74a8dadac1 ("net/mlx5e: Preparations for supporting larger number of channels")
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <Jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20260218072904.1764634-7-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In the mentioned "Fixes" commit, various work tasks triggering devlink
health reporter recovery were switched to use netdev_trylock to protect
against concurrent tear down of the channels being recovered. But this
had the side effect of introducing potential deadlocks because of
incorrect lock ordering.
The correct lock order is described by the init flow:
probe_one -> mlx5_init_one (acquires devlink lock)
-> mlx5_init_one_devl_locked -> mlx5_register_device
-> mlx5_rescan_drivers_locked -...-> mlx5e_probe -> _mlx5e_probe
-> register_netdev (acquires rtnl lock)
-> register_netdevice (acquires netdev lock)
=> devlink lock -> rtnl lock -> netdev lock.
But in the current recovery flow, the order is wrong:
mlx5e_tx_err_cqe_work (acquires netdev lock)
-> mlx5e_reporter_tx_err_cqe -> mlx5e_health_report
-> devlink_health_report (acquires devlink lock => boom!)
-> devlink_health_reporter_recover
-> mlx5e_tx_reporter_recover -> mlx5e_tx_reporter_recover_from_ctx
-> mlx5e_tx_reporter_err_cqe_recover
The same pattern exists in:
mlx5e_reporter_rx_timeout
mlx5e_reporter_tx_ptpsq_unhealthy
mlx5e_reporter_tx_timeout
Fix these by moving the netdev_trylock calls from the work handlers
lower in the call stack, in the respective recovery functions, where
they are actually necessary.
Fixes: 8f7b00307b ("net/mlx5e: Convert mlx5 netdevs to instance locking")
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <Jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20260218072904.1764634-6-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The macsec_aso_set_arm_event function calls mlx5_aso_poll_cq once
without a retry loop. If the CQE is not immediately available after
posting the WQE, the function fails unnecessarily.
Use read_poll_timeout() to poll 3-10 usecs for CQE, consistent with
other ASO polling code paths in the driver.
Fixes: 739cfa3451 ("net/mlx5: Make ASO poll CQ usable in atomic context")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <Jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20260218072904.1764634-5-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The write combining completion poll loop uses usleep_range() which can
sleep much longer than requested due to scheduler latency. Under load,
we witnessed a 20ms+ delay until the process was rescheduled, causing
the jiffies based timeout to expire while the thread is sleeping.
The original do-while loop structure (poll, sleep, check timeout) would
exit without a final poll when waking after timeout, missing a CQE that
arrived during sleep.
Instead of the open-coded while loop, use the kernel's poll_timeout_us()
which always performs an additional check after the sleep expiration,
and is less error-prone.
Note: poll_timeout_us() doesn't accept a sleep range, by passing 10
sleep_us the sleep range effectively changes from 2-10 to 3-10 usecs.
Fixes: d98995b4bf ("net/mlx5: Reimplement write combining test")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <Jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20260218072904.1764634-4-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The ASO completion poll loop uses usleep_range() which can sleep much
longer than requested due to scheduler latency. Under load, we witnessed
a 20ms+ delay until the process was rescheduled, causing the jiffies
based timeout to expire while the thread is sleeping.
The original do-while loop structure (poll, sleep, check timeout) would
exit without a final poll when waking after timeout, missing a CQE that
arrived during sleep.
Instead of the open-coded while loop, use the kernel's
read_poll_timeout() which always performs an additional check after the
sleep expiration, and is less error-prone.
Note: read_poll_timeout() doesn't accept a sleep range, by passing 10
sleep_us the sleep range effectively changes from 2-10 to 3-10 usecs.
Fixes: 739cfa3451 ("net/mlx5: Make ASO poll CQ usable in atomic context")
Fixes: 7e3fce82d9 ("net/mlx5e: Overcome slow response for first macsec ASO WQE")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Jianbo Liu <jianbol@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Jacob Keller <Jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20260218072904.1764634-3-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
XSK wakeup must use the async ICOSQ (with proper locking), as it is not
guaranteed to run on the same CPU as the channel.
The commit that converted the NAPI trigger path to use the sync ICOSQ
incorrectly applied the same change to XSK, causing XSK wakeups to use
the sync ICOSQ as well. Revert XSK flows to use the async ICOSQ.
XDP program attach/detach triggers channel reopen, while XSK pool
enable/disable can happen on-the-fly via NDOs without reopening
channels. As a result, xsk_pool state cannot be reliably used at
mlx5e_open_channel() time to decide whether an async ICOSQ is needed.
Update the async_icosq_needed logic to depend on the presence of an XDP
program rather than the xsk_pool, ensuring the async ICOSQ is available
when XSK wakeups are enabled.
This fixes multiple issues:
1. Illegal synchronize_rcu() in an RCU read- side critical section via
mlx5e_xsk_wakeup() -> mlx5e_trigger_napi_icosq() ->
synchronize_net(). The stack holds RCU read-lock in xsk_poll().
2. Hitting a NULL pointer dereference in mlx5e_xsk_wakeup():
[] BUG: kernel NULL pointer dereference, address: 0000000000000240
[] #PF: supervisor read access in kernel mode
[] #PF: error_code(0x0000) - not-present page
[] PGD 0 P4D 0
[] Oops: Oops: 0000 [#1] SMP
[] CPU: 0 UID: 0 PID: 2255 Comm: qemu-system-x86 Not tainted 6.19.0-rc5+ #229 PREEMPT(none)
[] Hardware name: [...]
[] RIP: 0010:mlx5e_xsk_wakeup+0x53/0x90 [mlx5_core]
Reported-by: Daniel Borkmann <daniel@iogearbox.net>
Closes: https://lore.kernel.org/all/20260123223916.361295-1-daniel@iogearbox.net/
Fixes: 56aca3e0f7 ("net/mlx5e: Use regular ICOSQ for triggering NAPI")
Tested-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Acked-by: Alice Mikityanska <alice.kernel@fastmail.im>
Link: https://patch.msgid.link/20260217074525.1761454-1-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Increasing the MTU beyond the HDS threshold causes the hardware to
fragment packets across multiple buffers. If a single-buffer XDP program
is attached, the driver will drop all multi-frag frames. While we can't
prevent a remote sender from sending non-TCP packets larger than the MTU,
this will prevent users from inadvertently breaking new TCP streams.
Traditionally, drivers supported XDP with MTU less than 4Kb
(packet per page). Fbnic currently prevents attaching XDP when MTU is too high.
But it does not prevent increasing MTU after XDP is attached.
Fixes: 1b0a3950db ("eth: fbnic: Add XDP pass, drop, abort support")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Dimitri Daskalakis <dimitri.daskalakis1@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
fbnic always advertises ETHTOOL_TCP_DATA_SPLIT_ENABLED via ethtool
.get_ringparam. To enable proper splitting for all flow types, even for
IP/Ethernet flows, this patch sets DMA_HINT_L4 unconditionally for all
RSS and NFC flow steering rules. According to the spec, L4 falls back to
L3 if no valid L4 is found, and L3 falls back to L2 if no L3 is found.
This makes sure that the correct header boundary is used regardless of
traffic type. This is important for zero-copy use cases where we must
ensure that all ZC packets are split correctly.
Fixes: 2b30fc01a6 ("eth: fbnic: Add support for HDS configuration")
Signed-off-by: Bobby Eshleman <bobbyeshleman@meta.com>
Link: https://patch.msgid.link/20260211-fbnic-tcp-hds-fixes-v1-3-55d050e6f606@meta.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
