Louis Scalbert [Wed, 3 Jun 2026 15:03:30 +0000 (17:03 +0200)]
bonding: 3ad: fix mux port state on oper down
When the bonding interface has carrier down due to the absence of
usable slaves and a slave transitions from down to up, the bonding
interface briefly goes carrier up, then down again, and finally up
once LACP negotiates collecting and distributing on the port.
When lacp_strict mode is on, the interface should not transition to
carrier up until LACP negotiation is complete.
This happens because the actor and partner port states remain in
Collecting_Distributing when the port goes down. When the port
comes back up, it temporarily remains in this state until LACP
renegotiation occurs.
Previously this was mostly cosmetic, but since the bonding carrier
state may depend on the LACP negotiation state, it causes the
interface to flap.
According to IEEE 802.3ad-2000 and IEEE 802.1ax-2014, Collecting and
Distributing should be reset when a port goes down:
- In the Receive state machine, port_enabled == FALSE causes a
transition to the PORT_DISABLED state, which is expected to clear
Partner_Oper_Port_State.Synchronization.
- In the Mux state machine, Partner_Oper_Port_State.Synchronization ==
FALSE causes a transition to the ATTACHED state, which disables
Collecting and Distributing.
However, Partner_Oper_Port_State.Synchronization is not cleared in the
PORT_DISABLED state.
Clear Partner_Oper_Port_State.Synchronization in the Receive
PORT_DISABLED state.
Louis Scalbert [Wed, 3 Jun 2026 15:03:28 +0000 (17:03 +0200)]
bonding: 3ad: add lacp_strict configuration knob
When an 802.3ad (LACP) bonding interface has no slaves in the
collecting/distributing state, the bonding master still reports
carrier as up as long as at least 'min_links' slaves have carrier.
In this situation, only one slave is effectively used for TX/RX,
while traffic received on other slaves is dropped. Upper-layer
daemons therefore consider the interface operational, even though
traffic may be blackholed if the lack of LACP negotiation means
the partner is not ready to deal with traffic.
Introduce a configuration knob to control this behavior. It allows
the bonding master to assert carrier only when at least 'min_links'
slaves are in Collecting_Distributing state.
The default mode preserves the existing behavior. This patch only
introduces the knob; its behavior is implemented in the subsequent
commit.
====================
net: dsa: realtek: rtl8365mb: bridge offloading and VLAN support
This series introduces bridge offloading, FDB management, and VLAN support
for the Realtek rtl8365mb DSA switch driver. The primary goal is to
enable hardware frame forwarding between bridge ports, reducing CPU
overhead and providing advanced features like VLAN and FDB isolation.
Some of these patches are based on original work by Alvin Šipraga,
subsequently adapted and updated for the current net-next state.
I attempted to reach Alvin for review of the final version but was
unable to establish contact. Any regressions in this version are my
responsibility.
====================
net: dsa: realtek: rtl8365mb: add bridge port flags
Implement support for bridge port flags to control learning and flooding
behavior. This patch maps hardware functionalities to the following
bridge flags:
Implement hardware offloading of bridge functionality. This is achieved
by using the per-port isolation registers, which contain a forwarding
port mask. The switch will refuse to forward packets ingressed on a
given port to a port which is not in its forwarding mask.
For each bridge that is offloaded, use the DSA-provided bridge number
for the Extended Filtering ID (EFID). When using Independent VLAN
Learning (IVL), the forwarding database is keyed with the tuple
{VID, MAC, EFID}. There are 8 EFIDs available (0~7), but we reserve the
default EFID 0 for standalone ports where learning is disabled. This
fits nicely because DSA indexes the bridge number starting from 1.
Because of the limited number of EFIDs, we have to set the
max_num_bridges property of our switch to 7: we can't offload more than
that or we will fail to offer IVL as at least two bridges would end up
having to share an EFID.
All ports start isolated, forwarding exclusively to CPU ports, and
with VLAN transparent, ignoring VLAN membership. Once a member in a
bridge, the port isolation is expanded to include the bridge members.
When that bridge enables VLAN filtering, the VLAN transparent feature is
disabled, letting the switch filter based on VLAN setup.
Alvin Šipraga [Sat, 6 Jun 2026 08:29:31 +0000 (05:29 -0300)]
net: dsa: realtek: rtl8365mb: add FDB support
Implement support for FDB and MDB management for the RTL8365MB series
switches.
The hardware supports IVL by keying the unicast forwarding database with
the {MAC, VID, EFID} tuple. The Extended Filtering ID (EFID) is 3 bits
wide, providing 8 unique filtering domains. This driver reserves EFID 0
for standalone ports, effectively limiting the hardware offload to a
maximum of 7 bridges. The multicast database uses a {MAC, VID} key, with
ports from different bridges sharing the same multicast group.
Introduce a mutex lock (l2_lock) to protect concurrent L2 table updates.
Add support for forwarding database operations, including unicast and
multicast entry handling as well as fast aging support.
Set DSA switch flags assisted_learning_on_cpu_port and fdb_isolation.
Alvin Šipraga [Sat, 6 Jun 2026 08:29:30 +0000 (05:29 -0300)]
net: dsa: realtek: rtl8365mb: add VLAN support
Realtek RTL8365MB switches (a.k.a. RTL8367C family) use two different
structures for VLANs:
- VLAN4K: A full table with 4096 entries defining port membership and
tagging.
- VLANMC: A smaller table with 32 entries used primarily for PVID
assignment.
In this hardware, a port's PVID must point to an index in the VLANMC
table rather than a VID directly. Since the VLANMC table is limited to
32 entries, the driver implements a dynamic allocation scheme to
maximize resource usage:
- VLAN4K is treated by the driver as the source of truth for membership.
- A VLANMC entry is only allocated when a port is configured to use a
specific VID as its PVID.
- VLANMC entries are deleted when no longer needed as a PVID by any port.
Although VLANMC has a members field, the switch only checks membership
in the VLAN4K table. This driver will use VLANMC members field as way to
track which ports are using that entry as PVID.
VLANMC index 0, although a valid entry, is reserved in this driver as a
neutral PVID value for ports not using a specific PVID.
In the subsequent RTL8367D switch family, VLANMC table was
removed and PVID assignment was delegated to a dedicated set of
registers.
The use of FIELD_PREP for reconstructing LO/HI values was suggested by
Yury Norov.
Fix for vlan_setup and vlan_filtering was suggested by Abdulkader
Alrezej.
net: dsa: realtek: rtl8365mb: use dsa helpers for port iteration
Convert open-coded port iteration loops to use the DSA helpers and
restructure rtl8365mb_setup() into clear blocking, user, and
CPU port phases.
As part of this refactoring, unused ports are explicitly placed into a
blocked, isolated state with learning disabled, ensuring safe default
hardware behavior. The driver also does not allocate a virtual IRQ
mapping for unused ports. To accommodate this, a guard check is added to
the interrupt handler (rtl8365mb_irq) to safely skip ports without a
valid IRQ mapping. The irq domain teardown, however, does clean all
ports as external PHYs may still map the IRQ.
Furthermore, since the new initialization loop starts with all ports
administratively isolated by default, CPU port forwarding and isolation
masks are explicitly configured at the end of the setup phase to prevent
egress traffic from being blocked.
Explicitly enforce the presence of a CPU port (-EINVAL) and reject DSA
cascade links (-EOPNOTSUPP) during setup to prevent silent failures.
These topologies were already non-functional. Without a CPU port, the
driver does not activate CPU tagging. Additionally, the switch hardware
was not designed to be cascaded, and DSA links never worked because
CPU tagging is not enabled for them.
Convert numeric error codes into human-readable strings by using %pe
together with ERR_PTR() in dev_err() messages. Also use dev_err_probe()
instead of checking for -EPROBE_DEFER.
Guangshuo Li [Sun, 7 Jun 2026 14:57:47 +0000 (22:57 +0800)]
net: lan966x: restore RX state on reload failure
lan966x_fdma_reload() backs up rx->page_pool and rx->fdma before
reallocating the RX resources for the new MTU. If the allocation fails,
the restore path puts these fields back before restarting RX.
However, the reload path also updates rx->page_order and rx->max_mtu
before calling lan966x_fdma_rx_alloc(). These fields are not restored on
failure, so RX can be restarted with the old pages, old FDMA state and
old page pool, but with the page geometry from the failed new MTU.
This can make the XDP path advertise a frame size derived from the new
page_order while the actual RX pages still come from the old allocation.
For example, after a failed reload to a jumbo MTU, xdp_init_buff() may be
called with a frame size larger than the restored RX pages.
lan966x_fdma_rx_alloc_page_pool() also registers the newly allocated page
pool with each port's XDP RXQ before fdma_alloc_coherent() is called. If
fdma_alloc_coherent() fails, the new page pool is destroyed, but the
rollback path does not restore the per-port XDP RXQ mem model
registration either.
Save and restore rx->page_order and rx->max_mtu, and restore the old page
pool registration for each port's XDP RXQ before RX is started again.
This keeps the restored RX state consistent after a failed reload.
====================
net: add retry mechanism to ndo_set_rx_mode_async
Original async ndo_set_rx_mode work left one place where we do netdev_WARN
in response to a ENOMEM. The intent was to see whether actual real
users can hit that (adding uc/mc under memory pressure seems like a
very unlikely thing to do). However, it was quickly triggered by
syzbot's failslab. Add a retry mechanism and downgrade netdev_WARN
to netdev_err. The retry logic is a typical exponential backoff:
1, 2, 4, 8 seconds, 15 in total, hopefully enough for a system to resolve
memory pressure.
====================
Remove the driver-specific BNXT_STATE_L2_FILTER_RETRY + timer + sp_task
retry mechanism and rely on the core stack's ndo_set_rx_mode_async retry
instead.
bnxt_cfg_rx_mode() now returns errors instead of swallowing them. The
PF-unavailable case (-ENODEV from HWRM on a VF) is normalized to
-EAGAIN at the boundary so callers can match on a single "retry me"
errno without re-implementing the VF/-ENODEV check. Other errors
propagate unchanged.
This removes:
- BNXT_STATE_L2_FILTER_RETRY state bit
- BNXT_RX_MASK_SP_EVENT sp_event bit
- Retry trigger from bnxt_timer()
- BNXT_RX_MASK_SP_EVENT handling from bnxt_sp_task()
bnxt_init_chip() still calls bnxt_cfg_rx_mode() directly during open.
On a fresh open dev->uc is empty and the call effectively cannot fail
on the unicast path. But on FW reset reopen (bnxt_fw_reset_task ->
bnxt_open) a VF may have a populated dev->uc and the PF may be
transiently unavailable; since that path doesn't go through
__dev_open(), the follow-up rx_mode call that would otherwise drive
the core retry doesn't fire. On -EAGAIN, swallow the error and call
netif_rx_mode_schedule_retry() explicitly. The unicast filter loop
truncates vnic->uc_filter_count on failure, so the retry's delta check
sees pending work and reinstalls.
Cc: Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Michael Chan <michael.chan@broadcom.com> Link: https://patch.msgid.link/20260608154014.227538-4-sdf@fomichev.me Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When ndo_set_rx_mode_async returns an error, schedule a retry with
exponential backoff (1s, 2s, 4s, 8s -- 15s total). Give up after the
4th retry and log an error via netdev_err().
This moves retry logic from individual drivers into the core stack.
Timer callback does not hold a ref on dev. Safe because the timer can
only be armed when dev is IFF_UP, and __dev_close_many runs
timer_delete_sync before clearing IFF_UP. Unregister always closes
IFF_UP devices first, so by the time dev can be freed the timer is
dead and cannot be re-armed.
net: change ndo_set_rx_mode_async return type to int
Change the return type of ndo_set_rx_mode_async from void to int to
allow drivers to report failures back to the core stack. This is a
prerequisite for adding retry logic in the core when drivers fail to
program RX filters (e.g. bnxt VF when PF is unavailable).
All existing implementations return 0 for now, maintaining current
behavior.
net: mana: Cache MANA_QUERY_LINK_CONFIG result to avoid repeated HWC queries
mana_query_link_cfg() sends an HWC command to firmware on every call,
but the link speed and QoS values it returns only change when the
driver explicitly calls mana_set_bw_clamp(). This function is called
not only by userspace via ethtool get_link_ksettings, but also
periodically by hv_netvsc through netvsc_get_link_ksettings and by
the sysfs speed_show attribute via dev_attr_show, resulting in
unnecessary HWC traffic every few minutes.
Add a link_cfg_error field to mana_port_context to cache the query
result. The field uses three states: 1 (not yet queried, initial
value set during mana_probe_port), 0 (success, speed/max_speed are
valid), or a negative errno for permanent errors like -EOPNOTSUPP
when the hardware does not support the command. Transient errors and
qos_unconfigured responses are not cached so that subsequent calls
will retry.
MANA is ops-locked because it implements net_shaper_ops, so the core
already takes netdev_lock() around all ethtool_ops and net_shaper_ops
entry points. Reuse that lock to serialize mana_query_link_cfg() and
mana_set_bw_clamp(). This prevents a concurrent mana_set_bw_clamp()
from racing with an in-flight query and publishing stale pre-clamp
speed/max_speed.
Invalidate the cache inside mana_set_bw_clamp() on success, so all
current and future callers that change the link configuration
automatically trigger a fresh query on the next mana_query_link_cfg()
call. Also reset link_cfg_error during resume in mana_probe() under
netdev_lock(), so that any query already in flight cannot later
store 0 and silently overwrite the post-resume invalidation.
David Yang [Sat, 6 Jun 2026 12:52:44 +0000 (20:52 +0800)]
net: mscc: ocelot: validate netdev belongs to switch in .netdev_to_port()
The .netdev_to_port() currently takes only a net_device and returns the
port index, without verifying the netdev actually belongs to the switch
being operated on. This can cause flower rule parsing to silently
resolve to a wrong port on the local hardware.
Update both implementations felix_netdev_to_port() and
ocelot_netdev_to_port() to validate ownership. Also update the callers
in ocelot_flower.c to pass through the ocelot context.
Haiyang Zhang [Fri, 5 Jun 2026 21:22:56 +0000 (14:22 -0700)]
net: mana: Add support for PF device 0x00C1
Update the device id table to include the new device id 0x00C1.
This device's BAR layout is similar to VF's, update the function,
mana_gd_init_registers(), accordingly.
net: dsa: qca8k: Add support for force mode for fixed link topology
A fixed link topology is commonly used to connect this switch (on port
0 or 6) to a SoC's MAC over SGMII. When inband negotiation is not used,
the switch needs to be configured to operate in force mode. As such,
enable support for force mode.
====================
Add motorcomm 8531s set ds func and 8522 driver
This patch is for Starfive JHB100 EVB board. JHB100 contain
1 RGMII/RMII and 1 RMII synopsys GMAC cores. In the EVB board, RGMII
interface connect with YT8531s Ethernet PHY. RMII interface connect
with YT8522 ethernet PHY. So patch 1-2 is for RGMII interface
patch 3 is RMII is for RMII interface.
JHB100 is a Starfive new RISC-V SoC for datacenter BMC (BaseBoard
Managent Controller). Similar with Aspeed 27x0.
The JHB100 minimal system upstream is in progress:
https://patchwork.kernel.org/project/linux-riscv/cover/20260508053632.818548-1-changhuang.liang@starfivetech.com/
====================
Minda Chen [Fri, 5 Jun 2026 06:02:11 +0000 (14:02 +0800)]
net: motorcomm: phy: set drive strength in YT8531s RGMII
Set RXD and RX CLK pin drive strength while in YT8531s connect
with RGMII. Need to check 8531s PHY ID because 8521 and 8531s
pin drive strength is different, 8521 can not call
yt8531_set_ds().
Minda Chen [Fri, 5 Jun 2026 06:02:10 +0000 (14:02 +0800)]
net: phy: motorcomm: move mdio lock out from yt8531_set_ds()
yt8531_set_ds() default set register with mdio lock and only called
with YT8531 PHY. But new type YT8531s support RGMII and has the same
pin strength setting with YT8531, YT8531s need to call yt8531_set_ds()
setting pin drive strength. But YT8531s config init function
yt8521_config_init() already get the mdio lock with phy_select_page().
If calling yt8521_config_init() with mdio lock will cause dead lock.
Need to get the lock before calling yt8531_set_ds() and move mdio
lock out from it for YT8531s.
Jakub Kicinski [Wed, 10 Jun 2026 00:23:42 +0000 (17:23 -0700)]
Merge branch 'mana-per-vport-eq'
Long Li says:
====================
net: mana: Per-vPort EQ and MSI-X management
This series moves EQ ownership from the shared mana_context to per-vPort
mana_port_context, enabling each vPort to have dedicated MSI-X vectors
when the hardware provides enough vectors. When vectors are limited, the
driver falls back to sharing MSI-X among vPorts.
The series introduces a GDMA IRQ Context (GIC) abstraction with reference
counting to manage interrupt context lifecycle. This allows both Ethernet
and RDMA EQs to dynamically acquire dedicated or shared MSI-X vectors at
vPort creation time rather than pre-allocating all vectors at probe time.
====================
Long Li [Fri, 5 Jun 2026 00:57:15 +0000 (17:57 -0700)]
RDMA/mana_ib: Allocate interrupt contexts on EQs
Use the GIC functions to allocate interrupt contexts for RDMA EQs. These
interrupt contexts may be shared with Ethernet EQs when MSI-X vectors
are limited.
The driver now supports allocating dedicated MSI-X for each EQ. Indicate
this capability through driver capability bits. The RDMA EQs pass
use_msi_bitmap=false to share MSI-X vectors with Ethernet, while the
capability flag advertises that the driver supports per-vPort EQ
separation when hardware has sufficient vectors.
Populate eq.irq on all RDMA EQs for consistency with the Ethernet path.
Also relocate the GDMA_DRV_CAP_FLAG_1_HW_VPORT_LINK_AWARE define to its
numeric BIT(6) position among the other capability flags.
Long Li [Fri, 5 Jun 2026 00:57:14 +0000 (17:57 -0700)]
net: mana: Allocate interrupt context for each EQ when creating vPort
Use GIC functions to create a dedicated interrupt context or acquire a
shared interrupt context for each EQ when setting up a vPort.
The caller now owns the GIC reference across the EQ create/destroy
lifecycle: mana_create_eq() calls mana_gd_get_gic() before creating
each EQ and mana_destroy_eq() calls mana_gd_put_gic() after destroying
it. The msix_index invalidation is moved from mana_gd_deregister_irq()
to the mana_gd_create_eq() error path so that mana_destroy_eq() can
read the index before teardown.
Long Li [Fri, 5 Jun 2026 00:57:13 +0000 (17:57 -0700)]
net: mana: Use GIC functions to allocate global EQs
Replace the GDMA global interrupt setup code with the new GIC allocation
and release functions for managing interrupt contexts.
This changes the per-queue interrupt names in /proc/interrupts from
mana_q0, mana_q1, ... to mana_msi1, mana_msi2, ... to reflect the
MSI-X index rather than a zero-based queue number. The HWC interrupt
name (mana_hwc) is unchanged.
Long Li [Fri, 5 Jun 2026 00:57:12 +0000 (17:57 -0700)]
net: mana: Introduce GIC context with refcounting for interrupt management
To allow Ethernet EQs to use dedicated or shared MSI-X vectors and RDMA
EQs to share the same MSI-X, introduce a GIC (GDMA IRQ Context) with
reference counting. This allows the driver to create an interrupt context
on an assigned or unassigned MSI-X vector and share it across multiple
EQ consumers.
Long Li [Fri, 5 Jun 2026 00:57:11 +0000 (17:57 -0700)]
net: mana: Query device capabilities and configure MSI-X sharing for EQs
When querying the device, adjust the max number of queues to allow
dedicated MSI-X vectors for each vPort. The per-vPort queue count is
clamped towards MANA_DEF_NUM_QUEUES but will not exceed the hardware
maximum reported by the device.
MSI-X sharing among vPorts is enabled when there are not enough MSI-X
vectors for dedicated allocation, or when the platform does not support
dynamic MSI-X allocation (in which case all vectors are pre-allocated
at probe time and sharing is always used). The msi_sharing flag is
reset at the top of mana_gd_query_max_resources() so it is recomputed
from current hardware state on each probe or resume cycle.
Clamp apc->max_queues to gc->max_num_queues_vport in mana_init_port()
so that on resume, if max_num_queues_vport has decreased due to fewer
MSI-X vectors, num_queues is reduced accordingly before EQ allocation.
A device reporting zero ports now results in a fatal probe error since
the per-vPort MSI-X math requires at least one port.
Rename mana_query_device_cfg() to mana_gd_query_device_cfg() as it is
used at GDMA device probe time for querying device capabilities.
Long Li [Fri, 5 Jun 2026 00:57:10 +0000 (17:57 -0700)]
net: mana: Create separate EQs for each vPort
To prepare for assigning vPorts to dedicated MSI-X vectors, remove EQ
sharing among the vPorts and create dedicated EQs for each vPort.
Move the EQ definition from struct mana_context to struct mana_port_context
and update related support functions. Export mana_create_eq() and
mana_destroy_eq() for use by the MANA RDMA driver.
RSS QPs now take a vport reference via pd->vport_use_count to ensure
EQs outlive all QP consumers. The vport must already be configured by
a raw QP before an RSS QP can be created. EQs are only destroyed when
the last QP (raw or RSS) on the PD releases its reference.
Restrict each vport to a single RSS QP. The hardware only supports one
steering configuration (indirection table / hash key) per vport, and
mana_disable_vport_rx() on QP destroy disables RX globally for the
vport. Previously, creating a second RSS QP would silently overwrite
the first QP's steering config and destroy would blackhole all traffic.
This is now explicitly rejected with -EBUSY. Existing applications
(DPDK being the primary RDMA consumer) always create one RSS QP per
vport, so no real-world flows are affected.
Reject cross-port PD sharing for both raw and RSS QPs. Since EQs and
vport configuration are per-port, a PD is bound to the port used by
its first raw QP. Subsequent QPs on the same PD must use the same
port or the creation fails with -EINVAL. Previously this was silently
broken: with shared EQs it appeared to work, but with per-vPort EQs
a cross-port PD would cause wrong-port EQ teardown and corruption.
DPDK creates one PD per port so no existing flows are affected.
Serialize mana_set_channels() and the async per-port queue reset
handler against RDMA vport configuration to prevent RDMA from claiming
the vport during the detach/attach window. A channel_changing flag is
set under apc->vport_mutex before detach and checked by
mana_cfg_vport() when called from the RDMA path, blocking RDMA from
grabbing the vport during the entire window. When the port is down
and RDMA already holds the vport, the channel change is rejected with
-EBUSY.
====================
Consolidate FCrypt and PCBC code into net/rxrpc/
The FCrypt "block cipher" and the PCBC mode of operation are obsolete
and insecure. Since their only user is net/rxrpc/, they belong there,
not in the crypto API.
Therefore, this series removes these algorithms from the crypto API and
replaces them with local implementations in net/rxrpc/.
The local implementations are simpler too, as they avoid the crypto API
boilerplate.
I don't know how to test all the code in net/rxrpc/, but everything
should still work. I added a KUnit test for the crypto functions.
====================
Eric Biggers [Fri, 22 May 2026 05:07:36 +0000 (00:07 -0500)]
crypto: pcbc - Remove support for PCBC mode
The only user of PCBC mode (Propagating Cipher Block Chaining mode) was
net/rxrpc/rxkad.c, which now uses local code instead.
While PCBC was an interesting cryptographic experiment, it has largely
been relegated to the history books and academic exercises. It is
non-parallelizable (i.e., very slow) and doesn't actually achieve the
integrity properties it was apparently intended to achieve.
Remove support for it from the crypto API.
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> # m68k Acked-by: David Howells <dhowells@redhat.com> Signed-off-by: Eric Biggers <ebiggers@kernel.org> Tested-by: Marc Dionne <marc.dionne@auristor.com> Link: https://patch.msgid.link/20260522050740.84561-6-ebiggers@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Biggers [Fri, 22 May 2026 05:07:35 +0000 (00:07 -0500)]
crypto: fcrypt - Remove support for FCrypt block cipher
Remove the insecure FCrypt block cipher from the crypto API. Its only
user was net/rxrpc/, but now net/rxrpc/ implements it locally. The
crypto API implementation is no longer needed.
For some additional context: FCrypt was designed in 1988 and is
essentially a weakened version of DES. It has the same 56-bit key size
as DES, which is easily brute forced. Moreover, it's cryptographically
weak and doesn't even provide the intended 56-bit security level. Its
author considers it to be a mistake, as well
(https://lists.openafs.org/pipermail/openafs-devel/2000-December/005320.html).
But fortunately this 1980s-era homebrew block cipher was never adopted
outside of net/rxrpc/. So its code can just be kept there.
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> # m68k Acked-by: David Howells <dhowells@redhat.com> Signed-off-by: Eric Biggers <ebiggers@kernel.org> Tested-by: Marc Dionne <marc.dionne@auristor.com> Link: https://patch.msgid.link/20260522050740.84561-5-ebiggers@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Biggers [Fri, 22 May 2026 05:07:34 +0000 (00:07 -0500)]
net/rxrpc: Reimplement DES-PCBC using DES library
Since the use of "pcbc(des)" in rxkad_decrypt_ticket() is the only
remaining user of the crypto API "pcbc" template, just implement
DES-PCBC by locally implementing PCBC mode on top of the DES library.
Note that only the decryption direction is needed.
This will allow support for the obsolete PCBC mode to be removed from
the crypto API.
Eric Biggers [Fri, 22 May 2026 05:07:33 +0000 (00:07 -0500)]
net/rxrpc: Use local FCrypt-PCBC implementation
Use the local implementation of FCrypt-PCBC instead of the crypto API
one. This will allow the crypto API one to be removed. It also
simplifies the code quite a bit.
The local FCrypt-PCBC implementation is also significantly faster than
the crypto API one, since the crypto API one had a lot of overhead. For
example, benchmarking on an x86_64 CPU, I see that FCrypt-PCBC
decryption throughput improved from 83 MB/s to 157 MB/s.
(Meanwhile, AES-256-GCM decryption is 8064 MB/s on the same CPU.
Clearly, anyone looking for good performance, or anything that is
actually secure for that matter, needs to look elsewhere anyway.)
Eric Biggers [Fri, 22 May 2026 05:07:32 +0000 (00:07 -0500)]
net/rxrpc: Add local FCrypt-PCBC implementation
Add a local implementation of FCrypt-PCBC encryption and decryption.
This will be used instead of the crypto API one, allowing the crypto API
one to be removed. It will also simplify rxkad.c quite a bit.
A KUnit test is included. The FCrypt-PCBC test vectors are borrowed
from the existing ones in crypto/testmgr.h. Note that this adds the
first KUnit test for net/rxrpc/, which previously had no KUnit tests.
The FCrypt code is based on crypto/fcrypt.c, but I simplified it a bit.
The PCBC part is straightforward and I just wrote it from scratch.
====================
net: ethtool: let ops locked drivers run without rtnl_lock
With the ethtool_get_link_ksettings() situation hopefully ironed out
the previous series (commit 6a5d837f0ce2) let's return to the main
part of the series.
We have been slowly moving towards removing the rtnl_lock dependency
in driver ops since the concept of "ops-locked" drivers have been
introduced last year. Since last year will take the netdev instance
lock before invoking any ndo or ethtool op of "ops-locked" drivers.
We dipped our toes into rtnl_lock-less ops with the queue binding API.
Queue stats, NAPI, and other netdev-netlink objects are also queried
without holding rtnl_lock already. It's time to take the next logical
step and lift the requirement from ethtool ops.
The direct motivation for this patchset is that ethtool ops often
involve communicating with device FW, and may take a long time
to complete. Aggressive polling of device state on machines
with 10+ NICs have been shown to significantly increase rtnl_lock
pressure.
There's a handful of areas which still need rtnl_lock (see below).
I decided to convert everything to rtnl_lock-less by default, and
add a set of flags which let the drivers request rtnl_lock to still
be taken. I don't love this, but I'm worried that opt-in would be
even more confusing.
Known issues / exclusions:
- qdiscs - qdisc configuration currently assumes rtnl_lock, this
is mostly impacting set_channels callback. qdisc config is probably
the easiest one of the exclusions to tackle, it's fairly self-contained.
- features - even tho feature changes are (correctly) plumbed to
the driver thru ndos they are part of ethtool uAPI. ethtool itself
calls netdev_features_change() if it has spotted device feature change
before vs after to the callback. Some drivers also call
netdev_features_change() directly in response to various changes,
e.g. setting priv flags.
Since features have to propagate to upper and lower devices anything
that touches features is quite hard to move from under rtnl_lock.
- phylink - phylink and SFP depend on rtnl_lock today, I suspect
that this is purely for historic reasons. I started poking at
it and don't really see a need for a global lock. But accessing
the netdev instance lock from the SFP entry points will require
some attention from the phylink folks.
- phydev - similar to phylink, looks quite doable. But no ops-locked
driver currently has a phydev (fbnic only uses phylink) so phydev
related paths retain a ASSERT_RTNL() for now.
Tested on mlx5, bnxt and fbnic.
====================
Jakub Kicinski [Fri, 5 Jun 2026 00:29:11 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock on IOCTL path
Convert the IOCTL path similarly to how we converted Netlink.
The device lookup gets a little hairy. We could take rtnl_lock
unconditionally and drop it before calling the driver (this would
avoid the reference + liveness check). But I think being able
to make progress even if rtnl is dead-locked is quite useful.
First extra concern is handling features. List all the cmds which
modify features and always take rtnl_lock. We could fold this list
into ethtool_ioctl_needs_rtnl() but seems cleaner to keep
ethtool_ioctl_needs_rtnl() driver-related. If a driver changed
features and we were not holding rtnl_lock - warn about it.
It can only happen on buggy ops locked drivers (buggy because
they should have set appropriate "I need rtnl for op X" bit).
Second wrinkle is the PHY ID hack which drops the locks while
sleeping. Convert its static "busy" variable which used to
be protected by rtnl_lock to a field in struct ethtool_netdev_state.
This feature is about identifying an adapter or a port within
a system, so being able to blink multiple LEDs at the same
time is likely not very useful in practice. But it's the simplest
fix, we can add a mutex if someone thinks a system should only
be ID'ing one port at a time.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-12-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:10 +0000 (17:29 -0700)]
net: ethtool: ioctl: concentrate the locking
Add another layer of helper functions to make upcoming locking
changes easier. Otherwise we'd need a pretty complex goto
structure. netdev instance lock is now taken slightly sooner
but that should not be an issue since rtnl_lock is already held,
anyway.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-11-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:09 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock in RSS context handlers
Skip rtnl_lock in RSS context handlers if device is ops-locked.
Fairly trivial conversion. bnxt needed rtnl_lock for changing
the main context but looks like additional contexts are fine
without it.
Note (for review bots?) that ethnl_ops_begin() checks whether
the device is still registered.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-10-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:08 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock in ethnl_act_module_fw_flash()
Module firmware flashing reads SFF-8024 identifier bytes via
.get_module_eeprom_by_page(). Other than that it modifies
a bit in the netdev->ethtool struct. Both should be ops-locked
at this point.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-9-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:07 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock in ethnl_tsinfo_dumpit()
ethnl_tsinfo_dumpit() iterates netdevs and per-netdev PHY topology
calling ops->get_ts_info(). Switch to the "ops compat locking"
helpers which take either rtnl_lock or instance lock, depending
on what the device needs.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-8-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:06 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock in cable test handlers
Skip rtnl_lock in cable test handlers. This is really a noop since
no ops locked device supports these.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-7-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:05 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock on Netlink path for SET ops
Make ethtool not take rtnl_lock for SET commands when operation
is performed on an ops-locked driver. cfg/cfg_pending are now
ops-locked, since only ethtool modifies them.
Some SET driver callbacks will still need rtnl_lock, most notably
those which may end up calling netdev_update_features() or the qdisc
layer (via netif_set_real_num_tx_queues()). Let drivers selectively
opt back into the rtnl_lock with a new bitfield in ops.
We need two helpers since Netlink and ioctl cmds have different
values. Keep the helpers side by side in common.h to make sure
they get updated together, even tho they will only get called
from ioctl.c and netlink.c.
SET commands which don't use ethnl_default_set_doit() are converted
by subsequent commits.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-6-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:04 +0000 (17:29 -0700)]
net: ethtool: optionally skip rtnl_lock on Netlink path for GET ops
ethnl_default_doit() and ethnl_default_dump_one() are both used
exclusively for GET callbacks (former to get info for a single
device or get global strings). ops-locked devices don't need
rtnl_lock for GET callbacks, stop taking it.
Introduce an opt-out mechanism for devices which use phylink (fbnic)
since phylink currently depends on rtnl_lock protection. Subsequent
patches will add more exceptions, anyway. Practically the new helpers
for judging if command needs rtnl_lock could also call
netdev_need_ops_lock() but I find that it makes the code in the callers
slightly less obvious.
Add a helper for IOCTLs already, even tho it's unused so that
we can keep them in sync as the series progresses.
This is the first user-visible step of moving ethtool ops out
from under rtnl. Subsequent patches do the same for SET ops,
as well as the ioctl path.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-5-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski [Fri, 5 Jun 2026 00:29:03 +0000 (17:29 -0700)]
net: ethtool: make dev->hwprov ops-protected
dev->hwprov tracks the active hwtstamp provider for the device.
Make it ops protected (instance lock if the netdev driver opts
into holding instance lock around callbacks, otherwise rtnl_lock).
hwprov is written and read in:
- drivers/net/phy/phy_device.c
phydev and ops protection don't currently mix, add a comment
- net/ethtool/
as of now holds both rtnl lock and ops lock, this one will
soon only hold one lock or the other
read in:
- net/core/dev_ioctl.c
holds both rtnl lock and ops lock
- net/core/timestamping.c
RCU reader
The new netdev_ops_lock_dereference() helper does not have
"compat" in the name. The name would be quite long and I think
in this case it should be obvious that we need _a_ lock.
netdev_lock_dereference() already exists and means dev->lock
is always expected.
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-4-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
phydev <> netdev linking and lifecycle depends on rtnl_lock.
We want to switch to instance locks for most ethtool ops.
Let's add an assert that ops locked devices don't use phydev
today. If one does we can either opt the phy ops out of
being purely ops locked, or do deeper surgery to make phy
locking ops-compatible. I don't think there's any fundamental
challenge to make that work.
Reviewed-by: Nicolai Buchwitz <nb@tipi-net.de> Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-3-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
ethnl_bcast_seq is a global counter stamped into the nlmsg_seq field
of every multicast notification, allowing userspace to detect dropped
messages. Today the ordering is achieved by using rtnl_lock().
Moving forward we will want ethtool ops to run under just the netdev
instance lock so to establish ordering we need a separate lock
for notifications. With the netdev instance locks operations on
different devices may bypass each other but the expectation is
that it should not matter. What we need to prevent is:
- notification IDs getting out of order
- operations on one device getting out of order
For simplicity defer allocating the ID of the notification right
before the notification is delivered. This removes the need for
special handling in ethnl_rss_create_send_ntf().
Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20260605002912.3456868-2-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Minxi Hou [Thu, 4 Jun 2026 16:30:16 +0000 (00:30 +0800)]
selftests/net/openvswitch: guard command substitutions against empty output
When ip-link output is unavailable, when the upcall daemon log has not
been written yet, or when pahole does not know the OVS drop subsystem
ID, the affected command substitutions silently produce empty strings.
The caller then passes empty sha= or pid= arguments to ovs_add_flow,
or matches against wrong drop reason codes, all without a diagnostic.
Add [ -z ] guards immediately after each assignment. For test_arp_ping,
also align the MAC extraction to use awk '/link\/ether/' as in
test_pop_vlan. The drop_reason guard returns ksft_skip because an
absent subsystem ID is an environment issue, not a test failure.
====================
ipv4: igmp: annotate diagnostic procfs data races
This patch series addresses several unannotated data races between lockless
RCU-protected diagnostic reads in /proc/net/igmp (igmp_mc_seq_show())
and concurrent writes in serialized paths (RTNL and group spinlocks).
Following the precedent in commit 061c0aa740d5 ("ipv4: igmp: annotate
data-races around im->users"), we annotate these intentional data races
using READ_ONCE() and WRITE_ONCE() macros.
- Patch 1 annotates races around `in_dev->mc_count` (interface-level joins).
- Patch 2 annotates races around active timer-related state tracking fields
(`tm_running`, `reporter`, `expires`) on individual multicast groups.
====================
Yuyang Huang [Fri, 5 Jun 2026 01:43:18 +0000 (10:43 +0900)]
ipv4: igmp: annotate data-races around timer-related fields
/proc/net/igmp walks the multicast list locklessly under RCU and reads
timer-related fields (im->tm_running, im->reporter, im->timer.expires)
to print the timer state of multicast memberships. Concurrently, these
fields are modified under im->lock spinlock in timer management paths
(igmp_stop_timer(), igmp_start_timer(), and igmp_timer_expire()). Fix this
intentional lockless snapshot by annotating the lockless reads with
READ_ONCE() and the updates with WRITE_ONCE().
Yuyang Huang [Fri, 5 Jun 2026 01:43:17 +0000 (10:43 +0900)]
ipv4: igmp: annotate data-races around in_dev->mc_count
/proc/net/igmp walks the multicast list for IPv4 interfaces locklessly
under RCU and prints state->in_dev->mc_count. Concurrently, device
init/destruction and multicast join/leave paths update the count
under the RTNL lock. Fix this intentional lockless snapshot by
annotating the read with READ_ONCE() and the updates with WRITE_ONCE().
These are some of the issues that LLM reported to netconsole, and they
are being addressed here before big refactors.
I was doing some big refactors, and got some "pre-existent-issues"
during LLM review of the refactor, that make them hard to guarantee that
refactor is not introducing any bug, so, let's clean these pre-existent
bugs first, and then submit the refactor.
The issues fixed in this patchset were reported during the review of
https://lore.kernel.org/all/20260524-netconsole_move_more-v1-0-909d1ab398b4@debian.org/
Not all of them got fixed, but, those that were easy to reason about.
Why net-next and not 'net' tree.
Most of the functions that are being fixed here moved from netpoll to
netconsole, thus, fixing this on net will cause merge conflicts from
'net' to 'net-next', thus I decided to fix it on 'net-next', given we
are on 7.1-rc6 already. Sorry if that is not the right approach.
====================
Breno Leitao [Thu, 4 Jun 2026 16:10:14 +0000 (09:10 -0700)]
netconsole: close netdevice unregister window during target resume
process_resume_target() removes the target from target_list before
calling resume_target() so that netpoll_setup() can run with interrupts
enabled, then re-adds it once setup completes. netpoll_setup() acquires a
net_device reference (netdev_hold()) and releases the RTNL before
returning.
While the target is off target_list and the RTNL is not held,
netconsole_netdev_event() cannot find it. If the egress device is
unregistered in that window, the NETDEV_UNREGISTER notifier walks
target_list, misses the resuming target, and never tears it down. The
target is then re-added in STATE_ENABLED still holding a reference to the
now-unregistered device, leaking it and hanging unregister_netdevice() in
netdev_wait_allrefs().
Re-check under RTNL before re-publishing the target: if the device left
NETREG_REGISTERED while we were off the list, run do_netpoll_cleanup() and
mark the target disabled. Taking the RTNL across the check and the
list_add() serialises against the NETDEV_UNREGISTER notifier, which also
runs under RTNL, so the device is either still registered (and the
notifier will find the re-added target later) or already unregistering
(and we drop the reference here). netdev_wait_allrefs() runs from
netdev_run_todo() outside the RTNL, so dropping the reference here cannot
deadlock against the pending unregister.
Breno Leitao [Thu, 4 Jun 2026 16:10:13 +0000 (09:10 -0700)]
netconsole: clean up deactivated targets dropped before the cleanup worker
drop_netconsole_target() downgrades a STATE_DEACTIVATED target to
STATE_DISABLED and then only calls netpoll_cleanup() when the target is
STATE_ENABLED. A target becomes STATE_DEACTIVATED when its underlying
interface is unregistered: netconsole_netdev_event() moves it to
target_cleanup_list, and netconsole_process_cleanups_core() is expected
to run do_netpoll_cleanup() on it.
Now that drop_netconsole_target() takes target_cleanup_list_lock around
the unlink, a configfs removal racing with NETDEV_UNREGISTER can pull the
target off target_cleanup_list before the cleanup worker processes it.
The notifier drops the lock before calling
netconsole_process_cleanups_core(), so the worker then iterates a list
that no longer contains the target and never runs do_netpoll_cleanup() on
it. Because drop_netconsole_target() has already rewritten the state to
STATE_DISABLED, its own STATE_ENABLED check is false and netpoll_cleanup()
is skipped too. The net_device reference taken by netpoll_setup() is then
leaked and unregister_netdevice() hangs forever in netdev_wait_allrefs().
Capture whether the target still owns a netpoll before the state is
downgraded and clean it up for both STATE_ENABLED and STATE_DEACTIVATED
targets. netpoll_cleanup() is idempotent -- it skips when np->dev is
already NULL -- so it is safe even when the cleanup worker won the race
and already tore the netpoll down.
Breno Leitao [Thu, 4 Jun 2026 16:10:12 +0000 (09:10 -0700)]
netconsole: take target_cleanup_list_lock in drop_netconsole_target()
drop_netconsole_target() unlinks the target while only holding
target_list_lock. However, when the underlying interface has been
unregistered, netconsole_netdev_event() moves the target from
target_list to target_cleanup_list, and netconsole_process_cleanups_core()
walks that list under target_cleanup_list_lock only.
If a user removes the configfs target at the same time the cleanup
worker is iterating target_cleanup_list, list_del() can corrupt the list
because the two paths take disjoint locks while operating on the same
list node.
Acquire target_cleanup_list_lock around the list_del() so the unlink is
serialised against netconsole_process_cleanups_core() regardless of
which list the target currently belongs to. The state transition that
downgrades STATE_DEACTIVATED to STATE_DISABLED is left intact and is
performed under the same combined locking, preserving the existing
ordering with resume_target().
Breno Leitao [Thu, 4 Jun 2026 16:10:11 +0000 (09:10 -0700)]
netconsole: do not dequeue pooled skbs that cannot satisfy len
find_skb() falls back to np->skb_pool when the GFP_ATOMIC alloc_skb()
fails. The pool is refilled by refill_skbs(), which always allocates
buffers of MAX_SKB_SIZE (ethhdr + iphdr + udphdr + MAX_UDP_CHUNK ==
1502 bytes).
netconsole, however, computes the requested length dynamically as
total_len + np->dev->needed_tailroom
If the egress device declares a non-zero needed_tailroom (e.g. some
tunnel or hardware accelerator devices), the required length can exceed
MAX_SKB_SIZE. The pooled skb is then handed back to the caller, which
immediately performs skb_put(skb, len), trips the tail > end check, and
triggers skb_over_panic().
Leave the normal alloc_skb(len, GFP_ATOMIC) path untouched -- the slab
allocator can still satisfy oversized requests when memory is available,
so senders to devices with non-zero needed_tailroom keep working in the
common case. Only the pool fallback is gated: when alloc_skb() failed
and len exceeds the pool buffer size, skip the skb_dequeue() instead of
burning a pre-allocated skb on a request that would later trip
skb_over_panic(). Reserving pool entries for requests they can actually
satisfy also keeps the panic path, which depends on the pool being
primed, intact.
When that drop happens, emit a rate-limited net_warn() so the user
notices that netconsole is unable to push messages on the egress device.
The warn is skipped under in_nmi() for the same reason schedule_work()
is: printk machinery taken by net_warn_ratelimited() is not NMI-safe and
would risk recursing into the same nbcon console we are servicing.
MAX_SKB_SIZE / MAX_UDP_CHUNK were private to net/core/netpoll.c. Move
them to include/linux/netpoll.h so netconsole can reference the same
definition that refill_skbs() uses, keeping the two in sync by
construction. The header now pulls in <linux/ip.h> and <linux/udp.h>
explicitly so MAX_SKB_SIZE remains self-contained for any future user.
Breno Leitao [Thu, 4 Jun 2026 16:10:10 +0000 (09:10 -0700)]
netconsole: do not schedule skb pool refill from NMI
When alloc_skb() fails in find_skb(), the fallback path dequeues an skb
from np->skb_pool and unconditionally calls schedule_work() to top the
pool back up. schedule_work() ends up taking the workqueue pool locks,
which are not NMI-safe.
netconsole_write() is registered as the nbcon write_atomic callback and
is explicitly marked CON_NBCON_ATOMIC_UNSAFE, meaning it is invoked from
emergency/panic contexts including NMIs. If the NMI interrupts a thread
already holding the workqueue pool lock, calling schedule_work()
self-deadlocks and the panic message that was being printed is lost.
Introduce netcons_skb_pop() to fold the pool dequeue and the refill
request into a single helper. The helper skips schedule_work() when
called from NMI context; the pool is best-effort, so the refill is simply
deferred to the next non-NMI find_skb() call that exhausts alloc_skb()
and hits the fallback again. This keeps the fast path untouched and the
locking rules around the fallback pool documented in one place.
Note this only removes the schedule_work() hazard from the NMI path. The
allocation itself is still not fully NMI-safe: the alloc_skb(GFP_ATOMIC)
attempted first may take slab locks, and the skb_dequeue() fallback takes
np->skb_pool.lock, so either can deadlock if the NMI interrupts a holder
of those locks. Closing those windows requires an NMI-safe (lockless) skb
pool and is left to a follow-up; this patch addresses the schedule_work()
deadlock, which is both the most likely and the easiest to trigger.
Jakub Kicinski [Sun, 7 Jun 2026 00:24:01 +0000 (17:24 -0700)]
selftests: drv-net: gro: signal over-coalescing more reliably
GRO test is very timing-sensitive, packets may be delayed
by the network or just sent slowly. Because of this we retry
each test case up to 6 times.
This makes perfect sense for positive cases, in which we want
to see coalescing. Negative test cases, which modify headers
and expect no coalescing should have opposite treatment.
We should really try 6 times and make sure that each time
the test failed. This would, however, require that we annotate
each test to indicate whether its positive or negative.
Let's start with a simpler improvement. Do not allow
retries if we detected over-coalescing. Previously the negative
case would have to get lucky at least once in 6 tries to pass.
Now the first failure breaks the retry loop.
For background - NICs tend to ignore the contents of the TCP
timestamp option, so that test case commonly fails. In NIPA
having 6 attempts, however, was enough for some NICs to get
multiple successful runs in a row, getting the test cases
auto-classified as expected to pass, even tho the NIC does
not comply with the expectations.
Victor Nogueira [Fri, 5 Jun 2026 21:22:39 +0000 (18:22 -0300)]
net/sched: Update function name in TCQ_F_NOPARENT comment
Commit 2ccccf5fb43f ("net_sched: update hierarchical backlog too") renamed
qdisc_tree_decrease_qlen() to qdisc_tree_reduce_backlog(), but this comment
was missed. Update it.
====================
tls: receive-path fixes and clean-ups
I'd like to encourage in-kernel kTLS consumers (NFSD, NVMe/TCP) to
coalesce on the use of read_sock. While auditing read_sock for that
purpose, Hannes and Sabrina flagged a few rough edges in the receive
paths.
This series is a set of clean-ups, not a performance series. Async
batch decryption and its submit/deliver scaffolding were dropped
during previous review: async_capable is always false for TLS 1.3,
the version NFSD and NVMe/TCP both require, so async-related
improvements were unreachable for the in-kernel consumers this
work targets.
A subsequent series will introduce infrastructure to support
KeyUpdate for in-kernel kTLS consumers, which need to handle TLS
Alert messages that trigger a tlshd upcall.
====================
Chuck Lever [Thu, 4 Jun 2026 17:48:29 +0000 (13:48 -0400)]
tls: Flush backlog before waiting for a new record
While lock_sock is held, incoming TCP segments land on
sk->sk_backlog rather than sk->sk_receive_queue.
tls_rx_rec_wait() inspects only sk_receive_queue, so backlog
data remains invisible. For non-blocking callers (read_sock,
and recvmsg or splice_read with MSG_DONTWAIT) this causes a
spurious -EAGAIN. For blocking callers it forces an
unnecessary sleep/wakeup cycle.
Flush the backlog inside tls_rx_rec_wait() before checking
sk_receive_queue so the strparser can parse newly-arrived
segments immediately. On the next loop iteration
tls_read_flush_backlog() may redundantly flush, but this
path is cold and the cost is negligible.
Backlog processing can run tcp_reset(), which calls
tcp_done_with_error() to set sk->sk_err = ECONNRESET and then
tcp_done() to set sk->sk_shutdown = SHUTDOWN_MASK. The pre-existing
top-of-loop sk_err check already ran before the flush, so the
freshly-set error would be masked by the next-line sk_shutdown test
returning 0 (EOF). Re-check sk_err immediately before the sk_shutdown
test so a connection abort surfaces as -ECONNRESET rather than a clean
EOF.
Commit f508262ae9f2 ("tls: Preserve sk_err across recvmsg() when
data has been copied") gave the top-of-loop sk_err check a
has_copied split. The recheck applies the same handling: when the
caller has already copied bytes, sk_err is reported but preserved
so the error surfaces on the next call; otherwise sock_error()
consumes it so the error is reported exactly once.
Chuck Lever [Thu, 4 Jun 2026 17:48:28 +0000 (13:48 -0400)]
tls: Suppress spurious saved_data_ready on all receive paths
Each record release via tls_strp_msg_done() triggered
tls_strp_check_rcv(), which called tls_rx_msg_ready() and
fired saved_data_ready(). During a multi-record receive, the
first N-1 wakeups are pure overhead: the caller is already
running and will pick up subsequent records on the next loop
iteration. The recvmsg and splice_read paths share this waste.
Suppress per-record notifications and emit a single one on
reader exit. tls_rx_rec_done() releases the current record
and parses the next without announcing; tls_strp_check_rcv()
gains a bool announce parameter so callers can request the
quiet form. tls_rx_reader_release() fires the deferred
announce on exit through tls_rx_msg_maybe_announce(), an
idempotent helper that calls saved_data_ready() only when a
record is parsed and has not yet been announced.
To keep the final notification idempotent against records that
the BH or the worker has already announced, tls_strparser gains
a msg_announced bit. tls_rx_msg_maybe_announce() sets the bit
when firing saved_data_ready(); the bit is cleared whenever
the parsed record is wiped, by tls_strp_msg_consume() on
consumption or by tls_strp_msg_load() when the lower socket
loses bytes from under the parse. A second call for the same
parsed record -- as when recvmsg() satisfies the request from
ctx->rx_list without touching the strparser -- becomes a
no-op.
With no remaining callers, tls_strp_msg_done() is removed.
Chuck Lever [Thu, 4 Jun 2026 17:48:27 +0000 (13:48 -0400)]
tls: Factor tls_strp_msg_consume() from tls_strp_msg_done()
tls_strp_msg_done() conflates releasing the current record with
checking for the next one via tls_strp_check_rcv(). A subsequent
patch needs to release a record without immediately triggering
that check, so the release step is separated into
tls_strp_msg_consume(). tls_strp_msg_done() is preserved as a
wrapper for existing callers.
Reviewed-by: Hannes Reinecke <hare@suse.de> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Reviewed-by: Sabrina Dubroca <sd@queasysnail.net> Link: https://patch.msgid.link/20260604-tls-read-sock-v12-4-b114efa6e3e2@oracle.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Chuck Lever [Thu, 4 Jun 2026 17:48:26 +0000 (13:48 -0400)]
tls: Move decrypt-failure abort into tls_rx_one_record()
Three receive paths -- recvmsg, read_sock, and splice_read --
each follow tls_rx_one_record() with the same tls_err_abort()
call. Consolidate the abort into tls_rx_one_record() so the
decrypt-and-abort sequence lives in one place.
A tls_check_pending_rekey() failure after successful
decryption no longer triggers tls_err_abort(). That path
fires only when skb_copy_bits() fails on a valid skb,
which is not a realistic scenario.
Suggested-by: Sabrina Dubroca <sd@queasysnail.net> Reviewed-by: Hannes Reinecke <hare@suse.de> Reviewed-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Link: https://patch.msgid.link/20260604-tls-read-sock-v12-3-b114efa6e3e2@oracle.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Chuck Lever [Thu, 4 Jun 2026 17:48:25 +0000 (13:48 -0400)]
tls: Re-present partially-consumed records in tls_sw_read_sock()
The tls_sw_read_sock() loop releases the current skb whether
read_actor() consumed the full record or only a prefix. When
the actor takes only part of the record and leaves desc->count
non-zero, the remainder is lost: skb is neither requeued nor
freed, and the next iteration overwrites it during dequeue or
tls_rx_rec_wait().
No mainline consumer reaches this path today. The only
in-tree TLS read_sock user is nvme/tcp, whose actor
nvme_tcp_recv_skb() loops until the input length is exhausted
and returns either the full length or a negative error.
The path becomes reachable with the upcoming NFSD svcsock
receive built on read_sock_cmsg. Its data actor,
svc_tcp_recv_actor(), parses an RPC fragment stream
incrementally and returns at fragment boundaries. When a TLS
record carries the tail of one RPC fragment plus the head of
the next, the actor returns fewer bytes than offered while
leaving desc->count non-zero, and without re-presentation the
trailing fragment header vanishes.
__tcp_read_sock() handles the equivalent case for plain TCP
by leaving the unread bytes available for the next iteration
to re-present, via sequence-number re-lookup. Adopt the same
loop-level behavior: when read_actor() consumes only part of
the record, update rxm->offset and rxm->full_len and requeue
the skb to the head of rx_list so the next iteration
re-presents the unread bytes. Switch the open-ended for-loop
to "while (desc->count)" so the partial- and full-consume
arms share a single exit check and read_actor() is not
re-invoked once desc->count is exhausted.
Cc: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Reviewed-by: Hannes Reinecke <hare@kernel.org> Reviewed-by: Sabrina Dubroca <sd@queasysnail.net> Link: https://patch.msgid.link/20260604-tls-read-sock-v12-2-b114efa6e3e2@oracle.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Chuck Lever [Thu, 4 Jun 2026 17:48:24 +0000 (13:48 -0400)]
tls: Avoid evaluating freed skb in tls_sw_read_sock() loop
tls_sw_read_sock() ends its receive loop with while (skb), but
the else branch in the body calls consume_skb(skb) before the
predicate is re-evaluated. A pointer becomes indeterminate when
the object it points to reaches end-of-lifetime (C2011 6.2.4p2),
and using an indeterminate value is undefined behavior (Annex
J.2). The pointer is not dereferenced today -- the predicate
either exits the loop or skb is overwritten at the top of the
next iteration -- but any future change that adds a dereference
between consume_skb() and the predicate would silently introduce
a use-after-free.
Replace the do/while form with an explicit for(;;) loop so
termination happens through a break statement rather than
predicate evaluation of a freed pointer.
Cc: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Reviewed-by: Hannes Reinecke <hare@kernel.org> Reviewed-by: Sabrina Dubroca <sd@queasysnail.net> Link: https://patch.msgid.link/20260604-tls-read-sock-v12-1-b114efa6e3e2@oracle.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Because LLC wasn't complicated/annoying enough, there's 2 more
"ethertypes" being used for it:
- 0x8870 is pretty "normal", it got standardized in
802.1AC-2016/Cor1-2018 for transporting LLC frames > 1500 bytes.
It simply replaces the length value (which is no longer encoded, and
must now be derived from the packet.) The actual value dates back to
2001; https://datatracker.ietf.org/doc/html/draft-ietf-isis-ext-eth-01
(it was used without "proper" standardization for a long time)
- 0x00fe is a doozy - actually "invalid" depending on how you look at
it; it's used in GRE (and possibly GENEVE) tunnels to transport the
IS-IS routing protocol. https://seclists.org/tcpdump/2002/q4/61 is
the best/oldest source I could find. It's inspired by the 0xfe SAP
value, a GRE packet with protocol 0x00fe is followed by a payload "as
if" it was Ethernet with "<length> 0xfe 0xfe 0x03". (Again the length
isn't encoded explicitly anymore.)
The 0x00fe value is quite close to other values the kernel is using
internally for various things (after all they "won't clash for 1500
types"). Except this one does clash, and if someone unknowingly starts
using it for something internal... we end up in a world of pain in
getting IS-IS running on GRE tunnels. Hence the "WARNING".
David Carlier [Thu, 4 Jun 2026 14:19:22 +0000 (15:19 +0100)]
net: garp: reload skb header pointers after pskb_may_pull()
garp_pdu_parse_attr() keeps a pointer into the skb linear area across
pskb_may_pull(skb, ga->len), and garp_pdu_parse_msg() dereferences gm
on every loop iteration even though the nested parse may pull again.
pskb_may_pull() can reallocate the skb head, which would leave those
pointers stale.
This is not reachable today: GARP PDUs arrive via the 802.2 LLC SAP
path, where llc_fixup_skb() already pulls and trims the whole payload
into the linear area, so the inner pulls never reallocate. Reload ga
after the pull and snapshot gm->attrtype into a local anyway, to harden
the parser and match the skb_header_pointer() discipline used by mrp.c.
Rosen Penev [Wed, 3 Jun 2026 23:08:21 +0000 (16:08 -0700)]
net: ibm: emac: mal: fix potential system hang in mal_remove()
napi_disable() is not idempotent and calling it on an already-disabled
or unenabled NAPI context will cause the kernel to spin indefinitely
waiting for the NAPI_STATE_SCHED bit to clear.
In mal_remove(), napi_disable() is called unconditionally. If no MACs were
registered, NAPI was never enabled. Also, if they were registered but
subsequently unregistered, NAPI was already disabled in
mal_unregister_commac(). In either case, calling napi_disable() causes
the kernel to hang upon module removal.
Fix this by only calling napi_disable() in mal_remove() if the commac list
is not empty (which implies NAPI is enabled).
Rosen Penev [Wed, 3 Jun 2026 23:07:53 +0000 (16:07 -0700)]
net: ibm: emac: Clear MAL descriptors without memset
Clear MAL descriptor rings with explicit field stores instead of
memset(). The descriptor rings are carved from MAL coherent DMA memory,
which may be mapped uncached on 32-bit powerpc. The optimized memset()
path can use dcbz there and trigger an alignment warning.
Use WRITE_ONCE() for each field to prevent the compiler from merging
the stores back into a memset() call.
The skb tracking arrays remain ordinary CPU memory and still use memset().
platform_get_irq() returns a negative errno on failure.
Commit c4f5d0454cab5 moved the platform_get_irq() calls and explicitly
removed the error checks that were previously present, claiming
devm_request_irq() can handle it. However, a negative IRQ number
passed to devm_request_irq() fails with -EINVAL instead of
propagating the real error from platform_get_irq().
Restore the missing error checks with proper errno propagation.
Nikhil P. Rao [Thu, 4 Jun 2026 21:36:37 +0000 (21:36 +0000)]
pds_core: quiesce DMA before freeing resources
pdsc_teardown() frees DMA buffers but does not disable bus mastering,
leaving the device able to perform DMA after the buffers are freed.
This can lead to use-after-free if the device writes to freed memory.
Add pci_clear_master() to pdsc_teardown() to disable bus mastering
before freeing resources, ensuring all DMA is quiesced.
Add pci_set_master() to pdsc_setup() to re-enable bus mastering,
which is needed for the firmware recovery path since pdsc_teardown()
now disables it.
ip6mr: Remove RTNL in ip6mr_rules_init() and ip6mr_net_init().
When ip6mr_free_table() is called from ip6mr_rules_init() or
ip6mr_net_init(), the netns is not yet published.
Thus, no device should have been registered, and
mroute_clean_tables() will not call mif6_delete(), so
unregister_netdevice_many() is unnecessary.
unregister_netdevice_many() does nothing if the list is empty,
but it requires RTNL due to the unconditional ASSERT_RTNL()
at the entry of unregister_netdevice_many_notify().
Let's remove unnecessary RTNL and ASSERT_RTNL() and instead
add WARN_ON_ONCE() in ip6mr_free_table().
Note that we use a local list for the new WARN_ON_ONCE() because
dev_kill_list passed from ip6mr_rules_exit_rtnl() may have some
devices when other ops->init() fails after ipmr durnig setup_net().
ip6mr: Convert ip6mr_net_exit_batch() to ->exit_rtnl().
ip6mr_net_ops uses ->exit_batch() to acquire RTNL only once
for dying network namespaces.
ip6mr does not depend on the ordering of ->exit_rtnl() and
->exit_batch() of other pernet_operations (unlike fib_net_ops).
Once ip6mr_free_table() is called and all devices are
queued for destruction in ->exit_rtnl(), later during
NETDEV_UNREGISTER, ip6mr_device_event() will not see anything
in vif table and just do nothing.
Let's convert ip6mr_net_exit_batch() to ->exit_rtnl().
We will remove RTNL and unregister_netdevice_many() in
ip6mr_rules_init().
Since default_device_exit_batch() is called after ->exit_rtnl(),
idev->mc_ifc_work could finally call mroute6_is_socket() under RCU
while ->exit_rtnl() is running. [0]
With CONFIG_IPV6_MROUTE_MULTIPLE_TABLES=n, ip6mr_fib_lookup() does
not check if net->ipv6.mrt6 is NULL. If ip6mr_net_exit_batch()
set net->ipv6.mrt6 to NULL and freed it, the mrt->mroute_sk access
could result in null-ptr-deref or use-after-free.
Let's prepare for that situation by applying RCU rule to ip6mr
table similarly.
!check_net(net) is added in ip6mr_cache_unresolved() and
mroute_clean_tables() to synchronise the two by mfc_unres_lock
so that ip6mr_cache_unresolved() will not queue skb after
mroute_clean_tables() purged &mrt->mfc_unres_queue.
rcu_read_lock() in reg_vif_xmit() is moved up to cover
ip6mr_fib_lookup() as with ipmr.
Commit 22e36ea9f5d7 ("inet: allow ip_valid_fib_dump_req() to
be called with RTNL or RCU") introduced the rtnl_held field in
struct fib_dump_filter to switch __dev_get_by_index() and
dev_get_by_index_rcu() depending on the caller's context.
This field served as an interim measure while we were incrementally
converting all callers of ip_valid_fib_dump_req() to RCU.
Now that all users (IPv4, IPv6, ipmr, ip6mr, and MPLS) have
been converted to RCU, the field is no longer necessary.
projects / thirdparty / kernel / stable.git / log
