This patch will allow to use a swnode/fwnode defined for a phy_device. The
MDIO bus (mii_bus) needs to contain nodes for the PHY devices, named
"ethernet-phy@i", with i being the MDIO address (0 .. PHY_MAX_ADDR - 1).
The fwnode is only attached to the phy_device if there isn't already an
fwnode attached.
fwnode_get_named_child_node will increase the usage counter of the fwnode.
However, no new code is needed to decrease the counter again, since this is
already implemented in the phy_device_release function.
Signed-off-by: Hans-Frieder Vogt <hfdevel@gmx.net>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250322-tn9510-v3a-v7-1-672a9a3d8628@gmx.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Joshua Washington says:
====================
Basic XDP Support for DQO RDA Queue Format
This patch series updates the GVE XDP infrastructure and introduces
XDP_PASS and XDP_DROP support for the DQO RDA queue format.
The infrastructure changes of note include an allocation path refactor
for XDP queues, and a unification of RX buffer sizes across queue
formats.
This patch series will be followed by more patch series to introduce
XDP_TX and XDP_REDIRECT support, as well as zero-copy and multi-buffer
support.
====================
Link: https://patch.msgid.link/20250321002910.1343422-1-hramamurthy@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In order to support installing an XDP program on DQ, RX buffers need to
be reposted using 4K buffers, which is larger than the default packet
buffer size of 2K. This is needed to accommodate the extra head and tail
that accompanies the data portion of an XDP buffer. Continuing to use 2K
buffers would mean that the packet buffer size for the NIC would have to
be restricted to 2048 - 320 - 256 = 1472B. However, this is problematic
for two reasons: first, 1472 is not a packet buffer size accepted by
GVE; second, at least 1474B of buffer space is needed to accommodate an
MTU of 1460, which is the default on GCP. As such, we allocate 4K
buffers, and post a 2K section of those 4K buffers (offset relative to
the XDP headroom) to the NIC for DMA to avoid a potential extra copy.
Because the GQ-QPL datapath requires copies regardless, this change was
not needed to support XDP in that case.
To capture this subtlety, a new field, packet_buffer_truesize, has been
added to the rx ring struct to represent size of the allocated buffer,
while packet_buffer_size has been left to represent the portion of the
buffer posted to the NIC.
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Praveen Kaligineedi <pkaligineedi@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Signed-off-by: Harshitha Ramamurthy <hramamurthy@google.com>
Link: https://patch.msgid.link/20250321002910.1343422-6-hramamurthy@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The data_buffer_size_dqo field in gve_priv and the packet_buffer_size
field in gve_rx_ring theoretically have the same meaning, but they are
defined in two different places and used in two separate contexts. There
is no good reason for this, so this change merges those fields into the
packet_buffer_size field in the RX ring.
This change also introduces a packet_buffer_size field to struct
gve_rx_queue_config to account for cases where queues are not allocated,
such as when the interface is down.
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Signed-off-by: Harshitha Ramamurthy <hramamurthy@google.com>
Link: https://patch.msgid.link/20250321002910.1343422-5-hramamurthy@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit ebdfae0d37 ("gve: adopt page pool for DQ RDA mode") introduced
a buf_size field to the gve_rx_slot_page_info struct, which can be used
in the datapath to take the place of the packet_buffer_size field, as it
will already be hot in the cache due to its extensive use. Using the
buf_size field in the datapath frees up the packet_buffer_size field in
the GQ-specific RX cacheline to be generalized for GQ and DQ (in the
next patch), as there is currently no common packet buffer size field
between the two queue formats.
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Signed-off-by: Harshitha Ramamurthy <hramamurthy@google.com>
Link: https://patch.msgid.link/20250321002910.1343422-4-hramamurthy@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
An earlier patch series[1] introduced RX/TX ring allocation configuration
structs which contained metadata used to allocate and configure new RX
and TX rings. This led to a much cleaner and safer allocation pattern
wherein queue resources were not deallocated until new queue resources
were successfully allocated.
Migrate the XDP allocation path to use the same pattern to allow for the
existence of a single allocation path instead of relying on XDP-specific
allocation methods. These extra allocation methods result in the
duplication of many existing behaviors while being prone to error when
configuration changes unrelated to XDP occur.
Link: https://lore.kernel.org/netdev/20240122182632.1102721-1-shailend@google.com/ [1]
Reviewed-by: Praveen Kaligineedi <pkaligineedi@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Signed-off-by: Harshitha Ramamurthy <hramamurthy@google.com>
Link: https://patch.msgid.link/20250321002910.1343422-3-hramamurthy@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Dumazet says:
====================
tcp/dccp: remove 16 bytes from icsk
icsk->icsk_timeout and icsk->icsk_ack.timeout can be removed.
They mirror existing fields in icsk->icsk_retransmit_timer and
icsk->icsk_retransmit_timer.
====================
Link: https://patch.msgid.link/20250324203607.703850-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
netdev netlink is the only reader of netdev_{,rx_}queue->napi,
and it already holds netdev->lock. Switch protection of
the writes to netdev->lock to "ops protected".
The expectation will be now that accessing queue->napi
will require netdev->lock for "ops locked" drivers, and
rtnl_lock for all other drivers.
Current "ops locked" drivers don't require any changes.
gve and netdevsim use _locked() helpers right next to
netif_queue_set_napi() so they must be holding the instance
lock. iavf doesn't call it. bnxt is a bit messy but all paths
seem locked.
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250324224537.248800-7-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Drivers which opt into instance lock protection of ops should
only call set_real_num_*_queues() under the instance lock.
This means that queue counts are double protected (writes
are under both rtnl_lock and instance lock, readers under
either).
Some readers may still be under the rtnl_lock, however, so for
now we need double protection of writers.
OTOH queue API paths are only under the protection of the instance
lock, so we need to validate that the instance is actually locking
ops, otherwise the input checks we do against queue count are racy.
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250324224537.248800-6-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Try to define some terminology for which fields are protected
by which lock and how. Some fields are protected by both rtnl_lock
and instance lock which is hard to talk about without having
a "key phrase" to refer to a particular protection scheme.
"ops protected" fields are defined later in the series, one by one.
Add ASSERT_RTNL() to netdev_ops_assert_locked() for drivers
not other instance protection of ops. Hopefully it's not too
confusion that netdev_lock_ops() does not match the lock which
netdev_ops_assert_locked() will assert, exactly. The noun "ops"
is in a different place in the name, so I think it's acceptable...
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250324224537.248800-5-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Since commit a953be53ce ("net-sysfs: add support for device-specific
rx queue sysfs attributes"), so for at least a decade now it is safe
to call net_rx_queue_update_kobjects() when SYSFS=n. That function
does its own ifdef-inery and will return 0. Remove the unnecessary
stub for netif_set_real_num_rx_queues().
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250324224537.248800-3-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A recent commit added taking the netdev instance lock
in netdev_nl_bind_rx_doit(), but didn't remove it in
net_devmem_unbind_dmabuf() which it calls from an error path.
Always expect the callers of net_devmem_unbind_dmabuf() to
hold the lock. This is consistent with net_devmem_bind_dmabuf().
(Not so) coincidentally this also protects mp_param with the instance
lock, which the rest of this series needs.
Fixes: 1d22d3060b ("net: drop rtnl_lock for queue_mgmt operations")
Reviewed-by: Mina Almasry <almasrymina@google.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250324224537.248800-2-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Revert "udp_tunnel: use static call for GRO hooks when possible"
This reverts commit 311b36574c.
Revert "udp_tunnel: create a fastpath GRO lookup."
This reverts commit 8d4880db37.
There are multiple small issues with the series. In the interest
of unblocking the merge window let's opt for a revert.
Link: https://lore.kernel.org/cover.1742557254.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
PHYs that are within copper SFP modules have their MDIO bus accessible
through address 0x56 (usually) on the i2c bus. The MDIO-I2C bridge is
desgned for 16 bits accesses, but we can also perform 8bits accesses by
reading/writing the high and low bytes sequentially.
This commit adds support for this type of accesses, thus supporting
smbus controllers such as the one in the VSC8552.
This was only tested on Copper SFP modules that embed a Marvell 88e1111
PHY.
Tested-by: Sean Anderson <sean.anderson@linux.dev>
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250322075745.120831-3-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The SFP module's eeprom and internals are accessible through an i2c bus.
It is possible that the SFP might be connected to an SMBus-only
controller, such as the one found in some PHY devices in the VSC85xx
family.
Introduce a set of sfp read/write ops that are going to be used if the
i2c bus is only capable of doing smbus byte accesses.
As Single-byte SMBus transaction go against SFF-8472 and breaks the
atomicity for diagnostics data access, hwmon is disabled in the case
of SMBus access.
Moreover, as this may cause other instabilities, print a warning at
probe time to indicate that the setup may be unreliable because of the
hardware design.
As hwmon may be disabled for both broken EEPROM and smbus, the warnings
are udpated accordingly.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250322075745.120831-2-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Steffen Klassert says:
====================
pull request (net-next): ipsec-next 2025-03-24
1) Prevent setting high order sequence number bits input in
non-ESN mode. From Leon Romanovsky.
2) Support PMTU handling in tunnel mode for packet offload.
From Leon Romanovsky.
3) Make xfrm_state_lookup_byaddr lockless.
From Florian Westphal.
4) Remove unnecessary NULL check in xfrm_lookup_with_ifid().
From Dan Carpenter.
* tag 'ipsec-next-2025-03-24' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next:
xfrm: Remove unnecessary NULL check in xfrm_lookup_with_ifid()
xfrm: state: make xfrm_state_lookup_byaddr lockless
xfrm: check for PMTU in tunnel mode for packet offload
xfrm: provide common xdo_dev_offload_ok callback implementation
xfrm: rely on XFRM offload
xfrm: simplify SA initialization routine
xfrm: delay initialization of offload path till its actually requested
xfrm: prevent high SEQ input in non-ESN mode
====================
Link: https://patch.msgid.link/20250324061855.4116819-1-steffen.klassert@secunet.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following batch contains Netfilter updates for net-next:
1) Use kvmalloc in xt_hashlimit, from Denis Kirjanov.
2) Tighten nf_conntrack sysctl accepted values for nf_conntrack_max
and nf_ct_expect_max, from Nicolas Bouchinet.
3) Avoid lookup in nft_fib if socket is available, from Florian Westphal.
4) Initialize struct lsm_context in nfnetlink_queue to avoid
hypothetical ENOMEM errors, Chenyuan Yang.
5) Use strscpy() instead of _pad when initializing xtables table name,
kzalloc is already used to initialized the table memory area.
From Thorsten Blum.
6) Missing socket lookup by conntrack information for IPv6 traffic
in nft_socket, there is a similar chunk in IPv4, this was never
added when IPv6 NAT was introduced. From Maxim Mikityanskiy.
7) Fix clang issues with nf_tables CONFIG_MITIGATION_RETPOLINE,
from WangYuli.
* tag 'nf-next-25-03-23' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next:
netfilter: nf_tables: Only use nf_skip_indirect_calls() when MITIGATION_RETPOLINE
netfilter: socket: Lookup orig tuple for IPv6 SNAT
netfilter: xtables: Use strscpy() instead of strscpy_pad()
netfilter: nfnetlink_queue: Initialize ctx to avoid memory allocation error
netfilter: fib: avoid lookup if socket is available
netfilter: conntrack: Bound nf_conntrack sysctl writes
netfilter: xt_hashlimit: replace vmalloc calls with kvmalloc
====================
Link: https://patch.msgid.link/20250323100922.59983-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
RFS is using two kinds of hash tables.
First one is controlled by /proc/sys/net/core/rps_sock_flow_entries = 2^N
and using the N low order bits of the l4 hash is good enough.
Then each RX queue has its own hash table, controlled by
/sys/class/net/eth1/queues/rx-$q/rps_flow_cnt = 2^X
Current hash function, using the X low order bits is suboptimal,
because RSS is usually using Func(hash) = (hash % power_of_two);
For example, with 32 RX queues, 6 low order bits have no entropy
for a given queue.
Switch this hash function to hash_32(hash, log) to increase
chances to use all possible slots and reduce collisions.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250321171309.634100-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Johannes Berg says:
====================
More features for 6.15, major changes:
* cfg80211/mac80211: fix and enable link reconfiguration
* rtw88: support RTL8814AE/RTL8814AU
* mt7996: preparations for MLO
* ath12k: continued work on MLO
* iwlwifi: add new iwlmld sub-driver/op-mode for
some current and future devices
* wfx: wowlan support
* tag 'wireless-next-2025-03-20' of https://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (311 commits)
wifi: mt76: mt7996: fix locking in mt7996_mac_sta_rc_work()
wifi: mt76: mt76x2u: add TP-Link TL-WDN6200 ID to device table
wifi: mt76: mt792x: re-register CHANCTX_STA_CSA only for the mt7921 series
wifi: mt76: mt7996: Update mt7996_tx to MLO support
wifi: mt76: mt7996: rework mt7996_ampdu_action to support MLO
wifi: mt76: mt7996: rework set/get_tsf callabcks to support MLO
wifi: mt76: mt7996: set vif default link_id adding/removing vif links
wifi: mt76: mt7996: rework mt7996_mcu_beacon_inband_discov to support MLO
wifi: mt76: mt7996: rework mt7996_mcu_add_obss_spr to support MLO
wifi: mt76: mt7996: rework mt7996_net_fill_forward_path to support MLO
wifi: mt76: mt7996: rework mt7996_update_mu_group to support MLO
wifi: mt76: mt7996: rework mt7996_mac_sta_poll to support MLO
wifi: mt76: mt7996: rework mt7996_mac_sta_rc_work to support MLO
wifi: mt76: mt7996: remove mt7996_mac_enable_rtscts()
wifi: mt76: mt7996: rework mt7996_sta_hw_queue_read to support MLO
wifi: mt76: mt7996: rework mt7996_set_hw_key to support MLO
wifi: mt76: mt7996: Add mt7996_sta_link to mt7996_mcu_add_bss_info signature
wifi: mt76: mt7996: rework mt7996_sta_set_4addr and mt7996_sta_set_decap_offload to support MLO
wifi: mt76: mt7996: rework mt7996_rx_get_wcid to support MLO
wifi: mt76: mt7996: Rely on wcid_to_sta in mt7996_mac_add_txs_skb()
...
====================
Link: https://patch.msgid.link/20250320131106.33266-3-johannes@sipsolutions.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Rockchip RK3528 (and RV1106) has a different integrated PHY compared to
the integrated PHY on RK3228/RK3328. Current powerup/down operation is
not compatible with the integrated PHY found in these newer SoCs.
Add operations to powerup/down the integrated PHY found in RK3528.
Use helpers that can be used by other GMAC variants in the future.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250319214415.3086027-6-jonas@kwiboo.se
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Rockchip RK3528 (and RV1106) has a different integrated PHY compared to
the integrated PHY on RK3228/RK3328. Current powerup/down operation is
not compatible with the integrated PHY found in these newer SoCs.
Add a new integrated_phy_powerdown operation and change the call chain
for integrated_phy_powerup to prepare support for the integrated PHY
found in these newer SoCs.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250319214415.3086027-5-jonas@kwiboo.se
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Rockchip RK3528 (and RV1106) has a different integrated PHY compared to
the integrated PHY on RK3228/RK3328. Current powerup/down operation is
not compatible with the integrated PHY found in these SoCs.
Move the rk_gmac_integrated_phy_powerup/down functions to top of the
file to prepare for them to be called directly by a GMAC variant
specific powerup/down operation.
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250319214415.3086027-4-jonas@kwiboo.se
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Russell King says:
====================
net: improve stmmac resume rx clocking
stmmac has had a long history of problems with resuming, illustrated by
reset failure due to the receive clock not running.
Several attempts have been attempted over the years to address this
issue, such as moving phylink_start() (now phylink_resume()) super
early in stmmac_resume() in commit 90702dcd19 ("net: stmmac: fix MAC
not working when system resume back with WoL a ctive.") However, this
has the downside that stmmac_mac_link_up() can (and demonstrably is)
called before or during the driver initialisation in another thread.
This can cause issues as packets could begin to be queued, and the
transmit/receive enable bits will be set before any initialisation has
been done.
Another attempt is used by dwmac-socfpga.c in commit 2d871aa071 ("net:
stmmac: add platform init/exit for Altera's ARM socfpga") which
pre-dates the above commit.
Neither of these two approaches consider the effect of EEE with a PHY
that supports receive clock-stop and has that feature enabled (which
the stmmac driver does enable). If the link is up, then there is the
possibility for the receive path to be in low-power mode, and the PHY
may stop its receive clock.
This series addresses these issues by (each is not necessarily a
separate patch):
1) introducing phylink_prepare_resume(), which can be used by MAC
drivers to ensure that the PHY is resumed prior to doing any
re-initialisation work. This call is added to stmmac_resume().
2) moving phylink_resume() after all re-initialisation has completed,
thereby ensuring that the hardware is ready to be enabled for
packet reception/transmission.
3) with (1) and (2) addressed, the need for socfpga to have a private
work-around is no longer necessary, so it is removed.
4) introducing phylink functions to block/unblock the receive clock-
stop at the PHY. As these require PHY access over the MDIO bus,
they can sleep, so are not suitable for atomic access.
5) the stmmac hardware requires the receive clock to be running for
reset to complete. Depending on synthesis options, this requirement
may also extend to writing various registers as well, e.g. setting
the MAC address, writing some of the vlan registers, etc. Full
details are in the databook.
We add blocking/unblocking of the PHY receive clock-stop around
parts of the main stmmac driver where we have a context that we
can sleep. These are wrapped with the new phylink functions.
However, depending on synthesis options, there could be other
places where the net core calls the driver with a BH-disabled
context where we can't sleep, and thus can't block the PHY from
disabling its receive clock. These are documented with FIXME
comments.
Given the last paragraph above, I am wondering whether a better
approach would be to ensure that receive clock-stop is always disabled
at the PHY with stmmac. From what I can see, implementations do not
document to this level of detail, which makes it difficult to tell
which registers require the receive clock to be running to behave
correctly.
This patch series has been tested on the Tegra194 Jetson Xavier NX
board kindly donated by NVidia, with two additional patches that are
pending in patchwork - the first is required to have EEE's LPI mode
passed through to the MAC on this platform to allow testing under
PHY clock-stop scenarios. The second is a bug fix for PHYLIB and
makes "eee off" functional, but should not affect this series.
All patches on top of net-next commit f749448ce9 ("Merge branch
'net-mlx5-hw-steering-cleanups'")
https://patchwork.kernel.org/project/netdevbpf/patch/E1ttnHW-00785s-Uq@rmk-PC.armlinux.org.uk/https://patchwork.kernel.org/project/netdevbpf/patch/E1ttmWN-0077Mb-Q6@rmk-PC.armlinux.org.uk/
====================
Link: https://patch.msgid.link/Z9ySeo61VYTClIJJ@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The DesignWare core requires the receive clock to be running during
certain operations. Ensure that we block PHY RXC clock-stop during
these operations.
This is a best-efforts change - not everywhere can be covered by this
because of net's core locking, which means we can't access the MDIO
bus to configure the PHY to disable RXC clock-stop in certain areas.
These are marked with FIXME comments.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/E1tvO6p-008Vjz-Qy@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
