linux.git/net/dsa/dsa.c, branch for-next

net: dsa: remove obsolete phylink dsa_switch operations

2024-10-07T23:23:10+00:00

No driver now uses the DSA switch phylink members, so we can now remove
the method pointers, but we need to leave empty shim functions to allow
those drivers that do not provide phylink MAC operations structure to
continue functioning.

Signed-off-by: Russell King (oracle) 
Reviewed-by: Vladimir Oltean 
Tested-by: Vladimir Oltean  # sja1105, felix, dsa_loop
Link: https://patch.msgid.link/E1swKNV-0060oN-1b@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski

net: dsa: improve shutdown sequence

2024-10-01T07:56:36+00:00

Alexander Sverdlin presents 2 problems during shutdown with the
lan9303 driver. One is specific to lan9303 and the other just happens
to reproduce there.

The first problem is that lan9303 is unique among DSA drivers in that it
calls dev_get_drvdata() at "arbitrary runtime" (not probe, not shutdown,
not remove):

phy_state_machine()
-> ...
   -> dsa_user_phy_read()
      -> ds->ops->phy_read()
         -> lan9303_phy_read()
            -> chip->ops->phy_read()
               -> lan9303_mdio_phy_read()
                  -> dev_get_drvdata()

But we never stop the phy_state_machine(), so it may continue to run
after dsa_switch_shutdown(). Our common pattern in all DSA drivers is
to set drvdata to NULL to suppress the remove() method that may come
afterwards. But in this case it will result in an NPD.

The second problem is that the way in which we set
dp->conduit->dsa_ptr = NULL; is concurrent with receive packet
processing. dsa_switch_rcv() checks once whether dev->dsa_ptr is NULL,
but afterwards, rather than continuing to use that non-NULL value,
dev->dsa_ptr is dereferenced again and again without NULL checks:
dsa_conduit_find_user() and many other places. In between dereferences,
there is no locking to ensure that what was valid once continues to be
valid.

Both problems have the common aspect that closing the conduit interface
solves them.

In the first case, dev_close(conduit) triggers the NETDEV_GOING_DOWN
event in dsa_user_netdevice_event() which closes user ports as well.
dsa_port_disable_rt() calls phylink_stop(), which synchronously stops
the phylink state machine, and ds->ops->phy_read() will thus no longer
call into the driver after this point.

In the second case, dev_close(conduit) should do this, as per
Documentation/networking/driver.rst:

| Quiescence
| ----------
|
| After the ndo_stop routine has been called, the hardware must
| not receive or transmit any data.  All in flight packets must
| be aborted. If necessary, poll or wait for completion of
| any reset commands.

So it should be sufficient to ensure that later, when we zeroize
conduit->dsa_ptr, there will be no concurrent dsa_switch_rcv() call
on this conduit.

The addition of the netif_device_detach() function is to ensure that
ioctls, rtnetlinks and ethtool requests on the user ports no longer
propagate down to the driver - we're no longer prepared to handle them.

The race condition actually did not exist when commit 0650bf52b31f
("net: dsa: be compatible with masters which unregister on shutdown")
first introduced dsa_switch_shutdown(). It was created later, when we
stopped unregistering the user interfaces from a bad spot, and we just
replaced that sequence with a racy zeroization of conduit->dsa_ptr
(one which doesn't ensure that the interfaces aren't up).

Reported-by: Alexander Sverdlin 
Closes: https://lore.kernel.org/netdev/2d2e3bba17203c14a5ffdabc174e3b6bbb9ad438.camel@siemens.com/
Closes: https://lore.kernel.org/netdev/c1bf4de54e829111e0e4a70e7bd1cf523c9550ff.camel@siemens.com/
Fixes: ee534378f005 ("net: dsa: fix panic when DSA master device unbinds on shutdown")
Reviewed-by: Alexander Sverdlin 
Tested-by: Alexander Sverdlin 
Signed-off-by: Vladimir Oltean 
Link: https://patch.msgid.link/20240913203549.3081071-1-vladimir.oltean@nxp.com
Signed-off-by: Paolo Abeni

net: dsa: remove mac_prepare()/mac_finish() shims

2024-05-30T01:41:15+00:00

No DSA driver makes use of the mac_prepare()/mac_finish() shimmed
operations anymore, so we can remove these.

Signed-off-by: Russell King (Oracle) 
Reviewed-by: Florian Fainelli 
Link: https://lore.kernel.org/r/E1sByNx-00ELW1-Vp@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski

net: dsa: Remove adjust_link paths

2024-05-02T02:06:32+00:00

Now that we no longer any drivers using PHYLIB's adjust_link callback,
remove all paths that made use of adjust_link as well as the associated
functions.

Signed-off-by: Florian Fainelli 
Reviewed-by: Russell King (Oracle) 
Link: https://lore.kernel.org/r/20240430164816.2400606-3-florian.fainelli@broadcom.com
Signed-off-by: Jakub Kicinski

net: dsa: allow DSA switch drivers to provide their own phylink mac ops

2024-04-12T03:01:14+00:00

Rather than having a shim for each and every phylink MAC operation,
allow DSA switch drivers to provide their own ops structure. When a
DSA driver provides the phylink MAC operations, the shimmed ops must
not be provided, so fail an attempt to register a switch with both
the phylink_mac_ops in struct dsa_switch and the phylink_mac_*
operations populated in dsa_switch_ops populated.

Signed-off-by: Russell King (Oracle) 
Reviewed-by: Vladimir Oltean 
Reviewed-by: Florian Fainelli 
Link: https://lore.kernel.org/r/E1rudqF-006K9H-Cc@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski

net: dsa: remove OF-based MDIO bus registration from DSA core

2024-02-16T05:30:02+00:00

The code block under the "!ds->user_mii_bus && ds->ops->phy_read" check
under dsa_switch_setup() populates ds->user_mii_bus. The use of
ds->user_mii_bus is inappropriate when the MDIO bus of the switch is
described on the device tree [1].

For this reason, use this code block only for switches [with MDIO bus]
probed on platform_data, and OF which the switch MDIO bus isn't described
on the device tree. Therefore, remove OF-based MDIO bus registration as
it's useless for these cases.

These subdrivers which control switches [with MDIO bus] probed on OF, will
lose the ability to register the MDIO bus OF-based:

drivers/net/dsa/b53/b53_common.c
drivers/net/dsa/lan9303-core.c
drivers/net/dsa/vitesse-vsc73xx-core.c

These subdrivers let the DSA core driver register the bus:
- ds->ops->phy_read() and ds->ops->phy_write() are present.
- ds->user_mii_bus is not populated.

The commit fe7324b93222 ("net: dsa: OF-ware slave_mii_bus") which brought
OF-based MDIO bus registration on the DSA core driver is reasonably recent
and, in this time frame, there have been no device trees in the Linux
repository that started describing the MDIO bus, or dt-bindings defining
the MDIO bus for the switches these subdrivers control. So I don't expect
any devices to be affected.

The logic we encourage is that all subdrivers should register the switch
MDIO bus on their own [2]. And, for subdrivers which control switches [with
MDIO bus] probed on OF, this logic must be followed to support all cases
properly:

No switch MDIO bus defined: Populate ds->user_mii_bus, register the MDIO
bus, set the interrupts for PHYs if "interrupt-controller" is defined at
the switch node. This case should only be covered for the switches which
their dt-bindings documentation didn't document the MDIO bus from the
start. This is to keep supporting the device trees that do not describe the
MDIO bus on the device tree but the MDIO bus is being used nonetheless.

Switch MDIO bus defined: Don't populate ds->user_mii_bus, register the MDIO
bus, set the interrupts for PHYs if ["interrupt-controller" is defined at
the switch node and "interrupts" is defined at the PHY nodes under the
switch MDIO bus node].

Switch MDIO bus defined but explicitly disabled: If the device tree says
status = "disabled" for the MDIO bus, we shouldn't need an MDIO bus at all.
Instead, just exit as early as possible and do not call any MDIO API.

After all subdrivers that control switches with MDIO buses are made to
register the MDIO buses on their own, we will be able to get rid of
dsa_switch_ops :: phy_read() and :: phy_write(), and the code block for
registering the MDIO bus on the DSA core driver.

Link: https://lore.kernel.org/netdev/20231213120656.x46fyad6ls7sqyzv@skbuf/ [1]
Link: https://lore.kernel.org/netdev/20240103184459.dcbh57wdnlox6w7d@skbuf/ [2]
Suggested-by: Luiz Angelo Daros de Luca
Acked-by: Luiz Angelo Daros de Luca
Signed-off-by: Arınç ÜNAL
Reviewed-by: Vladimir Oltean
Link: https://lore.kernel.org/r/20240213-for-netnext-dsa-mdio-bus-v2-1-0ff6f4823a9e@arinc9.com
Signed-off-by: Jakub Kicinski

net: dsa: Use conduit and user terms

2023-10-24T20:08:14+00:00

Use more inclusive terms throughout the DSA subsystem by moving away
from "master" which is replaced by "conduit" and "slave" which is
replaced by "user". No functional changes.

Acked-by: Rob Herring 
Acked-by: Stephen Hemminger 
Reviewed-by: Vladimir Oltean 
Signed-off-by: Florian Fainelli 
Link: https://lore.kernel.org/r/20231023181729.1191071-2-florian.fainelli@broadcom.com
Signed-off-by: Jakub Kicinski

net: dsa: avoid suspicious RCU usage for synced VLAN-aware MAC addresses

2023-06-27T16:37:41+00:00

When using the felix driver (the only one which supports UC filtering
and MC filtering) as a DSA master for a random other DSA switch, one can
see the following stack trace when the downstream switch ports join a
VLAN-aware bridge:

=============================
WARNING: suspicious RCU usage
-----------------------------
net/8021q/vlan_core.c:238 suspicious rcu_dereference_protected() usage!

stack backtrace:
Workqueue: dsa_ordered dsa_slave_switchdev_event_work
Call trace:
 lockdep_rcu_suspicious+0x170/0x210
 vlan_for_each+0x8c/0x188
 dsa_slave_sync_uc+0x128/0x178
 __hw_addr_sync_dev+0x138/0x158
 dsa_slave_set_rx_mode+0x58/0x70
 __dev_set_rx_mode+0x88/0xa8
 dev_uc_add+0x74/0xa0
 dsa_port_bridge_host_fdb_add+0xec/0x180
 dsa_slave_switchdev_event_work+0x7c/0x1c8
 process_one_work+0x290/0x568

What it's saying is that vlan_for_each() expects rtnl_lock() context and
it's not getting it, when it's called from the DSA master's ndo_set_rx_mode().

The caller of that - dsa_slave_set_rx_mode() - is the slave DSA
interface's dsa_port_bridge_host_fdb_add() which comes from the deferred
dsa_slave_switchdev_event_work().

We went to great lengths to avoid the rtnl_lock() context in that call
path in commit 0faf890fc519 ("net: dsa: drop rtnl_lock from
dsa_slave_switchdev_event_work"), and calling rtnl_lock() is simply not
an option due to the possibility of deadlocking when calling
dsa_flush_workqueue() from the call paths that do hold rtnl_lock() -
basically all of them.

So, when the DSA master calls vlan_for_each() from its ndo_set_rx_mode(),
the state of the 8021q driver on this device is really not protected
from concurrent access by anything.

Looking at net/8021q/, I don't think that vlan_info->vid_list was
particularly designed with RCU traversal in mind, so introducing an RCU
read-side form of vlan_for_each() - vlan_for_each_rcu() - won't be so
easy, and it also wouldn't be exactly what we need anyway.

In general I believe that the solution isn't in net/8021q/ anyway;
vlan_for_each() is not cut out for this task. DSA doesn't need rtnl_lock()
to be held per se - since it's not a netdev state change that we're
blocking, but rather, just concurrent additions/removals to a VLAN list.
We don't even need sleepable context - the callback of vlan_for_each()
just schedules deferred work.

The proposed escape is to remove the dependency on vlan_for_each() and
to open-code a non-sleepable, rtnl-free alternative to that, based on
copies of the VLAN list modified from .ndo_vlan_rx_add_vid() and
.ndo_vlan_rx_kill_vid().

Fixes: 64fdc5f341db ("net: dsa: sync unicast and multicast addresses for VLAN filters too")
Signed-off-by: Vladimir Oltean 
Link: https://lore.kernel.org/r/20230626154402.3154454-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski

net: dsa: introduce preferred_default_local_cpu_port and use on MT7530

2023-06-20T08:40:26+00:00

Since the introduction of the OF bindings, DSA has always had a policy that
in case multiple CPU ports are present in the device tree, the numerically
smallest one is always chosen.

The MT7530 switch family, except the switch on the MT7988 SoC, has 2 CPU
ports, 5 and 6, where port 6 is preferable on the MT7531BE switch because
it has higher bandwidth.

The MT7530 driver developers had 3 options:
- to modify DSA when the MT7531 switch support was introduced, such as to
  prefer the better port
- to declare both CPU ports in device trees as CPU ports, and live with the
  sub-optimal performance resulting from not preferring the better port
- to declare just port 6 in the device tree as a CPU port

Of course they chose the path of least resistance (3rd option), kicking the
can down the road. The hardware description in the device tree is supposed
to be stable - developers are not supposed to adopt the strategy of
piecemeal hardware description, where the device tree is updated in
lockstep with the features that the kernel currently supports.

Now, as a result of the fact that they did that, any attempts to modify the
device tree and describe both CPU ports as CPU ports would make DSA change
its default selection from port 6 to 5, effectively resulting in a
performance degradation visible to users with the MT7531BE switch as can be
seen below.

Without preferring port 6:

[ ID][Role] Interval           Transfer     Bitrate         Retr
[  5][TX-C]   0.00-20.00  sec   374 MBytes   157 Mbits/sec  734    sender
[  5][TX-C]   0.00-20.00  sec   373 MBytes   156 Mbits/sec    receiver
[  7][RX-C]   0.00-20.00  sec  1.81 GBytes   778 Mbits/sec    0    sender
[  7][RX-C]   0.00-20.00  sec  1.81 GBytes   777 Mbits/sec    receiver

With preferring port 6:

[ ID][Role] Interval           Transfer     Bitrate         Retr
[  5][TX-C]   0.00-20.00  sec  1.99 GBytes   856 Mbits/sec  273    sender
[  5][TX-C]   0.00-20.00  sec  1.99 GBytes   855 Mbits/sec    receiver
[  7][RX-C]   0.00-20.00  sec  1.72 GBytes   737 Mbits/sec   15    sender
[  7][RX-C]   0.00-20.00  sec  1.71 GBytes   736 Mbits/sec    receiver

Using one port for WAN and the other ports for LAN is a very popular use
case which is what this test emulates.

As such, this change proposes that we retroactively modify stable kernels
(which don't support the modification of the CPU port assignments, so as to
let user space fix the problem and restore the throughput) to keep the
mt7530 driver preferring port 6 even with device trees where the hardware
is more fully described.

Fixes: c288575f7810 ("net: dsa: mt7530: Add the support of MT7531 switch")
Signed-off-by: Vladimir Oltean 
Signed-off-by: Arınç ÜNAL 
Reviewed-by: Russell King (Oracle) 
Reviewed-by: Florian Fainelli 
Signed-off-by: David S. Miller

net: dsa: replace NETDEV_PRE_CHANGE_HWTSTAMP notifier with a stub

2023-04-09T14:35:49+00:00

There was a sort of rush surrounding commit 88c0a6b503b7 ("net: create a
netdev notifier for DSA to reject PTP on DSA master"), due to a desire
to convert DSA's attempt to deny TX timestamping on a DSA master to
something that doesn't block the kernel-wide API conversion from
ndo_eth_ioctl() to ndo_hwtstamp_set().

What was required was a mechanism that did not depend on ndo_eth_ioctl(),
and what was provided was a mechanism that did not depend on
ndo_eth_ioctl(), while at the same time introducing something that
wasn't absolutely necessary - a new netdev notifier.

There have been objections from Jakub Kicinski that using notifiers in
general when they are not absolutely necessary creates complications to
the control flow and difficulties to maintainers who look at the code.
So there is a desire to not use notifiers.

In addition to that, the notifier chain gets called even if there is no
DSA in the system and no one is interested in applying any restriction.

Take the model of udp_tunnel_nic_ops and introduce a stub mechanism,
through which net/core/dev_ioctl.c can call into DSA even when
CONFIG_NET_DSA=m.

Compared to the code that existed prior to the notifier conversion, aka
what was added in commits:
- 4cfab3566710 ("net: dsa: Add wrappers for overloaded ndo_ops")
- 3369afba1e46 ("net: Call into DSA netdevice_ops wrappers")

this is different because we are not overloading any struct
net_device_ops of the DSA master anymore, but rather, we are exposing a
rather specific functionality which is orthogonal to which API is used
to enable it - ndo_eth_ioctl() or ndo_hwtstamp_set().

Also, what is similar is that both approaches use function pointers to
get from built-in code to DSA.

There is no point in replicating the function pointers towards
__dsa_master_hwtstamp_validate() once for every CPU port (dev->dsa_ptr).
Instead, it is sufficient to introduce a singleton struct dsa_stubs,
built into the kernel, which contains a single function pointer to
__dsa_master_hwtstamp_validate().

I find this approach preferable to what we had originally, because
dev->dsa_ptr->netdev_ops->ndo_do_ioctl() used to require going through
struct dsa_port (dev->dsa_ptr), and so, this was incompatible with any
attempts to add any data encapsulation and hide DSA data structures from
the outside world.

Link: https://lore.kernel.org/netdev/20230403083019.120b72fd@kernel.org/
Signed-off-by: Vladimir Oltean 
Reviewed-by: Florian Fainelli 
Signed-off-by: David S. Miller