Florian Westphal says:
====================
netfilter: updates for net-next
The following patchset contains Netfilter updates for *net-next*:
1) Fix net-next-only use-after-free bug in nf_tables rbtree set:
Expired elements cannot be released right away after unlink anymore
because there is no guarantee that the binary-search blob is going to
be updated. Spotted by syzkaller.
2) Fix esoteric bug in nf_queue with udp fraglist gro, broken since
6.11. Patch 3 adds extends the nfqueue selftest for this.
4) Use dedicated slab for flowtable entries, currently the -512 cache
is used, which is wasteful. From Qingfang Deng.
5) Recent net-next update extended existing test for ip6ip6 tunnels, add
the required /config entry. Test still passed by accident because the
previous tests network setup gets re-used, so also update the test so
it will fail in case the ip6ip6 tunnel interface cannot be added.
6) Fix 'nft get element mytable myset { 1.2.3.4 }' on big endian
platforms, this was broken since code was added in v5.1.
7) Fix nf_tables counter reset support on 32bit platforms, where counter
reset may cause huge values to appear due to wraparound.
Broken since reset feature was added in v6.11. From Anders Grahn.
8-11) update nf_tables rbtree set type to detect partial
operlaps. This will eventually speed up nftables userspace: at this
time userspace does a netlink dump of the set content which slows down
incremental updates on interval sets. From Pablo Neira Ayuso.
* tag 'nf-next-26-02-06' of https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next:
netfilter: nft_set_rbtree: validate open interval overlap
netfilter: nft_set_rbtree: validate element belonging to interval
netfilter: nft_set_rbtree: check for partial overlaps in anonymous sets
netfilter: nft_set_rbtree: fix bogus EEXIST with NLM_F_CREATE with null interval
netfilter: nft_counter: fix reset of counters on 32bit archs
netfilter: nft_set_hash: fix get operation on big endian
selftests: netfilter: add IPV6_TUNNEL to config
netfilter: flowtable: dedicated slab for flow entry
selftests: netfilter: nft_queue.sh: add udp fraglist gro test case
netfilter: nfnetlink_queue: do shared-unconfirmed check before segmentation
netfilter: nft_set_rbtree: don't gc elements on insert
====================
Link: https://patch.msgid.link/20260206153048.17570-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The mentioned struct has an hole and uses unnecessary wide type to
store MAC length and indexes of very small arrays.
It's also embedded into the skb_extensions, and the latter, due
to recent CAN changes, may exceeds the 192 bytes mark (3 cachelines
on x86_64 arch) on some reasonable configurations.
Reordering and the sec_path fields, shrinking xfrm_offload.orig_mac_len
to 16 bits and xfrm_offload.{len,olen,verified_cnt} to u8, we can save
16 bytes and keep skb_extensions size under control.
Before:
struct sec_path {
int len;
int olen;
int verified_cnt;
/* XXX 4 bytes hole, try to pack */$
struct xfrm_state * xvec[6];
struct xfrm_offload ovec[1];
/* size: 88, cachelines: 2, members: 5 */
/* sum members: 84, holes: 1, sum holes: 4 */
/* last cacheline: 24 bytes */
};
After:
struct sec_path {
struct xfrm_state * xvec[6];
struct xfrm_offload ovec[1];
/* typedef u8 -> __u8 */ unsigned char len;
/* typedef u8 -> __u8 */ unsigned char olen;
/* typedef u8 -> __u8 */ unsigned char verified_cnt;
/* size: 72, cachelines: 2, members: 5 */
/* padding: 1 */
/* last cacheline: 8 bytes */
};
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Steffen Klassert <steffen.klassert@secunet.com>
Link: https://patch.msgid.link/83846bd2e3fa08899bd0162e41bfadfec95e82ef.1770398071.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Michael Chan says:
====================
bnxt_en: Add RSS context resource check
Add missing logic to check that we have enough RSS contexts. This
will make the recent change to increase the use of RSS contexts for
a larger RSS indirection table more complete.
====================
Link: https://patch.msgid.link/20260207235118.1987301-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
bnxt_need_reserve_rings() checks 6 ring resources against the reserved
values to determine if a new reservation is needed. Factor out the code
to collect the total resources into a new helper function
bnxt_get_total_resources() to make the code cleaner and easier to read.
Instead of individual scalar variables, use the struct bnxt_hw_rings to
hold all the ring resources. Using the struct, hwr.cp replaces the nq
variable and the chip specific hwr.cp_p5 replaces cp on newer chips.
There is no change in behavior. This will make it easier to check the
RSS context resource in the next patch.
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Joe Damato <joe@dama.to>
Link: https://patch.msgid.link/20260207235118.1987301-2-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Future extensions with psock will override their own sk->sk_write_space
callback. This patch ensures that the overridden sk_write_space can be
invoked by MPTCP.
INDIRECT_CALL is used to keep the default path optimised.
Note that sk->sk_write_space was never called directly with MPTCP
sockets, so changing it to sk_stream_write_space in the init, and using
it from mptcp_write_space() is not supposed to change the current
behaviour.
This patch is shared early to ease discussions around future RFC and
avoid confusions with this "fix" that is needed for different future
extensions.
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Co-developed-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Gang Yan <yangang@kylinos.cn>
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260206-net-next-mptcp-write_space-override-v2-1-e0b12be818c6@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Breno Leitao says:
====================
net: netconsole: convert to NBCON console infrastructure
This series adds support for the nbcon (new buffer console) infrastructure
to netconsole, enabling lock-free, priority-based console operations that
are safer in crash scenarios.
The implementation is introduced in three steps:
0) Extend printk to expose CPU and taskname (task->comm) where the
printk originated from. (Thanks John and Petr for the support in
getting this done)
1) Refactor the message fragmentation logic into a reusable helper function
2) Extend nbcon support to non-extended (basic) consoles using the same
infrastructure.
The initial discussion about it appeared a while ago in [1], in order to
solve Mike's HARDIRQ-safe -> HARDIRQ-unsafe lock order warning, and the root
cause is that some hosts were calling IRQ unsafe locks from inside console
lock.
At that time, we didn't have the CON_NBCON_ATOMIC_UNSAFE yet. John
kindly implemented CON_NBCON_ATOMIC_UNSAFE in 187de7c212 ("printk:
nbcon: Allow unsafe write_atomic() for panic"), and now we can
implement netconsole on top of nbcon.
Important to note that netconsole continues to call netpoll and the
network TX helpers with interrupt disable, given the TX are called with
target_list_lock.
====================
Link: https://patch.msgid.link/20260206-nbcon-v7-0-62bda69b1b41@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Use the CPU and task name captured at printk() time from
nbcon_write_context instead of querying the current execution context.
This provides accurate information about where the message originated,
rather than where netconsole happens to be running.
For CPU, use wctxt->cpu instead of raw_smp_processor_id().
For taskname, use wctxt->comm directly which contains the task
name captured at printk time.
This change ensures netconsole outputs reflect the actual context that
generated the log message, which is especially important when the
console driver runs asynchronously in a dedicated thread.
Reviewed-by: John Ogness <john.ogness@linutronix.de>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://patch.msgid.link/20260206-nbcon-v7-4-62bda69b1b41@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Convert netconsole from the legacy console API to the NBCON framework.
NBCON provides threaded printing which unblocks printk()s and flushes in
a thread, decoupling network TX from printk() when netconsole is
in use.
Since netconsole relies on the network stack which cannot safely operate
from all atomic contexts, mark both consoles with
CON_NBCON_ATOMIC_UNSAFE. (See discussion in [1])
CON_NBCON_ATOMIC_UNSAFE restricts write_atomic() usage to emergency
scenarios (panic) where regular messages are sent in threaded mode.
Implementation changes:
- Unify write_ext_msg() and write_msg() into netconsole_write()
- Add device_lock/device_unlock callbacks to manage target_list_lock
- Use nbcon_enter_unsafe()/nbcon_exit_unsafe() around network
operations.
- If nbcon_enter_unsafe() fails, just return given netconsole lost
the ownership of the console.
- Set write_thread and write_atomic callbacks (both use same function)
Link: https://lore.kernel.org/all/b2qps3uywhmjaym4mht2wpxul4yqtuuayeoq4iv4k3zf5wdgh3@tocu6c7mj4lt/ [1]
Reviewed-by: John Ogness <john.ogness@linutronix.de>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://patch.msgid.link/20260206-nbcon-v7-3-62bda69b1b41@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Extract the message fragmentation logic from write_msg() into a
dedicated send_msg_udp() function. This improves code readability
and prepares for future enhancements.
The new send_msg_udp() function handles splitting messages that
exceed MAX_PRINT_CHUNK into smaller fragments and sending them
sequentially. This function is placed before send_ext_msg_udp()
to maintain a logical ordering of related functions.
No functional changes - this is purely a refactoring commit.
Reviewed-by: Petr Mladek <pmladek@suse.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: John Ogness <john.ogness@linutronix.de>
Signed-off-by: Breno Leitao <leitao@debian.org>
Link: https://patch.msgid.link/20260206-nbcon-v7-2-62bda69b1b41@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Extend struct printk_info to include the task name, pid, and CPU
number where printk messages originate. This information is captured
at vprintk_store() time and propagated through printk_message to
nbcon_write_context, making it available to nbcon console drivers.
This is useful for consoles like netconsole that want to include
execution context in their output, allowing correlation of messages
with specific tasks and CPUs regardless of where the console driver
actually runs.
The feature is controlled by CONFIG_PRINTK_EXECUTION_CTX, which is
automatically selected by CONFIG_NETCONSOLE_DYNAMIC. When disabled,
the helper functions compile to no-ops with no overhead.
Suggested-by: John Ogness <john.ogness@linutronix.de>
Signed-off-by: Petr Mladek <pmladek@suse.com>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: John Ogness <john.ogness@linutronix.de>
Link: https://patch.msgid.link/20260206-nbcon-v7-1-62bda69b1b41@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
David Yang is saying that struct flow_action_entry in
include/net/flow_offload.h has gained new fields and DSA's struct
dsa_mall_policer_tc_entry, derived from that, isn't keeping up.
This structure is passed to drivers and they are completely oblivious to
the values of fields they don't see.
This has happened before, and almost always the solution was to make the
DSA layer thinner and use the upstream data structures. Here, the reason
why we didn't do that is because struct flow_action_entry :: police is
an anonymous structure.
That is easily enough fixable, just name those fields "struct
flow_action_police" and reference them from DSA.
Make the according transformations to the two users (sja1105 and felix):
"rate_bytes_per_sec" -> "rate_bytes_ps".
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Co-developed-by: David Yang <mmyangfl@gmail.com>
Signed-off-by: David Yang <mmyangfl@gmail.com>
Link: https://patch.msgid.link/20260206075427.44733-1-mmyangfl@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Jacky Chou says:
====================
net: ftgmac100: Various probe cleanups
The probe function of the ftgmac100 is rather complex, due to the way
it has evolved over time, dealing with poor DT descriptions, and new
variants of the MAC.
Make use of DT match data to identify the MAC variant, rather than
looking at the compatible string all the time.
Make use of devm_ calls to simplify cleanup. This indirectly fixes
inconsistent goto label names.
Always probe the MDIO bus, when it exists. This simplifies the logic a
bit.
Move code into helpers to simply probe.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jacky Chou <jacky_chou@aspeedtech.com>
====================
Link: https://patch.msgid.link/20260206-ftgmac-cleanup-v5-0-ad28a9067ea7@aspeedtech.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Both the Aspeed 2400 and 2500 and the original faraday version of the
MAC have MDIO bus controllers as part of the MAC. Since it exists,
always registering it makes the code simpler, and causes no harm. If
there is no mdio node in device tree, of_mdiobus_register() will fall
back to mdiobus_register(), making it safe.
AST2600 uses an external MDIO controller and does not have an embedded
MDIO bus in the MAC. For such configurations, the legacy MII probe path
must not be entered without a registered mii_bus.
Add an explicit check to fail gracefully when no MDIO bus is present,
preventing a NULL pointer dereference while keeping the intended
behavior for platforms without embedded MDIO.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Jacky Chou <jacky_chou@aspeedtech.com>
Link: https://patch.msgid.link/20260206-ftgmac-cleanup-v5-9-ad28a9067ea7@aspeedtech.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Felix Maurer says:
====================
hsr: Implement more robust duplicate discard algorithm
The duplicate discard algorithms for PRP and HSR do not work reliably
with certain link faults. Especially with packet loss on one link, the
duplicate discard algorithms drop valid packets. For a more thorough
description see patches 4 (for PRP) and 6 (for HSR).
This patchset replaces the current algorithms (based on a drop window
for PRP and highest seen sequence number for HSR) with a single new one
that tracks the received sequence numbers individually (descriptions
again in patches 4 and 6).
The changes will lead to higher memory usage and more work to do for
each packet. But I argue that this is an acceptable trade-off to make
for a more robust PRP and HSR behavior with faulty links. After all,
both protocols are to be used in environments where redundancy is needed
and people are willing to setup special network topologies to achieve
that.
Some more reasoning on the overhead and expected scale of the deployment
from the RFC discussion:
> As for the expected scale, there are two dimensions: the number of nodes
> in the network and the data rate with which they send.
>
> The number of nodes in the network affect the memory usage because each
> node now has the block buffer. For PRP that's 64 blocks * 32 byte =
> 2kbyte for each node in the node table. A PRP network doesn't have an
> explicit limit for the number of nodes. However, the whole network is a
> single layer-2 segment which shouldn't grow too large anyways. Even if
> one really tries to put 1000 nodes into the PRP network, the memory
> overhead (2Mbyte) is acceptable in my opinion.
>
> For HSR, the blocks would be larger because we need to track the
> sequence numbers per port. I expect 64 blocks * 80 byte = 5kbyte per
> node in the node table. There is no explicit limit for the size of an
> HSR ring either. But I expect them to be of limited size because the
> forwarding delays add up throughout the ring. I've seen vendors limiting
> the ring size to 50 nodes with 100Mbit/s links and 300 with 1Gbit/s
> links. In both cases I consider the memory overhead acceptable.
>
> The data rates are harder to reason about. In general, the data rates
> for HSR and PRP are limited because too high packet rates would lead to
> very fast re-use of the 16bit sequence numbers. The IEC 62439-3:2021
> mentions 100Mbit/s links and 1Gbit/s links. I don't expect HSR or PRP
> networks to scale out to, e.g., 10Gbit/s links with the current
> specification as this would mean that sequence numbers could repeat as
> often as every ~4ms. The default constants in the IEC standard, which we
> also use, are oriented at a 100Mbit/s network.
>
> In my tests with veth pairs, the CPU overhead didn't lead to
> significantly lower data rates. The main factor limiting the data rate
> at the moment, I assume, is the per-node spinlock that is taken for each
> received packet. IMHO, there is a lot more to gain in terms of CPU
> overhead from making this lock smaller or getting rid of it, than we
> loose with the more accurate duplicate discard algorithm in this patchset.
>
> The CPU overhead of the algorithm benefits from the fact that in high
> packet rate scenarios (where it really matters) many packets will have
> sequence numbers in already initialized blocks. These packets just have
> additionally: one xarray lookup, one comparison, and one bit setting. If
> a block needs to be initialized (once every 128 packets plus their 128
> duplicates if all sequence numbers are seen), we will have: one
> xa_erase, a bunch of memory writes, and one xa_store.
>
> In theory, all packets could end up in the slow path if a node sends
> every 128th packet to us. If this is sent from a well behaving node, the
> packet rate wouldn't be an issue anymore, though.
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
====================
Link: https://patch.msgid.link/cover.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Run the packet loss and reordering tests also for both HSR versions. Now
they can be removed from the hsr_ping tests completely. The timeout needs
to be increased because there are 15 link fault test cases now, with each
of them taking 5-6sec for the test and at most 5sec for the HSR node tables
to get merged and we also want some room to make the test runs stable.
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://patch.msgid.link/eb6f667d3804ce63d86f0ee3fbc0e0ac9e1a209a.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The HSR duplicate discard algorithm had even more basic problems than the
described for PRP in the previous patch. It relied only on the last
received sequence number to decide if a new frame should be forwarded to
any port. This does not work correctly in any case where frames are
received out of order. The linked bug report claims that this can even
happen with perfectly fine links due to the order in which incoming frames
are processed (which can be unexpected on multi-core systems). The issue
also occasionally shows up in the HSR selftests. The main reason is that
the sequence number that was last forwarded to the master port may have
skipped a number which will in turn never be delivered to the host.
As the problem (we accidentally skip over a sequence number that has not
been received but will be received in the future) is similar to PRP, we can
apply a similar solution. The duplicate discard algorithm based on the
"sparse bitmap" works well for HSR if it is extended to track one bitmap
for each port (A, B, master, interlink). To do this, change the sequence
number blocks to contain a flexible array member as the last member that
can keep chunks for as many bitmaps as we need. This design makes it easy
to reuse the same algorithm in a potential PRP RedBox implementation.
The duplicate discard algorithm functions are modified to deal with
sequence number blocks of different sizes and to correctly use the array of
bitmap chunks. There is a notable speciality for HSR: the port type has a
special port type NONE with value 0. This leads to the number of port types
being 5 instead of actually 4. To save memory, remove the NONE port from
the bitmap (by subtracting 1) when setting up the block buffer and when
accessing the bitmap chunks in the array.
Removing the old algorithm allows us to get rid of a few fields that are
not needed any more: time_out and seq_out for each port. We can also remove
some functions that were only necessary for the previous duplicate discard
algorithm.
The removal of seq_out is possible despite its previous usage in
hsr_register_frame_in: it was used to prevent updates to time_in when
"invalid" sequence numbers were received. With the new duplicate discard
algorithm, time_in has no relevance for the expiry of sequence numbers
anymore. They will expire based on the timestamps in the sequence number
blocks after at most 400ms. There is no need that a node "re-registers" to
"resume communication": after 400ms, all sequence numbers are accepted
again. Also, according to the IEC 62439-3:2021, all nodes are supposed to
send no traffic for 500ms after boot to lead exactly to this expiry of seen
sequence numbers. time_in is still used for pruning nodes from the node
table after no traffic has been received for 60sec. Pruning is only needed
if the node is really gone and has not been sending any traffic for that
period.
seq_out was also used to report the last incoming sequence number from a
node through netlink. I am not sure how useful this value is to userspace
at all, but added getting it from the sequence number blocks. This number
can be outdated after node merging until a new block has been added.
Update the KUnit test for the PRP duplicate discard so that the node
allocation matches and expectations on the removed fields are removed.
Reported-by: Yoann Congal <yoann.congal@smile.fr>
Closes: https://lore.kernel.org/netdev/7d221a07-8358-4c0b-a09c-3b029c052245@smile.fr/
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://patch.msgid.link/36dc3bc5bdb7e68b70bb5ef86f53ca95a3f35418.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Add tests where one link has different rates of packet loss or reorders
packets. PRP should still be able to recover from these link faults and
show no packet loss. However, it is acceptable to receive some level of
duplicate packets. This matches the current specification (IEC
62439-3:2021) of the duplicate discard algorithm that requires it to be
"designed such that it never rejects a legitimate frame, while occasional
acceptance of a duplicate can be tolerated." The rate of acceptable
duplicates in this test is intentionally high (10%) to make the test
stable, the values I observed in the worst test cases (20% loss) are around
5% duplicates.
The duplicates occur because of the 10ms ping interval in the test. As
blocks expire after 400ms based on the timestamp of the first received
sequence number in the block, every approx. 40th will lead to a new, clean
block being used where the sequence number hasn't been seen before. As this
occurs on both nodes in the test (for requests and replies), we observe
around 20 duplicate frames.
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://patch.msgid.link/7b36506d3a80e53786fe56526cf6046c74dfeee1.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The PRP duplicate discard algorithm does not work reliably with certain
link faults. Especially with packet loss on one link, the duplicate discard
algorithm drops valid packets which leads to packet loss on the PRP
interface where the link fault should in theory be perfectly recoverable by
PRP. This happens because the algorithm opens the drop window on the lossy
link, covering received and lost sequence numbers. If the other, non-lossy
link receives the duplicate for a lost frame, it is within the drop window
of the lossy link and therefore dropped.
Since IEC 62439-3:2012, a node has one sequence number counter for frames
it sends, instead of one sequence number counter for each destination.
Therefore, a node can not expect to receive contiguous sequence numbers
from a sender. A missing sequence number can be totally normal (if the
sender intermittently communicates with another node) or mean a frame was
lost.
The algorithm, as previously implemented in commit 05fd00e5e7 ("net: hsr:
Fix PRP duplicate detection"), was part of IEC 62439-3:2010 (HSRv0/PRPv0)
but was removed with IEC 62439-3:2012 (HSRv1/PRPv1). Since that, no
algorithm is specified but up to implementers. It should be "designed such
that it never rejects a legitimate frame, while occasional acceptance of a
duplicate can be tolerated" (IEC 62439-3:2021).
For the duplicate discard algorithm, this means that 1) we need to track
the sequence numbers individually to account for non-contiguous sequence
numbers, and 2) we should always err on the side of accepting a duplicate
than dropping a valid frame.
The idea of the new algorithm is to store the seen sequence numbers in a
bitmap. To keep the size of the bitmap in control, we store it as a "sparse
bitmap" where the bitmap is split into blocks and not all blocks exist at
the same time. The sparse bitmap is implemented using an xarray that keeps
the references to the individual blocks and a backing ring buffer that
stores the actual blocks. New blocks are initialized in the buffer and
added to the xarray as needed when new frames arrive. Existing blocks are
removed in two conditions:
1. The block found for an arriving sequence number is old and therefore not
relevant to the duplicate discard algorithm anymore, i.e., it has been
added more than the entry forget time ago. In this case, the block is
removed from the xarray and marked as forgotten (by setting its
timestamp to 0).
2. Space is needed in the ring buffer for a new block. In this case, the
block is removed from the xarray, if it hasn't already been forgotten
(by 1.). Afterwards, the new block is initialized in its place.
This has the nice property that we can reliably track sequence numbers on
low traffic situations (where they expire based on their timestamp) and
more quickly forget sequence numbers in high traffic situations before they
potentially wrap over and repeat before they are expired.
When nodes are merged, the blocks are merged as well. The timestamp of a
merged block is set to the minimum of the two timestamps to never keep
around a seen sequence number for too long. The bitmaps are or'd to mark
all seen sequence numbers as seen.
All of this still happens under seq_out_lock, to prevent concurrent
access to the blocks.
The KUnit test for the algorithm is updated as well. The updates are done
in a way to match the original intends pretty closely. Currently, there is
much knowledge about the actual algorithm baked into the tests (especially
the expectations) which may need some redesign in the future.
Reported-by: Steffen Lindner <steffen.lindner@de.abb.com>
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Tested-by: Steffen Lindner <steffen.lindner@de.abb.com>
Link: https://patch.msgid.link/8ce15a996099df2df5b700969a39e7df400e8dbb.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Add a test case that can support different types of faulty links for all
protocol versions (HSRv0, HSRv1, PRPv1). It starts with a baseline with
fully functional links. The first faulty case is one link being cut during
the ping. This test uses a different function for ping that sends more
packets in shorter intervals to stress the duplicate detection algorithms a
bit more and allow for future tests with other link faults (packet loss,
reordering, etc.).
As the link fault tests now cover the cut link for HSR and PRP, it can be
removed from the hsr_ping test. Note that the removed cut link test did not
really test the fault because do_ping_long takes about 1sec while the link
is only cut after a 3sec sleep.
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://patch.msgid.link/dad52276e2c349ecb96168bef7e3001bf7becc81.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Previously the hsr_ping test only checked that all nodes in a VLAN are
reachable (using do_ping). Update the test to also check that there is no
packet loss and no duplicate packets by running the same tests for VLANs as
without VLANs (including using do_ping_long). This also adds tests for IPv6
over VLAN. To unify the test code, the topology without VLANs now uses IP
addresses from dead:beef:0::/64 to align with the 100.64.0.0/24 range for
IPv4. Error messages are updated across the board to make it easier to find
what actually failed.
Also update the VLAN test to only run in VLAN 2, as there is no need to
check if ping really works with VLAN IDs 2, 3, 4, and 5. This lowers the
number of long ping tests on VLANs to keep the overall test runtime in
bounds.
It's still necessary to bump the test timeout a bit, though: a ping long
tests takes 1sec, do_ping_tests performs 12 of them, do_link_problem_tests
6, and the VLAN tests again 12. With some buffer for setup and waiting and
for two protocol versions, 90sec timeout seems reasonable.
Signed-off-by: Felix Maurer <fmaurer@redhat.com>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://patch.msgid.link/e3ded0e2547b5f720524b62fabeb96debc579697.1770299429.git.fmaurer@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Wei Fang says:
====================
net: fec: improve XDP copy mode and add AF_XDP zero-copy support
This patch set optimizes the XDP copy mode logic as follows.
1. Separate the processing of RX XDP frames from fec_enet_rx_queue(),
and adds a separate function fec_enet_rx_queue_xdp() for handling XDP
frames.
2. For TX XDP packets, using the batch sending method to avoid frequent
MMIO writes.
3. Use the switch statement to check the tx_buf type instead of the
if...else... statement, making the cleanup logic of TX BD ring cleared
and more efficient.
We compared the performance of XDP copy mode before and after applying
this patch set, and the results show that the performance has improved.
Before applying this patch set.
root@imx93evk:~# ./xdp-bench tx eth0
Summary 396,868 rx/s 0 err,drop/s
Summary 396,024 rx/s 0 err,drop/s
root@imx93evk:~# ./xdp-bench drop eth0
Summary 684,781 rx/s 0 err/s
Summary 675,746 rx/s 0 err/s
root@imx93evk:~# ./xdp-bench pass eth0
Summary 208,552 rx/s 0 err,drop/s
Summary 208,654 rx/s 0 err,drop/s
root@imx93evk:~# ./xdp-bench redirect eth0 eth0
eth0->eth0 311,210 rx/s 0 err,drop/s 311,208 xmit/s
eth0->eth0 310,808 rx/s 0 err,drop/s 310,809 xmit/s
After applying this patch set.
root@imx93evk:~# ./xdp-bench tx eth0
Summary 425,778 rx/s 0 err,drop/s
Summary 426,042 rx/s 0 err,drop/s
root@imx93evk:~# ./xdp-bench drop eth0
Summary 698,351 rx/s 0 err/s
Summary 701,882 rx/s 0 err/s
root@imx93evk:~# ./xdp-bench pass eth0
Summary 210,348 rx/s 0 err,drop/s
Summary 210,016 rx/s 0 err,drop/s
root@imx93evk:~# ./xdp-bench redirect eth0 eth0
eth0->eth0 354,407 rx/s 0 err,drop/s 354,401 xmit/s
eth0->eth0 350,381 rx/s 0 err,drop/s 350,389 xmit/s
This patch set also addes the AF_XDP zero-copy support, and we tested
the performance on i.MX93 platform with xdpsock tool. The following is
the performance comparison of copy mode and zero-copy mode. It can be
seen that the performance of zero-copy mode is better than that of copy
mode.
1. MAC swap L2 forwarding
1.1 Zero-copy mode
root@imx93evk:~# ./xdpsock -i eth0 -l -z
sock0@eth0:0 l2fwd xdp-drv
pps pkts 1.00
rx 414715 415455
tx 414715 415455
1.2 Copy mode
root@imx93evk:~# ./xdpsock -i eth0 -l -c
sock0@eth0:0 l2fwd xdp-drv
pps pkts 1.00
rx 356396 356609
tx 356396 356609
2. TX only
2.1 Zero-copy mode
root@imx93evk:~# ./xdpsock -i eth0 -t -s 64 -z
sock0@eth0:0 txonly xdp-drv
pps pkts 1.00
rx 0 0
tx 1119573 1126720
2.2 Copy mode
root@imx93evk:~# ./xdpsock -i eth0 -t -s 64 -c
sock0@eth0:0 txonly xdp-drv
pps pkts 1.00
rx 0 0
tx 406864 407616
====================
Link: https://patch.msgid.link/20260205085742.2685134-1-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
This patch adds AF_XDP zero-copy support for both TX and RX on the FEC
driver. It introduces new functions for XSK buffer allocation, RX/TX
queue processing in zero-copy mode, and XSK pool setup/teardown.
For RX, fec_alloc_rxq_buffers_zc() is added to allocate RX buffers from
XSK pool. And fec_enet_rx_queue_xsk() is used to process the frames from
the RX queue which is bound to the AF_XDP socket. Similar to the copy
mode, the zero-copy mode also supports XDP_TX, XDP_PASS, XDP_DROP and
XDP_REDIRECT actions. In addition, fec_enet_xsk_tx_xmit() is similar to
fec_enet_xdp_tx_xmit() and is used to handle XDP_TX action in zero-copy
mode.
For TX, there are two cases, one is the frames from the AF_XDP socket,
so fec_enet_xsk_xmit() is added to directly transmit the frames from
the socket and the buffer type is marked as FEC_TXBUF_T_XSK_XMIT. The
other one is the frames from the RX queue (XDP_TX action), the buffer
type is marked as FEC_TXBUF_T_XSK_TX. Therefore, fec_enet_tx_queue()
could correctly clean the TX queue base on the buffer type.
Also, some tests have been done on the i.MX93-EVK board with the xdpsock
tool, the following are the results.
Env: i.MX93 connects to a packet generator, the link speed is 1Gbps, and
flow-control is off. The RX packet size is 64 bytes including FCS. Only
one RX queue (CPU) is used to receive frames.
1. MAC swap L2 forwarding
1.1 Zero-copy mode
root@imx93evk:~# ./xdpsock -i eth0 -l -z
sock0@eth0:0 l2fwd xdp-drv
pps pkts 1.00
rx 414715 415455
tx 414715 415455
1.2 Copy mode
root@imx93evk:~# ./xdpsock -i eth0 -l -c
sock0@eth0:0 l2fwd xdp-drv
pps pkts 1.00
rx 356396 356609
tx 356396 356609
2. TX only
2.1 Zero-copy mode
root@imx93evk:~# ./xdpsock -i eth0 -t -s 64 -z
sock0@eth0:0 txonly xdp-drv
pps pkts 1.00
rx 0 0
tx 1119573 1126720
2.2 Copy mode
root@imx93evk:~# ./xdpsock -i eth0 -t -s 64 -c
sock0@eth0:0 txonly xdp-drv
pps pkts 1.00
rx 0 0
tx 406864 407616
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Link: https://patch.msgid.link/20260205085742.2685134-16-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
To support AF_XDP zero-copy mode in the subsequent patch, the following
adjustments have been made to fec_tx_queue().
1. Change the parameters of fec_tx_queue().
2. Some variables are initialized at the time of declaration, and the
order of local variables is updated to follow the reverse xmas tree
style.
3. Remove the variable xdpf and add the variable tx_buf.
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260205085742.2685134-15-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Currently, the buffers of RX queue are allocated from the page pool. In
the subsequent patches to support XDP zero copy, the RX buffers will be
allocated from the UMEM. Therefore, extract fec_alloc_rxq_buffers_pp()
from fec_enet_alloc_rxq_buffers() and we will add another helper to
allocate RX buffers from UMEM for the XDP zero copy mode. In addition,
fec_alloc_rxq_buffers_pp() only initializes bdp->bufaddr and does not
initialize other fields of bdp, because these will be initialized in
fec_enet_bd_init().
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Link: https://patch.msgid.link/20260205085742.2685134-14-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The tx_buf has three types: FEC_TXBUF_T_SKB, FEC_TXBUF_T_XDP_NDO and
FEC_TXBUF_T_XDP_TX. Currently, the driver uses 'if...else...' statements
to check the type and perform the corresponding processing. This is very
detrimental to future expansion. To support AF_XDP zero-copy mode, two
new types will be added in the future, continuing to use 'if...else...'
would be a very bad coding style. So the 'if...else...' statements in
the current driver are replaced with switch statements.
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Link: https://patch.msgid.link/20260205085742.2685134-11-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The tx_buf pointer will not NULL when its type is FEC_TXBUF_T_XDP_NDO or
FEC_TXBUF_T_XDP_TX. If the type is FEC_TXBUF_T_SKB, dev_kfree_skb_any()
will do NULL pointer check. So it is unnecessary to do NULL pointer check
in fec_enet_bd_init() and fec_enet_tx_queue().
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260205085742.2685134-10-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
