| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Benefit from the recent conversion of the driver to NAPI and enable GRO
support through the use of napi_gro_receive(). Pass the NAPI pointer
from the bus driver (mlxsw_pci) to the switch driver (mlxsw_spectrum)
through the skb control block where various packet metadata is already
encoded.
The main motivation is to improve forwarding performance through the use
of GRO fraglist [1]. In my testing, when the forwarding data path is
simple (routing between two ports) there is not much difference in
forwarding performance between GRO disabled and GRO enabled with
fraglist.
The improvement becomes more noticeable as the data path becomes more
complex since it is traversed less times with GRO enabled. For example,
with 10 ingress and 10 egress flower filters with different priorities
on the two ports between which routing is performed, there is an
improvement of about 140% in forwarded bandwidth.
[1] https://lore.kernel.org/netdev/20200125102645.4782-1-steffen.klassert@secunet.com/
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Amit Cohen <amcohen@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://patch.msgid.link/21258fe55f608ccf1ee2783a5a4534220af28903.1734354812.git.petrm@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Eric Dumazet says:
====================
inetpeer: reduce false sharing and atomic operations
After commit 8c2bd38b95f7 ("icmp: change the order of rate limits"),
there is a risk that a host receiving packets from an unique
source targeting closed ports is using a common inet_peer structure
from many cpus.
All these cpus have to acquire/release a refcount and update
the inet_peer timestamp (p->dtime)
Switch to pure RCU to avoid changing the refcount, and update
p->dtime only once per jiffy.
Tested:
DUT : 128 cores, 32 hw rx queues.
receiving 8,400,000 UDP packets per second, targeting closed ports.
Before the series:
- napi poll can not keep up, NIC drops 1,200,000 packets
per second.
- We use 20 % of cpu cycles
After this series:
- All packets are received (no more hw drops)
- We use 12 % of cpu cycles.
v1: https://lore.kernel.org/20241213130212.1783302-1-edumazet@google.com
====================
Link: https://patch.msgid.link/20241215175629.1248773-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
All inet_getpeer() callers except ip4_frag_init() don't need
to acquire a permanent refcount on the inetpeer.
They can switch to full RCU protection.
Move the refcount_inc_not_zero() into ip4_frag_init(),
so that all the other callers no longer have to
perform a pair of expensive atomic operations on
a possibly contended cache line.
inet_putpeer() no longer needs to be exported.
After this patch, my DUT can receive 8,400,000 UDP packets
per second targeting closed ports, using 50% less cpu cycles
than before.
Also change two calls to l3mdev_master_ifindex() by
l3mdev_master_ifindex_rcu() (Ido ideas)
Fixes: 8c2bd38b95f7 ("icmp: change the order of rate limits")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241215175629.1248773-5-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
inet_putpeer() will be removed in the following patch,
because we will no longer use refcounts.
Update inetpeer timestamp (p->dtime) at lookup time.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241215175629.1248773-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
All callers of inet_getpeer() want to create an inetpeer.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241215175629.1248773-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
All callers of inet_getpeer_v4() and inet_getpeer_v6()
want to create an inetpeer.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241215175629.1248773-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Thomas Weißschuh says:
====================
net: constify 'struct bin_attribute'
The sysfs core now allows instances of 'struct bin_attribute' to be
moved into read-only memory. Make use of that to protect them against
accidental or malicious modifications.
====================
Link: https://patch.msgid.link/20241216-sysfs-const-bin_attr-net-v1-0-ec460b91f274@weissschuh.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The sysfs core now allows instances of 'struct bin_attribute' to be
moved into read-only memory. Make use of that to protect them against
accidental or malicious modifications.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20241216-sysfs-const-bin_attr-net-v1-4-ec460b91f274@weissschuh.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The sysfs core now allows instances of 'struct bin_attribute' to be
moved into read-only memory. Make use of that to protect them against
accidental or malicious modifications.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20241216-sysfs-const-bin_attr-net-v1-2-ec460b91f274@weissschuh.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The sysfs core now allows instances of 'struct bin_attribute' to be
moved into read-only memory. Make use of that to protect them against
accidental or malicious modifications.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Reviewed-by: Simon Horman <horms@kernel.org>
Acked-by: Nikolay Aleksandrov <razor@blackwall.org>
Link: https://patch.msgid.link/20241216-sysfs-const-bin_attr-net-v1-1-ec460b91f274@weissschuh.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
page_pool_is_last_ref() releases a reference while the name,
to me at least, suggests it just checks if the refcount is 1.
The semantics of the function are the same as those of
atomic_dec_and_test() and refcount_dec_and_test(), so just
use the _and_test() suffix.
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://patch.msgid.link/20241215212938.99210-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Similar to the use of $crate::Module, ThisModule should be referred to as
$crate::ThisModule in the macro evaluation. The reason the macro previously
did not cause any errors is because all the users of the macro would use
kernel::prelude::*, bringing ThisModule into scope.
Signed-off-by: Rahul Rameshbabu <sergeantsagara@protonmail.com>
Reviewed-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Link: https://patch.msgid.link/20241214194242.19505-1-sergeantsagara@protonmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Now that we have generic QDISC_CONGESTED and QDISC_OVERLIMIT drop
reasons, let's have all the qdiscs that contain an AQM apply them
consistently when dropping packets.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://patch.msgid.link/20241214-fq-codel-drop-reasons-v1-1-2a814e884c37@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Kuniyuki Iwashima says:
====================
af_unix: Prepare for skb drop reason.
This is a prep series and cleans up error paths in the following
functions
* unix_stream_connect()
* unix_stream_sendmsg()
* unix_dgram_sendmsg()
to make it easy to add skb drop reason for AF_UNIX, which seems to
have a potential user.
https://lore.kernel.org/netdev/CAAf2ycmZHti95WaBR3s+L5Epm1q7sXmvZ-EqCK=-oZj=45tOwQ@mail.gmail.com/
v1: https://lore.kernel.org/netdev/20241206052607.1197-1-kuniyu@amazon.com/
====================
Link: https://patch.msgid.link/20241213110850.25453-1-kuniyu@amazon.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
unix_our_peer() is used only in unix_may_send().
Let's inline it in unix_may_send().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The error path is complicated in unix_dgram_sendmsg() because there
are two timings when other could be non-NULL: when it's fetched from
unix_peer_get() and when it's looked up by unix_find_other().
Let's move unix_peer_get() to the else branch for unix_find_other()
and clean up the error paths.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
When other has SOCK_DEAD in unix_dgram_sendmsg(), we hold
unix_state_lock() for the sender socket first.
However, we do not need it for sk->sk_type.
Let's move the lock down a bit.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
When other has SOCK_DEAD in unix_dgram_sendmsg(), we call sock_put() for
it first and then set NULL to other before jumping to the error path.
This is to skip sock_put() in the error path.
Let's not set NULL to other and defer the sock_put() to the error path
to clean up the labels later.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
There are two paths jumping to the restart label in unix_dgram_sendmsg().
One requires another lookup and sk_filter(), but the other doesn't.
Let's split the label to make each flow more straightforward.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In unix_dgram_sendmsg(), we use a local variable sunaddr pointing
NULL or msg->msg_name based on msg->msg_namelen.
Let's remove sunaddr and simplify the usage.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
When other is NULL in unix_dgram_sendmsg(), we check if sunaddr
is NULL before looking up a receiver socket.
There are three paths going through the check, but it's always
false for 2 out of the 3 paths: the first socket lookup and the
second 'goto restart'.
The condition can be true for the first 'goto restart' only when
SOCK_DEAD is flagged for the socket found with msg->msg_name.
Let's move the check to the single appropriate path.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
We will introduce skb drop reason for AF_UNIX, then we need to
set an errno and a drop reason for each path.
Let's set an error only when it's needed in unix_dgram_sendmsg().
Then, we need not (re)set 0 to err.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
If we move send_sig() to the SEND_SHUTDOWN check before
the while loop, then we can reuse the same kfree_skb()
after the pipe_err_free label.
Let's gather the scattered kfree_skb()s in error paths.
While at it, some style issues are fixed, and the pipe_err_free
label is renamed to out_pipe to match other label names.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
We will introduce skb drop reason for AF_UNIX, then we need to
set an errno and a drop reason for each path.
Let's set an error only when it's needed in unix_stream_sendmsg().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The label order is weird in unix_stream_connect(), and all NULL checks
are unnecessary if reordered.
Let's clean up the error paths to make it easy to set a drop reason
for each path.
While at it, a comment with the old style is updated.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
We will introduce skb drop reason for AF_UNIX, then we need to
set an errno and a drop reason for each path.
Let's set an error only when it's needed in unix_stream_connect().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Heiner Kallweit says:
====================
r8169: add support for RTL8125D rev.b
Add support for RTL8125D rev.b. Its XID is 0x689. It is basically
based on the one with XID 0x688, but with different firmware file.
To avoid a mess with the version numbering, adjust it first.
====================
Link: https://patch.msgid.link/15c4a9fd-a653-4b09-825d-751964832a7a@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add support for RTL8125D rev.b. Its XID is 0x689. It is basically
based on the one with XID 0x688, but with different firmware file.
Signed-off-by: ChunHao Lin <hau@realtek.com>
[hkallweit1@gmail.com: rebased after adjusted version numbering]
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/75e5e9ec-d01f-43ac-b0f4-e7456baf18d1@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Adjust version numbering for RTL8126, so that it doesn't overlap with
new RTL8125 versions.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/6a354364-20e9-48ad-a198-468264288757@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Anna Emese Nyiri says:
====================
Add support for SO_PRIORITY cmsg
Introduce a new helper function, `sk_set_prio_allowed`,
to centralize the logic for validating priority settings.
Add support for the `SO_PRIORITY` control message,
enabling user-space applications to set socket priority
via control messages (cmsg).
====================
Link: https://patch.msgid.link/20241213084457.45120-1-annaemesenyiri@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add new socket option, SO_RCVPRIORITY, to include SO_PRIORITY in the
ancillary data returned by recvmsg().
This is analogous to the existing support for SO_RCVMARK,
as implemented in commit 6fd1d51cfa253 ("net: SO_RCVMARK socket option
for SO_MARK with recvmsg()").
Reviewed-by: Willem de Bruijn <willemb@google.com>
Suggested-by: Ferenc Fejes <fejes@inf.elte.hu>
Signed-off-by: Anna Emese Nyiri <annaemesenyiri@gmail.com>
Link: https://patch.msgid.link/20241213084457.45120-5-annaemesenyiri@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Extend cmsg_sender.c with a new option '-Q' to send SO_PRIORITY
ancillary data.
cmsg_so_priority.sh script added to validate SO_PRIORITY behavior
by creating VLAN device with egress QoS mapping and testing packet
priorities using flower filters. Verify that packets with different
priorities are correctly matched and counted by filters for multiple
protocols and IP versions.
Reviewed-by: Willem de Bruijn <willemb@google.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Suggested-by: Ido Schimmel <idosch@idosch.org>
Signed-off-by: Anna Emese Nyiri <annaemesenyiri@gmail.com>
Link: https://patch.msgid.link/20241213084457.45120-4-annaemesenyiri@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The Linux socket API currently allows setting SO_PRIORITY at the
socket level, applying a uniform priority to all packets sent through
that socket. The exception to this is IP_TOS, when the priority value
is calculated during the handling of
ancillary data, as implemented in commit f02db315b8d8 ("ipv4: IP_TOS
and IP_TTL can be specified as ancillary data").
However, this is a computed
value, and there is currently no mechanism to set a custom priority
via control messages prior to this patch.
According to this patch, if SO_PRIORITY is specified as ancillary data,
the packet is sent with the priority value set through
sockc->priority, overriding the socket-level values
set via the traditional setsockopt() method. This is analogous to
the existing support for SO_MARK, as implemented in
commit c6af0c227a22 ("ip: support SO_MARK cmsg").
If both cmsg SO_PRIORITY and IP_TOS are passed, then the one that
takes precedence is the last one in the cmsg list.
This patch has the side effect that raw_send_hdrinc now interprets cmsg
IP_TOS.
Reviewed-by: Willem de Bruijn <willemb@google.com>
Suggested-by: Ferenc Fejes <fejes@inf.elte.hu>
Signed-off-by: Anna Emese Nyiri <annaemesenyiri@gmail.com>
Link: https://patch.msgid.link/20241213084457.45120-3-annaemesenyiri@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Simplify priority setting permissions with the 'sk_set_prio_allowed'
function, centralizing the validation logic. This change is made in
anticipation of a second caller in a following patch.
No functional changes.
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Suggested-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Anna Emese Nyiri <annaemesenyiri@gmail.com>
Link: https://patch.msgid.link/20241213084457.45120-2-annaemesenyiri@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add the missing phys-binding attr to the mctp-attrs in the rt_link spec.
This fixes commit 580db513b4a9 ("net: mctp: Expose transport binding
identifier via IFLA attribute").
Note that enum mctp_phys_binding is not currently uapi, but perhaps it
should be?
Signed-off-by: Donald Hunter <donald.hunter@gmail.com>
Link: https://patch.msgid.link/20241213112551.33557-1-donald.hunter@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
When userspace is adding data to an RPC call for transmission, it must pass
MSG_MORE to sendmsg() if it intends to add more data in future calls to
sendmsg(). Calling sendmsg() without MSG_MORE being asserted closes the
transmission phase of the call (assuming sendmsg() adds all the data
presented) and further attempts to add more data should be rejected.
However, this is no longer the case. The change of call state that was
previously the guard got bumped over to the I/O thread, which leaves a
window for a repeat sendmsg() to insert more data. This previously went
unnoticed, but the more recent patch that changed the structures behind the
Tx queue added a warning:
WARNING: CPU: 3 PID: 6639 at net/rxrpc/sendmsg.c:296 rxrpc_send_data+0x3f2/0x860
and rejected the additional data, returning error EPROTO.
Fix this by adding a guard flag to the call, setting the flag when we queue
the final packet and then rejecting further attempts to add data with
EPROTO.
Fixes: 2d689424b618 ("rxrpc: Move call state changes from sendmsg to I/O thread")
Reported-by: syzbot+ff11be94dfcd7a5af8da@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/r/6757fb68.050a0220.2477f.005f.GAE@google.com/
Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: syzbot+ff11be94dfcd7a5af8da@syzkaller.appspotmail.com
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://patch.msgid.link/2870480.1734037462@warthog.procyon.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Use spin_lock_irq(), not spin_lock_bh() to take the lock when accessing the
->attend_link() to stop a delay in the I/O thread due to an interrupt being
taken in the app thread whilst that holds the lock and vice versa.
Fixes: a2ea9a907260 ("rxrpc: Use irq-disabling spinlocks between app and I/O thread")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://patch.msgid.link/2870146.1734037095@warthog.procyon.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux
Tariq Toukan says:
====================
mlx5-next 2024-12-16
The following pull-request contains mlx5 IFC updates.
* 'mlx5-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux:
net/mlx5: Add device cap abs_native_port_num
net/mlx5: qos: Add ifc support for cross-esw scheduling
net/mlx5: Add support for new scheduling elements
net/mlx5: Add ConnectX-8 device to ifc
net/mlx5: ifc: Reorganize mlx5_ifc_flow_table_context_bits
====================
Link: https://patch.msgid.link/20241216124028.973763-1-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Kory Maincent says:
====================
net: Make timestamping selectable
Up until now, there was no way to let the user select the hardware
PTP provider at which time stamping occurs. The stack assumed that PHY time
stamping is always preferred, but some MAC/PHY combinations were buggy.
This series updates the default MAC/PHY default timestamping and aims to
allow the user to select the desired hwtstamp provider administratively.
Here is few netlink spec usage examples:
./ynl/cli.py --spec netlink/specs/ethtool.yaml --no-schema
--dump tsinfo-get
--json '{"header":{"dev-name":"eth0"}}'
[{'header': {'dev-index': 3, 'dev-name': 'eth0'},
'hwtst-provider': {'index': 0, 'qualifier': 0},
'phc-index': 0,
'rx-filters': {'bits': {'bit': [{'index': 0, 'name': 'none'},
{'index': 2, 'name': 'some'}]},
'nomask': True,
'size': 16},
'timestamping': {'bits': {'bit': [{'index': 0, 'name': 'hardware-transmit'},
{'index': 2, 'name': 'hardware-receive'},
{'index': 6,
'name': 'hardware-raw-clock'}]},
'nomask': True,
'size': 17},
'tx-types': {'bits': {'bit': [{'index': 0, 'name': 'off'},
{'index': 1, 'name': 'on'}]},
'nomask': True,
'size': 4}},
{'header': {'dev-index': 3, 'dev-name': 'eth0'},
'hwtst-provider': {'index': 2, 'qualifier': 0},
'phc-index': 2,
'rx-filters': {'bits': {'bit': [{'index': 0, 'name': 'none'},
{'index': 1, 'name': 'all'}]},
'nomask': True,
'size': 16},
'timestamping': {'bits': {'bit': [{'index': 0, 'name': 'hardware-transmit'},
{'index': 1, 'name': 'software-transmit'},
{'index': 2, 'name': 'hardware-receive'},
{'index': 3, 'name': 'software-receive'},
{'index': 4,
'name': 'software-system-clock'},
{'index': 6,
'name': 'hardware-raw-clock'}]},
'nomask': True,
'size': 17},
'tx-types': {'bits': {'bit': [{'index': 0, 'name': 'off'},
{'index': 1, 'name': 'on'},
{'index': 2, 'name': 'onestep-sync'}]},
'nomask': True,
'size': 4}}]
./ynl/cli.py --spec netlink/specs/ethtool.yaml --no-schema --do tsinfo-get
--json '{"header":{"dev-name":"eth0"},
"hwtst-provider":{"index":0, "qualifier":0 }
}'
{'header': {'dev-index': 3, 'dev-name': 'eth0'},
'hwtst-provider': {'index': 0, 'qualifier': 0},
'phc-index': 0,
'rx-filters': {'bits': {'bit': [{'index': 0, 'name': 'none'},
{'index': 2, 'name': 'some'}]},
'nomask': True,
'size': 16},
'timestamping': {'bits': {'bit': [{'index': 0, 'name': 'hardware-transmit'},
{'index': 2, 'name': 'hardware-receive'},
{'index': 6, 'name': 'hardware-raw-clock'}]},
'nomask': True,
'size': 17},
'tx-types': {'bits': {'bit': [{'index': 0, 'name': 'off'},
{'index': 1, 'name': 'on'}]},
'nomask': True,
'size': 4}}
./ynl/cli.py --spec netlink/specs/ethtool.yaml --no-schema --do tsinfo-set
--json '{"header":{"dev-name":"eth0"},
"hwtst-provider":{"index":2, "qualifier":0}}'
None
./ynl/cli.py --spec netlink/specs/ethtool.yaml --no-schema --do tsconfig-get
--json '{"header":{"dev-name":"eth0"}}'
{'header': {'dev-index': 3, 'dev-name': 'eth0'},
'hwtstamp-flags': 1,
'hwtstamp-provider': {'index': 1, 'qualifier': 0},
'rx-filters': {'bits': {'bit': [{'index': 12, 'name': 'ptpv2-event'}]},
'nomask': True,
'size': 16},
'tx-types': {'bits': {'bit': [{'index': 1, 'name': 'on'}]},
'nomask': True,
'size': 4}}
./ynl/cli.py --spec netlink/specs/ethtool.yaml --no-schema --do tsconfig-set
--json '{"header":{"dev-name":"eth0"},
"hwtstamp-provider":{"index":1, "qualifier":0 },
"rx-filters":{"bits": {"bit": {"name":"ptpv2-l4-event"}},
"nomask": 1},
"tx-types":{"bits": {"bit": {"name":"on"}},
"nomask": 1}}'
{'header': {'dev-index': 3, 'dev-name': 'eth0'},
'hwtstamp-flags': 1,
'hwtstamp-provider': {'index': 1, 'qualifier': 0},
'rx-filters': {'bits': {'bit': [{'index': 12, 'name': 'ptpv2-event'}]},
'nomask': True,
'size': 16},
'tx-types': {'bits': {'bit': [{'index': 1, 'name': 'on'}]},
'nomask': True,
'size': 4}}
Changes in v21:
- NIT fixes.
- Link to v20: https://lore.kernel.org/r/20241204-feature_ptp_netnext-v20-0-9bd99dc8a867@bootlin.com
Changes in v20:
- Change hwtstamp provider design to avoid saving "user" (phy or net) in
the ptp clock structure.
- Link to v19: https://lore.kernel.org/r/20241030-feature_ptp_netnext-v19-0-94f8aadc9d5c@bootlin.com
Changes in v19:
- Rebase on net-next
- Link to v18: https://lore.kernel.org/r/20241023-feature_ptp_netnext-v18-0-ed948f3b6887@bootlin.com
Changes in v18:
- Few changes in the tsconfig-set ethtool command.
- Add tsconfig-set-reply ethtool netlink socket.
- Add missing netlink tsconfig documentation
- Link to v17: https://lore.kernel.org/r/20240709-feature_ptp_netnext-v17-0-b5317f50df2a@bootlin.com
Changes in v17:
- Fix a documentation nit.
- Add a missing kernel_ethtool_tsinfo update from a new MAC driver.
- Link to v16: https://lore.kernel.org/r/20240705-feature_ptp_netnext-v16-0-5d7153914052@bootlin.com
Changes in v16:
- Add a new patch to separate tsinfo into a new tsconfig command to get
and set the hwtstamp config.
- Used call_rcu() instead of synchronize_rcu() to free the hwtstamp_provider
- Moved net core changes of patch 12 directly to patch 8.
- Link to v15: https://lore.kernel.org/r/20240612-feature_ptp_netnext-v15-0-b2a086257b63@bootlin.com
Changes in v15:
- Fix uninitialized ethtool_ts_info structure.
- Link to v14: https://lore.kernel.org/r/20240604-feature_ptp_netnext-v14-0-77b6f6efea40@bootlin.com
Changes in v14:
- Add back an EXPORT_SYMBOL() missing.
- Link to v13: https://lore.kernel.org/r/20240529-feature_ptp_netnext-v13-0-6eda4d40fa4f@bootlin.com
Changes in v13:
- Add PTP builtin code to fix build errors when building PTP as a module.
- Fix error spotted by smatch and sparse.
- Link to v12: https://lore.kernel.org/r/20240430-feature_ptp_netnext-v12-0-2c5f24b6a914@bootlin.com
Changes in v12:
- Add missing return description in the kdoc.
- Fix few nit.
- Link to v11: https://lore.kernel.org/r/20240422-feature_ptp_netnext-v11-0-f14441f2a1d8@bootlin.com
Changes in v11:
- Add netlink examples.
- Remove a change of my out of tree marvell_ptp patch in the patch series.
- Remove useless extern.
- Link to v10: https://lore.kernel.org/r/20240409-feature_ptp_netnext-v10-0-0fa2ea5c89a9@bootlin.com
Changes in v10:
- Move declarations to net/core/dev.h instead of netdevice.h
- Add netlink documentation.
- Add ETHTOOL_A_TSINFO_GHWTSTAMP netlink attributes instead of a bit in
ETHTOOL_A_TSINFO_TIMESTAMPING bitset.
- Send "Move from simple ida to xarray" patch standalone.
- Add tsinfo ntf command.
- Add rcu_lock protection mechanism to avoid memory leak.
- Fixed doc and kdoc issue.
- Link to v9: https://lore.kernel.org/r/20240226-feature_ptp_netnext-v9-0-455611549f21@bootlin.com
Changes in v9:
- Remove the RFC prefix.
- Correct few NIT fixes.
- Link to v8: https://lore.kernel.org/r/20240216-feature_ptp_netnext-v8-0-510f42f444fb@bootlin.com
Changes in v8:
- Drop the 6 first patch as they are now merged.
- Change the full implementation to not be based on the hwtstamp layer
(MAC/PHY) but on the hwtstamp provider which mean a ptp clock and a
phc qualifier.
- Made some patch to prepare the new implementation.
- Expand netlink tsinfo instead of a new ts command for new hwtstamp
configuration uAPI and for dumping tsinfo of specific hwtstamp provider.
- Link to v7: https://lore.kernel.org/r/20231114-feature_ptp_netnext-v7-0-472e77951e40@bootlin.com
Changes in v7:
- Fix a temporary build error.
- Link to v6: https://lore.kernel.org/r/20231019-feature_ptp_netnext-v6-0-71affc27b0e5@bootlin.com
Changes in v6:
- Few fixes from the reviews.
- Replace the allowlist to default_timestamp flag to know which phy is
using old API behavior.
- Rename the timestamping layer enum values.
- Move to a simple enum instead of the mix between enum and bitfield.
- Update ts_info and ts-set in software timestamping case.
Changes in v5:
- Update to ndo_hwstamp_get/set. This bring several new patches.
- Add few patches to make the glue.
- Convert macb to ndo_hwstamp_get/set.
- Add netlink specs description of new ethtool commands.
- Removed netdev notifier.
- Split the patches that expose the timestamping to userspace to separate
the core and ethtool development.
- Add description of software timestamping.
- Convert PHYs hwtstamp callback to use kernel_hwtstamp_config.
Changes in v4:
- Move on to ethtool netlink instead of ioctl.
- Add a netdev notifier to allow packet trapping by the MAC in case of PHY
time stamping.
- Add a PHY whitelist to not break the old PHY default time-stamping
preference API.
Changes in v3:
- Expose the PTP choice to ethtool instead of sysfs.
You can test it with the ethtool source on branch feature_ptp of:
https://github.com/kmaincent/ethtool
- Added a devicetree binding to select the preferred timestamp.
Changes in v2:
- Move selected_timestamping_layer variable of the concerned patch.
- Use sysfs_streq instead of strmcmp.
- Use the PHY timestamp only if available.
====================
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Introduce support for ETHTOOL_MSG_TSCONFIG_GET/SET ethtool netlink socket
to read and configure hwtstamp configuration of a PHC provider. Note that
simultaneous hwtstamp isn't supported; configuring a new one disables the
previous setting.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Either the MAC or the PHY can provide hwtstamp, so we should be able to
read the tsinfo for any hwtstamp provider.
Enhance 'get' command to retrieve tsinfo of hwtstamp providers within a
network topology.
Add support for a specific dump command to retrieve all hwtstamp
providers within the network topology, with added functionality for
filtered dump to target a single interface.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Introduce the description of a hwtstamp provider, mainly defined with a
the hwtstamp source and the phydev pointer.
Add a hwtstamp provider description within the netdev structure to
allow saving the hwtstamp we want to use. This prepares for future
support of an ethtool netlink command to select the desired hwtstamp
provider. By default, the old API that does not support hwtstamp
selectability is used, meaning the hwtstamp provider pointer is unset.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Make the net_hwtstamp_validate function accessible in prevision to use
it from ethtool to validate the hwtstamp configuration before setting it.
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Make the dev_get_hwtstamp_phylib function accessible in prevision to use
it from ethtool to read the hwtstamp current configuration.
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Sabrina Dubroca says:
====================
tls: implement key updates for TLS1.3
This adds support for receiving KeyUpdate messages (RFC 8446, 4.6.3
[1]). A sender transmits a KeyUpdate message and then changes its TX
key. The receiver should react by updating its RX key before
processing the next message.
This patchset implements key updates by:
1. pausing decryption when a KeyUpdate message is received, to avoid
attempting to use the old key to decrypt a record encrypted with
the new key
2. returning -EKEYEXPIRED to syscalls that cannot receive the
KeyUpdate message, until the rekey has been performed by userspace
3. passing the KeyUpdate message to userspace as a control message
4. allowing updates of the crypto_info via the TLS_TX/TLS_RX
setsockopts
This API has been tested with gnutls to make sure that it allows
userspace libraries to implement key updates [2]. Thanks to Frantisek
Krenzelok <fkrenzel@redhat.com> for providing the implementation in
gnutls and testing the kernel patches.
=======================================================================
Discussions around v2 of this patchset focused on how HW offload would
interact with rekey.
RX
- The existing SW path will handle all records between the KeyUpdate
message signaling the change of key and the new key becoming known
to the kernel -- those will be queued encrypted, and decrypted in
SW as they are read by userspace (once the key is provided, ie same
as this patchset)
- Call ->tls_dev_del + ->tls_dev_add immediately during
setsockopt(TLS_RX)
TX
- After setsockopt(TLS_TX), switch to the existing SW path (not the
current device_fallback) until we're able to re-enable HW offload
- tls_device_sendmsg will call into tls_sw_sendmsg under lock_sock
to avoid changing socket ops during the rekey while another
thread might be waiting on the lock
- We only re-enable HW offload (call ->tls_dev_add to install the new
key in HW) once all records sent with the old key have been
ACKed. At this point, all unacked records are SW-encrypted with the
new key, and the old key is unused by both HW and retransmissions.
- If there are no unacked records when userspace does
setsockopt(TLS_TX), we can (try to) install the new key in HW
immediately.
- If yet another key has been provided via setsockopt(TLS_TX), we
don't install intermediate keys, only the latest.
- TCP notifies ktls of ACKs via the icsk_clean_acked callback. In
case of a rekey, tls_icsk_clean_acked will record when all data
sent with the most recent past key has been sent. The next call
to sendmsg will install the new key in HW.
- We close and push the current SW record before reenabling
offload.
If ->tls_dev_add fails to install the new key in HW, we stay in SW
mode. We can add a counter to keep track of this.
In addition:
Because we can't change socket ops during a rekey, we'll also have to
modify do_tls_setsockopt_conf to check ctx->tx_conf and only call
either tls_set_device_offload or tls_set_sw_offload. RX already uses
the same ops for both TLS_HW and TLS_SW, so we could switch between HW
and SW mode on rekey.
An alternative would be to have a common sendmsg which locks
the socket and then calls the correct implementation. We'll need that
anyway for the offload under rekey case, so that would only add a test
to the SW path's ops (compared to the current code). That should allow
us to simplify build_protos a bit, but might have a performance
impact - we'll need to check it if we want to go that route.
=======================================================================
Changes since v4:
- add counter for received KeyUpdate messages
- improve wording in the documentation
- improve handling of bogus messages when looking for KeyUpdate's
- some coding style clean ups
Changes since v3:
- rebase on top of net-next
- rework tls_check_pending_rekey according to Jakub's feedback
- add statistics for rekey: {RX,TX}REKEY{OK,ERROR}
- some coding style clean ups
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Test the kernel's ability to:
- update the key (but not the version or cipher), only for TLS1.3
- pause decryption after receiving a KeyUpdate message, until a new
RX key has been provided
- reflect the pause/non-readable socket in poll()
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This allows us to generate different keys, so that we can test that
rekey is using the correct one.
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Document the kernel's behavior and userspace expectations.
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This introduces 5 counters to keep track of key updates:
Tls{Rx,Tx}Rekey{Ok,Error} and TlsRxRekeyReceived.
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This adds the possibility to change the key and IV when using
TLS1.3. Changing the cipher or TLS version is not supported.
Once we have updated the RX key, we can unblock the receive side. If
the rekey fails, the context is unmodified and userspace is free to
retry the update or close the socket.
This change only affects tls_sw, since 1.3 offload isn't supported.
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
