For ACKs generated inside the I/O thread, transmit the ACK at the point of
generation.  Where the ACK is generated outside of the I/O thread, it's
offloaded to the I/O thread to transmit it.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Fold __rxrpc_unuse_local() into rxrpc_unuse_local() as the latter is now
the only user of the former.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

When we've gone for >1RTT without transmitting a packet, we should reduce
the ssthresh and cut the cwnd by half (as suggested in RFC2861 sec 3.1).

However, we may receive ACK packets in a batch and the first of these may
cut the cwnd, preventing further transmission, and each subsequent one cuts
the cwnd yet further, reducing it to the floor and killing performance.

Fix this by moving the cwnd reset to after doing the transmission and
resetting the base time such that we don't cut the cwnd by half again for
at least another RTT.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Add a tracepoint to log when a cwnd reset occurs due to lack of
transmission on a call.

Add stat counters to count transmission underflows (ie. when we have tx
window space, but sendmsg doesn't manage to keep up), cwnd resets and
transmission failures.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

None of the spinlocks in rxrpc need a _bh annotation now as the RCU
callback routines no longer take spinlocks and the bulk of the packet
wrangling code is now run in the I/O thread, not softirq context.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Move the functions from the call->processor and local->processor work items
into the domain of the I/O thread.

The call event processor, now called from the I/O thread, then takes over
the job of cranking the call state machine, processing incoming packets and
transmitting DATA, ACK and ABORT packets.  In a future patch,
rxrpc_send_ACK() will transmit the ACK on the spot rather than queuing it
for later transmission.

The call event processor becomes purely received-skb driven.  It only
transmits things in response to events.  We use "pokes" to queue a dummy
skb to make it do things like start/resume transmitting data.  Timer expiry
also results in pokes.

The connection event processor, becomes similar, though crypto events, such
as dealing with CHALLENGE and RESPONSE packets is offloaded to a work item
to avoid doing crypto in the I/O thread.

The local event processor is removed and VERSION response packets are
generated directly from the packet parser.  Similarly, ABORTs generated in
response to protocol errors will be transmitted immediately rather than
being pushed onto a queue for later transmission.

Changes:
========
ver #2)
 - Fix a couple of introduced lock context imbalances.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Extract the peer address from an incoming packet earlier, at the beginning
of rxrpc_input_packet() and thence pass a pointer to it to various
functions that use it as part of the lookup rather than doing it on several
separate paths.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Shrink the region of rxrpc_input_packet() that is covered by the RCU read
lock so that it only covers the connection and call lookup.  This means
that the bits now outside of that can call sleepable functions such as
kmalloc and sendmsg.

Also take a ref on the conn or call we're going to use before we drop the
RCU read lock.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

A received skbuff needs a ref when it gets put on a call data queue or conn
packet queue, and rxrpc_input_packet() and co. jump through a lot of hoops
to avoid double-dropping the skbuff ref so that we can avoid getting a ref
when we queue the packet.

Change this so that the skbuff ref is unconditionally dropped by the caller
of rxrpc_input_packet().  An additional ref is then taken on the packet if
it is pushed onto a queue.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Remove the RCU requirements from the peer's list of error targets so that
the error distributor can call sleeping functions.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Move DATA transmission into the call processor work item.  In a future
patch, this will be called from the I/O thread rather than being itsown
work item.

This will allow DATA transmission to be driven directly by incoming ACKs,
pokes and timers as those are processed.

The Tx queue is also split: The queue of packets prepared by sendmsg is now
places in call->tx_sendmsg and the packet dispatcher decants the packets
into call->tx_buffer as space becomes available in the transmission
window.  This allows sendmsg to run ahead of the available space to try and
prevent an underflow in transmission.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Copy client call parameters into rxrpc_call earlier so that that can be
used to convey them to the connection code - which can then be offloaded to
the I/O thread.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Provide a means by which an event notification can be sent to a call such
that the I/O thread can process it rather than it being done in a separate
workqueue.  This will allow a lot of locking to be removed.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Don't use sk->sk_receive_queue.lock to guard socket state changes as the
socket mutex is sufficient.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Remove call->input_lock as it was only necessary to serialise access to the
state stored in the rxrpc_call struct by simultaneous softirq handlers
presenting received packets.  They now dump the packets in a queue and a
single process-context handler now processes them.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Move the processing of error packets into the local endpoint I/O thread,
leaving the handover from UDP to merely transfer them into the local
endpoint queue.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Split the packet input handler to make the softirq side just dump the
received packet into the local endpoint receive queue and then call the
remainder of the input function from the I/O thread.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Create a per-local receive queue to which, in a future patch, all incoming
packets will be directed and an I/O thread that will process those packets
and perform all transmission of packets.

Destruction of the local endpoint is also moved from the local processor
work item (which will be absorbed) to the thread.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Split the code that handles packet reception in softirq mode as a prelude
to moving all the packet processing beyond routing to the appropriate call
and setting up of a new call out into process context.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Currently, rxrpc gives the connection's work item a ref on the connection
when it queues it - and this is called from the timer expiration function.
The problem comes when queue_work() fails (ie. the work item is already
queued): the timer routine must put the ref - but this may cause the
cleanup code to run.

This has the unfortunate effect that the cleanup code may then be run in
softirq context - which means that any spinlocks it might need to touch
have to be guarded to disable softirqs (ie. they need a "_bh" suffix).

 (1) Don't give a ref to the work item.

 (2) Simplify handling of service connections by adding a separate active
     count so that the refcount isn't also used for this.

 (3) Connection destruction for both client and service connections can
     then be cleaned up by putting rxrpc_put_connection() out of line and
     making a tidy progression through the destruction code (offloaded to a
     workqueue if put from softirq or processor function context).  The RCU
     part of the cleanup then only deals with the freeing at the end.

 (4) Make rxrpc_queue_conn() return immediately if it sees the active count
     is -1 rather then queuing the connection.

 (5) Make sure that the cleanup routine waits for the work item to
     complete.

 (6) Stash the rxrpc_net pointer in the conn struct so that the rcu free
     routine can use it, even if the local endpoint has been freed.

Unfortunately, neither the timer nor the work item can simply get around
the problem by just using refcount_inc_not_zero() as the waits would still
have to be done, and there would still be the possibility of having to put
the ref in the expiration function.

Note the connection work item is mostly going to go away with the main
event work being transferred to the I/O thread, so the wait in (6) will
become obsolete.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Currently, rxrpc gives the call timer a ref on the call when it starts it
and this is passed along to the workqueue by the timer expiration function.
The problem comes when queue_work() fails (ie. the work item is already
queued): the timer routine must put the ref - but this may cause the
cleanup code to run.

This has the unfortunate effect that the cleanup code may then be run in
softirq context - which means that any spinlocks it might need to touch
have to be guarded to disable softirqs (ie. they need a "_bh" suffix).

Fix this by:

 (1) Don't give a ref to the timer.

 (2) Making the expiration function not do anything if the refcount is 0.
     Note that this is more of an optimisation.

 (3) Make sure that the cleanup routine waits for timer to complete.

However, this has a consequence that timer cannot give a ref to the work
item.  Therefore the following fixes are also necessary:

 (4) Don't give a ref to the work item.

 (5) Make the work item return asap if it sees the ref count is 0.

 (6) Make sure that the cleanup routine waits for the work item to
     complete.

Unfortunately, neither the timer nor the work item can simply get around
the problem by just using refcount_inc_not_zero() as the waits would still
have to be done, and there would still be the possibility of having to put
the ref in the expiration function.

Note the call work item is going to go away with the work being transferred
to the I/O thread, so the wait in (6) will become obsolete.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

In rxrpc tracing, use enums to generate lists of points of interest rather
than __builtin_return_address() for the sk_buff tracepoint.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Add a tracepoint for the rxrpc_bundle refcounting.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

In rxrpc tracing, use enums to generate lists of points of interest rather
than __builtin_return_address() for the rxrpc_call tracepoint

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

In rxrpc tracing, use enums to generate lists of points of interest rather
than __builtin_return_address() for the rxrpc_conn tracepoint

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

In rxrpc tracing, use enums to generate lists of points of interest rather
than __builtin_return_address() for the rxrpc_peer tracepoint

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

In rxrpc tracing, use enums to generate lists of points of interest rather
than __builtin_return_address() for the rxrpc_local tracepoint

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Extract the code from a received rx ABORT packet much earlier and in a
single place and harmonise the responses to malformed ABORT packets.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Remove the rxrpc_conn_parameters struct from the rxrpc_connection and
rxrpc_bundle structs and emplace the members directly.  These are going to
get filled in from the rxrpc_call struct in future.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Remove the _net() and knet() debugging macros in favour of tracepoints.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Remove the kproto() and _proto() debugging macros in preference to using
tracepoints for this.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

We should not now see duplicate packets in the recvmsg_queue.  At one
point, jumbo packets that overlapped with already queued data would be
added to the queue and dealt with in recvmsg rather than in the softirq
input code, but now jumbo packets are split/cloned before being processed
by the input code and the subpackets can be discarded individually.

So remove the recvmsg-side code for handling this.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

rxrpc_kernel_call_is_complete() has been removed, so remove its declaration
too.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

When retransmitting a packet, rxrpc_resend() shouldn't be attaching a ref
to the call to the txbuf as that pins the call and prevents the call from
clearing the packet buffer.

Signed-off-by: David Howells <dhowells@redhat.com>
Fixes: d57a3a15 ("rxrpc: Save last ACK's SACK table rather than marking txbufs")
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Implement an in-kernel rxperf server to allow kernel-based rxrpc services
to be tested directly, unlike with AFS where they're accessed by the
fileserver when the latter decides it wants to.

This is implemented as a module that, if loaded, opens UDP port 7009
(afs3-rmtsys) and listens on it for incoming calls.  Calls can be generated
using the rxperf command shipped with OpenAFS, for example.

Changes
=======
ver #2)
 - Use min_t() instead of min().

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
cc: Jakub Kicinski <kuba@kernel.org>

Fix the following checker warning:

../net/rxrpc/key.c:692:9: error: subtraction of different types can't work (different address spaces)

Checker is wrong in this case, but cast the pointers to unsigned long to
avoid the warning.

Whilst we're at it, reduce the assertions to WARN_ON() and return an error.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org

Test NIC hardware checksum offload:

- Rx + Tx
- IPv4 + IPv6
- TCP + UDP

Optional features:

- zero checksum 0xFFFF
- checksum disable 0x0000
- transport encap headers
- randomization

See file header for detailed comments.

Expected results differ depending on NIC features:

- CHECKSUM_UNNECESSARY vs CHECKSUM_COMPLETE
- NETIF_F_HW_CSUM (csum_start/csum_off) vs NETIF_F_IP(V6)_CSUM

Signed-off-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20221128140210.553391-1-willemdebruijn.kernel@gmail.com


Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Steffen Klassert says:

====================
ipsec-next 2022-11-26

1) Remove redundant variable in esp6.
   From Colin Ian King.

2) Update x->lastused for every packet. It was used only for
   outgoing mobile IPv6 packets, but showed to be usefull
   to check if the a SA is still in use in general.
   From Antony Antony.

3) Remove unused variable in xfrm_byidx_resize.
   From Leon Romanovsky.

4) Finalize extack support for xfrm.
   From Sabrina Dubroca.

* 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next:
  xfrm: add extack to xfrm_set_spdinfo
  xfrm: add extack to xfrm_alloc_userspi
  xfrm: add extack to xfrm_do_migrate
  xfrm: add extack to xfrm_new_ae and xfrm_replay_verify_len
  xfrm: add extack to xfrm_del_sa
  xfrm: add extack to xfrm_add_sa_expire
  xfrm: a few coding style clean ups
  xfrm: Remove not-used total variable
  xfrm: update x->lastused for every packet
  esp6: remove redundant variable err
====================

Link: https://lore.kernel.org/r/20221126110303.1859238-1-steffen.klassert@secunet.com


Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Maxime Chevallier says:

====================
net: pcs: altera-tse: simplify and clean-up the driver

This small series does a bit of code cleanup in the altera TSE pcs
driver, removing unused register definitions, handling 1000BaseX speed
configuration correctly according to the datasheet, and making use of
proper poll_timeout helpers.

No functional change is introduced.
====================

Link: https://lore.kernel.org/r/20221125131801.64234-1-maxime.chevallier@bootlin.com


Signed-off-by: Jakub Kicinski <kuba@kernel.org>

remove unused register definitions, left from the split with the
altera-tse mac driver.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>