linux/net/rxrpc
Tejun Heo 78ef970385 rxrpc: Use alloc_ordered_workqueue() to create ordered workqueues
BACKGROUND
==========

When multiple work items are queued to a workqueue, their execution order
doesn't match the queueing order. They may get executed in any order and
simultaneously. When fully serialized execution - one by one in the queueing
order - is needed, an ordered workqueue should be used which can be created
with alloc_ordered_workqueue().

However, alloc_ordered_workqueue() was a later addition. Before it, an
ordered workqueue could be obtained by creating an UNBOUND workqueue with
@max_active==1. This originally was an implementation side-effect which was
broken by 4c16bd327c ("workqueue: restore WQ_UNBOUND/max_active==1 to be
ordered"). Because there were users that depended on the ordered execution,
5c0338c687 ("workqueue: restore WQ_UNBOUND/max_active==1 to be ordered")
made workqueue allocation path to implicitly promote UNBOUND workqueues w/
@max_active==1 to ordered workqueues.

While this has worked okay, overloading the UNBOUND allocation interface
this way creates other issues. It's difficult to tell whether a given
workqueue actually needs to be ordered and users that legitimately want a
min concurrency level wq unexpectedly gets an ordered one instead. With
planned UNBOUND workqueue updates to improve execution locality and more
prevalence of chiplet designs which can benefit from such improvements, this
isn't a state we wanna be in forever.

This patch series audits all callsites that create an UNBOUND workqueue w/
@max_active==1 and converts them to alloc_ordered_workqueue() as necessary.

WHAT TO LOOK FOR
================

The conversions are from

  alloc_workqueue(WQ_UNBOUND | flags, 1, args..)

to

  alloc_ordered_workqueue(flags, args...)

which don't cause any functional changes. If you know that fully ordered
execution is not necessary, please let me know. I'll drop the conversion and
instead add a comment noting the fact to reduce confusion while conversion
is in progress.

If you aren't fully sure, it's completely fine to let the conversion
through. The behavior will stay exactly the same and we can always
reconsider later.

As there are follow-up workqueue core changes, I'd really appreciate if the
patch can be routed through the workqueue tree w/ your acks. Thanks.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Marc Dionne <marc.dionne@auristor.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: linux-afs@lists.infradead.org
Cc: netdev@vger.kernel.org
2023-05-25 12:15:36 -10:00
..
af_rxrpc.c rxrpc: Use alloc_ordered_workqueue() to create ordered workqueues 2023-05-25 12:15:36 -10:00
ar-internal.h rxrpc: Fix timeout of a call that hasn't yet been granted a channel 2023-05-01 07:43:19 +01:00
call_accept.c rxrpc: Fix trace string 2023-01-30 14:13:29 +00:00
call_event.c rxrpc: De-atomic call->ackr_window and call->ackr_nr_unacked 2023-01-31 16:38:26 +00:00
call_object.c Including fixes from netfilter. 2023-05-05 19:12:01 -07:00
call_state.c rxrpc: Move client call connection to the I/O thread 2023-01-06 09:43:33 +00:00
conn_client.c rxrpc: Move client call connection to the I/O thread 2023-01-06 09:43:33 +00:00
conn_event.c rxrpc: Trace ack.rwind 2023-02-07 23:11:21 +00:00
conn_object.c rxrpc: Move client call connection to the I/O thread 2023-01-06 09:43:33 +00:00
conn_service.c rxrpc: Kill service bundle 2023-01-31 16:38:35 +00:00
input.c rxrpc: Fix overproduction of wakeups to recvmsg() 2023-02-20 08:33:25 +01:00
insecure.c rxrpc: Tidy up abort generation infrastructure 2023-01-06 09:43:32 +00:00
io_thread.c rxrpc: Allow a delay to be injected into packet reception 2023-01-31 16:38:09 +00:00
Kconfig rxrpc: Allow a delay to be injected into packet reception 2023-01-31 16:38:09 +00:00
key.c rxrpc: Fix error when reading rxrpc tokens 2023-04-23 13:38:28 +01:00
local_event.c rxrpc: Make the I/O thread take over the call and local processor work 2022-12-01 13:36:42 +00:00
local_object.c rxrpc: Remove local->defrag_sem 2023-01-31 16:38:35 +00:00
Makefile rxrpc: Split out the call state changing functions into their own file 2023-01-06 09:43:32 +00:00
misc.c rxrpc: Allow a delay to be injected into packet reception 2023-01-31 16:38:09 +00:00
net_ns.c rxrpc: Move the client conn cache management to the I/O thread 2023-01-06 09:43:33 +00:00
output.c rxrpc: Trace ack.rwind 2023-02-07 23:11:21 +00:00
peer_event.c rxrpc: Fix locking issues in rxrpc_put_peer_locked() 2022-12-19 09:51:31 +00:00
peer_object.c rxrpc: Stash the network namespace pointer in rxrpc_local 2023-01-06 09:43:31 +00:00
proc.c rxrpc: De-atomic call->ackr_window and call->ackr_nr_unacked 2023-01-31 16:38:26 +00:00
protocol.h rxrpc: Replace fake flex-array with flexible-array member 2023-04-23 13:36:05 +01:00
recvmsg.c rxrpc: Fix overproduction of wakeups to recvmsg() 2023-02-20 08:33:25 +01:00
rtt.c rxrpc: Fix _usecs_to_jiffies() by using usecs_to_jiffies() 2021-09-24 14:18:34 +01:00
rxkad.c rxrpc: Move client call connection to the I/O thread 2023-01-06 09:43:33 +00:00
rxperf.c rxrpc: Fix potential race in error handling in afs_make_call() 2023-04-22 15:16:39 +01:00
security.c rxrpc: Fix incoming call setup race 2023-01-07 09:30:26 +00:00
sendmsg.c rxrpc: Fix timeout of a call that hasn't yet been granted a channel 2023-05-01 07:43:19 +01:00
server_key.c rxrpc: Implement an in-kernel rxperf server for testing purposes 2022-12-01 13:36:37 +00:00
skbuff.c rxrpc: Use consume_skb() rather than kfree_skb_reason() 2023-02-07 23:11:20 +00:00
sysctl.c rxrpc: Allow a delay to be injected into packet reception 2023-01-31 16:38:09 +00:00
txbuf.c rxrpc: Don't lock call->tx_lock to access call->tx_buffer 2023-01-31 16:38:35 +00:00
utils.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 36 2019-05-24 17:27:11 +02:00