[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250324224537.248800-1-kuba@kernel.org>
Date: Mon, 24 Mar 2025 15:45:26 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: davem@...emloft.net
Cc: netdev@...r.kernel.org,
edumazet@...gle.com,
pabeni@...hat.com,
andrew+netdev@...n.ch,
horms@...nel.org,
sdf@...ichev.me,
Jakub Kicinski <kuba@...nel.org>
Subject: [PATCH net-next v2 00/11] net: skip taking rtnl_lock for queue GET
Skip taking rtnl_lock for queue GET ops on devices which opt
into running all ops under the instance lock.
This fixes and completes Stan's ops-locking work, so I think
for sanity / ease of backporting fixes we should merge it for
v6.15.
v2:
- rebase
- only clear XSK if netdev still set
v1: https://lore.kernel.org/20250312223507.805719-1-kuba@kernel.org
Jakub Kicinski (11):
net: bubble up taking netdev instance lock to callers of
net_devmem_unbind_dmabuf()
net: remove netif_set_real_num_rx_queues() helper for when SYSFS=n
net: constify dev pointer in misc instance lock helpers
net: explain "protection types" for the instance lock
net: designate queue counts as "double ops protected" by instance lock
net: designate queue -> napi linking as "ops protected"
net: protect rxq->mp_params with the instance lock
net: make NETDEV_UNREGISTER and instance lock more consistent
net: designate XSK pool pointers in queues as "ops protected"
netdev: add "ops compat locking" helpers
netdev: don't hold rtnl_lock over nl queue info get when possible
include/linux/netdevice.h | 41 ++++++++++++---------
include/net/netdev_lock.h | 36 ++++++++++++++++--
include/net/netdev_rx_queue.h | 6 +--
net/core/dev.h | 15 ++++++++
net/core/dev.c | 69 +++++++++++++++++++++++++++++++----
net/core/devmem.c | 2 -
net/core/net-sysfs.c | 2 +
net/core/netdev-genl.c | 27 ++++++++------
net/core/netdev_rx_queue.c | 3 ++
net/core/page_pool.c | 7 +---
net/xdp/xsk_buff_pool.c | 7 +++-
11 files changed, 165 insertions(+), 50 deletions(-)
--
2.49.0
Powered by blists - more mailing lists