From: Jakub Kicinski <kuba@kernel.org>
To: davem@davemloft.net
Cc: netdev@vger.kernel.org, edumazet@google.com, pabeni@redhat.com,
andrew+netdev@lunn.ch, horms@kernel.org, sdf@fomichev.me,
Jakub Kicinski <kuba@kernel.org>
Subject: [PATCH net-next v2 06/11] net: designate queue -> napi linking as "ops protected"
Date: Mon, 24 Mar 2025 15:45:32 -0700 [thread overview]
Message-ID: <20250324224537.248800-7-kuba@kernel.org> (raw)
In-Reply-To: <20250324224537.248800-1-kuba@kernel.org>
netdev netlink is the only reader of netdev_{,rx_}queue->napi,
and it already holds netdev->lock. Switch protection of
the writes to netdev->lock to "ops protected".
The expectation will be now that accessing queue->napi
will require netdev->lock for "ops locked" drivers, and
rtnl_lock for all other drivers.
Current "ops locked" drivers don't require any changes.
gve and netdevsim use _locked() helpers right next to
netif_queue_set_napi() so they must be holding the instance
lock. iavf doesn't call it. bnxt is a bit messy but all paths
seem locked.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
include/linux/netdevice.h | 5 +++--
include/net/netdev_lock.h | 8 ++++++++
include/net/netdev_rx_queue.h | 2 +-
net/core/dev.c | 3 +--
4 files changed, 13 insertions(+), 5 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 3dbce9cd3f5a..fd508e9f148c 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -710,7 +710,7 @@ struct netdev_queue {
* slow- / control-path part
*/
/* NAPI instance for the queue
- * Readers and writers must hold RTNL
+ * "ops protected", see comment about net_device::lock
*/
struct napi_struct *napi;
@@ -2522,7 +2522,8 @@ struct net_device {
* Double ops protects:
* @real_num_rx_queues, @real_num_tx_queues
*
- * Also protects some fields in struct napi_struct.
+ * Also protects some fields in:
+ * struct napi_struct, struct netdev_queue, struct netdev_rx_queue
*
* Ordering: take after rtnl_lock.
*/
diff --git a/include/net/netdev_lock.h b/include/net/netdev_lock.h
index efd302375ef2..1c0c9a94cc22 100644
--- a/include/net/netdev_lock.h
+++ b/include/net/netdev_lock.h
@@ -56,6 +56,14 @@ static inline void netdev_ops_assert_locked(const struct net_device *dev)
ASSERT_RTNL();
}
+static inline void
+netdev_ops_assert_locked_or_invisible(const struct net_device *dev)
+{
+ if (dev->reg_state == NETREG_REGISTERED ||
+ dev->reg_state == NETREG_UNREGISTERING)
+ netdev_ops_assert_locked(dev);
+}
+
static inline int netdev_lock_cmp_fn(const struct lockdep_map *a,
const struct lockdep_map *b)
{
diff --git a/include/net/netdev_rx_queue.h b/include/net/netdev_rx_queue.h
index af40842f229d..b2238b551dce 100644
--- a/include/net/netdev_rx_queue.h
+++ b/include/net/netdev_rx_queue.h
@@ -24,7 +24,7 @@ struct netdev_rx_queue {
struct xsk_buff_pool *pool;
#endif
/* NAPI instance for the queue
- * Readers and writers must hold RTNL
+ * "ops protected", see comment about net_device::lock
*/
struct napi_struct *napi;
struct pp_memory_provider_params mp_params;
diff --git a/net/core/dev.c b/net/core/dev.c
index aa99c91fd68f..690d46497b2f 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -6896,8 +6896,7 @@ void netif_queue_set_napi(struct net_device *dev, unsigned int queue_index,
if (WARN_ON_ONCE(napi && !napi->dev))
return;
- if (dev->reg_state >= NETREG_REGISTERED)
- ASSERT_RTNL();
+ netdev_ops_assert_locked_or_invisible(dev);
switch (type) {
case NETDEV_QUEUE_TYPE_RX:
--
2.49.0
next prev parent reply other threads:[~2025-03-24 22:46 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-24 22:45 [PATCH net-next v2 00/11] net: skip taking rtnl_lock for queue GET Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 01/11] net: bubble up taking netdev instance lock to callers of net_devmem_unbind_dmabuf() Jakub Kicinski
2025-03-25 5:19 ` Mina Almasry
2025-03-24 22:45 ` [PATCH net-next v2 02/11] net: remove netif_set_real_num_rx_queues() helper for when SYSFS=n Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 03/11] net: constify dev pointer in misc instance lock helpers Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 04/11] net: explain "protection types" for the instance lock Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 05/11] net: designate queue counts as "double ops protected" by " Jakub Kicinski
2025-03-24 22:45 ` Jakub Kicinski [this message]
2025-03-24 22:45 ` [PATCH net-next v2 07/11] net: protect rxq->mp_params with the " Jakub Kicinski
2025-03-25 5:34 ` Mina Almasry
2025-03-25 9:50 ` Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 08/11] net: make NETDEV_UNREGISTER and instance lock more consistent Jakub Kicinski
2025-03-25 12:17 ` Cosmin Ratiu
2025-03-25 17:04 ` Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 09/11] net: designate XSK pool pointers in queues as "ops protected" Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 10/11] netdev: add "ops compat locking" helpers Jakub Kicinski
2025-03-24 22:45 ` [PATCH net-next v2 11/11] netdev: don't hold rtnl_lock over nl queue info get when possible Jakub Kicinski
2025-03-25 4:04 ` [PATCH net-next v2 00/11] net: skip taking rtnl_lock for queue GET Stanislav Fomichev
2025-03-25 17:30 ` patchwork-bot+netdevbpf
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250324224537.248800-7-kuba@kernel.org \
--to=kuba@kernel.org \
--cc=andrew+netdev@lunn.ch \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=horms@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sdf@fomichev.me \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).