* [PATCH net-next] net: ethtool: don't take rtnl_lock for global string dump
@ 2026-05-27 16:25 Jakub Kicinski
2026-05-27 16:40 ` Vadim Fedorenko
0 siblings, 1 reply; 2+ messages in thread
From: Jakub Kicinski @ 2026-05-27 16:25 UTC (permalink / raw)
To: davem
Cc: netdev, edumazet, pabeni, andrew+netdev, horms, Jakub Kicinski,
andrew, maxime.chevallier
ETHTOOL_MSG_STRSET_GET is the only op which sets allow_nodev_do.
When no device is provided it dumps static tables, there's no
need to hold rtnl_lock for this.
Not taking rtnl_lock is a minor win in itself so I think this
patch stands on its own merits. Later on it will be useful
to do locking only in paths which have access to a netdev,
so that we can decide which locks to take per-netdev.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
CC: andrew@lunn.ch
CC: maxime.chevallier@bootlin.com
---
net/ethtool/netlink.h | 4 +++-
net/ethtool/netlink.c | 10 ++++++----
2 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/net/ethtool/netlink.h b/net/ethtool/netlink.h
index aaf6f2468768..b31de00fe683 100644
--- a/net/ethtool/netlink.h
+++ b/net/ethtool/netlink.h
@@ -333,7 +333,9 @@ int ethnl_sock_priv_set(struct sk_buff *skb, struct net_device *dev, u32 portid,
* @hdr_attr: attribute type for request header
* @req_info_size: size of request info
* @reply_data_size: size of reply data
- * @allow_nodev_do: allow non-dump request with no device identification
+ * @allow_nodev_do:
+ * Allow non-dump request with no device identification.
+ * Note that locks (rtnl_lock etc.) are only taken if device is set.
* @set_ntf_cmd: notification to generate on changes (SET)
* @parse_request:
* Parse request except common header (struct ethnl_req_info). Common
diff --git a/net/ethtool/netlink.c b/net/ethtool/netlink.c
index 5046023a30b1..a42bdd5ded51 100644
--- a/net/ethtool/netlink.c
+++ b/net/ethtool/netlink.c
@@ -526,13 +526,15 @@ static int ethnl_default_doit(struct sk_buff *skb, struct genl_info *info)
goto err_free;
ethnl_init_reply_data(reply_data, ops, req_info->dev);
- rtnl_lock();
- if (req_info->dev)
+ if (req_info->dev) {
+ rtnl_lock();
netdev_lock_ops(req_info->dev);
+ }
ret = ops->prepare_data(req_info, reply_data, info);
- if (req_info->dev)
+ if (req_info->dev) {
netdev_unlock_ops(req_info->dev);
- rtnl_unlock();
+ rtnl_unlock();
+ }
if (ret < 0)
goto err_dev;
ret = ops->reply_size(req_info, reply_data);
--
2.54.0
^ permalink raw reply related [flat|nested] 2+ messages in thread* Re: [PATCH net-next] net: ethtool: don't take rtnl_lock for global string dump
2026-05-27 16:25 [PATCH net-next] net: ethtool: don't take rtnl_lock for global string dump Jakub Kicinski
@ 2026-05-27 16:40 ` Vadim Fedorenko
0 siblings, 0 replies; 2+ messages in thread
From: Vadim Fedorenko @ 2026-05-27 16:40 UTC (permalink / raw)
To: Jakub Kicinski, davem
Cc: netdev, edumazet, pabeni, andrew+netdev, horms, andrew,
maxime.chevallier
On 27/05/2026 17:25, Jakub Kicinski wrote:
> ETHTOOL_MSG_STRSET_GET is the only op which sets allow_nodev_do.
> When no device is provided it dumps static tables, there's no
> need to hold rtnl_lock for this.
>
> Not taking rtnl_lock is a minor win in itself so I think this
> patch stands on its own merits. Later on it will be useful
> to do locking only in paths which have access to a netdev,
> so that we can decide which locks to take per-netdev.
>
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
> ---
> CC: andrew@lunn.ch
> CC: maxime.chevallier@bootlin.com
> ---
> net/ethtool/netlink.h | 4 +++-
> net/ethtool/netlink.c | 10 ++++++----
> 2 files changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/net/ethtool/netlink.h b/net/ethtool/netlink.h
> index aaf6f2468768..b31de00fe683 100644
> --- a/net/ethtool/netlink.h
> +++ b/net/ethtool/netlink.h
> @@ -333,7 +333,9 @@ int ethnl_sock_priv_set(struct sk_buff *skb, struct net_device *dev, u32 portid,
> * @hdr_attr: attribute type for request header
> * @req_info_size: size of request info
> * @reply_data_size: size of reply data
> - * @allow_nodev_do: allow non-dump request with no device identification
> + * @allow_nodev_do:
> + * Allow non-dump request with no device identification.
> + * Note that locks (rtnl_lock etc.) are only taken if device is set.
> * @set_ntf_cmd: notification to generate on changes (SET)
> * @parse_request:
> * Parse request except common header (struct ethnl_req_info). Common
> diff --git a/net/ethtool/netlink.c b/net/ethtool/netlink.c
> index 5046023a30b1..a42bdd5ded51 100644
> --- a/net/ethtool/netlink.c
> +++ b/net/ethtool/netlink.c
> @@ -526,13 +526,15 @@ static int ethnl_default_doit(struct sk_buff *skb, struct genl_info *info)
> goto err_free;
> ethnl_init_reply_data(reply_data, ops, req_info->dev);
>
> - rtnl_lock();
> - if (req_info->dev)
> + if (req_info->dev) {
> + rtnl_lock();
> netdev_lock_ops(req_info->dev);
> + }
> ret = ops->prepare_data(req_info, reply_data, info);
> - if (req_info->dev)
> + if (req_info->dev) {
> netdev_unlock_ops(req_info->dev);
> - rtnl_unlock();
> + rtnl_unlock();
> + }
> if (ret < 0)
> goto err_dev;
> ret = ops->reply_size(req_info, reply_data);
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-05-27 16:41 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-05-27 16:25 [PATCH net-next] net: ethtool: don't take rtnl_lock for global string dump Jakub Kicinski
2026-05-27 16:40 ` Vadim Fedorenko
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox