From: Kuniyuki Iwashima <kuniyu@amazon.com>
To: <shaw.leon@gmail.com>
Cc: <alex.aring@gmail.com>, <andrew+netdev@lunn.ch>,
<b.a.t.m.a.n@lists.open-mesh.org>, <bpf@vger.kernel.org>,
<bridge@lists.linux.dev>, <davem@davemloft.net>,
<donald.hunter@gmail.com>, <dsahern@kernel.org>,
<edumazet@google.com>, <herbert@gondor.apana.org.au>,
<horms@kernel.org>, <kuba@kernel.org>, <kuniyu@amazon.com>,
<linux-can@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<linux-kselftest@vger.kernel.org>, <linux-ppp@vger.kernel.org>,
<linux-rdma@vger.kernel.org>, <linux-wireless@vger.kernel.org>,
<linux-wpan@vger.kernel.org>, <miquel.raynal@bootlin.com>,
<netdev@vger.kernel.org>, <osmocom-net-gprs@lists.osmocom.org>,
<pabeni@redhat.com>, <shuah@kernel.org>,
<stefan@datenfreihafen.org>, <steffen.klassert@secunet.com>,
<wireguard@lists.zx2c4.com>
Subject: Re: [PATCH net-next v9 02/11] rtnetlink: Pack newlink() params into struct
Date: Thu, 13 Feb 2025 15:53:48 +0900 [thread overview]
Message-ID: <20250213065348.8507-1-kuniyu@amazon.com> (raw)
In-Reply-To: <20250210133002.883422-3-shaw.leon@gmail.com>
From: Xiao Liang <shaw.leon@gmail.com>
Date: Mon, 10 Feb 2025 21:29:53 +0800
> There are 4 net namespaces involved when creating links:
>
> - source netns - where the netlink socket resides,
> - target netns - where to put the device being created,
> - link netns - netns associated with the device (backend),
> - peer netns - netns of peer device.
>
> Currently, two nets are passed to newlink() callback - "src_net"
> parameter and "dev_net" (implicitly in net_device). They are set as
> follows, depending on netlink attributes in the request.
>
> +------------+-------------------+---------+---------+
> | peer netns | IFLA_LINK_NETNSID | src_net | dev_net |
> +------------+-------------------+---------+---------+
> | | absent | source | target |
> | absent +-------------------+---------+---------+
> | | present | link | link |
> +------------+-------------------+---------+---------+
> | | absent | peer | target |
> | present +-------------------+---------+---------+
> | | present | peer | link |
> +------------+-------------------+---------+---------+
>
> When IFLA_LINK_NETNSID is present, the device is created in link netns
> first and then moved to target netns. This has some side effects,
> including extra ifindex allocation, ifname validation and link events.
> These could be avoided if we create it in target netns from
> the beginning.
>
> On the other hand, the meaning of src_net parameter is ambiguous. It
> varies depending on how parameters are passed. It is the effective
> link (or peer netns) by design, but some drivers ignore it and use
> dev_net instead.
>
> This patch packs existing newlink() parameters, along with the source
> netns, link netns and peer netns, into a struct. The old "src_net"
> is renamed to "net" to avoid confusion with real source netns, and
> will be deprecated later. The use of src_net are converted to
> params->net trivially.
>
> To make the semantics more clear, two helper functions -
> rtnl_newlink_link_net() and rtnl_newlink_peer_net() - are provided
> for netns fallback logic. Peer netns falls back to link netns, and
> link netns falls back to source netns.
>
> In following patches, to prepare for creating link in target netns
> directly:
>
> - For device drivers that are aware of the old "src_net", the use of
> it are replace with one of the two helper functions.
> - And for those that takes dev_net() as link netns, we try
> params->link_net and then dev_net(), in order to maintain
> compatibility with the old behavior.
>
> Signed-off-by: Xiao Liang <shaw.leon@gmail.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
I left small comments below.
[...]
> diff --git a/include/linux/if_macvlan.h b/include/linux/if_macvlan.h
> index 523025106a64..0f7281e3e448 100644
> --- a/include/linux/if_macvlan.h
> +++ b/include/linux/if_macvlan.h
> @@ -59,8 +59,10 @@ static inline void macvlan_count_rx(const struct macvlan_dev *vlan,
>
> extern void macvlan_common_setup(struct net_device *dev);
>
> -extern int macvlan_common_newlink(struct net *src_net, struct net_device *dev,
> - struct nlattr *tb[], struct nlattr *data[],
> +struct rtnl_newlink_params;
You can just include <net/rtnetlink.h> and remove it from .c
files, then this forward declaration will be unnecessary.
> +
> +extern int macvlan_common_newlink(struct net_device *dev,
> + struct rtnl_newlink_params *params,
> struct netlink_ext_ack *extack);
>
> extern void macvlan_dellink(struct net_device *dev, struct list_head *head);
[...]
> diff --git a/include/net/rtnetlink.h b/include/net/rtnetlink.h
> index bc0069a8b6ea..00c086ca0c11 100644
> --- a/include/net/rtnetlink.h
> +++ b/include/net/rtnetlink.h
> @@ -69,6 +69,42 @@ static inline int rtnl_msg_family(const struct nlmsghdr *nlh)
> return AF_UNSPEC;
> }
>
> +/**
> + * struct rtnl_newlink_params - parameters of rtnl_link_ops::newlink()
The '\t' after '*' should be single '\s'.
Same for lines below.
> + *
> + * @net: Netns of interest
> + * @src_net: Source netns of rtnetlink socket
> + * @link_net: Link netns by IFLA_LINK_NETNSID, NULL if not specified
> + * @peer_net: Peer netns
> + * @tb: IFLA_* attributes
> + * @data: IFLA_INFO_DATA attributes
> + */
> +struct rtnl_newlink_params {
[...]
> +/* Get effective link netns from newlink params. Generally, this is link_net
> + * and falls back to src_net. But for compatibility, a driver may * choose to
> + * use dev_net(dev) instead.
> + */
> +static inline struct net *rtnl_newlink_link_net(struct rtnl_newlink_params *p)
> +{
> + return p->link_net ? : p->src_net;
> +}
> +
> +/* Get peer netns from newlink params. Fallback to link netns if peer netns is
> + * not specified explicitly.
> + */
> +static inline struct net *rtnl_newlink_peer_net(struct rtnl_newlink_params *p)
> +{
> + return p->peer_net ? : rtnl_newlink_link_net(p);
> +}
These helpers should belong to patch 2 ?
next prev parent reply other threads:[~2025-02-13 6:54 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-10 13:29 [PATCH net-next v9 00/11] net: Improve netns handling in rtnetlink Xiao Liang
2025-02-10 13:29 ` [PATCH net-next v9 01/11] rtnetlink: Lookup device in target netns when creating link Xiao Liang
2025-02-10 13:29 ` [PATCH net-next v9 02/11] rtnetlink: Pack newlink() params into struct Xiao Liang
2025-02-13 6:53 ` Kuniyuki Iwashima [this message]
2025-02-13 8:36 ` Xiao Liang
2025-02-13 9:17 ` Paolo Abeni
2025-02-13 10:01 ` Xiao Liang
2025-02-10 13:29 ` [PATCH net-next v9 03/11] net: Use link netns in newlink() of rtnl_link_ops Xiao Liang
2025-02-13 6:55 ` Kuniyuki Iwashima
2025-02-10 13:29 ` [PATCH net-next v9 04/11] ieee802154: 6lowpan: Validate " Xiao Liang
2025-02-13 5:48 ` Kuniyuki Iwashima
2025-02-10 13:29 ` [PATCH net-next v9 05/11] net: ip_tunnel: Use " Xiao Liang
2025-02-13 6:20 ` Kuniyuki Iwashima
2025-02-13 8:27 ` Xiao Liang
2025-02-10 13:29 ` [PATCH net-next v9 06/11] net: ipv6: " Xiao Liang
2025-02-13 7:05 ` Kuniyuki Iwashima
2025-02-13 8:37 ` Xiao Liang
2025-02-13 9:55 ` Xiao Liang
2025-02-13 11:00 ` Kuniyuki Iwashima
2025-02-14 9:22 ` Xiao Liang
2025-02-14 10:51 ` Kuniyuki Iwashima
2025-02-10 13:29 ` [PATCH net-next v9 07/11] net: xfrm: " Xiao Liang
2025-02-10 13:29 ` [PATCH net-next v9 08/11] rtnetlink: Remove "net" from newlink params Xiao Liang
2025-02-10 13:30 ` [PATCH net-next v9 09/11] rtnetlink: Create link directly in target net namespace Xiao Liang
2025-02-10 13:30 ` [PATCH net-next v9 10/11] selftests: net: Add python context manager for netns entering Xiao Liang
2025-02-10 13:30 ` [PATCH net-next v9 11/11] selftests: net: Add test cases for link and peer netns Xiao Liang
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=20250213065348.8507-1-kuniyu@amazon.com \
--to=kuniyu@amazon.com \
--cc=alex.aring@gmail.com \
--cc=andrew+netdev@lunn.ch \
--cc=b.a.t.m.a.n@lists.open-mesh.org \
--cc=bpf@vger.kernel.org \
--cc=bridge@lists.linux.dev \
--cc=davem@davemloft.net \
--cc=donald.hunter@gmail.com \
--cc=dsahern@kernel.org \
--cc=edumazet@google.com \
--cc=herbert@gondor.apana.org.au \
--cc=horms@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-can@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-ppp@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linux-wireless@vger.kernel.org \
--cc=linux-wpan@vger.kernel.org \
--cc=miquel.raynal@bootlin.com \
--cc=netdev@vger.kernel.org \
--cc=osmocom-net-gprs@lists.osmocom.org \
--cc=pabeni@redhat.com \
--cc=shaw.leon@gmail.com \
--cc=shuah@kernel.org \
--cc=stefan@datenfreihafen.org \
--cc=steffen.klassert@secunet.com \
--cc=wireguard@lists.zx2c4.com \
/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).