From: Maxime Chevallier <maxime.chevallier@bootlin.com>
To: "Björn Töpel" <bjorn@kernel.org>,
netdev@vger.kernel.org, "David S. Miller" <davem@davemloft.net>,
"Andrew Lunn" <andrew+netdev@lunn.ch>,
"Donald Hunter" <donald.hunter@gmail.com>,
"Eric Dumazet" <edumazet@google.com>,
"Jakub Kicinski" <kuba@kernel.org>,
"Naveen Mamindlapalli" <naveenm@marvell.com>,
"Paolo Abeni" <pabeni@redhat.com>,
"Simon Horman" <horms@kernel.org>
Cc: Danielle Ratson <danieller@nvidia.com>,
Hariprasad Kelam <hkelam@marvell.com>,
Ido Schimmel <idosch@nvidia.com>,
Kory Maincent <kory.maincent@bootlin.com>,
Leon Romanovsky <leon@kernel.org>,
Michael Chan <michael.chan@broadcom.com>,
Oleksij Rempel <o.rempel@pengutronix.de>,
Pavan Chebbi <pavan.chebbi@broadcom.com>,
Piergiorgio Beruto <piergiorgio.beruto@gmail.com>,
Russell King <linux@armlinux.org.uk>,
Saeed Mahameed <saeedm@nvidia.com>, Shuah Khan <shuah@kernel.org>,
Tariq Toukan <tariqt@nvidia.com>,
Willem de Bruijn <willemb@google.com>,
Kees Cook <kees@kernel.org>,
linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
linux-rdma@vger.kernel.org
Subject: Re: [PATCH net-next v2 01/12] ethtool: Add dump_one_dev callback for per-device sub-iteration
Date: Wed, 25 Mar 2026 19:20:31 +0100 [thread overview]
Message-ID: <267901bc-bde6-443b-bf78-8c73831fc886@bootlin.com> (raw)
In-Reply-To: <20260325145022.2607545-2-bjorn@kernel.org>
Hi Björn,
On 25/03/2026 15:50, Björn Töpel wrote:
> Add the dump_one_dev callback to ethnl_request_ops, allowing commands
> to provide custom per-device dump logic with sub-positioning. Extend
> ethnl_dump_ctx with ifindex and pos_sub fields.
>
> No functional change; no command uses dump_one_dev yet.
>
> Signed-off-by: Björn Töpel <bjorn@kernel.org>
Awesome :) I confirm that the phy dump still works just fine, and this
looks much simpler to me.
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Tested-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Maxime
> ---
> net/ethtool/netlink.c | 66 ++++++++++++++++++-------------------------
> net/ethtool/netlink.h | 31 ++++++++++++++++++++
> 2 files changed, 58 insertions(+), 39 deletions(-)
>
> diff --git a/net/ethtool/netlink.c b/net/ethtool/netlink.c
> index 5046023a30b1..8d161f0882d0 100644
> --- a/net/ethtool/netlink.c
> +++ b/net/ethtool/netlink.c
> @@ -346,36 +346,6 @@ int ethnl_multicast(struct sk_buff *skb, struct net_device *dev)
>
> /* GET request helpers */
>
> -/**
> - * struct ethnl_dump_ctx - context structure for generic dumpit() callback
> - * @ops: request ops of currently processed message type
> - * @req_info: parsed request header of processed request
> - * @reply_data: data needed to compose the reply
> - * @pos_ifindex: saved iteration position - ifindex
> - *
> - * These parameters are kept in struct netlink_callback as context preserved
> - * between iterations. They are initialized by ethnl_default_start() and used
> - * in ethnl_default_dumpit() and ethnl_default_done().
> - */
> -struct ethnl_dump_ctx {
> - const struct ethnl_request_ops *ops;
> - struct ethnl_req_info *req_info;
> - struct ethnl_reply_data *reply_data;
> - unsigned long pos_ifindex;
> -};
> -
> -/**
> - * struct ethnl_perphy_dump_ctx - context for dumpit() PHY-aware callbacks
> - * @ethnl_ctx: generic ethnl context
> - * @ifindex: For Filtered DUMP requests, the ifindex of the targeted netdev
> - * @pos_phyindex: iterator position for multi-msg DUMP
> - */
> -struct ethnl_perphy_dump_ctx {
> - struct ethnl_dump_ctx ethnl_ctx;
> - unsigned int ifindex;
> - unsigned long pos_phyindex;
> -};
> -
> static const struct ethnl_request_ops *
> ethnl_default_requests[__ETHTOOL_MSG_USER_CNT] = {
> [ETHTOOL_MSG_STRSET_GET] = ðnl_strset_request_ops,
> @@ -618,6 +588,7 @@ static int ethnl_default_dumpit(struct sk_buff *skb,
> struct netlink_callback *cb)
> {
> struct ethnl_dump_ctx *ctx = ethnl_dump_context(cb);
> + const struct genl_info *info = genl_info_dump(cb);
> struct net *net = sock_net(skb->sk);
> netdevice_tracker dev_tracker;
> struct net_device *dev;
> @@ -625,10 +596,20 @@ static int ethnl_default_dumpit(struct sk_buff *skb,
>
> rcu_read_lock();
> for_each_netdev_dump(net, dev, ctx->pos_ifindex) {
> + if (ctx->ifindex && ctx->ifindex != ctx->pos_ifindex)
> + break;
> +
> netdev_hold(dev, &dev_tracker, GFP_ATOMIC);
> rcu_read_unlock();
>
> - ret = ethnl_default_dump_one(skb, dev, ctx, genl_info_dump(cb));
> + if (ctx->ops->dump_one_dev) {
> + ctx->req_info->dev = dev;
> + ret = ctx->ops->dump_one_dev(skb, ctx, &ctx->pos_sub,
> + info);
> + ctx->req_info->dev = NULL;
> + } else {
> + ret = ethnl_default_dump_one(skb, dev, ctx, info);
> + }
>
> rcu_read_lock();
> netdev_put(dev, &dev_tracker);
> @@ -674,19 +655,26 @@ static int ethnl_default_start(struct netlink_callback *cb)
> ret = ethnl_default_parse(req_info, &info->info, ops, false);
> if (ret < 0)
> goto free_reply_data;
> - if (req_info->dev) {
> - /* We ignore device specification in dump requests but as the
> - * same parser as for non-dump (doit) requests is used, it
> - * would take reference to the device if it finds one
> - */
> - netdev_put(req_info->dev, &req_info->dev_tracker);
> - req_info->dev = NULL;
> - }
>
> ctx->ops = ops;
> ctx->req_info = req_info;
> ctx->reply_data = reply_data;
> ctx->pos_ifindex = 0;
> + ctx->ifindex = 0;
> + ctx->pos_sub = 0;
> +
> + if (req_info->dev) {
> + if (ops->dump_one_dev) {
> + /* Sub-iterator dumps keep track of the dev's ifindex
> + * so the dumpit handler can grab/release the netdev
> + * per iteration.
> + */
> + ctx->ifindex = req_info->dev->ifindex;
> + ctx->pos_ifindex = ctx->ifindex;
> + }
> + netdev_put(req_info->dev, &req_info->dev_tracker);
> + req_info->dev = NULL;
> + }
>
> return 0;
>
> diff --git a/net/ethtool/netlink.h b/net/ethtool/netlink.h
> index aaf6f2468768..e01adc5db02f 100644
> --- a/net/ethtool/netlink.h
> +++ b/net/ethtool/netlink.h
> @@ -10,6 +10,28 @@
>
> struct ethnl_req_info;
>
> +/**
> + * struct ethnl_dump_ctx - context structure for generic dumpit() callback
> + * @ops: request ops of currently processed message type
> + * @req_info: parsed request header of processed request
> + * @reply_data: data needed to compose the reply
> + * @pos_ifindex: saved iteration position - ifindex
> + * @ifindex: for filtered dump requests, the ifindex of the targeted netdev
> + * @pos_sub: iterator position for per-device iteration
> + *
> + * These parameters are kept in struct netlink_callback as context preserved
> + * between iterations. They are initialized by ethnl_default_start() and used
> + * in ethnl_default_dumpit() and ethnl_default_done().
> + */
> +struct ethnl_dump_ctx {
> + const struct ethnl_request_ops *ops;
> + struct ethnl_req_info *req_info;
> + struct ethnl_reply_data *reply_data;
> + unsigned long pos_ifindex;
> + unsigned int ifindex;
> + unsigned long pos_sub;
> +};
> +
> u32 ethnl_bcast_seq_next(void);
> int ethnl_parse_header_dev_get(struct ethnl_req_info *req_info,
> const struct nlattr *nest, struct net *net,
> @@ -365,6 +387,10 @@ int ethnl_sock_priv_set(struct sk_buff *skb, struct net_device *dev, u32 portid,
> * used e.g. to free any additional data structures outside the main
> * structure which were allocated by ->prepare_data(). When processing
> * dump requests, ->cleanup() is called for each message.
> + * @dump_one_dev:
> + * Optional callback for dumping data for a single device. When set,
> + * overrides the default dump behavior for GET requests, allowing
> + * per-device iteration with sub-positioning via @pos_sub.
> * @set_validate:
> * Check if set operation is supported for a given device, and perform
> * extra input checks. Expected return values:
> @@ -409,6 +435,11 @@ struct ethnl_request_ops {
> const struct ethnl_reply_data *reply_data);
> void (*cleanup_data)(struct ethnl_reply_data *reply_data);
>
> + int (*dump_one_dev)(struct sk_buff *skb,
> + struct ethnl_dump_ctx *ctx,
> + unsigned long *pos_sub,
> + const struct genl_info *info);
> +
> int (*set_validate)(struct ethnl_req_info *req_info,
> struct genl_info *info);
> int (*set)(struct ethnl_req_info *req_info,
next prev parent reply other threads:[~2026-03-25 18:20 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-25 14:50 [PATCH net-next v2 00/12] ethtool: Generic loopback support Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 01/12] ethtool: Add dump_one_dev callback for per-device sub-iteration Björn Töpel
2026-03-25 18:20 ` Maxime Chevallier [this message]
2026-03-25 14:50 ` [PATCH net-next v2 02/12] ethtool: Convert per-PHY commands to dump_one_dev Björn Töpel
2026-03-25 18:21 ` Maxime Chevallier
2026-03-25 14:50 ` [PATCH net-next v2 03/12] ethtool: Add loopback netlink UAPI definitions Björn Töpel
2026-03-26 8:10 ` Maxime Chevallier
2026-03-26 8:55 ` Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 04/12] ethtool: Add loopback GET/SET netlink implementation Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 05/12] ethtool: Add CMIS loopback helpers for module loopback control Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 06/12] selftests: drv-net: Add loopback driver test Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 07/12] ethtool: Add MAC loopback support via ethtool_ops Björn Töpel
2026-03-26 9:49 ` Breno Leitao
2026-03-25 14:50 ` [PATCH net-next v2 08/12] netdevsim: Add MAC loopback simulation Björn Töpel
2026-03-26 9:40 ` Breno Leitao
2026-03-25 14:50 ` [PATCH net-next v2 09/12] selftests: drv-net: Add MAC loopback netdevsim test Björn Töpel
2026-03-26 9:32 ` Breno Leitao
2026-03-26 9:44 ` Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 10/12] MAINTAINERS: Add entry for ethtool loopback Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 11/12] netdevsim: Add module EEPROM simulation via debugfs Björn Töpel
2026-03-25 14:50 ` [PATCH net-next v2 12/12] selftests: drv-net: Add CMIS loopback netdevsim test Björn Töpel
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=267901bc-bde6-443b-bf78-8c73831fc886@bootlin.com \
--to=maxime.chevallier@bootlin.com \
--cc=andrew+netdev@lunn.ch \
--cc=bjorn@kernel.org \
--cc=danieller@nvidia.com \
--cc=davem@davemloft.net \
--cc=donald.hunter@gmail.com \
--cc=edumazet@google.com \
--cc=hkelam@marvell.com \
--cc=horms@kernel.org \
--cc=idosch@nvidia.com \
--cc=kees@kernel.org \
--cc=kory.maincent@bootlin.com \
--cc=kuba@kernel.org \
--cc=leon@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linux@armlinux.org.uk \
--cc=michael.chan@broadcom.com \
--cc=naveenm@marvell.com \
--cc=netdev@vger.kernel.org \
--cc=o.rempel@pengutronix.de \
--cc=pabeni@redhat.com \
--cc=pavan.chebbi@broadcom.com \
--cc=piergiorgio.beruto@gmail.com \
--cc=saeedm@nvidia.com \
--cc=shuah@kernel.org \
--cc=tariqt@nvidia.com \
--cc=willemb@google.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