From: Jacob Keller <jacob.e.keller@intel.com>
To: Jiri Pirko <jiri@resnulli.us>, <netdev@vger.kernel.org>
Cc: <stephen@networkplumber.org>, <dsahern@gmail.com>,
<kuba@kernel.org>, <moshe@nvidia.com>, <saeedm@nvidia.com>
Subject: Re: [patch iproute2/net-next 2/4] devlink: get devlink port for ifname using RTNL get link command
Date: Thu, 8 Dec 2022 11:00:01 -0800 [thread overview]
Message-ID: <bfaf41d2-b70b-12ef-ca79-bd96807c8fad@intel.com> (raw)
In-Reply-To: <20221205122158.437522-3-jiri@resnulli.us>
On 12/5/2022 4:21 AM, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@nvidia.com>
>
> Currently, when user specifies ifname as a handle on command line of
> devlink, the related devlink port is looked-up in previously taken dump
> of all devlink ports on the system. There are 3 problems with that:
> 1) The dump iterates over all devlink instances in kernel and takes a
> devlink instance lock for each.
> 2) Dumping all devlink ports would not scale.
> 3) Alternative ifnames are not exposed by devlink netlink interface.
>
Oh neat I wasn't even aware we could pass netdev device names! That's slick.
> Instead, benefit from RTNL get link command extension and get the
> devlink port handle info from IFLA_DEVLINK_PORT attribute, if supported.
>
Makes sense.
Code looks ok.
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
> Signed-off-by: Jiri Pirko <jiri@nvidia.com>
> ---
> devlink/devlink.c | 96 ++++++++++++++++++++++++++++++++++++++++++++---
> 1 file changed, 91 insertions(+), 5 deletions(-)
>
> diff --git a/devlink/devlink.c b/devlink/devlink.c
> index d224655cd0e9..80c18d690c10 100644
> --- a/devlink/devlink.c
> +++ b/devlink/devlink.c
> @@ -43,6 +43,8 @@
> #include "json_print.h"
> #include "utils.h"
> #include "namespace.h"
> +#include "libnetlink.h"
> +#include "../ip/ip_common.h"
>
> #define ESWITCH_MODE_LEGACY "legacy"
> #define ESWITCH_MODE_SWITCHDEV "switchdev"
> @@ -797,6 +799,81 @@ static void ifname_map_del(struct ifname_map *ifname_map)
> ifname_map_free(ifname_map);
> }
>
> +static int ifname_map_rtnl_port_parse(struct dl *dl, const char *ifname,
> + struct rtattr *nest)
> +{
> + struct rtattr *tb[DEVLINK_ATTR_MAX + 1];
> + const char *bus_name;
> + const char *dev_name;
> + uint32_t port_index;
> +
> + parse_rtattr_nested(tb, DEVLINK_ATTR_MAX, nest);
> + if (!tb[DEVLINK_ATTR_BUS_NAME] || !tb[DEVLINK_ATTR_DEV_NAME] ||
> + !tb[DEVLINK_ATTR_PORT_INDEX])
> + return -ENOENT;
> +
> + bus_name = rta_getattr_str(tb[DEVLINK_ATTR_BUS_NAME]);
> + dev_name = rta_getattr_str(tb[DEVLINK_ATTR_DEV_NAME]);
> + port_index = rta_getattr_u32(tb[DEVLINK_ATTR_PORT_INDEX]);
> + return ifname_map_add(dl, ifname, bus_name, dev_name, port_index);
> +}
> +
> +static int ifname_map_rtnl_init(struct dl *dl, const char *ifname)
> +{
> + struct iplink_req req = {
> + .n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifinfomsg)),
> + .n.nlmsg_flags = NLM_F_REQUEST,
> + .n.nlmsg_type = RTM_GETLINK,
> + .i.ifi_family = AF_UNSPEC,
> + };
> + struct rtattr *tb[IFLA_MAX + 1];
> + struct rtnl_handle rth;
> + struct ifinfomsg *ifi;
> + struct nlmsghdr *n;
> + int len;
> + int err;
> +
> + if (rtnl_open(&rth, 0) < 0) {
> + pr_err("Cannot open rtnetlink\n");
> + return -EINVAL;
> + }
> +
> + addattr_l(&req.n, sizeof(req),
> + !check_ifname(ifname) ? IFLA_IFNAME : IFLA_ALT_IFNAME,
> + ifname, strlen(ifname) + 1);
> +
> + if (rtnl_talk(&rth, &req.n, &n) < 0) {
> + err = -EINVAL;
> + goto out;
> + }
> +
> + if (n->nlmsg_type != RTM_NEWLINK) {
> + err = -EINVAL;
> + goto out;
> + }
> +
> + ifi = NLMSG_DATA(n);
> + len = n->nlmsg_len;
> +
> + len -= NLMSG_LENGTH(sizeof(*ifi));
> + if (len < 0) {
> + err = -EINVAL;
> + goto out;
> + }
> +
> + parse_rtattr_flags(tb, IFLA_MAX, IFLA_RTA(ifi), len, NLA_F_NESTED);
> + if (!tb[IFLA_DEVLINK_PORT]) {
> + err = -EOPNOTSUPP;
> + goto out;
> + }
> +
> + err = ifname_map_rtnl_port_parse(dl, ifname, tb[IFLA_DEVLINK_PORT]);
> +
> +out:
> + rtnl_close(&rth);
> + return err;
> +}
> +
> static int ifname_map_cb(const struct nlmsghdr *nlh, void *data)
> {
> struct nlattr *tb[DEVLINK_ATTR_MAX + 1] = {};
> @@ -842,11 +919,20 @@ static void ifname_map_init(struct dl *dl)
> INIT_LIST_HEAD(&dl->ifname_map_list);
> }
>
> -static int ifname_map_load(struct dl *dl)
> +static int ifname_map_load(struct dl *dl, const char *ifname)
> {
> struct nlmsghdr *nlh;
> int err;
>
> + if (ifname) {
> + err = ifname_map_rtnl_init(dl, ifname);
> + if (!err)
> + return 0;
> + /* In case kernel does not support devlink port info passed over
> + * RT netlink, fall-back to ports dump.
> + */
> + }
> +
> nlh = mnlu_gen_socket_cmd_prepare(&dl->nlg, DEVLINK_CMD_PORT_GET,
> NLM_F_REQUEST | NLM_F_ACK | NLM_F_DUMP);
>
> @@ -858,14 +944,14 @@ static int ifname_map_load(struct dl *dl)
> return 0;
> }
>
> -static int ifname_map_check_load(struct dl *dl)
> +static int ifname_map_check_load(struct dl *dl, const char *ifname)
> {
> int err;
>
> if (dl->map_loaded)
> return 0;
>
> - err = ifname_map_load(dl);
> + err = ifname_map_load(dl, ifname);
> if (err) {
> pr_err("Failed to create index map\n");
> return err;
> @@ -882,7 +968,7 @@ static int ifname_map_lookup(struct dl *dl, const char *ifname,
> struct ifname_map *ifname_map;
> int err;
>
> - err = ifname_map_check_load(dl);
> + err = ifname_map_check_load(dl, ifname);
> if (err)
> return err;
>
> @@ -905,7 +991,7 @@ static int ifname_map_rev_lookup(struct dl *dl, const char *bus_name,
>
> int err;
>
> - err = ifname_map_check_load(dl);
> + err = ifname_map_check_load(dl, NULL);
> if (err)
> return err;
>
next prev parent reply other threads:[~2022-12-08 19:00 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-05 12:21 [patch iproute2/net-next 0/4] devlink: optimize ifname handling Jiri Pirko
2022-12-05 12:21 ` [patch iproute2/net-next 1/4] devlink: add ifname_map_add/del() helpers Jiri Pirko
2022-12-05 12:21 ` [patch iproute2/net-next 2/4] devlink: get devlink port for ifname using RTNL get link command Jiri Pirko
2022-12-08 19:00 ` Jacob Keller [this message]
2022-12-05 12:21 ` [patch iproute2/net-next 3/4] devlink: push common code to __pr_out_port_handle_start_tb() Jiri Pirko
2022-12-05 12:21 ` [patch iproute2/net-next 4/4] devlink: update ifname map when message contains DEVLINK_ATTR_PORT_NETDEV_NAME Jiri Pirko
2022-12-08 17:50 ` [patch iproute2/net-next 0/4] devlink: optimize ifname handling 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=bfaf41d2-b70b-12ef-ca79-bd96807c8fad@intel.com \
--to=jacob.e.keller@intel.com \
--cc=dsahern@gmail.com \
--cc=jiri@resnulli.us \
--cc=kuba@kernel.org \
--cc=moshe@nvidia.com \
--cc=netdev@vger.kernel.org \
--cc=saeedm@nvidia.com \
--cc=stephen@networkplumber.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.