Netdev List
 help / color / mirror / Atom feed
From: Ido Schimmel <idosch@idosch.org>
To: Aya Levin <ayal@mellanox.com>
Cc: "David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>, Jiri Pirko <jiri@mellanox.com>,
	netdev@vger.kernel.org, Moshe Shemesh <moshe@mellanox.com>,
	Eran Ben Elisha <eranbe@mellanox.com>,
	Ido Schimmel <idosch@mellanox.com>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH net-next RFC v1 3/4] devlink: Add hierarchy between traps in device level and port level
Date: Sun, 6 Sep 2020 18:58:58 +0300	[thread overview]
Message-ID: <20200906155858.GB2431016@shredder> (raw)
In-Reply-To: <1599060734-26617-4-git-send-email-ayal@mellanox.com>

On Wed, Sep 02, 2020 at 06:32:13PM +0300, Aya Levin wrote:
> Managing large scale port's traps may be complicated. This patch
> introduces a shortcut: when setting a trap on a device and this trap is
> not registered on this device, the action will take place on all related
> ports that did register this trap.

I'm not really a fan of this and I'm not sure there is precedent for
something similar. Also, it's an optimization, so I wouldn't put it as
part of the first submission before you gather some operational
experience with the initial interface.

In addition, I find it very unintuitive for users. When I do 'devlink
trap show' I will not see anything. I will only see the traps when I
issue 'devlink port trap show', yet 'devlink trap set ...' is expected
to work.

Lets assume that this is a valid change, it would be better implemented
with my suggestion from the previous patch: When devlink sees that a
trap is registered on all the ports it can auto-register a new
per-device trap and user space gets the appropriate notification.

> 
> Signed-off-by: Aya Levin <ayal@mellanox.com>
> ---
>  net/core/devlink.c | 43 +++++++++++++++++++++++++++++++++----------
>  1 file changed, 33 insertions(+), 10 deletions(-)
> 
> diff --git a/net/core/devlink.c b/net/core/devlink.c
> index b13e1b40bf1c..dea5482b2517 100644
> --- a/net/core/devlink.c
> +++ b/net/core/devlink.c
> @@ -6501,23 +6501,46 @@ static int devlink_nl_cmd_trap_set_doit(struct sk_buff *skb,
>  	struct devlink *devlink = info->user_ptr[0];
>  	struct devlink_trap_mngr *trap_mngr;
>  	struct devlink_trap_item *trap_item;
> +	struct devlink_port *devlink_port;
>  	int err;
>  
> -	trap_mngr = devlink_trap_get_trap_mngr_from_info(devlink, info);
> -	if (list_empty(&trap_mngr->trap_list))
> -		return -EOPNOTSUPP;
> +	devlink_port = devlink_port_get_from_attrs(devlink, info->attrs);
> +	if (IS_ERR(devlink_port)) {
> +		trap_mngr =  &devlink->trap_mngr;
> +		if (list_empty(&trap_mngr->trap_list))
> +			goto loop_over_ports;
>  
> -	trap_item = devlink_trap_item_get_from_info(trap_mngr, info);
> -	if (!trap_item) {
> -		NL_SET_ERR_MSG_MOD(extack, "Device did not register this trap");
> -		return -ENOENT;
> +		trap_item = devlink_trap_item_get_from_info(trap_mngr, info);
> +		if (!trap_item)
> +			goto loop_over_ports;
> +	} else {
> +		trap_mngr = &devlink_port->trap_mngr;
> +		if (list_empty(&trap_mngr->trap_list))
> +			return -EOPNOTSUPP;
> +
> +		trap_item = devlink_trap_item_get_from_info(trap_mngr, info);
> +		if (!trap_item) {
> +			NL_SET_ERR_MSG_MOD(extack, "Port did not register this trap");
> +			return -ENOENT;
> +		}
>  	}
>  	return devlink_trap_action_set(devlink, trap_mngr, trap_item, info);
>  
> -	err = devlink_trap_action_set(devlink, trap_mngr, trap_item, info);
> -	if (err)
> -		return err;
> +loop_over_ports:
> +	if (list_empty(&devlink->port_list))
> +		return -EOPNOTSUPP;
> +	list_for_each_entry(devlink_port, &devlink->port_list, list) {
> +		trap_mngr = &devlink_port->trap_mngr;
> +		if (list_empty(&trap_mngr->trap_list))
> +			continue;
>  
> +		trap_item = devlink_trap_item_get_from_info(trap_mngr, info);
> +		if (!trap_item)
> +			continue;
> +		err = devlink_trap_action_set(devlink, trap_mngr, trap_item, info);
> +		if (err)
> +			return err;
> +	}
>  	return 0;
>  }
>  
> -- 
> 2.14.1
> 

  reply	other threads:[~2020-09-06 15:59 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-02 15:32 [PATCH net-next RFC v1 0/4] Add devlink traps in devlink port context Aya Levin
2020-09-02 15:32 ` [PATCH net-next RFC v1 1/4] devlink: Wrap trap related lists and ops in trap_mngr Aya Levin
2020-09-08 14:12   ` Jiri Pirko
2020-09-02 15:32 ` [PATCH net-next RFC v1 2/4] devlink: Add devlink traps under devlink_ports context Aya Levin
2020-09-06 15:44   ` Ido Schimmel
2020-09-07 16:52     ` Aya Levin
2020-09-08 14:04     ` Jiri Pirko
2020-09-10  6:16       ` Aya Levin
2020-09-02 15:32 ` [PATCH net-next RFC v1 3/4] devlink: Add hierarchy between traps in device level and port level Aya Levin
2020-09-06 15:58   ` Ido Schimmel [this message]
2020-09-07 16:26     ` Aya Levin
2020-09-02 15:32 ` [PATCH net-next RFC v1 4/4] net/mlx5e: Add devlink trap to catch oversize packets Aya Levin

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=20200906155858.GB2431016@shredder \
    --to=idosch@idosch.org \
    --cc=ayal@mellanox.com \
    --cc=davem@davemloft.net \
    --cc=eranbe@mellanox.com \
    --cc=idosch@mellanox.com \
    --cc=jiri@mellanox.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=moshe@mellanox.com \
    --cc=netdev@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox