linux-kernel - Re: [PATCH] connector/cn_proc: cn_netlink_has_listeners replaces proc_event_num

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date: Thu, 08 Feb 2024 10:46:09 +0100
From: Paolo Abeni <pabeni@...hat.com>
To: Keqi Wang <wangkeqi_chris@....com>, davem@...emloft.net, 
	edumazet@...gle.com, kuba@...nel.org
Cc: netdev@...r.kernel.org, linux-kernel@...r.kernel.org, kernel test robot
	 <oliver.sang@...el.com>, Anjali Kulkarni <anjali.k.kulkarni@...cle.com>
Subject: Re: [PATCH] connector/cn_proc: cn_netlink_has_listeners replaces
 proc_event_num_listeners

On Sun, 2024-02-04 at 16:22 +0800, Keqi Wang wrote:
> It is not accurate to reset proc_event_num_listeners according to
> cn_netlink_send_mult() return value -ESRCH.
> 
> In the case of stress-ng netlink-proc, -ESRCH will always be returned,
> because netlink_broadcast_filtered will return -ESRCH,
> which may cause stress-ng netlink-proc performance degradation.
> 
> proc_event_num_listeners cannot accurately reflect whether
> the listener exists, so add cn_netlink_has_listeners() functon
> and use that instead of proc_event_num_listeners.

Adding Anjali for awareness, as the last active developer in this area.

This change looks really invasive for a net fix. It's not trivially
clear to me if the change of accounting schema will bring or not
unintended functional/behavioural changes.

What about a simple revert of the blamed commit, and target net-next
with this kind of change?

> @@ -396,10 +404,10 @@ static void cn_proc_ack(int err, int rcvd_seq, int rcvd_ack)
>  static void cn_proc_mcast_ctl(struct cn_msg *msg,
>  			      struct netlink_skb_parms *nsp)
>  {
> -	enum proc_cn_mcast_op mc_op = 0, prev_mc_op = 0;
> +	enum proc_cn_mcast_op mc_op = 0;
>  	struct proc_input *pinput = NULL;
>  	enum proc_cn_event ev_type = 0;
> -	int err = 0, initial = 0;
> +	int err = 0;
>  	struct sock *sk = NULL;

Minor nit: the above breaks the reverse xmas tree order

Cheers,

Paolo