linux-kernel - Re: [PATCH 2/2] cgroup: Fix reference count leak when cft->open is NULL

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]

Message-ID: <vlgk7nn72odfg4xk34yagrvtwlqb3qkq24wbsrvjctd4upa5vm@7jg2iy7pkqof>
Date: Fri, 18 Jul 2025 14:55:21 +0200
From: Michal Koutný <mkoutny@...e.com>
To: Zijiang Huang <huangzjsmile@...il.com>
Cc: tj@...nel.org, hannes@...xchg.org, cgroups@...r.kernel.org, 
	linux-kernel@...r.kernel.org, Zijiang Huang <kerayhuang@...cent.com>, 
	Hao Peng <flyingpeng@...cent.com>
Subject: Re: [PATCH 2/2] cgroup: Fix reference count leak when cft->open is
 NULL

On Fri, Jul 18, 2025 at 07:54:09PM +0800, Zijiang Huang <huangzjsmile@...il.com> wrote:
> @@ -4134,8 +4134,10 @@ static int cgroup_file_open(struct kernfs_open_file *of)
>  	get_cgroup_ns(ctx->ns);
>  	of->priv = ctx;
>  
> -	if (!cft->open)
> +	if (!cft->open) {
> +		get_cgroup_ns(ctx->ns);
>  		return 0;
> +	}
>  
>  	ret = cft->open(of);
>  	if (ret) {

1) You wanted to call put_cgroup_ns() instead of get_cgroup_ns()
2) The refernce needs to be kept during the whole lifetime of
   cgroup_file_ctx, this return path still leads to a valid ctx, so it's
   released in cgroup_file_release().

Or could you decribe more how could a release be missed?

Thanks,
Michal

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)