linux-kernel - Re: [PATCH 0/4] allow multiple kthreadd's

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

Message-ID: <20200501192057.GH5462@mtj.thefacebook.com>
Date:   Fri, 1 May 2020 15:20:57 -0400
From:   "tj@...nel.org" <tj@...nel.org>
To:     Trond Myklebust <trondmy@...merspace.com>
Cc:     "bfields@...hat.com" <bfields@...hat.com>,
        "torvalds@...ux-foundation.org" <torvalds@...ux-foundation.org>,
        "jlayton@...hat.com" <jlayton@...hat.com>,
        "linux-nfs@...r.kernel.org" <linux-nfs@...r.kernel.org>,
        "shli@...com" <shli@...com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "dhowells@...hat.com" <dhowells@...hat.com>,
        "oleg@...hat.com" <oleg@...hat.com>
Subject: Re: [PATCH 0/4] allow multiple kthreadd's

Hello,

On Fri, May 01, 2020 at 07:05:46PM +0000, Trond Myklebust wrote:
> Wen running an instance of knfsd from inside a container, you want to
> be able to have the knfsd kthreads be parented to the container init
> process so that they get killed off when the container is killed.
> 
> Right now, we can easily leak those kernel threads simply by killing
> the container.

Hmm... I'm not sure that'd be a lot easier because they're in their own
thread group. It's not like you can queue signal to the group leader cause
the other kthreads to automatically die. Each would have to handle the exit
explicitly. As long as there is a way to iterate the member kthreads (e.g.
list going through kthread_data or sth else hanging off there), just using
existing kthread interface might not be much different, or maybe even easier
given how hairy signal handling in kthreads can get.

Thanks.

-- 
tejun