linux-kernel - Re: [PATCH] softirq: fix memory corruption when freeing tasklet

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]

Date: Mon, 29 Jan 2024 07:06:46 -1000
From: Tejun Heo <tj@...nel.org>
To: Allen <allen.lkml@...il.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
	Mikulas Patocka <mpatocka@...hat.com>,
	Thomas Gleixner <tglx@...utronix.de>, linux-kernel@...r.kernel.org,
	dm-devel@...ts.linux.dev, Mike Snitzer <msnitzer@...hat.com>,
	Ignat Korchagin <ignat@...udflare.com>,
	Damien Le Moal <damien.lemoal@....com>,
	Bob Liu <bob.liu@...cle.com>, Hou Tao <houtao1@...wei.com>,
	Nathan Huckleberry <nhuck@...gle.com>,
	Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...e.hu>
Subject: Re: [PATCH] softirq: fix memory corruption when freeing
 tasklet_struct

Hello,

On Mon, Jan 29, 2024 at 09:00:38AM -0800, Allen wrote:
>     I rushed to reply to the draft patch you sent, I should have
> looked harder. My apologies.
> The idea that I have been working on is to completely move away from
> using tasklets.
> Essentially, "get rid of tasklets entirely in the kernel". So, the use
> of tasklet_schedule() & tasklet_hi_schedule()
> will have to go.

The idea is to take over the tasklet[_hi] softirqs once all users are
converted. Otherwise, we run into the problem of setting priorities between
tasklets and the atomic workqueue, which may be a theoretical problem.

>    I have a very hacky draft that is still wip. I am going to borrow
> many bits from your patch which makes
> the work I have better.
> 
>   Perhaps we should start a separate thread, thoughts?

Sure, please go ahead.

Thanks.

-- 
tejun