netdev - Re: Incomplete fix for recent bug in tc / hfsc

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: <aF847kk6H+kr5kIV@pop-os.localdomain>
Date: Fri, 27 Jun 2025 17:35:58 -0700
From: Cong Wang <xiyou.wangcong@...il.com>
To: Lion Ackermann <nnamrec@...il.com>
Cc: netdev@...r.kernel.org, Jamal Hadi Salim <jhs@...atatu.com>,
	Jiri Pirko <jiri@...nulli.us>
Subject: Re: Incomplete fix for recent bug in tc / hfsc

On Tue, Jun 24, 2025 at 12:43:27PM +0200, Lion Ackermann wrote:
> Actually I was intrigued, what do you think about addressing the root of the
> use-after-free only and ignore the backlog corruption (kind of). After the 
> recent patches where qlen_notify may get called multiple times, we could simply
> loosen qdisc_tree_reduce_backlog to always notify when the qdisc is empty.
> Since deletion of all qdiscs will run qdisc_reset / qdisc_purge_queue at one
> point or another, this should always catch left-overs. And we need not care
> about all the complexities involved of keeping the backlog right and / or
> prevent certain hierarchies which seems rather tedious.
> This requires some more testing, but I was imagining something like this:

I like your patch which looks really clean, in fact I still have
troubles to totally understand the cases you removed by your patch.

Could you tested it with all tdc test cases? If they all pass, we can
feel confident. Of course, also make sure it fixes the problem you
reported here.

Thanks!