netdev - Re: [PATCH v3] net/sched: sch_qfq: Fix race condition on qfq

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: <aHaaQ1aSVt6vSQlT@pop-os.localdomain>
Date: Tue, 15 Jul 2025 11:13:23 -0700
From: Cong Wang <xiyou.wangcong@...il.com>
To: Xiang Mei <xmei5@....edu>
Cc: Jakub Kicinski <kuba@...nel.org>, netdev@...r.kernel.org,
	gregkh@...uxfoundation.org, jhs@...atatu.com, jiri@...nulli.us,
	security@...nel.org
Subject: Re: [PATCH v3] net/sched: sch_qfq: Fix race condition on
 qfq_aggregate

On Mon, Jul 14, 2025 at 05:09:42PM -0700, Xiang Mei wrote:
> 
> Here is more information no how I tested:
> 
> 1) I ran `python3 ./tdc.py -f ./tc-tests/infra/qdiscs.json -e 5e6d` 100
> times
> 2) The KASAN is enabled, and my patch is on it
> 3) All 100 results show `ok 1 5e6d - Test QFQ's enqueue reentrant behaviour
> with netem` without any crashing in dmesg
> 
> I may need more information to trace this crash.

Now I figured out why... It is all because of I used a wrong vmlinux to
test this. Although I switched to vanilla -net branch, I forgot to
rebuild the vmlinux which was still the one with my netem patches. And I
just saw "netem duplicate 100%" in test case 5e6d, now it explains
everything.

Appologize for my stupid mistake here. I think it is clearly caused by
my netem duplication patch (although the fix is not necessarily there).

I will take care of this in my netem patchset.

Sorry for the noise.