linux-kernel - Re: WARNING: at kernel/workqueue.c:1473 __queue

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: <20200220090350.GA19858@Red>
Date:   Thu, 20 Feb 2020 10:03:50 +0100
From:   Corentin Labbe <clabbe.montjoie@...il.com>
To:     Daniel Jordan <daniel.m.jordan@...cle.com>
Cc:     tj@...nel.org, jiangshanlai@...il.com, will@...nel.org,
        mark.rutland@....com, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, linux-crypto@...r.kernel.org
Subject: Re: WARNING: at kernel/workqueue.c:1473 __queue_work+0x3b8/0x3d0

On Tue, Feb 18, 2020 at 11:35:04AM -0500, Daniel Jordan wrote:
> Hi Corentin,
> 
> On Mon, Feb 17, 2020 at 09:48:03PM +0100, Corentin Labbe wrote:
> > When running some CI test jobs (targeting crypto tests), I always get the following WARNING:
> 
> Can you be more specific about which test triggers this?  I used the config
> option you mention and failed to reproduce after doing the boot-time crypto
> tests and running various tcrypt incantations.
> 

Hello

It appears before any user space start. But according to the "Modules linked", probably ghash is already loaded and perhaps tested.

Removing GHASH lead to:
[    7.920931] ------------[ cut here ]------------
[    7.920955] WARNING: CPU: 1 PID: 120 at kernel/workqueue.c:1469 __queue_work+0x370/0x388
[    7.920960] Modules linked in: ccm

And removing CCM lead to
[    7.798877] ------------[ cut here ]------------
[    7.798902] WARNING: CPU: 2 PID: 127 at kernel/workqueue.c:1469 __queue_work+0x370/0x388
[    7.798907] Modules linked in: ctr

So it confirm that the problem is not related to the tested crypto algorithm.

> > [    7.886361] ------------[ cut here ]------------
> > [    7.886388] WARNING: CPU: 2 PID: 147 at kernel/workqueue.c:1473 __queue_work+0x3b8/0x3d0
> > [    7.886394] Modules linked in: ghash_generic
> > [    7.886409] CPU: 2 PID: 147 Comm: modprobe Not tainted 5.6.0-rc1-next-20200214-00068-g166c9264f0b1-dirty #545
> 
> I was using just plain next-20200214.  Can't find 166c9264f0b1, what tag/branch
> were you using exactly?
> 

The pasted example has some commit to try to debug it.
But I got the same with plain next (like yesterday 5.6.0-rc2-next-20200219 and tomorow 5.6.0-rc2-next-20200220) and master got the same issue.

But for reproductability on different hardware, I agree it is difficult.
For the moment, I got it only on Allwinner H5, A64, H6 SoCs and imx8q.
[    6.611449] ------------[ cut here ]------------
[    6.613234] WARNING: CPU: 1 PID: 157 at /srv/data/clabbe/linux-next/kernel/workqueue.c:1471 __queue_work+0x324/0x3b0
[    6.623809] Modules linked in: ghash_generic
[    6.628101] CPU: 1 PID: 157 Comm: modprobe Not tainted 5.6.0-rc2-next-20200220 #82
[    6.635710] Hardware name: NXP i.MX8MNano DDR4 EVK board (DT)

I tried amlogic boards and some qemu "virt" without success.

(I have added linux-crypto@...r.kernel.org to the CC)