linux-kernel - Re: New kernel warning after updating from LTS 5.15.110 to 5.15.112 (and 5.15.113)

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:   Wed, 7 Jun 2023 03:43:18 +0200
From:   Lino Sanfilippo <l.sanfilippo@...bus.com>
To:     Bagas Sanjaya <bagasdotme@...il.com>,
        Chris Packham <Chris.Packham@...iedtelesis.co.nz>,
        Jarkko Sakkinen <jarkko@...nel.org>,
        Sasha Levin <sashal@...nel.org>,
        Peter Huewe <peterhuewe@....de>, Jason Gunthorpe <jgg@...pe.ca>
Cc:     Linux Kernel Integrity <linux-integrity@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Linux Regressions <regressions@...ts.linux.dev>,
        Lukas Wunner <lukas@...ner.de>,
        Philipp Rosenberger <p.rosenberger@...bus.com>
Subject: Re: New kernel warning after updating from LTS 5.15.110 to 5.15.112
 (and 5.15.113)

Hi Bagas, hi Chris

On 06.06.23 04:00, Bagas Sanjaya wrote:

> On Tue, Jun 06, 2023 at 01:41:01AM +0000, Chris Packham wrote:
>>

>>
>> Bisecting between v5.15.110 and v5.15.112 points to
>>
>> 51162b05a44cb5d98fb0ae2519a860910a47fd4b is the first bad commit
> 
> Thanks for the bisection.
> 
> Lino, it looks like this regression is caused by (backported) commit of yours.
> Would you like to take a look on it?
> 

Before commit 51162b05a44c interrupt activation has failed since the concerning register was accessed
without holding the required locality.

Now with the commit applied the activation is successful and the interrupt handler is called as soon
as an interrupt fires. However the handler runs in interrupt context while the register accesses
are done via SPI which involves the SPI bus_lock_mutex. Calling the (sleepable) SPI functions in
interrupt context results in the observed warning.

To fix this additionally upstream commit 0c7e66e5fd69 ("tpm, tpm_tis: Request threaded interrupt handler") is
required, since it ensures that the handler runs in process context.

Note that even with this commit interrupts will eventually be disabled since the test for interrupts
still fails (for the test to succeed at least upstream commit e644b2f498d2 "tpm, tpm_tis: Enable interrupt test"
would be required).

Chris, could you test again with commit 0c7e66e5fd69 additionally applied and confirm that the warning is gone?

Regards,
Lino