linux-kernel - Re: [PATCH v2 0/4] ACPI: SBS: Fix various issues

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] [day] [month] [year] [list]

Message-ID: <1de16fdb-e6e2-57eb-b074-cdd5473172ba@gmx.de>
Date:   Wed, 15 Mar 2023 00:04:34 +0100
From:   Armin Wolf <W_Armin@....de>
To:     "Rafael J. Wysocki" <rafael@...nel.org>
Cc:     lenb@...nel.org, linux-acpi@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 0/4] ACPI: SBS: Fix various issues

Am 14.03.23 um 20:49 schrieb Rafael J. Wysocki:

> On Sun, Mar 12, 2023 at 6:15 PM Armin Wolf <W_Armin@....de> wrote:
>> Am 25.02.23 um 12:51 schrieb Armin Wolf:
>>
>>> On my Acer Travelmate 4002WLMi, the system locks up upon
>>> suspend/shutdown. After a lot of research, it turned out
>>> that the sbs module was the culprit. The driver would not
>>> correctly mask out the value used to select a battery using
>>> the "Smart Battery Selector" (subset of the "Smart Battery Manager").
>>> This accidentally caused a invalid power source to be selected,
>>> which was automatically corrected by the selector. Upon
>>> notifing the host about the corrected power source, some batteries
>>> would be selected for re-reading, causing a endless loop.
>>> This would lead to some workqueues filling up, which caused the
>>> lockup upon suspend/shutdown.
>>>
>>> The first three patches fix a stacktrace on module removal caused
>>> by some locking issues. The last patch finally fixes the
>>> suspend/shutdown issues.
>>>
>>> As a side note: This was the first machine on which i installed Linux,
>>> to finally fixing this took ~5 years of tinkering.
>> What is the status of this patchset? Should i use a SRCU notifier chain
>> for the query notifiers? I would really like to see this getting fixed,
>> as it prevents me from using linux on this machine.
> I'm not entirely convinced about the query notifiers idea TBH.

I already thought about just flushing the query workqueue, but acpi_ec_remove_query_handler()
would still also remove query handlers installed to handle _Qxx methods. This might cause errors,
as for example on the Acer Travelmate 4002WLMI, the SMBUS query handler (_Q20) resets the SMBUS
alert bit in case no EC SMBus driver is overriding it.
Is there any specific reason for your dislike of the query notifiers? I could turn them into
SRCU call chains to avoid performance issues.

Armin Wolf