linux-kernel - Re: [PATCH 1/2] thermal/core: Hardening the self-encapsulation

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: <ac2696c0-d4f5-4222-9b54-f574ec2049f4@arm.com>
Date:   Fri, 13 Oct 2023 09:03:56 +0100
From:   Lukasz Luba <lukasz.luba@....com>
To:     Daniel Lezcano <daniel.lezcano@...aro.org>
Cc:     Thierry Reding <thierry.reding@...il.com>, rafael@...nel.org,
        Amit Kucheria <amitk@...nel.org>,
        Zhang Rui <rui.zhang@...el.com>,
        "open list:THERMAL" <linux-pm@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] thermal/core: Hardening the self-encapsulation



On 10/12/23 14:14, Daniel Lezcano wrote:
> 
> Hi Lukasz,
> 
> On 12/10/2023 14:01, Lukasz Luba wrote:
>> Hi Daniel,
>>
>> On 10/12/23 11:26, Daniel Lezcano wrote:
>>> The thermal private header has leaked all around the drivers which
>>> interacted with the core internals. The thermal zone structure which
>>> was part of the exported header led also to a leakage of the fields
>>> into the different drivers, making very difficult to improve the core
>>> code without having to change the drivers.
>>>
>>> Now we mostly fixed how the thermal drivers were interacting with the
>>> thermal zones (actually fixed how they should not interact). The
>>> thermal zone structure will be moved to the private thermal core
>>> header. This header has been removed from the different drivers and
>>> must belong to the core code only. In order to prevent this private
>>> header to be included again in the drivers, make explicit only the
>>> core code can include this header by defining a THERMAL_CORE_SUBSYS
>>> macro. The private header will contain a check against this macro.
>>>
>>> The Tegra SoCtherm driver needs to access thermal_core.h to have the
>>> get_thermal_instance() function definition. It is the only one
>>> remaining driver which need to access the thermal_core.h header, so
>>> the check will emit a warning at compilation time.
>>>
>>> Thierry Reding is reworking the driver to get rid of this function [1]
>>> and thus when the changes will be merged, the compilation warning will
>>> be converted to a compilation error, closing definitively the door to
>>> the drivers willing to play with the thermal zone device internals.
>>
>> That looks like a good idea. Although, shouldn't we avoid the
>> compilation warnings and just first merge the fixes for drivers?
> 
> Yes, we should but there is the series for nvidia (pointed in the 
> changelog) which need a slight refresh for the bindings AFAIR. That 
> series is since March 2023 and Thierry seems busy [1]. I'm holding the 
> hardening since then.
> 
> So I don't know how to make progress on this? I was assuming we can 
> merge this series and let the compiler recall what has to be fixed.
> 
> [1] https://lore.kernel.org/all/ZK14edZUih1kH_sZ@orome/
> 
> and as soon as it is fixed, we convert the WARNING to ERROR :P
> 
> 
> 
> 

OK, so that should be good for possible distro configs IMO.
The chicken egg problem should be addressed and we can start from
this patch approach.

Reviewed-by: Lukasz Luba <lukasz.luba@....com>