linux-kernel - Re: [PATCH] acpi: video: fix reversed indexed BQC

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: <CAMP44s2PhrgRH9vsn9r89A4+qwD3PZku35NUh6x18cNme7Go=w@mail.gmail.com>
Date:	Fri, 2 Aug 2013 06:20:12 -0500
From:	Felipe Contreras <felipe.contreras@...il.com>
To:	Aaron Lu <aaron.lu@...el.com>
Cc:	linux-kernel@...r.kernel.org, linux-acpi@...r.kernel.org,
	"Rafael J. Wysocki" <rjw@...k.pl>, Len Brown <lenb@...nel.org>,
	Zhang Rui <rui.zhang@...el.com>
Subject: Re: [PATCH] acpi: video: fix reversed indexed BQC

On Fri, Aug 2, 2013 at 3:25 AM, Aaron Lu <aaron.lu@...el.com> wrote:
> On 08/02/2013 04:14 PM, Felipe Contreras wrote:
>> On Fri, Aug 2, 2013 at 3:06 AM, Aaron Lu <aaron.lu@...el.com> wrote:
>>> On 08/02/2013 03:59 PM, Felipe Contreras wrote:
>>>> On Fri, Aug 2, 2013 at 1:56 AM, Aaron Lu <aaron.lu@...el.com> wrote:
>>>>> On 08/02/2013 02:44 PM, Felipe Contreras wrote:
>>>>
>>>>>> The initial _BCM commands don't work, so the level remains at 100%.
>>>>>> Since the level is max_level, acpi_video_bqc_quirk() tries with the
>>>>>> first level, which is 0, and 0 happens to be the index of 100.
>>>>>>
>>>>>> So _BQC is returning 100, which is not the index of 0 (what we tested
>>>>>> for), but actually 100.
>>>>>>
>>>>>> I think the current code is correct, but acpi_video_bqc_quirk() should
>>>>>> be testing br->levels[3], or anything other than 0/100 which can be
>>>>>> easily confused.
>>>>>>
>>>>>> If so, the code would find that _BQC doesn't work on this machine (in
>>>>>> win8 mode)... at least initially. My guess is that it only starts to
>>>>>> work after acpi_video_bus_start_devices() is called.
>>>>>>
>>>>>> Forcing br->flags._BQC_use_index = 0 seems to work.
>>>>>
>>>>> Seems ASUS machines tend to have this issue:
>>>>> https://bugzilla.kernel.org/show_bug.cgi?id=52951
>>>>> https://bugzilla.kernel.org/show_bug.cgi?id=56711
>>>>
>>>> I don't see any real solution for the ACPI driver.
>>>>
>>>>> I have a patch to enhance the quirk some time ago:
>>>>> https://github.com/aaronlu/linux/commit/0a3d2c5b59caf80ae5bb1ca1fda0f7bf448b38c9
>>>>
>>>> I think this is unnecessarily complicated; the comment makes it clear
>>>
>>> For your system, yes it is unnecessarily complicated. But since this is
>>> a quirk, it better solves as many potential problems as possible, or we
>>> would simply use a DMI entry to do the quirk.
>>
>> The only difference between my patch and yours is that your patch
>> checks that br->level[i] is not the current level, but that check is
>> not necessary. If _BQC always returns the max level, all we need to do
>
> _BQC does not always returns the max level.
>
>> is pick another value, any other value, and br->level[3] works just
>> fine.
>
> For a _BCL only having 4 elements { 100, 40, 40, 100 }, the br->levels[3]
> will be the max level. The example here may be too crazy to be true, but
> since we are dealing with firmware, I tend to believe anything could
> happen.

That can be fixed easily by checking test_level == current_level and
do the same your patch does, but I actually don't think we should do
that. It might make sense to test two values instead of only one, that
way we can we properly test _BQC, while your patch simply assumes it
doesn't work, even though it might.

-- 
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/