linux-kernel - Re: [PATCH v4 0/2] hid-asus: asus-wmi: refactor Ally suspend/resume

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: <41bf9d87-be11-4814-bc3e-c6c9297e0cc4@ljones.dev>
Date: Mon, 24 Mar 2025 23:33:52 +1300
From: "Luke D. Jones" <luke@...nes.dev>
To: Antheas Kapenekakis <lkml@...heas.dev>
Cc: linux-kernel@...r.kernel.org, hdegoede@...hat.com,
 ilpo.jarvinen@...ux.intel.com, platform-driver-x86@...r.kernel.org,
 linux-input@...r.kernel.org, bentiss@...nel.org, jikos@...nel.org,
 mario.limonciello@....com
Subject: Re: [PATCH v4 0/2] hid-asus: asus-wmi: refactor Ally suspend/resume

On 24/03/25 21:11, Antheas Kapenekakis wrote:
> On Mon, 24 Mar 2025 at 02:41, Luke D. Jones <luke@...nes.dev> wrote:
>>
>> On 24/03/25 00:41, Antheas Kapenekakis wrote:
>>> On Sun, 23 Mar 2025 at 03:34, Luke Jones <luke@...nes.dev> wrote:
>>>>
>>>> This short series refactors the Ally suspend/resume functionality in the
>>>> asus-wmi driver along with adding support for ROG Ally MCU version checking.
>>>>
>>>> The version checking is then used to toggle the use of older CSEE call hacks
>>>> that were initially used to combat Ally suspend/wake issues arising from the MCU
>>>> not clearing a particular flag on resume. ASUS have since corrected this
>>>> especially for Linux in newer firmware versions.
>>>>
>>>> - hid-asus requests the MCU version and displays a warning if the version is
>>>>     older than the one that fixes the issue.
>>>> - hid-asus awill also toggle the CSEE hack off, and mcu_powersave to on if the
>>>> version is high enough.
>>>>
>>>> *Note: In review it was requested by Mario that I try strsep() for parsing
>>>> the version. I did try this and a few variations but the result was much
>>>> more code due to having to check more edge cases due to the input being
>>>> raw bytes. In the end the cleaned up while loop proved more robust.
>>>>
>>>> - Changelog:
>>>>     + V2: https://lore.kernel.org/platform-driver-x86/20250226010129.32043-1-luke@ljones.dev/T/#t
>>>>       - Adjust warning message to explicitly mention suspend issues
>>>
>>> How did the testing go with this one, especially with mcu_powersave 0?
>>
>> Appears to be good. Checked a few reboots with powersave off - it is
>> setting on as I expect every time. Did modules unload/load also. And
>> tested with it set off after boot plus suspend resumes.
> 
> Did you test suspends with mcu_powersave to 0 and rgb on? I had a few
> issues you can reference the previous version for and I want to see if
> you have them.
> 
> Even with powersave set to 1, the RGB does not fade anymore without the quirk

Yes I tested every scenario I could think of. I don't think the fade is 
something to worry about - seems like it happening at all previously was 
just due to suspend being held up for a bit longer and now that the hack 
is disabled for new FW, it relies fully on Linux suspend (async? 
Honestly it's never been fully clear how async it really is).

I'd rather the faster suspend/resume. And so far I've heard no 
complaints (although my userbase is smaller than bazzites).

Cheers,
Luke.

> Antheas
> 
>> Very much hope this is the end of that particular saga, and with
>> bazzites help we can hopefully get everyone on November MCU FW or later,
>> then finally remove the hack completely this year.
>>
>> A small side note - I expect ASUS to fully reuse the X hardware, or at
>> least the bios/acpi/mcu-fw for that new windows handheld they've doing,
>> so fingers crossed that they actually do, and there will be nomore
>> suspend issues with current kernels plus this patch.
>>
>> Cheers,
>> Luke.
>>
>>>>       - Use switch/case block to set min_version
>>>>         - Set min_version to 0 by default and toggle hacks off
>>>>     + V3
>>>>       - Remove noise (excess pr_info)
>>>>       - Use kstrtoint, not kstrtolong
>>>>       - Use __free(kfree) for allocated mem and drop goto + logging
>>>>       - Use print_hex_dump() to show failed data after pr_err in mcu_request_version()
>>>>       - Use pr_debug in set_ally_mcu_hack() and set_ally_mcu_powersave() plus
>>>>         correct the message.
>>>>     + V4
>>>>       - Change use_ally_mcu_hack var to enum to track init state and
>>>>         prevent a race condition
>>>>
>>>> Luke D. Jones (2):
>>>>     hid-asus: check ROG Ally MCU version and warn
>>>>     platform/x86: asus-wmi: Refactor Ally suspend/resume
>>>>
>>>>    drivers/hid/hid-asus.c                     | 111 ++++++++++++++++-
>>>>    drivers/platform/x86/asus-wmi.c            | 133 +++++++++++++++------
>>>>    include/linux/platform_data/x86/asus-wmi.h |  19 +++
>>>>    3 files changed, 222 insertions(+), 41 deletions(-)
>>>>
>>>> --
>>>> 2.49.0
>>>>
>>