linux-kernel - [PATCH 0/3] pinctrl-amd powerbtn handling regression

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-next>] [day] [month] [year] [list]

Message-ID: <20230829165627.156542-1-mario.limonciello@amd.com>
Date:   Tue, 29 Aug 2023 11:56:24 -0500
From:   Mario Limonciello <mario.limonciello@....com>
To:     <linus.walleij@...aro.org>
CC:     <Shyam-sundar.S-k@....com>, <Basavaraj.Natikar@....com>,
        <linux-gpio@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <regressions@...ts.linux.dev>, <lucapgl2001@...il.com>,
        Mario Limonciello <mario.limonciello@....com>
Subject: [PATCH 0/3] pinctrl-amd powerbtn handling regression

AMD GPIO controllers have a bit in one of the registers that sets debounce
to meet Windows certification requirements.

As many platforms ship Windows they enable this by default.
Linux was not using it until a bug fix commit a855724dc08b ("pinctrl: amd:
Fix mistake in handling clearing pins at startup") aligned the register
with the intended values.

On systems that program GPIOs through _AEI and handle interrupts in the
GPIO controller driver this makes them behave like Windows.

On systems that don't program GPIOs in _AEI, the interrupts aren't handled
by the GPIO controller driver this causes the GPIO to get "stuck".

It's stuck because according to the spec the interrupt is supposed to be
cleared by driver when the button is pressed less than 2s.

However as the GPIO doesn't trigger an interrupt, it can't be cleared
until another GPIO that is handled by the driver (such as touchpad).

This series adds handling behavior from the spec, but as it's not ideal
on the reported platform from the lack of interrupt to handle set a quirk
to revert the debounce behavior back to old behavior.

A module parameter is also added to let anyone else affected by this debug
it.

Mario Limonciello (3):
  pinctrl: amd: Clear `Less2secSts` and `Less10secSts` for GPIO0
  pinctrl: amd: Add a module parameter to configure power button
    behavior
  pinctrl: amd: Add a quirk for Lenovo Ideapad 5

 drivers/pinctrl/pinctrl-amd.c | 70 +++++++++++++++++++++++++++++++++++
 drivers/pinctrl/pinctrl-amd.h |  2 +
 2 files changed, 72 insertions(+)

base-commit: b4e880a8d840e2b64937ab47ad518185c07747e3
-- 
2.34.1