linux-kernel - Re: Regression: Dell XPS 13 9360 keyboard no longer works

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: <20180222195000.GE19843@fury>
Date:   Thu, 22 Feb 2018 11:50:00 -0800
From:   Darren Hart <dvhart@...radead.org>
To:     Mario.Limonciello@...l.com
Cc:     jeremy@...ine.org, notmart@...il.com, pali.rohar@...il.com,
        andriy.shevchenko@...ux.intel.com, mjg59@...f.ucam.org,
        platform-driver-x86@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: Regression: Dell XPS 13 9360 keyboard no longer works

On Thu, Feb 22, 2018 at 05:05:39PM +0000, Mario.Limonciello@...l.com wrote:
...
> OK thanks for confirming.  Here's what I've concluded:
> 
> * So looking through the ACPI tables on the 9360 it initializes that status
> (slate vs laptop mode) bit to "slate" mode.  The 9360 isn't a 2-in1- so that
> seems wrong to me, but that's what it does.
> It only gets updated based on dock status.
> 
> The 9365 (which is a 2 in 1) however seems to initialize the status properly.
> 
> So that's an impasse of what to do.
> It's not clear to me what is really happening:
> a) We're missing something else in this driver (eg something else that 
> indicates whether to trust VGBS output)
> 
> b) Mis-interpreting the results from it (we shouldn't report the switch for 
> tablet mode based on what we do)
> 
> c) 9360 has a BIOS bug (seems unlikely since Windows doesn't freak out and
> show virtual keyboard at wrong time)
> 
> I'm leaning on it's probably <a>.
> We should check for tablet mode should only be run if chassis type 
> matches 2-in-1 (chassis type 0x1F).  

This sounds like a good approach to me.

> I believe that should fix the problem on the 9360, let it continue to work on
> the 9365 (and other 2-in-1's).
> 
> Can you guys please test this?  If that works I'll split up the routine and submit
> it.

Queueing this up as well.
-- 
Darren Hart
VMware Open Source Technology Center