Received-SPF: pass (google.com: domain of linux-kernel+bounces-92578-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Date: Tue, 5 Mar 2024 16:08:29 +0100
From: Johan Hovold <johan@kernel.org>
To: Linux regressions mailing list <regressions@lists.linux.dev>
Cc: LKML <linux-kernel@vger.kernel.org>, Vinod Koul <vinod.koul@linaro.org>,
	Bjorn Andersson <quic_bjorande@quicinc.com>,
	Abhinav Kumar <quic_abhinavk@quicinc.com>
Subject: Re: Lenovo ThinkPad X13s regerssions (was Re: Linux regressions
 report for mainline [2024-02-25])
Message-ID: <Zec1bSx4H0VRmzfe@hovoldconsulting.com>
References: <170886726066.1516351.14377022830495300698@leemhuis.info>
 <Zd23ZKlNnDKPaU9I@hovoldconsulting.com>
 <83c6019f-9c6f-4ad1-87d5-e4f1bdaca93c@leemhuis.info>
 <ZecjcApvdZSGRysT@hovoldconsulting.com>
 <1726bbb2-2cdc-4cad-b6b1-295f6338e771@leemhuis.info>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1726bbb2-2cdc-4cad-b6b1-295f6338e771@leemhuis.info>

On Tue, Mar 05, 2024 at 03:50:13PM +0100, Linux regression tracking (Thorsten Leemhuis) wrote:
> On 05.03.24 14:51, Johan Hovold wrote:
> > On Tue, Mar 05, 2024 at 10:33:39AM +0100, Linux regression tracking (Thorsten Leemhuis) wrote:
> >> [dropping Linus from CC, we can add him back later when needed]
> >>
> >> On 27.02.24 11:20, Johan Hovold wrote:

> >>> But also with these fixes, there are still a couple of regressions
> >>> related to the Qualcomm DRM runtime PM rework in 6.8-rc1. I'll send
> >>> separate reports to track those.
> >
> >> Any decision yet if they are going to be reverted for now?
> >>
> >> Am I right assuming those would fix
> >> https://lore.kernel.org/lkml/Zd3kvD02Qvsh2Sid@hovoldconsulting.com/
> >> which did not get even a single reply?
> > 
> > That was the hope, but I've managed to trigger a reset on disconnect
> > once also with the runtime PM series reverted.
> 
> Ohh. So did the PM series increase the chance of hitting this? Because
> if not, then...

What we know is that some change in 6.8-rc1 either introduced or
increased the chances of hitting the disconnect resets, while the
runtime PM series (and patch which has now been reverted) broke hotplug
detect.

> > One of the patches from that series has already been reverted (to fix
> > the VT console hotplug regression) and there is some indication that
> > that was sufficient to address the issue with hotplug not being detected
> > in X/Wayland too. I'm waiting for confirmation from some users that have
> > not been able to use their external displays at all since 6.8-rc1, but
> > it does seem to fix the X/Wayland issues I could reproduce here.
> > 
> > But either way, the reset on disconnect is still there, and have since
> > been reproduced by Bjorn also on another Qualcomm platform without a
> > hypervisor so that we've now got a call stack. I've heard that Abhinav
> > is looking into that, but I don't know if there's any chance to have a
> > fix ready this week.
> 
> ...this sounds (please correct me if I'm wrong) like on Sunday the
> situation likely will be "the problem is basically in 6.7.y already, so
> there is not much we can do for 6.8 and reverting or delaying the
> release is unneeded" -- unless of course a fix comes in reach during
> this week.

Yes, unless Abhinav and Bjorn can pinpoint the change that makes us hit
this since 6.8-rc1 and revert that change (or come up with some
temporary band-aid).

It is also possible that we're dealing more than one bug here, since
we're seeing resets both on disconnect and when stopping X some time
after a disconnect.

Johan