05:41 <JensGlathe[m]> he mentioned the msm hotplug desaster, yes

06:44 <jhovold> clover[m]: internal display not showing up after suspend with external display was fixed in 6.8

06:45 <jhovold> see https://github.com/jhovold/linux/wiki/X13s#suspend

06:47 <jhovold> clover[m], steev: there's no need to add iommu.passthrough=0 iommu.strict=0 after the underlying display bug that killed usb thoughput was fixed

06:49 <jhovold> the required parameters are listed in the wiki: https://github.com/jhovold/linux/wiki/X13s#kernel-command-line

06:51 <jhovold> abby: I dropped the workaround that effectively set 'arm64.nopauth' workaround to minimise the delta to mainline

06:53 <jhovold> so it's needed until lenovo manages to integrate the fix for underlying fw bug (qcom has supposedly fixed it)

06:53 <abby> i see

07:27 <steev> jhovold: it seems to be back with 6.9.1 :D or some weird subset of it - the external display came back but the internal did not. unfortunately, i was in the middle of something and couldn't grab anything and the fastest fix was to just unplug and replug the adapter

07:37 <jhovold> steev: ok, thanks, we know the hotplug implementation is broken but it seemed that particular symptom had been resolved

07:37 <jhovold> how reproducable is it?

07:41 <jhovold> I have seen the internal display PHY fail to resume a few times, but it happens very rarely and results in a garbled screen, but perhaps you hit something like that

07:43 <abby> would garbled include offset of normal and wrapped around (e.g. some of the top of the screen is at the bottom)?

07:44 <abby> and the same left/right

07:44 <jhovold> no, when the PHY failed to resume the screen was completly garbled, random noise/patterns

07:45 <jhovold> there's been reports of artefacts like similar to what you describe with some high-res modes, but that affected the far right part of the display IIRC

07:45 <jhovold> https://gitlab.freedesktop.org/drm/msm/-/issues/43

07:46 <abby> what i saw was much more significant, and on the internal display

07:46 <abby> i've only seen it a couple times

07:47 <jglathe_x13s> if I get a truncated (right part missing) screen, xrandr says its twice the desired resolution. There seems to be a connection to fractional scaling though.

07:48 <jhovold> abby: ok, try to take a photo next time and save dmesg, was it after resume?

07:48 <abby> will do

07:49 <jglathe_x13s> Since my eyes are not good enough for 4k res @ 100%, I had it at 125%. Now switched 2560x1440 @ 100%. No effects, resume appears to work.

07:49 <abby> yes, on lid open, and went away after closing/Opening

07:50 <jhovold> abby: ok, sounds similar in that sense, but you should see an error in the log if it's the PHY timing out on resume

07:51 <jhovold> jglathe_x13s: interesting, sound like another bug indeed, could you file a bug on freedesktop?

07:51 <jhovold> I assume this is with 6.9?

07:53 <jhovold> (without experimental stuff on top, so either mainline 6.9 or my wip branch)

07:57 <jglathe_x13s> jhovold: I had it with 6.9 and earlier. Truncated screen was the effect when using fractional scaling with pop-os, X11 and gnome 42.9. your WIP branch, but my config. Will look to get some info for a bug report

07:58 <jhovold> sounds good, thanks

08:24 <clover[m]> jglathe do you have a popos iso for the x13s just curious

08:40 <jglathe_x13s> no complete iso yet, I booted via grub from Ubuntu (its 2.12, this didn't seem to work with 2.06)

08:40 <jglathe_x13s> I was planning on completing a bootable image, yes - not far away

08:41 <clover[m]> Nice nice

08:43 <jglathe_x13s> I guess I have an idea now how to accomplish this - its a hack, but hey