00:03 <abc123bac2[m]> recently got a ideapad 5 2 in 1 (x1p42100) with a broken screen, intending to use it has a headless desktop. this doesn't seem to be a linux specific issue (happens on windows too) but it looks like this laptop doesn't output on USB-C/DP when the internal display's edp cable is unplugged from the motherboard.

00:03 <abc123bac2[m]> not sure if it extends to all snapdragon x laptops, or just mine. wondering if anyone here knows anything

00:13 <robclark> you might approximately be the first one to test this scenario, but I assume it is similar to the devkit/nuc thing which has no eDP?

00:18 <robclark> got dmesg.. and `cat /sys/kernel/debug/devices_deferred`? Could be something about panel probe is blocking the drm driver (but lumag had an interesting idea that might solve that)

01:59 <abc123bac2[m]> will post in a bit, need to open up the device to unplug the edp cable

02:27 <abc123bac2[m]> eDP plugged in: https://pastebin.com/7vxMM8aK... (full message at <https://matrix.org/oftc/media/v1/media/download/AZJyDjv8ijmob1bKi5RRzt-zQLTnB9opRxWaT_pQLQAgTUzaPRJ9EGAZ86egeFqJ1-GcbzpAaUxtQkX6a_aJle5CeXtFxUjgAG1hdHJpeC5vcmcvWVNtV2JoeW1MQm1QRFdxcnhSaVlRZnlV>)

02:27 <abc123bac2[m]> seems like matrix screwed this up, let me resend

02:27 <abc123bac2[m]> eDP plugged in: https://pastebin.com/7vxMM8aK

02:27 <abc123bac2[m]> eDP unplugged: https://pastebin.com/uPR3SVYg

02:28 <abc123bac2[m]> In both cases, this is the same cat /sys/kernel/debug/devices_deferred -> c42d000.spmi:pmic@0:rtc@6100

02:28 <abc123bac2[m]> I'm actually able to see grub and a small duration of the kernel boot logs in the 2nd case. It all goes black after that.

02:28 <abc123bac2[m]> In the first case, it goes black for a moment, and then GDM appears (sadly with the main parts of the UI on the connected but non functioning eDP display). I can login to gnome, open settings and disable the internal display in that case.

02:51 <abc123bac2[m]> the system as it is (without edp plugged in), stays online for 5 min and reboots on it's own. Been doing that for the last 5-10 min in a loop. On one of the reboots, the external display came up at a super low resolution and before I could check anything, it rebooted again

02:53 <abc123bac2[m]> These are the last lines in journalctl -b -1 before the reboot

02:53 <abc123bac2[m]> kernel: [drm:dpu_encoder_phys_vid_wait_for_commit_done:537] [dpu error]vblank timeout: 80022041

02:53 <abc123bac2[m]> kernel: [drm:dpu_kms_wait_for_commit_done:524] [dpu error]wait for commit done returned -110

03:43 <JensGlathe[m]> abc123bac2: which dt do you use

03:44 <abc123bac2[m]> I used the image you put on google drive for ideapad 5 and that dtb (which is a slightly different device, yes)

03:44 <abc123bac2[m]> I dumped the DSDT on mine, but haven't looked at the diff between the two yet

03:46 <JensGlathe[m]> Would be enough to remove the mdss_dp3 node to not have internal display, not enabling it (and the phy) should be sufficient

03:47 <JensGlathe[m]> The type-c dp altmode works in only one orientation

03:48 <JensGlathe[m]> (From my current status on the tb16)

03:49 <abc123bac2[m]> I'll try that. I also wanna try using a newer kernel with the patches for the GPU on the X1P42, I saw you have a branch with that

03:51 <abc123bac2[m]> Is there a reason why the DP out works for GRUB but not after booting?

03:58 <JensGlathe[m]> 6.15.0-jg-6 should work nice enough, but orientation issue.

03:59 <JensGlathe[m]> Don‘t have the x1p gpu firmware files yet

04:03 <JensGlathe[m]> Bios sets up dp (probably) but driver init very likely kills it. On the dev kit you get grub, after selecting no screen, eventually output again with windows or linux

04:05 <abc123bac2[m]> Thanks for the suggestion. I deleted the mdss_dp3 node and am able to boot. The behavior is the same as what you described (grub, no screen, output with linux)

04:07 <abc123bac2[m]> Hope someone can figure out what the error is though, and fix upstream

05:22 <steev> so, i've applied https://lore.kernel.org/all/20250613124839.2197945-8-cassel@kernel.org/ this pci related patch, and i've not seen the aer correction stuff (yet)

05:22 <steev> the rockchip portion didn't apply (to .14 or .15) but since i only care about the qcom portion, i'm okay with dropping that one

05:51 <JensGlathe[m]> Nice

06:16 <JensGlathe[m]> abc123bac2: can you proved some more data on the device for distinguishing them (if required)

07:24 <valpackett> oh huh, went to check for bios updates, learned that it's common between Dell Inspiron_5441_7441_Latitude_5455_7455 and the codename is Thena

07:31 <jdb[m]> abc123bac2: robclark: I can confirm a similar experience

07:33 <jdb[m]> Whenever I was trying to bring up the built-in display for the Surface Pro 9, each time I tried a non-working config, it was blocking the external output via USB-C

07:34 <jdb[m]> Whereas I had a known working config for external display...

07:35 <jdb[m]> It was annoying as I had to connect remotely via ssh from another computer to see the dmesg output

07:37 <jdb[m]> If the probe could be non-blocking, that would be a great improvement

07:38 <jdb[m]> And it would make the Linux config via devicetree more robust to hardware failures

08:00 <valpackett> just had another random hang after 1.5 days of uptime, during light load, not doing anything, just moving the cursor. it gets so frozen it doesn't respond to magic sysrq. could it be something power management related?

12:44 <deathmist> I also had a random full system hang on X1E Vivobook yesterday, running a 6.15.0-jg-5 tree based kernel myself

13:20 <abc123bac2[m]> Jens Glatheyeah sure let me know what you need

19:03 <steev> ah, nope, still get the aer

19:05 <JensGlathe[m]> SN770M, right?

19:11 <konradybcio> valpackett: "computer goes bad" is unfortunately not enough to take any meaningful conclusions..

19:12 <konradybcio> any chance systemd saved the dmesg?

19:32 <steev> yeah jens

19:32 <steev> WD BLACK SN770M 2TB

19:34 <JensGlathe[m]> I was about to put in my SN770 for test, too.

19:37 <steev> they're correctable errors so...

19:37 <steev> i did just update the firmware to latest though (i think)

19:38 <steev> yeah, 731130WD is now the latest firmware for mine

19:38 <steev> https://gist.github.com/sorend/38aa32b1b07124575026918e5201e299

19:38 <JensGlathe[m]> hmm need to take a look

19:39 <JensGlathe[m]> The low level errors result in btrfs read-only mode eventually, and this is bad for a VM host

19:40 <JensGlathe[m]> To avoid it I used an USB external drive for a while, the Silicon Power 2TB 2230 which doesn't seem to have that issue

19:53 <Jasper[m]> <steev> "https://gist.github.com/sorend/3..."; <- Thanks!

20:05 <zetam> About <valpackett>'s “[PATCH 0/4] Add support for Dell Latitude 7455 (X1E-80-100)”: I have a Latitude 7455 that still won’t boot, and I’d like to try these patches, could you suggest which kernel version or tree is best suited for them? linux-next, or another?

20:26 <JensGlathe[m]> I have it on my tree as well

20:28 <JensGlathe[m]> Question for the Yoga Slim7 torturers: Is the wcn7850 a m-2 card on it or soldered-on?

20:38 <HdkR> JensGlathe[m]: Yea

20:38 <HdkR> https://media.discordapp.net/attachments/702130548419723294/1263624858995265628/IMG_1532.jpg?ex=684ea19b&is=684d501b&hm=9b9c1c535f7aee638c4a6d3adf761598a04e772f6ed78115e15a6fde2b6f97c7&=&format=webp&width=1750&height=1313

20:38 <HdkR> Right in the center of the board

20:39 <JensGlathe[m]> That looks soldered on

20:39 <JensGlathe[m]> Thanks

20:39 <HdkR> Oh, hah. I partially read the sentence. Yes, soldered :D

20:40 <JensGlathe[m]> Was about to compose a spicy answer already 😎

20:40 <HdkR> Brain rats getting to me today.

20:41 <JensGlathe[m]> 2 fans, wow

20:41 <JensGlathe[m]> and they don't behave?

20:41 <HdkR> Yea, I haven't been able to get the thing to thermally throttle even throwing 60W+ of load at it.

20:42 <JensGlathe[m]> good

20:42 <JensGlathe[m]> but ig they don't throttle on Linux, they switch off (seen on HP Omnibook X14)

20:43 <HdkR> Not sure there.

20:45 <JensGlathe[m]> The X14 makes a lot of noise, and if the airflow is sufficient it will power through - sucking the battery dry since the 65 PSU is not enough

20:46 <JensGlathe[m]> The tb16 will also make a lot of noise, but no high pitched whine - haven't been able to throttle / switch it off yet

22:35 <steev> theres still something funky with soundwire

22:35 <steev> https://www.irccloud.com/pastebin/oMxY5uBm/

23:18 <anthony25> JensGlathe[m]: it throttles, just the frequency as shown by the governor is lying

23:19 <anthony25> someone here explained it to me, I don't remember whom, when I was complaining that the laptop shutdowns when too hot, and that I didn't see it throttling

23:19 <anthony25> the shutdown part I think has been fixed, it hasn't happened to me for a while

23:48 <valpackett> zetam: linux-next is what i use, but the 7455 dt patch is already in jglathe's tree for example

23:48 <valpackett> we really need to unify latitude/inspiron tbh because they're such close variants of the same thing that they share the bios update file

23:52 <valpackett> konradybcio: nothing interesting in journalctl -b-1, the last message is a systemd-run scope for a terminal starting. it's an entirely sudden hard freeze, so hard that magic sysrq doesn't respond, seemingly not correlated to anything in particular (though we thought it was slightly more correlated to audio playback but eeehhh) and it happens for many people on different laptops

23:53 <valpackett> if someone with internal debug equipment could reproduce and investigate it, that would really save the day ;)