13:21 <pengyu[m]> hi, I set ufs-qcom hc and ufs phy driver builtin, when I boot up, I get `ufshcd-qcom 1d84000.ufshc: deferred probe timeout, ignoring dependency`, I put the kernel dir to ramdisk too, this doesn't help. can anyone point out what I had missed? I am working on a sm8650-mtp based retail device.

14:13 <bamse> Pengyu[m]: "deferred probe timeout, ignoring dependency" implies that something that the ufshc depends on didn't show up... not sure what you're missing though, seems that the necessary dependencies are in the defconfig already

14:13 <bamse> Pengyu[m]: i don't know what "put the kernel dir to ramdisk" implies, it doesn't sound like the right thing to do

14:14 <bamse> Pengyu[m]: https://www.linaro.org/blog/let-s-boot-the-mainline-linux-kernel-on-qualcomm-devices/ is a pretty good resource, but i see the mkbootimg step wouldn't work on 8650 as documented

14:31 <pengyu[m]> <bamse> "Pengyu: i don't know what "put..." <- I mean, all modules are accessible without real root mounted, it should have solved dependency issues.

14:34 <pengyu[m]> <bamse> "Pengyu: https://www.linaro.org/..."; <- it is outdated, we should append more to mkbootimg

14:43 <pengyu[m]> I use mkbootimg \... (full message at <https://matrix.org/oftc/media/v1/media/download/ARI49YxrIiSKrg1mpjYSGavgRVSG4OA8lHtcPC8ZUCFamQSsg1lCaz1VOuW8Yn3yetTVTJveDMfL-JrJHyiXwzRCeWNMD0OwAG1hdHJpeC5vcmcvWmFXTFJOSUZEVmp3R1ZTSm1IeWJhSnhv>)

14:43 <pengyu[m]> it varies from devices

14:44 <pengyu[m]> Pengyu[m]: > <@pengyu:matrix.org> I use mkbootimg \... (full message at <https://matrix.org/oftc/media/v1/media/download/AfNO1ltnZqfgBD2AHtepnbJVUZx7CnUlDZ64au8DCYZooOOJMAPW-3kjIREDokymkv20n8J5Fmren4FLzMkMt4hCeWNMK0RQAG1hdHJpeC5vcmcvQ1NIcUJ3TFNHRXRVYWFlb2RoUVZuRXlx>)

14:46 <krzk> Pengyu[m]: above messages are not readable. This is IRC, not matrix, so none of above work.

14:47 <krzk> anyway, you should understand which resource defers till timeout and then check whether this device driver is actually loaded

14:47 <krzk> I assume you use mainline (linux-next) tree, right?

14:48 <krzk> Although I am not aware of any laptops based oon sm8650 and sm8650-mtp is not a laptop, so this is a bit confusing (that's laptop's channel)

14:49 <pengyu[m]> krzk: sorry, I will try to not use client app features.

14:50 <pengyu[m]> yes, next tree. This is a tablet, I want to make it a laptop

17:57 <kuruczgy[m]> I am seeing an issue where sometimes the display goes black when booting (slim 7x). It's inconsistent, here are two kernel logs booting with the exact same config, one good and one bad: https://gist.github.com/kuruczgy/59006541df93ab529f143974c8a38654

17:57 <kuruczgy[m]> I am hitting this very rarely, but there are apparently two users of my NixOS config who are hitting it much more consistently: https://github.com/kuruczgy/x1e-nixos-config/issues/100

17:57 <kuruczgy[m]> It's clear from diffing the logs that the graphics stack fails to initialize (e.g. the "[drm] Initialized msm 1.12.0 for ae01000.display-controller on minor 1" message is missing in the bad log), but I don't know what the root cause could be.

17:57 <kuruczgy[m]> Bad/good boot diff from another user, it looks pretty similar to mine, though for me a bunch of "deferred probe" messages are missing in the bad boot: https://gist.github.com/LunNova/53fff2b356afc57c8d55db57af006842

18:06 <steev> on the bad boots is there anything in devices_deferred?

20:41 <jhovold> kuruczgy[m]: just skimmed the log, a retimer fails to probe, that in turns causes the probe deferrals and broken display

20:42 <jhovold> Dec 19 22:27:51 cobalt kernel: ps883x_retimer 3-0008: failed to read conn_status_0: -6

20:42 <jhovold> Dec 19 22:27:51 cobalt kernel: ps883x_retimer 3-0008: probe with driver ps883x_retimer failed with error -5

20:47 <JensGlathe[m]> kuruczgy: what is your hardware

20:47 <JensGlathe[m]> Yoga 7?

20:47 <steev> slim7x

20:47 <kuruczgy[m]> jhovold: Yep, I just checked and that difference also exists in logs from other two persons. Thanks.

20:47 <kuruczgy[m]> Any idea for what could be causing it? One person reported that "It's only happening if I boot while plugged in to power, on battery it seems to be ok.", could the charger somehow be messing with the retimer through the usb-c port or something? Sounds quite far fetched though...

22:18 <HdkR> I haven't been paying much attention, but I saw something about HDMI audio tunneling in the ML. Does this mean X1E->DP->HDMI audio is going to work soon?