09:01 <lumag> z3ntu, could you please remind me, what was the outcome of configurable SID discussion with krzk? We should drop #ifdef's or something else?

09:02 <z3ntu> lumag: yes, don't hide a default value behind #ifndef , so this patch is ok https://lore.kernel.org/linux-arm-msm/20230919-fp5-initial-v2-3-14bb7cedadf5@fairphone.com/

09:02 <lumag> z3ntu, ack, thanks!

09:02 <z3ntu> one thing to kind of look out (not totally related) for is that some spmi-adc7 headers already have an ifndef for the SID so it might leak into dts also..

09:06 <lumag> z3ntu, we should probably fix them

09:06 <lumag> patch would be appreciated ;-)

10:08 <lumag> z3ntu, regarding UCSI on qcm6490. Do you know if it needs the GET_PDOS workaround? In other way, have you tried enabling UCSI on FP5 beforehand?

10:10 <z3ntu> Without the workaround I very quickly see "sync write returned: -16" and "UCSI_GET_PDOS failed (-5)" errors in dmesg. Very seldom it works once or so, so something plugged in actually works

10:10 <lumag> ack, thanks.

10:10 <z3ntu> With the workaround I didn't see it (but as mentioned didn't test too much but quite confident it improved things). But still role swiching back from host to gadget didn't seem to work, or maybe in postmarketOS this doesn't work properly, not sure

10:11 <lumag> Hmm. this is strange.

10:11 <z3ntu> Meaning when I boot the phone, rndis gadget gets set up and I have ssh access to the phone, I then unplug, plug in a USB stick to the phone which works, but when I plug the phone back into the computer the usb network doesn't re-appear

10:11 <z3ntu> Again, take it with a grain of salt but I think I saw this behavior yesterday

10:16 <lumag> z3ntu, reproduced here.

10:17 <lumag> And I got the mysterious [ 40.013494] ucsi_glink.pmic_glink_ucsi pmic_glink.ucsi.0: GET_CONNECTOR_STATUS failed (-95)

10:21 <lumag> z3ntu, however restarting the gadget (I'm using adb) fixes the issue.

11:02 <z3ntu> lumag: any clue if this a qcom/ucsi bug or something maybe user space handles wrong? But I imagine if you see the same with your environment, at least it's not caused by whatever we have in postmarketOS :D

11:53 <lumag> z3ntu, either a kernel, or a firmware issue. I'll try sorting it out later.

11:58 <z3ntu> thanks!

12:28 <z3ntu> lumag: Also any idea what pmic-glink compatible I should use on this qcm6490 phone? Either qcom,sc7280-pmic-glink or qcom,qcm6490-pmic-glink make some sense I guess

12:29 <z3ntu> Currently in my tree I added qcom,sm7325-pmic-glink but this makes no sense really, it would be "Snapdragon" variant of this chip

12:38 <narmstrong> I get GET_CONNECTOR_STATUS on recent SoCs aswell, but it may be a different issue from the GET_PDOS workaround

12:41 <lumag> narmstrong, I have yet to debug it, because I see it only rarely and it returns -EOPNOTSUPP.

12:42 <lumag> z3ntu, well, since it is a firmware thing, you might want to emphasise that. So from my POV qcm6490 makes more sense

12:42 <lumag> In the end, typical sc7280 has different firmware.

12:47 <z3ntu> sc7280 with qualcomm firmware (non-cros) will probably behave the same though as qcm6490 ^^ but I'll change it to qcm6490 compatible in my tree then, let's see if anyone in review will say something different

14:41 <strongtz[m]> iirc I also randomly get "UCSI_GET_PDOS failed" on qcs8550 (sm8550)

14:42 <strongtz[m]> unaware of "GET_CONNECTOR_STATUS" though

16:04 <lumag> strongtz[m], strange. It was supposed to be fixed with newer firmware releases.

16:04 <lumag> bamse, ^^

16:05 <lumag> strongtz[m], do you have a log of ucsi failures?

23:15 <strongtz[m]> lumag: Sure, but the device isn't at hand now. I'll probably get a log later today or tomorrow.