| T42 | <edp_17> mal: Do you remember the fix for hybris-12.1 was being too large? | 12:10 |
|---|---|---|
| T42 | <edp_17> I've checked out hybris-12.1-minimize and did a repo sync. | 12:10 |
| T42 | <edp_17> The size of local repo went down from 74GB to 34GB. I got 40Gb back. :) Thanks. | 12:11 |
| T42 | <edp_17> I have not tried whether I can build anything with the shrinked repo though. :) | 12:12 |
| mal | good | 12:22 |
| mal | I thought those were merged already, it seems not | 12:23 |
| T42 | <edp_17> mal: Could you please delete the hybris-12.1-minimize branch once it got merged into hybris-12.1? | 12:25 |
| mal | sure | 12:26 |
| T42 | <edp_17> Thanks! | 12:27 |
| deathmist | mal: what questions did you have about mainline stuff btw? | 12:41 |
| mal | deathmist: just wondering about modem on your device, is it working at all? | 12:44 |
| deathmist | mal: yeah, I've not done any SMS/MMS/call testing recently though but it does at least start and it shows signal info etc. for an old prepaid SIM which I can also unlock | 12:47 |
| deathmist | I really should test it again | 12:47 |
| deathmist | it doesn't work via ofono-qrtr fork but ModemManager is fine | 12:48 |
| mal | deathmist: I was trying to use fp4 mainline with ofono and it gets quite far but fails to set the modem to online state | 12:56 |
| mal | qmi error code says device not ready | 12:56 |
| mal | deathmist: I had to patch ofono qrtr fork a bit | 12:57 |
| mal | same issue on fp4 using modemmanager so it's not an ofono issue | 12:57 |
| mal | deathmist: do you have the udev rules etc for that device | 12:57 |
| deathmist | mal: it looks like I've just got /usr/lib/udev/rules.d/77-mm-qcom-soc.rules which ships as standard on ModemManager when the qcom-soc plugin is enabled during configure: https://pastebin.com/raw/VbccpJEB | 13:03 |
| deathmist | as for ofono, I never got any signs of life from that on MSM8998, nothing at all in debug logs and test scripts did nothing as far as I could tell | 13:03 |
| deathmist | didn't bother to debug it further considering most distros just use ModemManager anyway and it has a lot more active development (and things actually getting merged) | 13:04 |
| mal | deathmist: the patch I had to make for ofono was to fix it select correct qrtr node, it was selecting completely wrong one | 13:05 |
| mal | that depends on device and in which order it reports things and what kind of services | 13:05 |
| deathmist | bleh, so I guess it was hardcoded for MSM8974 or whatever | 13:08 |
| deathmist | ah 8953 | 13:09 |
| mal | the issue with that selection was that qrtr support 32-bit ids and qmi only 8-bit, it blindly assigned the 32-bit value to qmi packet and that caused in some cases for it to select incorrect one for example 4098 instead of 2 | 13:10 |
| deathmist | mal: do you have WLAN working? usually when that works the modem should be fine and all required daemons running | 13:11 |
| mal | deathmist: wlan works, bt doesn't | 13:11 |
| deathmist | yeah BT is separate though, runs it's own firmware and can work even if modem doesn't | 13:11 |
| mal | yeah | 13:12 |
| mal | something is causing modem to think it's not ready, it does successfully report imei and stuff to ofono but final startup fails | 13:12 |
| mal | as mentioned earlier modemmanager fails the same way | 13:12 |
| deathmist | the issue could be in libqmi if the test commands (via qmicli) for that cannot unlock the SIM either and show carrier etc. | 13:14 |
| deathmist | I can dig up the old stuff I used to test early modem things | 13:14 |
| deathmist | not sure if ofono uses libqmi though | 13:14 |
| mal | it doesn't use it | 13:14 |
| mal | but I can see if I can do something with libqmi as a test | 13:15 |
| deathmist | mal: here's the commands of interest that I just tested as working on 8998: https://pastebin.com/raw/j2ppn0r1 | 13:28 |
| deathmist | qmicli has a handy -v flag for verbose details if you're interested in the QMI communication packets | 13:28 |
| mal | thanks | 13:29 |
| mal | deathmist: does gps work on your device? | 13:29 |
| mal | for me it seems like it starts but I never get any incoming nmea messages | 13:30 |
| deathmist | mal: yep, but requires a patch in GPSD for example to support the PDS protocol which a running service provides access to: https://gitlab.com/gpsd/gpsd/-/merge_requests/139 | 13:30 |
| deathmist | I've actually not tested it in a while but at least it used to work | 13:30 |
| mal | wondering how different that is from the gpsd I used, I used rebased this one https://github.com/andersson/gpsd/ | 13:31 |
| deathmist | it has the same patch. though I don't know if the one in the MR is different | 13:32 |
| deathmist | I've also just rebased the patch but in the MR instead | 13:32 |
| mal | deathmist: that "qmicli -d qrtr://0 --dms-set-operating-mode=online" returns the same DeviceNotReady error | 14:40 |
| deathmist | mal: yeah I figured :/ something could be missing from both libqmi and ofono-qrtr for your SoC potentially | 14:42 |
| deathmist | you could also try these before the previuosly sent ones (if you have a dual-sim device): https://gitlab.com/postmarketOS/pmaports/-/blob/master/modem/msm-modem/msm-modem-uim-selection.initd | 14:42 |
| deathmist | you could make an issue on libqmi's GitLab to track progress perhaps and maybe get others to chime in if they have the same (or similar gen) SoC | 14:46 |
| mal | deathmist: those didn't help | 15:15 |
| T42 | <adampigg> https://irc.thaodan.de/.imgstore/9603ec6d/file_3479.jpg | 16:36 |
| T42 | <adampigg> https://irc.thaodan.de/.imgstore/23945dd5/file_3480.jpg | 16:36 |
| T42 | <adampigg> ^^ this is the result of biktori from the pinephone modem fw, debugging with ofono, and fixing some things to work | 16:38 |
| poetaster | phone's talking to themselves via dbus prompts. be afraid. | 17:02 |
| T42 | <adampigg> its a firmware feature that allows you to talk to the modem :) | 17:04 |
| lotheac | is the code used to build the sailfish platform SDK public somewhere? eg. if I wanted to fix some annoyance in the ubu-chroot script or something | 17:11 |
| lotheac | (having googled a certain comment from that script and getting 0 results I kind of know the answer, but asking anyway) | 17:11 |
| poetaster | T42<adampigg> it is kind of cool in a 'dave, what are you doing dave' sort of way. | 17:47 |
| T42 | <elros34> lotheac: maybe this one https://github.com/sailfishos/android-tools-hadk? | 17:51 |
| T42 | <adampigg> poetester: the key really is that the fw dev is now testing/fixing bugs that exist on ofono, making it better all round | 17:53 |
| lotheac | elros34: oh, that's the one. thanks! | 17:57 |
| Thaodan | lotheac: What is it? I'm working on that right now. Feel free to ask. | 18:53 |
| Thaodan | You can also try zypper se --provides <file> if you want to look from which package which file is | 18:58 |
| Thaodan | zypper info <pkg> tells you the srcpackage which often mathes the git repo name of the project on github | 18:59 |
| T42 | <b100dian> lotheac: google or any general purpose search engine is no match for gihub, in my experience (or any git hosting, for that matter). If google would follow the link on git hosting services without implementing git itself it would run into some scalability issues:) | 19:42 |
| T42 | <b100dian> Of course the same goes for many scms, hg, svn, you name em | 19:43 |
| voidanix[m] | is it normal for `rpm/dhd/helpers/build_packages.sh --mw=https://github.com/mer-hybris/droid-system-sony-template` in the 10 III build guide to fail? | 19:58 |
| voidanix[m] | says there are unstaged changes, probably because of `copy_tree.sh` | 19:58 |
| voidanix[m] | probably even worse is that mic spat out a home.img and root.img that i have no idea how to flash :D | 21:03 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
