Google Pixel 3a (sargo)



  • @dieharddan I think the order matters here. Not familiar with ofonod, but it looks like the help page suggests that the -P is for noplugin, followed by a list of plugins not to use. So reading through here, provision,udev... and the others need to follow -P directly. I'd edit it to reflect this instead:

    exec ofonod -P provision,udev*,dun*,smart*,hfp_bluez5,stktest,sap - n - d

    I was able to get the log... now I'm trying to get it off my Pixel 😆



  • @dsimmons lol nice, i will give this a try thanks again



  • hah, yeah I am having a rough time of it. Default SSH config isn't working, trying to sort that so I can use SCP, if I try to email the file or copy over from the /home/phablet/ folders, it seems to think the files empty, even though when I look at it in nano it's not.



  • @fredldotme hey, I was able to get the ofono.log (I also have the SMS issue, but airplane mode on/off doesn't resolve)

    I'm not sure about pastebining this whole thing, is there anything in particular you're looking for, or a section I can parse out? It looks like there is a lot of identifying information, and I don't want to just slap it all in a public pastebin.



  • @dieharddan / @fredldotme looks like I may have found something interesting, I think short enough to drop here:

    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_list() 6 cell(s):
    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_cell_lte() [lte] reg=1,mcc=311,mnc=480,ci=78151704,pci=166,tac=1280,strength=18,rsrp=107,rsrq=11
    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_cell_lte() [lte] reg=0,pci=172,strength=12,rsrp=107,rsrq=10
    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_cell_lte() [lte] reg=0,pci=166,strength=13,rsrp=107,rsrq=11
    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_cell_lte() [lte] reg=0,pci=172,strength=13,rsrp=111,rsrq=16
    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_cell_lte() [lte] reg=0,pci=167,strength=13,rsrp=114,rsrq=15
    ofonod[2048]: drivers/ril/ril_cell_info.c:ril_cell_info_parse_cell_lte() [lte] reg=0,pci=200,strength=13,rsrp=116,rsrq=19
    ofonod[2048]: plugins/sailfish_manager/sailfish_cell_info_dbus.c:sailfish_cell_info_dbus_cells_changed_cb()
    2020-11-19 20:55:18 slot1< [00000146] RESPONSE_ACK
    2020-11-19 20:55:18 slot1< 0000: 20 03 00 00 46 01 00 00 ...F...
    2020-11-19 20:55:22 [gbinder-radio] WARNING: /dev/hwbinder:slot1 died
    2020-11-19 20:55:22 [grilio-binder] ERROR: slot1 radio died

    It looks like right after ACK, the radio hardware binder up and fails. Possibly related to the issue? Maybe driver failure?



  • @dsimmons that's interesting, thanks for finding that! I have built updated versions of some components and would like to see if they work for your use case:

    sudo mount -o remount,rw /
    sudo cp ~/Downloads/libgbinder-radio.so.1.0.11 /usr/lib/libgbinder-radio.so.1.0.8
    sudo cp ~/Downloads/libgbinder.so.1.0.47 /usr/lib/libgbinder.so.1.0.37
    sudo reboot
    

    It's certainly interesting that the rilmodem would just disappear from binder, so I'm eager to find out if that somewhat fixes your issue.

    EDIT: Also, could you please verify with dmesg -w | grep -i binder in a separate terminal window if there are binder transaction failures coming up when your issue appears?



  • @fredldotme Duped the two files to the /usr/lib/ location and rebooted. Here's the grep... ouch!

    [    3.842925] binder: 1661:1661 transaction failed 29189/-22, size 0-0 line 3047
    [    3.962031] binder: 1680:1680 ioctl 4018620d 7ff1452850 returned -22
    [    4.552391] binder: undelivered death notification, 0000000000000000
    [   10.773000] binder: release 2586:2586 transaction 6162 out, still active
    [   10.773011] binder: undelivered TRANSACTION_COMPLETE
    [   10.792496] binder: 2554:3673 transaction failed 29189/0, size 164-32 line 2987
    [   10.792508] binder: send failed reply for transaction 6162, target dead
    [   10.851370] init: bluebinder post-start process (3720) terminated with status 1
    [   13.790649] audit: type=1400 audit(1605921619.678:71): apparmor="DENIED" operation="open" profile="/usr/bin/media-hub-server" name="/proc/4541/timerslack_ns" pid=4415 comm="Binder:4415_1" requested_mask="w" denied_mask="w" fsuid=32011 ouid=32011
    

    As an aside - do you have any pointers on if this solution can be dual-bootable with an Android build? I'm excited about trying to push UBports forward, but the SMS/GPS issue is a major deal breaker as this is my daily driver now. I think I'm going to have to roll it back to the default Android build, but would like to keep helping.

    Else, I might make a bad habit of toggling back and forth between the two when I don't need my Pixel at full health 😄

    Are you not seeing any of the mentioned issues on your 3a? Maybe it's a carrier problem? I'm on Verizon (US).



  • @dsimmons i have the same issue on at&t



  • @dsimmons @dieharddan No, I'm not seeing those issues on my phone (T-Mobile/Magenta, Austria). What I've noticed is a different issue where ofono (UT-side) and rild (Android-side) don't communicate with each other anymore because rild runs out of IPC memory. Hard to fix since it's the proprietary blob acting up.

    Your problem, it could be either differences in operators, but also there are specific models with different modem settings. In the case it's the modem settings, could you please post the output of getprop ro.boot.hardware.sku? Mine is G020F.

    We have one known US citizen getting his compatible device soon, so at least we can verify and potentially fix those issues.

    What I would propose for now, since it's your daily driver, is to revert to Android until we have a solution. Those issues are harder to fix than expected because most developers/porters who get into contact with ofono live in Europe. We will surely ping you in the thread as soon as we have something for ya.



  • @fredldotme G020C



  • @fredldotme Just out of curiosity, are you going to possibly do a port for the Pixel 4a as well?



  • @mendevi_0 4a is very pricy, even as used device. Nothing below 250-300 € - thats far too much for a private porter. So either a device is donated in some form or lend out.



  • Not even 300,--- its more like 350-400,- € here. So thats definitely too much to expense for us atm.


Log in to reply