rinigus | @b100dian: I have seen it while testing bt during phonecall, but never looked further | 03:58 |
---|---|---|
rinigus | Will try to test here | 03:58 |
rinigus | ... And indeed, having bt on kills suspend | 04:28 |
Mister_Magister | rinigus: don't try to open advanced camera | 04:33 |
T42 | <b100dian> rinugus: and lukewarm phone the whole day | 05:06 |
T42 | <b100dian> Thanks for repro, I am now doing a day-driving w/o BT to get some morale boost:) | 05:07 |
rinigus | @b100dian: in my case suspend failures with bt are "only" once per second or so. but I realized that I have updated kernel and carried over one hack from tama: https://github.com/sony-nagara/android_kernel_sony_sm8450/commit/e8af2f1baa1f2b3b0856df908b2e17a0514c6fe6 | 07:45 |
rinigus | maybe you have a touch older kernel and don't have it included? | 07:45 |
rinigus | in general, sfos is using kernel auto suspend which, as far as I know, not used in android for a long time already. and it was an issue in tama days :) | 07:46 |
rinigus | instead, some userspace daemon is used. I suspect that bt service interacts with that daemon on android and this results in way better interaction. maybe mce should use that too instead | 07:47 |
rinigus | @b100dian: that kernel patch (included in los20 branch) will at least limit autosuspend rate to something reasonable. that would prevent your device getting hot in this case | 07:48 |
T42 | <b100dian> I should have that commit built | 08:15 |
rinigus | I am trying to drop volte on nagara as I currently cannot receive sms. which I find is critical functionality. later, with volte work done, hopefully it will be possible to apply on nagara too | 15:08 |
rinigus | on nagara/los21, we have android.hardware.radio@1.5::IRadio on hidl and vendor.qti.hardware.radio.ims.IImsRadio/imsradio0 on aidl. such mix, as far as I can see from libgbinder-radio ofono-binder-plugin ofono-binder-plugin-ext-qti sources is not supported. volte basically works by chance, without any control on sfos side right now. | 15:08 |
rinigus | so, I tried to disable vendor.imsdaemon and vendor.ims_rtp_daemon as well as drop ofono qti plugin and set radioInterface to default. calls work (non volte), receiving sms doesn't | 15:09 |
rinigus | what am I missing? | 15:09 |
mal | rinigus: x10iv and v also have same mix, it's possible to handle it, I have done it in the jolla volte-plugin, still not fully ready but mostly implemented | 15:19 |
mal | of course doing that aidl interface was not very fun | 15:20 |
rinigus | mal: I guess you didn't try to drop volte as it would be waste of time and instead went for the real solution. | 15:22 |
mal | yeah | 15:22 |
mal | is there any android side property for volte? | 15:22 |
mal | afaik aidl volte is in the ril daemon | 15:22 |
rinigus | mal: not sure. which property you have in mind? | 15:23 |
rinigus | persist.dbg.volte_avail_ovr is the only one I can see with volte | 15:23 |
mal | need to boot one of the devices and check if I see anything | 15:23 |
mal | in my aidl volte tests I have trouble getting it to register to volte | 15:24 |
rinigus | mal: are you by chance allowed to tell whether jolla's volte will be one of these components that are planned to be released? | 15:24 |
rinigus | ... or whether you even know it :) | 15:25 |
mal | I haven't been part of the discussions about those | 15:26 |
rinigus | I see. then we will wait. | 15:26 |
rinigus | back to tests - in qti volte oss module we have RadioInstance and ->dev is used for binder, if I read it correctly. I guess that's where we will have to switch from hidl to aidl. somehow. | 15:27 |
mal | but personally I would like to see the plugin somehow available to community, either as closed binary package or opensource | 15:28 |
mal | the interface is a bit different and determining the content of the binder messages was a bit of work | 15:28 |
mal | rinigus: does changing that persist.dbg.volte_avail_ovr work? | 15:29 |
mal | it's in /vendor/build.prop | 15:29 |
rinigus | mal: I hope they will open it. :) | 15:30 |
rinigus | re property - never tried. will try. what should I look for? | 15:30 |
mal | just change that in the file I mentioned and reboot | 15:30 |
mal | assuming you have that as read-write | 15:31 |
rinigus | not rw, but will have to find it first. its somewhere else on my device | 15:33 |
mal | also other possible names for volte things are those ims ones | 15:33 |
rinigus | its in system/build.prop | 15:33 |
rinigus | and few other ovr props in the same file next to ims | 15:35 |
rinigus | its write protected, so it will take a bit of time. I'll probably bind mount it over | 15:41 |
rinigus | ok, after making sure that persist properties are not there (commented them out), some changes. with all ims services disabled, phone switches from 4g to 2g when receiving a call, but call gets cancelled. if I force 2g, call goes through | 15:58 |
rinigus | sms does not | 15:58 |
rinigus | even if forced to 2g. | 15:58 |
rinigus | will try to set these properties to 0 instead and reboot again | 15:59 |
rinigus | ok, this time calls went through even with transition from 4g down happening in time. but sms not received | 16:03 |
rinigus | once call was dropped again. OK, from these tests it seems that spending time on longer time solution would make more sense. | 16:10 |
T42 | <edp_17> mal, Mister_Magister, I am trying to follow the instruction and got this error: https://paste.ubuntu.com/p/JHqmyG4S6z/ | 21:34 |
T42 | <edp_17> Hmm, I guess should apply those three patches first. | 21:34 |
Mister_Magister | your fixup_mountpoints is wrong | 21:34 |
Mister_Magister | /boot appears to live on ERROR: *fstab* not found | 21:35 |
T42 | <edp_17> Hmm, probably because the device is booting with multiboot | 21:38 |
T42 | <edp_17> However, when I tried to apply the first from those three patches I got this: https://paste.ubuntu.com/p/K9T6vnFwnf/ | 21:39 |
Mister_Magister | you have ongoing rebase | 21:52 |
Mister_Magister | cancel it or finish it | 21:52 |
Mister_Magister | or as it says | 21:52 |
Mister_Magister | Use "git am --abort" to remove it. | 21:52 |
Mister_Magister | also | 21:52 |
Mister_Magister | you're applying it in wrong directory | 21:53 |
Mister_Magister | those are kernel patches | 21:53 |
Mister_Magister | you have to be in kernel directory | 21:53 |
T42 | <edp_17> I did git am —abort and managed to apply the three patches. | 21:54 |
T42 | <edp_17> The instruction says, apply those patches, set those flags and build kernel. I applied patches, set flags and building kernel gave this: https://paste.ubuntu.com/p/RtxZr2rxJz/ | 21:55 |
T42 | <edp_17> I keep trying. | 21:55 |
Mister_Magister | are you building for hammerhead? | 21:56 |
T42 | <edp_17> Yep | 21:59 |
Mister_Magister | you don't even have device repo | 22:00 |
Mister_Magister | so how are you building it? | 22:00 |
T42 | <edp_17> The same way how I did it about 5 times. :) | 22:00 |
Mister_Magister | well then you probably removed something very badly | 22:00 |
Mister_Magister | do repo sync | 22:00 |
T42 | <edp_17> Probably. Yeah, I'll start over. | 22:01 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!