| entil | no lineageos for fp6 yet :/ any good news for the fp5? waydroid? | 05:37 |
|---|---|---|
| T42 | <TheVancedGamer> mal: it seems to start charging again for me | 05:45 |
| T42 | <TheVancedGamer> I also tested it in android where 500mA charging works with data transfer too | 05:45 |
| T42 | <edp_17> @elros, defconfig-bluetooth: https://paste.ubuntu.com/p/JPPSVxmtBv/ | 07:45 |
| T42 | <edp_17> and .config: https://paste.ubuntu.com/p/GbSDXHQ7xn/ | 07:45 |
| T42 | <elros34> around 30 modules and you have built just few. with my limited understanding I guess something must have failed during build | 07:56 |
| T42 | <edp_17> Here is the full log for building hybris-hal, and those two backport makes https://paste.ubuntu.com/p/zbF3H2w599/ | 08:10 |
| T42 | <edp_17> As you see all three finished successfully. | 08:10 |
| T42 | <elros34> yeah but you didn't have copied all of them in device. for example in previous paste compat.ko was missing | 08:47 |
| T42 | <edp_17> Where to I should copy them exactly? | 09:02 |
| T42 | <edp_17> I only copied the one from backports-bluetooth/drivers/bluetooth. Nothing else. | 09:04 |
| T42 | <edp_17> Okay, so I guess need to copy all of these onto device: https://paste.ubuntu.com/p/QvKQwSbhHQ/ | 09:07 |
| T42 | <edp_17> Let me try that. | 09:08 |
| T42 | <elros34> maybe not all are required, but compat which has defined missing symbol for sure | 09:12 |
| T42 | <elros34> where? /lib/modules/$(uname -r) | 09:13 |
| T42 | <edp_17> I compied them all to device and now hci_smd module is loaded: https://paste.ubuntu.com/p/rsdT4dd75v/ | 09:16 |
| T42 | <edp_17> Now when I tried to turn BT on, it disconnected wifi and I cannot turn wifi back on. | 09:20 |
| rinigus | mal: do you have any issues with reading in CallInfo from AIDL? I seem to be able to get it in except the last 3 elements: additionalCallInfo, audioQuality, or modemCallId. maybe some of them are vectors (additionalCallInfo, audioQuality). although, smali code does save them as objects, not lists. | 09:52 |
| rinigus | current broken code (just to show to you) at https://github.com/rinigus/ofono-binder-plugin-ext-qti/blob/aidl/src/qti_radio_ext.c#L573 and below | 09:52 |
| T42 | <edp_17> mal: I am still puzzled how that hci_smd module (and others from backports) should go into the kernel-modules*.rpm | 09:55 |
| T42 | <edp_17> The backports has successfully ran and completed for the hammerhead but this module is not in the *kernel-modules*.rpm. | 09:56 |
| T42 | <edp_17> Okay, I think got it now. The make backports copies the modules into /home/nexus5/mer/android/droid/out/target/product/$DEVICE/system/lib/modules/ | 10:08 |
| T42 | <edp_17> As the i9505 and hammerhead has different android base and different kernel, the make backports builds different modules. | 10:09 |
| T42 | <edp_17> So, it seems, on the i9505, I have to copy those .ko files into /home/nexus5/mer/android/droid/out/target/product/$DEVICE/system/lib/modules/ and not into where I copied. | 10:10 |
| T42 | <edp_17> Yep! After I copied them into the right place, they did go into *kernel-modules*.rpm. Great. | 10:13 |
| mal | rinigus: I got that package_native working nicely from droidmedia | 14:18 |
| mal | rinigus: you can always skip parcelables which you don't need, like using gbinder_reader_read_parcelable(&reader, NULL); | 14:21 |
| rinigus | mal: great to hear! let me know when you push for droidmedia PR, so I could test it | 14:34 |
| rinigus | mal: gbinder_reader_read_parcelable was exactly what I used. but looks like something is off towards the end of the package (although we don't use that data in the plugin) | 14:35 |
| mal | I created a function called binder_skip_remaining_parcelable :) | 14:40 |
| rinigus | mal: I see. should probably do the same - tested so far with the single call and it could just ignore the end. wouldn't work for concurrent calls. good tip! | 14:42 |
| rinigus | but still, have you tried to deserialize it in full? or just went for a skip to keep it simple? | 14:42 |
| mal | which stores the parcelable size and initial size (how much has been read after reading parcelable size (gbinder_reader_bytes_read)), then just loops through until we have read everything from the parcelable | 14:43 |
| mal | I mean those parcelable size and initial size are of course read outside the function much earlier | 14:44 |
| mal | so the function just does a while loop where it does gbinder_reader_read_uint32(reader, NULL); | 14:45 |
| rinigus | I wonder if we can use that reader pointer somehow instead. but that approach would work. although, alternative is then to get whole parcelable and read from that memory object directly. | 14:49 |
| mal | rinigus: like this https://github.com/mer-hybris/ofono-binder-plugin/blob/master/src/binder_network.c#L587 | 14:49 |
| mal | not sure why I didn't add a function in there | 14:49 |
| mal | yeah, there might be other way also | 14:50 |
| rinigus | mal: thanks! will look into it tonight | 14:50 |
| mal | I was thinking if I should do some cleanup of services files | 14:51 |
| mal | in droidmedia | 14:51 |
| mal | it's such a mess | 14:51 |
| mal | but maybe in separate PR | 14:52 |
| rinigus | :) | 14:52 |
| mal | it hurts every time I see those files :) | 14:53 |
| mal | the Android.mk is even more terrible | 14:54 |
| rinigus | Android.mk did make me flip and generate this: https://github.com/rinigus/fake-package-native-android/blob/main/Android.bp . but that's not for droidmedia | 14:57 |
| mal | heh | 15:02 |
| mal | rinigus: https://github.com/sailfishos/droidmedia/tree/package_native | 15:18 |
| mal | I'll do some changes to that still | 15:37 |
| mal | should be ok now, the a13 part was broken earlier | 15:49 |
| rinigus | mal: thanks! will test | 16:14 |
| rinigus | mal: works as it should, thanks! | 17:49 |
| rinigus | b100dian: as mal added package_native to droidmedia, remove fake-package-native after you update. new configs don't require it anymore and droidmedia is also bumped at OBS | 17:50 |
| mal | rinigus: https://github.com/sailfishos/droidmedia/pull/131 make PR of that | 18:11 |
| rinigus | nice! | 18:13 |
| *** amccarthy is now known as Guest22735 | 19:28 | |
| *** amccarthy_ is now known as amccarthy | 19:28 | |
| T42 | <edp_17> mal: Thank you. You were right about that obexd. I had bluez4 related configs in droid-config. Once I removed it, OBS built fine. I am not sure whether it boots, but it's built. :) | 21:25 |
| T42 | <edp_17> @elros34 , mal : I've packaged an image with (I think) backported bluez5. The device booted but there is no Bluetooth option in Settings/Bluetooth menu. | 22:55 |
| mal | is the smd kernel module loaded? | 22:58 |
| T42 | <edp_17> mal: I don't have hci_smd module. The backports didn't make it. | 23:00 |
| mal | you have it in some defconfig? | 23:01 |
| T42 | <elros34> what are talking about? you literally modprobe it previously | 23:04 |
| T42 | <edp_17> @elros34 : That was for the S4 (i9505). This is for the Nexus 5 (hammerhead) | 23:05 |
| T42 | <edp_17> mal: I've ran the make backports and these are the modules it built: https://paste.ubuntu.com/p/W2h9kjwDgk/ | 23:05 |
| T42 | <elros34> does hammerhead even use qualcomm and hci_smd? | 23:05 |
| mal | good question, that should be seen in kernel defconfig | 23:07 |
| T42 | <edp_17> Let me check kernel defconfig. | 23:07 |
| T42 | <edp_17> Here is the kernel config from device: https://paste.ubuntu.com/p/XcMx2mWBKw/ | 23:09 |
| T42 | <edp_17> I this what I need CONFIG_BT_RFCOMM ? | 23:11 |
| T42 | <edp_17> or this CONFIG_BT_HCIUART ? | 23:12 |
| T42 | <elros34> this is broadcom bt | 23:15 |
| T42 | <edp_17> Okay. Does it mean I cannot use that backports? | 23:16 |
| T42 | <elros34> no idea, do you even need them? mal will probably know | 23:19 |
| T42 | <edp_17> I had to use bluez4 for that device until sfos 4.4, so I would need somekind of backport for sure. But I have no idea. | 23:20 |
| mal | so which kernel driver does it use for bt hardware? | 23:21 |
| T42 | <elros34> I think CONFIG_BCM4339/BCMDHD and HCIUART | 23:23 |
| T42 | <edp_17> Yes, correct. Those are in defconfig | 23:23 |
| mal | did it build hciuart module? | 23:24 |
| T42 | <edp_17> Yes, I got hci_uart.ko | 23:25 |
| T42 | <edp_17> I've loaded it with modprobe | 23:25 |
| mal | and no bluetooth device appears? | 23:26 |
| T42 | <edp_17> Nope. | 23:26 |
| T42 | <elros34> isn't it from your kernel instead from backports? | 23:26 |
| T42 | <edp_17> I have these modules loaded: https://paste.ubuntu.com/p/3y2htYhrDW/ | 23:26 |
| T42 | <edp_17> How can I check whether it is from kernel or backport? | 23:27 |
| T42 | <edp_17> According to this log, it is from kernel, not backport : https://paste.ubuntu.com/p/W2h9kjwDgk/ | 23:28 |
| mal | so is bt a module in kernel? | 23:29 |
| mal | i.e. CONFIG_BT=m | 23:29 |
| T42 | <edp_17> Yes | 23:29 |
| mal | have you checked what the commandline bluetooth tools say? | 23:30 |
| T42 | <edp_17> Commandline bluetooth tools? | 23:31 |
| mal | from like bluez5-tools or -utils or whatever it was called | 23:31 |
| T42 | <edp_17> I've installed bluez5-tools. Which one should I run? https://paste.ubuntu.com/p/8J9mnnPFh4/ | 23:34 |
| mal | bluetoothctl or hcitool might give something | 23:36 |
| mal | also do you see anything in bluez5 log in journal | 23:37 |
| T42 | <edp_17> bluetoothctl gives me: "Waiting to connect to bluetoothd..." | 23:39 |
| T42 | <edp_17> Bluetooth service is crashing. Journal: https://paste.ubuntu.com/p/5nyM3sHpcd/ | 23:39 |
| T42 | <edp_17> bluetoothd[14060]: src/adapter.c:adapter_init() Failed to access management interface | 23:40 |
| mal | and you are sure the bluetooth module is from backports? | 23:40 |
| T42 | <edp_17> I think the hci_uart is from the kernel, not from backport. | 23:41 |
| T42 | <edp_17> What packages do I need to install exactly? Here are the bluez* ones: https://paste.ubuntu.com/p/9G4kFyfqxZ/ | 23:46 |
| mal | the problem is in kernel side now | 23:48 |
| T42 | <edp_17> So, do I need to build your backports? I guess not. | 23:49 |
| T42 | <edp_17> I just need to apply those three patches into kernel, right? | 23:49 |
| T42 | <edp_17> strace bluetoothd: https://paste.ubuntu.com/p/BBqsT6tsbz/ | 23:50 |
| mal | you need to make sure you really have correct bluetooth modules loaded | 23:56 |
| T42 | <edp_17> This BT and backport are so confusing. | 23:57 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!