| T42 | <edp_17> @elros34 , Hi, I think when I built backports and hybris-hal as I specified them on the same make command, an error was suppressed. Now, I built hybris-hal first, then backports and the process took me through a series of question about what driver I wanted as built in or moduile (as the attached log shows). Then, it errored. I think that's why I don't have backported drivers. Log: https://paste.ubuntu.com/p/kf4PGXzSY8/ | 07:03 |
|---|---|---|
| T42 | <edp_17> mal, the error is "external/backports-bluetooth/backport-include/linux/acpi.h:60:56: warning: 'struct acpi_device' declared inside parameter list [enabled by default] | 07:11 |
| T42 | <edp_17> error, forbidden warning: acpi.h:60" | 07:11 |
| T42 | <edp_17> Do you know how to resolve this? | 07:11 |
| T42 | <elros34> I am not sure you are supposed to built all at once. Look at docs: "IMPORTANT: if you rerun make hybris-hal at any time you will always have to rerun make backports after that" | 07:15 |
| T42 | <edp_17> Yep, that was the mistake I made. | 07:16 |
| T42 | <elros34> now you enabled to many options. How about you remove .config and try again make backports | 07:22 |
| T42 | <elros34> you don't need hci_inter driver | 07:24 |
| T42 | <elros34> intel* | 07:24 |
| T42 | <edp_17> I've deleted .config from external/backports-bluetooth and re-run make backports. This time I didn't get those questions and the build completed without error. Log: https://paste.ubuntu.com/p/dvF43hd4wx/ | 07:26 |
| T42 | <elros34> as expected and you have even hci_uart | 07:27 |
| T42 | <edp_17> Now I have hci_smd.jo in out/.../system/lib/modules and yes, hci_art.ko is there too. | 07:28 |
| T42 | <edp_17> Yay, and those are in the droid-hal*-kernel-modules*.rpm too. :) Looks promising. | 07:29 |
| T42 | <edp_17> Okay, I've installed the rpms on device and now I can load hci_smd module and bluetooth service starts and looks like happy (not crashing). However, the Bluetooth option in Settings/bluetooth is still missing. | 07:38 |
| T42 | <edp_17> That page shows me an empty screen | 07:39 |
| T42 | <edp_17> What package am I missing? https://paste.ubuntu.com/p/9HF9MSQnCH/ | 07:45 |
| T42 | <elros34> smd not uart? | 07:47 |
| T42 | <edp_17> Well, I have both driver and they both load fine. | 07:50 |
| T42 | <elros34> but IIRC we know that this is broadcom device with uart interface not smd or I am mixing devices? Didn't you already use uart module and and all these hciattch stuff? | 07:52 |
| T42 | <edp_17> Yes, you are right. I change to load hci_uart and install bluetooth-rfkill-event-hciattach & broadcom-bluetooth-bluez5 packages | 07:53 |
| T42 | <edp_17> I am still puzzled why I don't have Bluetooth option in Settings. | 07:54 |
| T42 | <elros34> that is less important than checking logs especially these related to hciattach, rfkill | 07:55 |
| T42 | <edp_17> rfkill runs but hciattach gives an error in log: Failed to start Run hciattach when HCI UART device becomes available. | 08:02 |
| T42 | <edp_17> It seems /usr/sbin/hciattach is missing. | 08:04 |
| T42 | <edp_17> Hmm, the hciattach service try to execute: /usr/sbin/hciattach -f /system/vendor/firmware ttyHS99 bcm43xx 4000000 flow sleep /data/property/persist.service.bdroid.bdaddr | 08:15 |
| T42 | <edp_17> However, the option 'f' is invalid: /usr/bin/hciattach: invalid option -- 'f' | 08:15 |
| T42 | <edp_17> Usage: | 08:17 |
| T42 | <edp_17> hciattach [-n] [-p] [-b] [-r] [-t timeout] [-s initial_speed] [-d] [-m] <tty> <type | id> [speed] [flow|noflow] [sleep|nosleep] [bdaddr] | 08:17 |
| T42 | <edp_17> hciattach -l | 08:17 |
| T42 | <elros34> I wonder why you didn't have all these packages installed. How did you had working bluez4 without all of them | 08:26 |
| T42 | <fredldotme> if it helps, there’s also brcm-patchram-plus for firmware uploads on Broadcom BT chips: https://github.com/fredldotme/brcm-patchram-plus/tree/master | 08:27 |
| T42 | <elros34> and he should have something similar already installed for older bluez and sfos | 08:28 |
| T42 | <elros34> there were also patched rfkill-event but not sure if for this device | 08:34 |
| T42 | <edp_17> I am not sure how bluez4 worked on there. | 09:13 |
| T42 | <edp_17> Let me reboot into sfos4.4 to check. | 09:14 |
| T42 | <edp_17> The device on sfos 4.4 is not using brcm-patchram-plus | 09:19 |
| T42 | <edp_17> Haha, there are no bluetooth-rfkill-event-hciattach, bluetooth-rfkill-event nor broadcom-bluetooth-bluez4 | 09:20 |
| T42 | <edp_17> However, I do have hciattach service and it is using that option 'f' https://paste.ubuntu.com/p/wqFN75p7cc/ | 09:23 |
| T42 | <edp_17> @elros34 , On sfos4.4 for hammerhead, the hciattach binary is provided by your bluez4 pretty hostname package. | 11:15 |
| T42 | <edp_17> It also provided some other stuff that I don't have on sfos 4.5 (example: /lib/udev/hid2hci and /lib/udev/rules.d/97-bluetooth-hid2hci.rules, etc.) | 11:16 |
| T42 | <edp_17> I've replaced the -f option to -s in hciattach service and Bluetooth option has appeared in Settings. | 14:33 |
| T42 | <edp_17> Hmm, "/usr/bin/bt-sleep-monitor: line 24: hcitool: command not found" | 14:37 |
| T42 | <edp_17> where from can I get that hcitool? | 14:37 |
| T42 | <edp_17> Okay, I found it, the bluez5-tools rpm provided it. | 14:40 |
| T42 | <edp_17> I get this error now: "Bluetooth device /org/bluez/hci0 not enabled" | 14:41 |
| mal | check rfkill | 14:41 |
| T42 | <edp_17> mal, it is running. Or how do you mean check? | 14:42 |
| mal | I mean check using rfkill command if bluetooth is blocked and unblock if needed | 14:43 |
| T42 | <edp_17> https://paste.ubuntu.com/p/Q6W8DWvRfs/ | 14:44 |
| T42 | <edp_17> bluetoth hci0 is blocked | 14:44 |
| T42 | <edp_17> When I turn BT on, both the SOFT and HARD are unblocked | 14:44 |
| T42 | <edp_17> When I turn BT off, the SOFT goes blocked the HARD remains unblocked. | 14:46 |
| T42 | <edp_17> BT now turns on/off, but device cannot find any other devices and it cannot be found either. Something is still missing. Journal when I turn BT on: https://paste.ubuntu.com/p/KKYrCzRYdS/ | 15:01 |
| T42 | <elros34> maybe start with fixing hciattach warnings. checking kernel logs is obviously next step | 18:00 |
| T42 | <edp_17> Will try that. | 18:02 |
| T42 | <elros34> if you are going to use this hciattach then check all arguments. I have impression they are wrong like -s /system/vendor/firmware | 18:12 |
| T42 | <edp_17> It was -f /system/vendor/firmware but I got an error saying option f is not allowed. | 19:12 |
| T42 | <edp_17> That error is interesting because exactly the same command is used on sfos 4.4 where bluetooth works. | 19:13 |
| T42 | <elros34> you can't change -f <firmware path> to -s <firmware path> just like that where -s is some speed. You are using now different hciattch | 19:22 |
| T42 | <elros34> so is not really same command | 19:22 |
| T42 | <edp_17> On sfos 4.4 I used the one from your repo (pretty hostname) on this version, sfos 4.5, I use the one came with bluez5. | 19:35 |
| T42 | <edp_17> Is it possible for you to update your fork? | 19:35 |
| T42 | <elros34> what for? IIRC this had just one tiny irrelevant | 19:37 |
| T42 | <elros34> now chang | 19:37 |
| T42 | <elros34> e | 19:37 |
| T42 | <elros34> just search for your driver name + hciattch in web and for sure you will find correct command for initialization | 19:38 |
| mal | rinigus: I merged the droidmedia PR for package_native, so the custom service can be dropped | 19:41 |
| rinigus | mal: way ahead of you - done already before merge by using your PR :) | 19:42 |
| rinigus | today recompiled with new tagged version | 19:42 |
| mal | good | 19:43 |
| mal | I disabled avc hw codec for now on fp4 and fp5, need to debug it more some other time | 19:44 |
| T42 | <edp_17> Thanks @elros34 | 19:47 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!