Sunday, 2020-05-17

T42<elros34> @edp_17 I don't know why this symbol is not exported. I wonder what  happen if you write nvram/firmware_path manually Another  way could be: copy nvram_net and bcmdhd_sta.txt to /lib/firmware and change default path for them in defconfig.00:22
T42<edp_17> Yeah, I am not sure either. I have tried to comment out #ifdef CUSTOMER_HW4, then together with #ifdef SUPPORT_LINKDOWN_RECOVERY, then all three with #if defined(CONFIG_MACH_UNIVERSAL5433) || defined(CONFIG_MACH_UNIVERSAL7420)00:24
T42<edp_17> but didn't make any difference.00:24
T42<edp_17> Now I am checking whether I can force to write the mac address back to /efs/wifi/.mac.info00:25
T42<edp_17> But I don't really like to idea to hardcoding the mac address.00:25
T42<edp_17> I check what you sent.00:25
T42<elros34> I built this kernel but for some reason I coudn't export any symbol from dhd_linux.c00:26
T42<edp_17> What should be the value to write in nvram/firmware_path?00:29
T42<elros34> but with /system prefix00:30
T42<elros34> same as default values in defoncig00:31
T42<edp_17> Can I change these on the device?00:32
T42<edp_17> And should I change WIFI_DRIVER_FW_PATH_AP too?00:33
T42<edp_17> You are right, on device /etc/wifi/ doesn't exist but /syste/etc/wifi/ does.00:34
T42<elros34> What do you mean? changing it in doesnt make sense. Use: or copy these files to / nad change default paths in kernel defconfig00:35
T42<edp_17> I have done this on the device:
T42<edp_17> And I changed the CONFIG_BCMDHD_FW_PATH="/sys/module/dhd/parameters/firmware_path" and CONFIG_BCMDHD_NVRAM_PATH="/sys/module/dhd/parameters/nvram_path"00:42
T42<edp_17> I recompile the kernel.00:43
T42<elros34> I assume these paths are just a typo00:46
T42<edp_17> Oops, the device is kept rebooting pretty much instantly.00:46
T42<edp_17> I have pushed the defconfig to github (branch cm-14.1-treltexx-v2)00:48
T42秦蒋豪焰 %lastname% was added by: 秦蒋豪焰 %lastname%00:50
T42<elros34> yeah so not a type, basically you tell dhd modules to load firmware from itself.00:51
T42<elros34> typo*00:51
T42<elros34> so again: CONFIG_BCMDHD_FW_PATH is default path to firmware which is used when your wifi driver is loaded. If its build as module then /system partition is mounted and it can find firmware there but if it's builtin /system is not yet mounted00:53
T42<edp_17> I've flashed the previous boot.img and now the device started.00:53
T42<edp_17> I understand, thanks for explaining.00:54
T42<elros34> I am not sure, your changes are not correct. Copy /system/etc/wifi/bcmdhd_sta.bin to /lib/firmware/  and set CONFIG_BCMDHD_FW_PATH=/system/etc/wifi/bcmdhd_sta.bin (simillar for nvram)00:55
T42<edp_17> Okay, I try this.00:56
T42<elros34> to CONFIG_BCMDHD_FW_PATH=/lib/firmware/bcmdhd_sta.bin00:56
T42<ankaos> I ask to learn. how will you write?00:57
T42<edp_17> @elros34 : Nope, the same result. The device is rebooting constantly.01:01
T42<elros34> show your changes, did you revert config_modules and your other changes?01:03
T42<edp_17> I pushed the latest defconfig to github.01:04
T42<edp_17> I didn't remove config_modules=y01:05
T42<edp_17> What is the /preload partition? Can I try to use these files from?01:11
T42<edp_17> Or do you know the first partition what is mounted before the wifi starts?01:12
T42<edp_17> I see that is /. That's why you suggested to put the files into /lib which is on /.01:14
T42<elros34> first is initramfs then /data. All other partitions are mounted by systemd which is probably too late. BTW It's unlikley that changin these 2 paths cause kernel reboot01:14
T42<edp_17> Hmm. then there is something else what caused that.01:16
T42<edp_17> What about if I put those files somewhere in /data?01:16
T42<edp_17> I see.01:18
T42<edp_17> Okay, I go back to an earlier commit and re-built the kernel, once it works, I'll try to change those two paths and see whether it reboots.01:19
T42<edp_17> @elros34 : You were right, those two lines didn't cause reboot. The root cause was my stupidity. I flashed the boot.img instead of hybris-boot.img.01:36
T42<edp_17> However the wifi doesn't work.01:36
T42<edp_17> With the files in /lib/firmware/ and with this in defconfig, it doesn't find any network. Only the circle is going round and round.01:37
T42<edp_17> I need some sleep now. Thanks for you help! Good night.01:41
T42<Amber %lastname%> hey guys i am having issues with sdk manage command can you please check below and help me about this03:47
T42<Amber %lastname%>
T42<Rodrigosolari> @Amber %lastname% [], Try in the sailfish-sdk03:49
T42<Rodrigosolari> Not in the Systems root03:49
T42<Amber %lastname%> how do i run sailfish sdk03:50
T42<birdzhang> follow the HADK03:50
piggzrinigus: reckon we can port this to sfos?
riniguspiggz: that's for fingerprint, right?09:21
piggzit is09:21
piggzubports support for FP09:21
riniguspiggz: interesting.09:24
piggzthat gives us an example implementation, ust need to figue how to mash it with sfos09:25
riniguspiggz: ideally, we should join them in its development. if it is a daemon then it could run probably as it is on sfos09:26
rinigusand then we have to see how to add it into lockscreen unlocking...09:26
piggzrinigus: looking at the commit history, nothing since 2018, then added FP support09:31
piggzreally, no development since 201709:32
riniguspiggz: but is developed now. like you could take popcorn, sit back, and follow09:33
riniguswe could ask in ubports channel regarding it.09:34
piggzyup ... all the new stuff is for FP support, just need to figure out how to utilise it09:34
piggzi guess we start with trying to build09:35
rinigus(hitting enter before message)09:35
riniguspiggz: that's a good start. also need to figure out how to plug any kind of extra authentification into sfos.09:36
rinigusbut I guess some developed patches can help - like maze login...09:37
rinigus... I'll still continue with compass. want to use rotation_vector (or its geo form) as provided by android devices. changed sensorfwd accordingly, but something is still struggling with it. debugging / working on it09:38
T42<Sid127> I have a question - how do I work with a prebuilt version of Lineage OS for my device?10:11
T42RILuse was added by: RILuse10:36
T42SilY@n %lastname% was added by: SilY@n %lastname%10:37
T42<SilY@n %lastname%> Good afternoon! Please help me understand. I can't fix these configurations.
rinigusmal: just got compass working using SENSOR_TYPE_ROTATION_VECTOR and going to test SENSOR_TYPE_GEOMAGNETIC_ROTATION_VECTOR11:25
rinigusthat way it is using android-provided sensors, possibly done in hardware. sensorfw will get a new orientationadaptor(s) that use the rotation vector to calculate azimuth11:26
rinigusthen the path is similar to what we have: orientationadaptor=hybrisrotationadaptor in sensorfw; qt Sensors.conf will use sensorfw.compass.11:27
rinigusso, we could add generic.compass for devices that don't have rotation vectors provided and use rotation vectors in hybris devices11:27
rinigusI presume that the rotation vector and its geo variant should be there for newer devices11:28
malrinigus: I did try rotation vector before but had some issues11:33
rinigusmal: it looks to be working very well on tama11:33
rinigusgeo option haven't tried yet, its building now11:34
rinigusmal: do you remember which issues were there?11:34
malthe values didn't make sense, maybe I did something wrong11:35
malplease push some code somewhere so I can test it11:35
rinigusmal: as I have to build at obs (some tech issues @home), its all out there, with every typo I made:
rinigusadaptor at
rinigusif all works, I will clean up and make new branch with merged commits11:37
rinigusmal: ^11:37
malI will need to implement the binder side because that is used on x1011:38
rinigusbut, as you mentioned earlier, we need generic.compass for native ports11:38
rinigusmal: that will just require the corresponding data vector access.11:38
malyes, I still think having sensor fusion sensorfw might also make sense, probably easier to maintain11:40
rinigusmal: there is info regarding calibration (see
rinigusbut its not used currently. will have to figure that out too11:40
maldo you mean the values[4] is -1?11:41
malto determine the accuracy11:41
rinigusmal: I am not using that value at all. as d->level_ is int, not sure how to convert that and what's expected radian accuracy to pass levels test in csd11:44
malI think we can figure out some guess what is a good accuracy in degrees11:45
rinigusexactly! right now, its a dirty implementation and have to clean up a lot. but it is good for testing.11:45
maldid it give better values than generic.compass?11:46
malrinigus: I think sensorfw has some filtering already for compass value so it might give ok-ish reaction11:47
malof course assuming the values are correct11:47
rinigusmal: the end values were very close on tama. but note that the values were already before close to what physical compass would show11:48
rinigusmal: but response time with rotation_vector was very good. I could move phone fast and it would be able to follow with rather small delay11:49
malyeah, well I consider 20 degree error a quite big one11:49
malrinigus: generic.compass had a bit slow response, maybe the coefficient you modified affected that11:50
rinigusmal: for me, the angle was <10 degrees11:51
rinigusmal: generic.compass should be faster with smaller filter_coeff, at least with gyro. well, will have to test it later11:53
rinigusmal: but maybe I wrong. ignore it regarding generic compass.11:55
T42<SilY@n %lastname%> WARNING: CONFIG_AUDIT is invalid11:55
T42<SilY@n %lastname%> Value is: y11:55
T42<SilY@n %lastname%> Allowed values : n, !11:55
T42<SilY@n %lastname%> Comment says: This will disable SELinux! That's ok, because hybris adaptations must not have SELinux, but if your device needs its support in kernel, set AUDIT=y and SELINUX_BOOTPARAM=y. Then disable them via kernel cmdline: audit=0 selinux=0. You can also leave audit enabled, if you don't plan to use systemd's containers: h11:55
T42<SilY@n %lastname%> WARNING: CONFIG_MODULES is invalid11:55
T42<SilY@n %lastname%> It is unset11:55
T42<SilY@n %lastname%> Allowed values : y, !11:55
T42<SilY@n %lastname%> Comment says: optional, required for module support (Such as WLAN for example)11:55
T42<SilY@n %lastname%> WARNING: CONFIG_CGROUP_MEM_RES_CTLR is invalid11:55
T42<SilY@n %lastname%> It is unset11:55
T42<SilY@n %lastname%> Allowed values : y, !, <=3.511:55
T42<SilY@n %lastname%> Comment says: systemd (optional):, only valid if kernel version <= 3.511:55
T42<SilY@n %lastname%> WARNING: CONFIG_CGROUP_MEM_RES_CTLR_SWAP is invalid11:55
rinigusmal: SENSOR_TYPE_GEOMAGNETIC_ROTATION_VECTOR also works quite well and should be fine for compass.12:02
T42<ankaos> Fwd from ankaos: 426/427] including vendor/xiaomi/msm8953-common/ ...12:02
T42<ankaos> build/make/core/ error: vendor/xiaomi/msm8953-common: MODULE.TARGET.SHARED_LIBRARIES.libantradio already defined by external/ant-wireless/ant_native.12:02
T42<ankaos> 11:55:16 ckati failed with: exit status 112:02
rinigusI guess there would be Geo and plain rotation_vector adaptors allowing porters / users to select the favorite one12:02
T42<ankaos> how fix this error?12:02
malrinigus: there is some note in sensor docs about geomagnetic being more noisy in some cases12:16
malrinigus: yes, I think both should be supported and porters can select which one works better12:17
malrinigus: please make a branch with both so I can do some testing easily12:17
rinigusmal: will do. question is now how to reuse the code. looks like I will have to make separate folders, one per adapter.12:19
malrinigus: good question, wondering if there could be some way to reuse things12:20
rinigusmal: the code is the same, just constructor will have to get rotation vector source. I'll see how to do that taking into account the build system12:21
T42<ankaos> ı used to
T42<ankaos> what else?12:22
mal@ankaos remove the conflicting sources (maybe the one in external) and see if it builds12:24
malrinigus: those hybris adaptors are all quite the same so maybe just different folders12:26
malrinigus: some redesign of that might make sense, although the whole sensorfw is a bit overly complicated system12:27
malso maybe doesn't make sense to do it now12:27
T42<ankaos> okey testing your say12:27
rinigusmal: OK, will put something together...12:28
malrinigus: so just different folders for now and we can think of other ways later12:29
T42<ankaos> ninja: error: '/home/enes/hadk/out/target/product/tissot/obj/STATIC_LIBRARIES/libawk_main_intermediates/export_includes', needed by '/home/enes/hadk/out/target/product/tissot/obj/EXECUTABLES/recovery_intermediates/import_includes', missing and no known rule to make it12:29
rinigusmal: OK, will do12:29
T42<ankaos> O.o12:29
maladd the missing dependency to manifest, check the lineage manifest for the line you need12:30
T42<elros34> you had one mount unit (systemd.mount) to fix to get droid-hal-init working but instead you switched whole base to 16?12:43
T42<ankaos> @elros34 [you had one mount unit (systemd.mount) to fix …], System.mount error not fix. I dont fint answer12:48
T42<elros34> what do you mean? Does it fails to bind mount or droid-hal-init still fail?12:51
rinigusmal: has the both adaptors.12:58
rinigusmal: names are hybrisgeorotationadaptor and hybrisrotationadaptor12:58
T42<ankaos> @elros34 [what do you mean? Does it fails to bind mount …], both13:28
T42<elros34> you need to be more verbose: show journal, cat system.mount system_root.mount13:29
T42<elros34> in a way that is clear which file do you show, not like in your previous paste13:33
rinigusmal: that constant in generic.compass should play a role iff you have gyro available. to me, it looks like the value which I used earlier is a reasonable balance.13:41
T42<ankaos> @elros34 [you need to be more verbose: show journal, cat …],
T42<ankaos> and cat system.mount system_root.mount14:24
T42<elros34> yeah that is wrong, you should not edit system_root.mount at all14:27
T42<elros34> not sure if I gave you correct options earlier but system.mount should be like:
T42<ankaos> @elros34 [yeah that is wrong, you should not edit system …],
T42<ankaos> this is system_root14:32
T42<elros34> yes that looks correct14:32
T42<ankaos> @ankaos [], this is  journal14:40
T42<ankaos> system.mount correct?14:40
T42<elros34> no it's not I just gave you the correct one few minutes ago14:41
T42<ankaos> @elros34 [no it's not I just gave you the correct one fe …],
T42<ankaos> @elros34 [no it's not I just gave you the correct one fe …], this is system.mount14:43
T42<ankaos> and only system14:44
T42<elros34> like I said it's incorrect, use this:
T42<ankaos> you sait this
T42<elros34> no, I gave 2 days ago wrong one with swapped Type and Options but for sure not that14:47
T42<ankaos> @elros34 [no, I gave 2 days ago wrong one with swapped T …],
T42<elros34> focus, still incorrect14:49
T42<ankaos> line?14:49
T42<elros34> 6 ...14:50
T42<ankaos> What=/system/system -> What=/system_root/system14:50
T42<ankaos> ?14:50
T42<ankaos> (Photo, 960x1280)
T42<ankaos> Than devire reboot14:53
T42<elros34> after it reboot itself, try cat /proc/last_kmsg, and journalctl -f14:54
T42<ankaos> I can't enter the terminal.14:57
T42<elros34> in TWRP: 'touch /data/.stowaways/sailfishos/init_enter_debug2'  Next time it will stop booting system until you use: 'echo continue > /init-ctl/stdin'15:01
rinigusmal: cleaned up and added level calculations. on tama, it returns heading accuracy as zero. adjusted level calculations to make csd happy if the accuracy is smaller than 10 degrees15:19
rinigusmal: all changes squashed at
rinigusnow we need binder part and then it should be ready for upstream15:20
rinigusor do I miss something?15:20
malI'll have a look15:41
T42<ankaos> telnet not open15:53
T42<ankaos> slot a install sailfish os b slot install recovery.16:02
T42<ankaos> but not see sailfish os.16:02
T42<ankaos> in twrp 'touch /data/.stowaways/sailfishos/init_enter_debug2' than reboot system.16:39
T42<ankaos> @elros34 all right?16:42
T42<ankaos> `'echo continue > /init-ctl/stdin'` in twrp or telnet?16:45
malrinigus: seems like x10 has the same issue on rotation sensors than it has on orientation sensor, both only get couple of values and then it gets stuck17:04
T42<ankaos> @ankaos ['echo continue > /init-ctl/stdin' in twrp or t …], because ı used this touch /data/.stowaways/sailfishos/init_enter_debug2 command . telnet not open Elros.17:05
rinigusmal: and it is not due to the binder, right? as it is expected to be the same as hal17:05
malI can test hal version also17:06
rinigusmal: probably it is some hw issue then... on tama it works nicely17:06
malor some bug in firmware17:06
malor blobs17:07
rinigusyes, that's most probable17:07
malrinigus: same with hal plugin17:12
rinigusmal: which is good as we can trust results from any of them...17:13
malrinigus: the direction I get looks good but the problem is that it just won't update after couple of values17:13
rinigusmal: xperia10 can then use the generic.compass17:13
malwe had similar issue in the past with compass but that was fixed, not sure if this is related or not17:14
rinigusmal: I can keep it on for a while, no problem with that here17:14
malrinigus: for example in orienteering compass app if I minimize the app and toggle the pause button in cover it updates the value17:15
rinigusmal: do you know whether xperia 10 compass works in aosp?17:15
malnot sure, would need some testing17:15
malbased on some info it might be broken in aosp also17:16
rinigusmal: I think that aosp apps are using rotation vectors for it.17:16
rinigusso it could be an indicator. well, if generic works, it should be ok as well17:16
rinigusbut it is strange that it just stops and needs restart17:17
malyes, the sensor does send data though, it just sends the same value over and over again17:17
malwhich does suggest blob/firmware issue17:18
rinigusmal: what should we do now with the code? do I submit sensorfw for merge and you add binder part later? also I would like to submit qtsensors generic compass as well17:19
rinigusis there some specific MER# or JOLLA# bug for compass?17:20
malyou should cleanup the code a bit, some whitespace issue and some unused commented code17:24
malalso the binder codes are ok so just fix those17:24
rinigusmal: will do.17:29
rinigusmal: cleaned up and submitted for review - sensorsfw and qtsensors17:56
malrinigus: a tiny nitpick, in other parts of the sensorfw code the { after if is in the same line, not on the next line18:07
malrinigus: also please squash the commits into one in sensorfw18:08
T42<ankaos> @elros34 [in TWRP: 'touch /data/.stowaways/sailfishos/in …], not work18:21
rinigusmal: re { - those seem to be mixed up in the code. I used the same as in hybrisorientationadaptor. looking at it, I don't think I have added any { }. but I don't mind to change it in that `if`18:23
malrinigus: ah, interesting, I just happened to check the hybrisadaptor.cpp18:24
malI'll grep the code a bit18:25
riniguswell, its a small thing, I have changed it already18:25
mallooks like the code is a bit of a mess18:27
piggzmal: ive installed the latest tooling and target locally.......but18:28
piggzwhen I try and install dbus-cpp-devel from you anbox repo, it fails with18:28
piggzProblem: nothing provides needed by libdbus-cpp5-5.0.0-1.1.15.jolla.armv7hl18:28
piggzhave the hadk tagets not been updated yet with the new compiler etc?18:29
rinigusmal: change submitted and its rebuilding for testing at OBS18:30
piggzmal: yes, the latest is Device release is currently:
piggzatleast the target exists18:43
piggzjust the lastes is pointing to the wrong one18:43
malseems like latest is not updated yet18:45
T42<elros34> @ankaos this might be useful if you can't telnet:
T42<ankaos> @elros34 [@ankaos this might be useful if you can't teln …], use19:23
T42<ankaos> and ~ # echo continue > /init-ctl/stdin19:24
T42<ankaos> /sbin/sh: can't create /init-ctl/stdin: nonexistent directory19:24
T42<elros34> I don't get, so now you were able to telnet? Which port?19:28
T42<ankaos> telnet 232319:28
T42<elros34> so now grab logs ask previous. last_kmsg will be available only after kernel crash so it might be missing19:30
T42<ankaos> @elros34 [so now grab logs ask previous. last_kmsg will …], telnet not work19:37
T42<ankaos> `touch /data/.stowaways/sailfishos/init_enter_debug2` not open telnet19:37
T42<elros34> sorry I don't get you at all. You should already learn how to check whether sailfish starts at all(host dmesg)19:41
T42<ankaos> dmesg
kalube@ankaos are you building on an external drive?19:47
T42<elros34> don't show me it, just tell whether it 's correct then use instruction from hallium:  enp*: renamed from usb0  in dmesg19:49
T42<ankaos> Building sailfish for a119:51
piggzmal: string link error in dhd build regardin img2simg...20:36
piggzthe function is in the included sources20:37
piggzso not sure why its not linking20:37
T42<elros34> @ankaos could you answer kalube question. It could be quite important because external drive can be mounted with nosuid,nodev options. Did you have any /dev/ errrors when creating sfos image?20:44
T42<ankaos> where will I look at it?20:44
malpiggz: why didn't you have the issue before?20:52
T42<elros34> @ankaos in mount output, mic output20:52
piggzmal: i dont know ... all ive done is update the tooling/target20:52
T42<ankaos> @elros34 [@ankaos in mount output, mic output], is this
malpiggz: could it be that you have some old leftover things in system/core repo from the previous builds20:59
piggzi already tried deleting some object files, ill try some more21:00
piggzmal: yes, deleting obejcts from base/ helped21:04
T42<ankaos> Sorry for not knowing as much as you.21:06
piggzmal: why would a new target create these errors tho!
malpiggz: I think the detection of unpackaged files was fixed21:15
T42<elros34> rpm fix was finally merged in 3.3.021:15
malpiggz: it used to be broken for some time21:15
T42<ankaos> @ankaos [is this], @elros34 Did you see it?21:18
T42<elros34> sure but how this is related to question whether you use external drive or 'mount' command output, mic errors?21:19
T42<ankaos> I still couldn't figure out how to take care of it.21:19
T42<elros34> just type mount21:20
T42<ankaos> @elros34 [just type mount],
T42<elros34> on your host system21:22
T42<ankaos> @elros34 [on your host system], device or pc?21:23
kalube> host21:30
kalubethat would be the one with the keyboard on it21:30
piggzunless its a pro1 :D21:31
kalubeI don't think we need any more complication here -_- :P21:32
T42<ankaos> @elros34 [on your host system], @elros34 and Kalube21:35
T42<elros34> does 'df $ANDROID_ROOT' returns /dev/sda3?21:38
T42<ankaos> @elros34 [does 'df $ANDROID_ROOT' returns /dev/sda3?],
T42<elros34> that variable is valid in sdk, so just answer whether you are using /media/enes/1E8C40128C3FE2C71 to build sailfish?21:41
T42<ankaos> maybe.21:44
T42<elros34> ok so you are not using external drive so forget about it21:47
T42<ankaos> there are only internal drive21:48
T42<ankaos> Another reason may be due to restart?21:54
T42<elros34> no more guesses, you need to get logs21:55
T42<ankaos> how?21:55
T42<ankaos> journalctl -f etc.21:56
kalube@ankaos you should be reading and re-reading the hadk and any other documentation you can find to get an understand of what's going on before asking so many questions over here, we can only help so much but if you can't figure it out for yourself you'll end up getting stuck21:56
T42<ankaos> then i will wait. If someone says, I use it.22:08

Generated by 2.17.1 by Marius Gedminas - find it at!