| T42 | <duhjokersinhell> im at the repo init stage of the HADK and im building for android 9 pie which hybris is that please | 03:46 |
|---|---|---|
| T42 | <eriki73> hybris-16.0 | 03:49 |
| T42 | <eriki73> @duhjokersinhell | 03:49 |
| T42 | <duhjokersinhell> ty | 03:49 |
| T42 | <eriki73> Found out why my kernel was not booting... | 03:49 |
| T42 | <eriki73> https://github.com/fredldotme/android_kernel_google_bonito/commit/ef593d74f7f60996f7d1abff1fb3022b4f18e098 | 03:49 |
| T42 | <duhjokersinhell> @eriki73 are you building for pixel 3a? | 03:50 |
| T42 | <eriki73> Yes? | 03:50 |
| T42 | <duhjokersinhell> thats what im starting | 03:51 |
| T42 | <duhjokersinhell> i know that for UT the halium-boot is not used and instead you make a boot.img where jumper-cable takes over does that also apply with hybris | 03:56 |
| T42 | <eriki73> My kernel does not take in the command line option from boot.img... It does not print them in the journal... | 04:17 |
| rinigus | @eugenio_g7: no, I haven't tried it. Don't know what that tls padding is, but could search the logs. Sounds like it was mentioned over here earlier | 04:26 |
| T42 | <eriki73> Nvm, cat /proc/cmdline displays those... | 05:25 |
| T42 | <eriki73> Unable to read config file '/product/etc/init': open() failed: No such file or directory | 05:25 |
| T42 | <eriki73> This might be interesting tho... | 05:25 |
| erikinkinen | Testing | 05:30 |
| T42 | <eriki73> Ok, now I found the correct chat... | 05:31 |
| erikinkinen | Somebody should update the tg description... It contains wrong server... | 05:33 |
| T42 | <eugenio_g7> rinigus: had a similar problem to you in droidian, although unrelated to flatpak - reserving more space in the tls area has helped. Will try on flatpak this weekend :) | 06:19 |
| rinigus | @eugenio_g7: I wrote down in the issue your comment and will come back to it later, after encryption will be sorted out. so, looking forward to your experiments with flatpak | 06:20 |
| rinigus | fingers crossed that it will work | 06:20 |
| rinigus | I wonder whether droidian would require something like flatpak-runner to make flatpaks usable... as getting keyboard input and screen rotation taken care of | 06:21 |
| rinigus | @eugenio_g7 ^ | 06:21 |
| T42 | <eugenio_g7> flatpaks (gtk ones at least) work fine already, including keyboard and rotation - of course without the hybris extensions, I suspect something like flatpak-runner would be necessary to properly set the environment when using hybris | 06:32 |
| T42 | <adampigg> @eugenio_g7 rinigus: tls-padding is packaged in yggdrasil:los17 if you need it | 06:40 |
| T42 | <adampigg> An amazing piece of code which on the face of it does nothing 😂 | 06:41 |
| rinigus | @eugenio_g7: gtk for us is no go due to old wayland composer. and without hw acceleration, qml apps are not working well either (missing toolbar on angelfish). so, looking forward to what you can do on the weekend | 06:42 |
| rinigus | @adampigg: it must be record lib if you calculate (lines of comments + lines of packaging + misc other things) / (line of real code) | 06:43 |
| T42 | <eugenio_g7> haha | 06:44 |
| T42 | <eugenio_g7> rinigus: will let you know :) | 06:44 |
| rinigus | thanks! | 06:44 |
| T42 | <eriki73> got droid-hal-init starting but now every possible service crashes | 07:04 |
| T42 | <adampigg> @eriki73 completely normal early porting :) | 07:06 |
| T42 | <eriki73> I guess this is the reason: Opening '/dev/hwbinder' failed: Permission denied | 07:07 |
| erikinkinen | https://paste.ubuntu.com/p/rSf8zNxr6P/ | 07:57 |
| erikinkinen | Got it booting this far... | 07:58 |
| T42 | <elros34> @eriki73 you can find solution for it in hadk-faq hybris-16 section but put service in /usr/lib/systemd instead /lib/systemd/ | 09:01 |
| T42 | <eriki73> @elros34 , did that but had a typo... | 09:02 |
| T42 | <eriki73> Hey, got gui | 09:02 |
| T42 | <eriki73> Wallpaper did not load, @elros34 | 09:08 |
| T42 | <elros34> IIRC solution should be in notes | 09:10 |
| T42 | <eriki73> @duhjokersinhell , sorry for ping, but got Pixel 3a to GUI. | 09:27 |
| T42 | <eriki73> Wlan isn't working although it is built inside the kernel... | 09:47 |
| T42 | <amyosx> /notes (re @elros34: IIRC solution should...) | 09:59 |
| T42 | <amyosx> Ahh | 10:00 |
| T42 | <elros34> @amyosx unofficial notes: http://collabedit.com/kr9xx | 10:01 |
| T42 | <amyosx> Ah | 10:01 |
| T42 | <eriki73> Yeah, and got the wallpaper fixed... | 10:06 |
| T42 | <eriki73> But WLAN is a different story... All documentation is about WLAN modules and I have WLAN baked in the kernel | 10:07 |
| T42 | <elros34> it's usually builtin but also it's usually not in issue to make it as module. If you want to keep it builtin then you need to figure out whether it has some interface for deferred initialization | 10:09 |
| T42 | <eriki73> I did. | 10:13 |
| T42 | <eriki73> echo 1 > /dev/wcnss_wlan | 10:13 |
| T42 | <eriki73> echo sta > /sys/module/wlan/parameters/fwpath | 10:13 |
| T42 | <eriki73> But it didn't do anything | 10:13 |
| T42 | <eriki73> It did execute successfully... | 10:13 |
| T42 | <eriki73> @elros34 | 10:14 |
| T42 | <b100dian> Do you see firmware loading for it? | 10:14 |
| T42 | <elros34> good point, did you follow advice about fw_loader kernel config? | 10:14 |
| T42 | <eriki73> Advice from kernel checker? | 10:15 |
| T42 | <elros34> yeah it's wrong, you should enable all fw_loader* options | 10:16 |
| T42 | <eriki73> I did not get any advice but I guess the default was off in my defconfig. | 10:17 |
| T42 | <eriki73> + there aren't any in the defconfig | 10:17 |
| T42 | <elros34> I was talking about it yesterday and put it in notes also | 10:18 |
| T42 | <eriki73> What are the exact options? | 10:18 |
| T42 | <eriki73> My defconfig doesn't contain them | 10:18 |
| T42 | <elros34> zcat /proc/config.gz | grep -i fw_loader | 10:19 |
| T42 | <eriki73> Ah, of course... | 10:19 |
| mighty17[m] | `Product: Failed to boot init in real rootfs` | 10:22 |
| mighty17[m] | i can `telnet 192.168.2.15 23` | 10:22 |
| mighty17[m] | https://paste.debian.net/1209261/ my init.log | 10:23 |
| mighty17[m] | i think its a mounting/fixup-mountpoint issue but can anyone confirm? | 10:24 |
| T42 | <elros34> if you are unsure then check diagnosis.log | 10:25 |
| T42 | <elros34> line 156 | 10:26 |
| mighty17[m] | so it is fixup mountpoint issue isnt it? | 10:27 |
| T42 | <elros34> yes | 10:27 |
| mighty17[m] | `CONFIG_AUTOFS4_FS=y not found in /proc/config.gz` I dont think 3.4 kernel has that config | 10:27 |
| T42 | <adampigg> Mal: reported bug? | 10:28 |
| T42 | <edp_17> Interesting, I got that black background issue but once lipstick restarted, it was gone. (Without applying the workaround mentiond in the notes.) | 10:32 |
| mighty17[m] | @elros34 i can just remake hybris boot and flash it with proper mountpoints right? | 10:39 |
| T42 | <eriki73> @elros34 , nope, https://paste.ubuntu.com/p/gpq9RFg2cp/ | 10:45 |
| mighty17[m] | `Product: init-debug in real rootfs` now whats this? | 10:46 |
| mighty17[m] | i cannot telnet | 10:46 |
| T42 | <TheVancedGamer> It booted (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>i canno...) | 10:47 |
| mighty17[m] | no ui | 10:47 |
| T42 | <TheVancedGamer> Ssh time | 10:47 |
| T42 | <TheVancedGamer> It's in debug in rootfs | 10:47 |
| mighty17[m] | ??? | 10:47 |
| T42 | <TheVancedGamer> It's in debug mode in the rootfs | 10:47 |
| mighty17[m] | even ssh doesnt seem to work | 10:48 |
| T42 | <TheVancedGamer> Wh~ (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>even ss...) | 10:48 |
| T42 | <TheVancedGamer> How | 10:48 |
| T42 | <TheVancedGamer> Check ifconfig | 10:48 |
| mighty17[m] | https://paste.debian.net/1209269/ | 10:48 |
| T42 | <TheVancedGamer> Hm | 10:48 |
| T42 | <TheVancedGamer> "Mer Boat Loader" | 10:49 |
| T42 | <TheVancedGamer> Try running ifconfig on host | 10:49 |
| mighty17[m] | i did | 10:49 |
| T42 | <TheVancedGamer> Wats output | 10:50 |
| mighty17[m] | https://paste.debian.net/1209270/ | 10:50 |
| T42 | <TheVancedGamer> ip addr (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>https:/...) | 10:51 |
| mighty17[m] | https://paste.debian.net/1209271/ | 10:51 |
| T42 | <TheVancedGamer> Looks like no IP was assigned (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>https:/...) | 10:52 |
| T42 | <TheVancedGamer> Assign an IP, check hadk and halium | 10:52 |
| mighty17[m] | halium? | 10:53 |
| T42 | <TheVancedGamer> I thought you remembered (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>halium?) | 10:54 |
| mighty17[m] | `SerialNumber: Mer Debug telnet on port 2323 on rndis0 192.168.2.15 - also running udhcpd` it does say this in dmesg now | 10:54 |
| T42 | <TheVancedGamer> wat (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>`Serial...) | 10:54 |
| mighty17[m] | nah like hows halum related to sfos | 10:54 |
| T42 | <TheVancedGamer> wtf happened here | 10:54 |
| T42 | <TheVancedGamer> They have guide on how to assign IP to hybris-boot (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>nah lik...) | 10:54 |
| mighty17[m] | oh like that | 10:54 |
| mighty17[m] | i dont think thats related | 10:55 |
| T42 | <TheVancedGamer> Yeah now it's in telnet | 10:55 |
| T42 | <TheVancedGamer> Weird | 10:55 |
| T42 | <TheVancedGamer> Still porting that tab? | 10:55 |
| mighty17[m] | this is a new port | 10:56 |
| mighty17[m] | with oreo and 3.4 kernel | 10:56 |
| T42 | <edp_17> @elros34 : I have an interesting behaviour of BT on my S2 (i9100). I can turn BT on and off twice but when I turn it off second time, it cannot be turned on again. At this point I need to restart bluetooth-rfkill-event.service, then I can turn BT on again: https://paste.ubuntu.com/p/ysfjVycXbx/ | 10:56 |
| T42 | <edp_17> Do you have idea what the problem can be? | 10:56 |
| mighty17[m] | `telnet: Unable to connect to remote host: No route to host` i ran `telnet 192.168.2.15 2323` | 10:56 |
| T42 | <TheVancedGamer> Try just 23 (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>`telnet...) | 10:57 |
| mighty17[m] | that doesnt work | 10:57 |
| mighty17[m] | ie no output | 10:57 |
| T42 | <TheVancedGamer> Hm | 10:57 |
| T42 | <TheVancedGamer> Run ifconfig again | 10:57 |
| T42 | <TheVancedGamer> And check what IP usb0 has | 10:58 |
| T42 | <edp_17> mighty17: You may need to bring up the network manually: https://docs.halium.org/en/latest/porting/debug-build/early-init.html | 10:58 |
| mighty17[m] | that halium docs worked | 11:01 |
| mighty17[m] | `systemd[1]: Failed to determine whether /sys is a mount point: Bad file descriptor` in dmesg | 11:02 |
| mighty17[m] | https://paste.debian.net/1209273/ | 11:03 |
| T42 | <elros34> @eriki73 boardconfig file sometimes have some tips how to enable wifi otherwise check dmesg for any errors | 11:05 |
| T42 | <elros34> @mighty17[m] it's probably new systemd issue. Your kernel probably lacks some patches for IIRC stat.c and statfs.c . You should be able to find them in logs | 11:07 |
| T42 | <elros34> @edp_17 silly idea do you use gnu-bash? | 11:09 |
| T42 | <edp_17> Yes, I do. | 11:09 |
| mal | piggz: I have no knowledge of evdevtablet | 11:12 |
| T42 | <elros34> @edp_17 check dmesg, for some reason brcm_patchram_plus fail | 11:13 |
| mal | mighty17[m]: which kernel is that? | 11:19 |
| T42 | <elros34> @edp_17 do you have up to date droid-config? | 11:22 |
| T42 | <edp_17> I updated all sub-modules yesterday. | 11:23 |
| T42 | <elros34> I meant do you have your changes on github? | 11:23 |
| T42 | <edp_17> Yes, the only thing that is not there that I moved stuff from /lib/systemd/system to /usr/lib/systemd/system and I added that 99-qtscenegraph.conf into /var/lib/environment/nemo | 11:28 |
| T42 | <edp_17> https://github.com/edp17/droid-config-i9100/tree/master/sparse | 11:28 |
| T42 | <elros34> so I do not see gnu-bash there, show ls -al /usr/bin/bash | 11:29 |
| T42 | <edp_17> -rwxr-xr-x 1 root root 744992 Mar 17 12:52 /usr/bin/bash | 11:29 |
| T42 | <elros34> ok | 11:29 |
| T42 | <elros34> so execute patchram command and see why it fails | 11:30 |
| T42 | <edp_17> Oh, yes I forgot to mention patterns: I have Requires: gnu-bash in one and Requires: gmp-droid in the other .inc file. | 11:31 |
| T42 | <edp_17> About that patchram in bcm4330b1.conf. I needed to replace that to fw_patch and remove the no2bytes because got reported: | 11:33 |
| T42 | <edp_17> check_config(bluetooth_rfkill_event.c:600): Unknown key patchram in main.conf | 11:33 |
| T42 | <edp_17> free_hci(bluetooth_rfkill_event.c:908): No brcm_patchram_plus process to be found | 11:33 |
| T42 | <edp_17> I've checked bluetooth_rfkill_event.c and seen it is using fw_patch, so I defined that in the config. | 11:34 |
| T42 | <edp_17> However in the log it still shows both of these params (patchram and no2bytes) so I don't know where they are from. | 11:35 |
| T42 | <elros34> are you sure you needs bluetooth-rfkill-event-configs-mer if you have own config in sparse? | 11:37 |
| T42 | <edp_17> Nope, I am not. | 11:38 |
| T42 | <edp_17> Shoud I remove? | 11:38 |
| T42 | <elros34> compare what files that package and your droid-config provide and decide | 11:39 |
| T42 | <edp_17> Okay, thanks. Can it be the reason why the service fails? | 11:40 |
| T42 | <elros34> no idea what command you used previously and what errors if eny you have in dmesg, it's easy to just patchram execute command | 11:41 |
| T42 | <edp_17> Do you mean manually run the brcm_patchram_plus command with proper parameter? I can do that. | 11:42 |
| T42 | <elros34> Failed to execute brcm_patchram_plus ... so I would try to execute it manually (strace, dmesg) to see why it fail | 11:44 |
| T42 | <edp_17> When I manually run rcm_patchram_plus, it doesn't crash. It stays running until I turn BT off. | 11:48 |
| T42 | <edp_17> However, it behaves the same way, I need to start/stop BT twice to get the BT on in Settings app. Weird. | 11:49 |
| T42 | <elros34> I see you have same strange udev kill bluetooth rule, remove it | 11:50 |
| T42 | <edp_17> The one that kills brcm_patchram_plus? | 11:51 |
| T42 | <elros34> yeah. Also I hope you masked service when you were testing manually | 11:51 |
| T42 | <edp_17> Didn't need becasue the service was not running as it crashed before. | 11:52 |
| T42 | <edp_17> It still in failed status | 11:52 |
| T42 | <edp_17> I removed that udev rule and now the BT cannot be turned on once it is turned off. Because brcm_patchram_plus remains running. I need that udev rule to kill it once I toggle BT off. | 11:55 |
| T42 | <elros34> fork = 0 > 1? | 11:56 |
| T42 | <edp_17> With or without udev rule? | 11:56 |
| T42 | <elros34> without, I see not reason to keep it | 11:57 |
| T42 | <edp_17> Okay. | 11:57 |
| T42 | <adampigg> Mal: i use evdevtablet on PP as seems to work better | 11:58 |
| mal | @adampigg ok, need to check sources for differences | 11:58 |
| T42 | <elros34> @edp_17 there is already killall command in rfkilevent, do you have psmisc-tools installed? | 11:58 |
| T42 | <edp_17> @elros34 : Yes, Requires: psmisc-tools is in the patterns. | 12:00 |
| T42 | <elros34> also you are missing hfp patch for bluez4 in sparse | 12:03 |
| T42 | <edp_17> @elros34 : That udev rule removed fork = 1 wasn't making any difference. BT cannot BT cannot be turned on once it is off because brcm* stays running. | 12:03 |
| T42 | <elros34> hard to tell without log | 12:04 |
| T42 | <edp_17> hfp patch? | 12:04 |
| T42 | <elros34> https://github.com/elros34/droid-config-moto_msm8960_jbbl/commit/977dc0ca9fc8e03ef4bfa812b5891bf85a932537 | 12:04 |
| T42 | <edp_17> Yep, this is missing. | 12:06 |
| T42 | <edp_17> I am applying it on device. | 12:06 |
| T42 | <eriki73> /vendor/etc/init/hw/init.sdm670.rc: 807: Invalid keyword 'interface_start' | 12:08 |
| T42 | <eriki73> Hmmm... | 12:08 |
| T42 | <eriki73> This might be the reason why connectivity drivers fail to start... | 12:08 |
| T42 | <edp_17> @elros34 : I reverted fork to 0 because it caused to turn BT off a seconds after it is turned on. | 12:15 |
| T42 | <edp_17> I'll try to find out why bluetooth-rfkill-event service is crashing. Fixing that will probably fix the issue. Thanks! | 12:16 |
| T42 | <elros34> it fails because brcm_patchram_plus does not return with status 0 | 12:17 |
| T42 | <elros34> that is why I asked for logs from manual execution: starce and dmesg | 12:19 |
| T42 | <elros34> strace* | 12:19 |
| T42 | <edp_17> Because the udev kills it or without that udev, it remains running. | 12:19 |
| T42 | <elros34> so without that rule there is no "bluetooth_rfkill_event[2709]: attach_hci(bluetooth_rfkill_event.c:926): Failed to execute brcm_patchram_plus"? | 12:20 |
| T42 | <edp_17> strace: https://paste.ubuntu.com/p/TWpvdfhsHB/ | 12:22 |
| T42 | <edp_17> When I toggle the BT it doesn't do anything. | 12:22 |
| T42 | <edp_17> strace again but this time BT turned on: https://paste.ubuntu.com/p/ymgdqPRNfP/ | 12:24 |
| T42 | <edp_17> But when I toggled off, nothing changed in strace | 12:24 |
| T42 | <elros34> so If I get it right whole issue is udev rule, remove it then figure out how to fix rfkill-event | 12:26 |
| T42 | <edp_17> Okay. | 12:39 |
| T42 | <edp_17> Why do you think the issue is the udev rule? From the log I shared? What part in there? | 12:40 |
| T42 | <elros34> maybe I am missing something because I do not know when you enable/disable bluetooth without debug option added to rfkill so I only see failing rfkill | 12:43 |
| T42 | <edp_17> How can I add that debug option? | 12:44 |
| T42 | <elros34> search how it's started and add -d: https://github.com/mer-hybris/bluetooth-rfkill-event/blob/master/bluetooth-rfkill-event/bluetooth_rfkill_event.c#L1133 | 12:45 |
| T42 | <edp_17> I've added -d into /etc/sysconfig/bluetooth-rfkill-event-hciattach like: CONFIGFILE="-d -c /etc/firmware/bcm4330b1.conf" | 12:48 |
| T42 | <edp_17> Should I run brcm* manually with strace or restart bluetooth-rfkill-event service? | 12:49 |
| T42 | <eriki73> @fredldotme , as you ported sargo on Halium, I'd like to ask if you have any tips on the wifi... It seems I cannot start it on SFOS... | 12:50 |
| T42 | <elros34> no strace, just logs form rfkill. | 12:50 |
| T42 | <fredldotme> WIFI was working right out of the box here. Have you checked whether the WIFI firmware has been loaded? (re @eriki73: @fredldotme , as you...) | 12:51 |
| T42 | <eriki73> @fredldotme , I had the same on Manjaro. I am no expert on porting but how do you check if a firmware is loaded... I cannot see anything in dmesg, journal, or logcat | 12:52 |
| T42 | <fredldotme> Nothing from qcacld popping up? | 12:53 |
| T42 | <edp_17> @elros34 : Not much in the log: https://paste.ubuntu.com/p/f8S3dzgG3k/ | 12:53 |
| T42 | <edp_17> The 'system_timeout(bluetooth_rfkill_event.c:492)' repeated until I toggled BT off. Then it stopped, then started again when I toggled BT on again. | 12:54 |
| T42 | <eriki73> @fredldotme , at least in dmesg no... | 12:55 |
| T42 | <eriki73> I did also try writing sta to fwpath... | 12:55 |
| T42 | <fredldotme> Is vendor.wlan_sh started? | 12:57 |
| T42 | <eriki73> @fredldotme , how should I check that? | 13:04 |
| T42 | <eriki73> It says disabled in the vendor files if that helps... | 13:04 |
| T42 | <fredldotme> Should be printed in dmesg. I don't know if it's the case for SfOS but running setprop ctl.start vendor.wlan_sh works on UT | 13:05 |
| T42 | <eriki73> @fredldotme , it was not on but it started now... | 13:06 |
| T42 | <fredldotme> So something must be stuck during Android init | 13:07 |
| T42 | <elros34> @edp_17 have you tired with default fork = true? | 13:09 |
| T42 | <eriki73> @fredldotme , looking at rc files it starts only when wlan.driver.status=ok | 13:09 |
| T42 | <eriki73> But I do not have that prop... | 13:09 |
| T42 | <fredldotme> I'd try to find the proprietary blob that sets this property, strings should be helpful | 13:13 |
| T42 | <elros34> if that prop makes your wifi working then set it in systemd service, some android services are disabled in sfos | 13:15 |
| T42 | <eriki73> @fredldotme , the wlan_sh search for a nonexistant file /sys/kernel/wifi/wlan/driver_version | 13:20 |
| T42 | <eriki73> And sets its contents as a prop... | 13:20 |
| T42 | <eriki73> So, I guess it isn't doing anything useful... | 13:21 |
| T42 | <fredldotme> I have that file on UT (re @eriki73: @fredldotme , the wl...) | 13:21 |
| T42 | <eriki73> @fredldotme , maybe the driver did not get loader? | 13:22 |
| T42 | <fredldotme> Yeah | 13:22 |
| T42 | <eriki73> *loaded | 13:22 |
| T42 | <eriki73> I use almost the same defconfig... | 13:23 |
| T42 | <fredldotme> As UT or Droidian? | 13:23 |
| T42 | <eriki73> As UT and droidian... | 13:23 |
| T42 | <eriki73> The same happened on Manjaro Hybris on Droidian kernel... | 13:24 |
| T42 | <fredldotme> Check your kernel build tree if the object files are actually built | 13:25 |
| T42 | <eriki73> out/target/product/sargo/obj/KERNEL_OBJ/drivers/staging/qcacld-3.0/wlan.o exists, @fredldotme | 13:27 |
| T42 | <eriki73> And I did load the driver (success message in dmesg= | 13:28 |
| T42 | <eriki73> And I did load the driver (success message in dmesg) (edited) | 13:28 |
| T42 | <fredldotme> That's odd then | 13:28 |
| T42 | <elros34> is there any specific reason why you do not want to build it as module like everybody else? | 13:28 |
| T42 | <fredldotme> I would not say "like everybody else", works fine built in on UT and others | 13:29 |
| T42 | <elros34> I meant in sfos ports | 13:29 |
| T42 | <eriki73> @fredldotme , About the firmware, can I check it it exists on my device? | 13:31 |
| T42 | <fredldotme> Sure, should be in /vendor/firmware | 13:34 |
| T42 | <eriki73> @fredldotme , do you happen to know its filename? | 13:35 |
| T42 | <fredldotme> Not specifically, I’d check the kernel driver for request_firmware calls to get the correct name. | 13:36 |
| T42 | <eriki73> @fredldotme , It seems to be looking for qwlan*.bin but I do not have that... | 13:44 |
| T42 | <elros34> do you have something like boot_wlan? | 13:44 |
| T42 | <elros34> in sysfs | 13:44 |
| T42 | <eriki73> Yeah, it works but doesn't make the wlan work... | 13:45 |
| T42 | <eriki73> @elros34 , it loads the driver successfully tho... | 13:45 |
| T42 | <fredldotme> There’s a chance it needs a trigger from userspace to load the firmware, as is the case with the ADSP and similar, as Elros said. | 13:46 |
| T42 | <eriki73> @fredldotme , I tried anything qcom devices could have to trigger the firmware load... | 13:47 |
| T42 | <eriki73> Do you happen to have a sargo near you? | 13:48 |
| T42 | <fredldotme> sure, | 13:48 |
| T42 | <eriki73> What is in your /vendor/firmware/ | 13:48 |
| T42 | <eriki73> I guess I could check if I am missing something? | 13:49 |
| T42 | <fredldotme> A mess, I played around with it during ADSP/voice call volume debugging | 13:49 |
| T42 | <fredldotme> Would have to flash stock firmware back | 13:49 |
| T42 | <elros34> sfos writes only to /data so your /vendor is from lineage | 13:49 |
| T42 | <eriki73> In my case it is from stock rom... @elros34 | 13:50 |
| T42 | <elros34> ok | 13:50 |
| T42 | <fredldotme> (btw, is there anything new regarding voice call volume on 64bit devices? I noticed SfOS has the same ADSP_EFAILED issue as sargo on UT) | 13:51 |
| T42 | <eriki73> Works on droidian + volla + pa, @fredldotme | 13:52 |
| T42 | <eriki73> I mean setting the volume... | 13:52 |
| T42 | <fredldotme> The issue I’m describing is QCOM-specific | 13:52 |
| T42 | <eriki73> Ah, ok! Can't remember if I tested on sargo. | 13:53 |
| T42 | <fredldotme> If PulseAudio is 64bit on SailfishOS, this patch might help: https://github.com/Halium/hybris-patches/commit/7f8bd678daec01489d0b70aeebd88fbd39288436 | 14:03 |
| T42 | <fredldotme> It allows to send data that the 64bit snd control actually can parse, fixing voice volume setting half way. It’s not enough on UT but maybe it works for you guys. | 14:03 |
| mighty17[m] | mal its kernel 3.4 | 14:38 |
| mighty17[m] | Fwiw a 3.0 kernel worked normally | 14:38 |
| mighty17[m] | @elros34 wdym by logs? | 14:38 |
| mal | I had to patch my 3.4 kernel to make it work with sfos 4.x | 14:39 |
| mighty17[m] | Could you share the patch please? | 14:39 |
| mighty17[m] | (or is it in hadk-faq?) | 14:39 |
| T42 | <eriki73> @elros34 , I'll try building wlan as a module and see if that changes anything... | 14:47 |
| mal | mighty17[m]: check the commit about namespaces and vfs and proc on july and after https://github.com/mlehtima/android_kernel_fairphone_fp2/commits/hybris-21.05.0-rel | 14:48 |
| mal | *commits | 14:48 |
| T42 | <elros34> @might17[m] by logs I mean journalctl, dmesg, logcat. Which one depends on situation, in your case I guess you had error in dmesg | 14:48 |
| T42 | <eriki73> @elros34 , building as module had no effect. I can insert the module but it seems that firmware doesn't load... | 14:56 |
| T42 | <elros34> strange... by insert do you mean modprobe? Have you tried now all these sysfs control files? | 14:58 |
| T42 | <eriki73> @elros34 , using insmod... https://paste.ubuntu.com/p/D4JjFmTGjr/ | 14:59 |
| T42 | <eriki73> In sysfs that fails, but nothing else... | 14:59 |
| T42 | <eriki73> as before... | 14:59 |
| T42 | <eriki73> And I did try loading a wrong one first... Ignore that... | 15:00 |
| T42 | <elros34> yeah use modprobe | 15:00 |
| T42 | <fredldotme> You’d have to check the interplay of wlan_hdd_state_ctrl_param_write and what causes the write it expects. I’m pretty sure building as a module won’t solve your problems. | 15:00 |
| T42 | <b100dian> @eriki73 try also adding printk.devkmsg=on to your kernel params - you might miss some kmsgs | 15:06 |
| T42 | <eriki73> @b100dian , I have that already. | 15:07 |
| T42 | <b100dian> ok | 15:07 |
| T42 | <eriki73> @fredldotme , the main action in the function is init_completion(&wlan_start_comp); | 15:07 |
| T42 | <eriki73> Could not figure out what does it do... But the error comes if it times out... | 15:08 |
| T42 | <eriki73> That is directly the write handler of the file I wrote to... | 15:09 |
| T42 | <eriki73> /dev/wlan | 15:09 |
| T42 | <edp_17> Not yet but will try that too. (re @elros34: @edp_17 have you tir...) | 15:15 |
| poetaster | anyone notice that ffmpeg/ffprobe are no longer installed as of 4.1? | 15:15 |
| T42 | <elros34> @eriki73 so only 1 module was built? do you have some kernel/module mismatch? | 15:22 |
| T42 | <eriki73> No, I might actually know now what is the problem... | 15:22 |
| T42 | <eriki73> Some qcom stuff should be in /persist on some devices but it is not mounted... | 15:23 |
| T42 | <eriki73> @elros34 | 15:23 |
| T42 | <eriki73> Or it is a broken symlink... | 15:24 |
| T42 | <edp_17> @elros34 : I've tried fork = 1. Result, bt can be turned on once but few seconds later it runs off and then never turns on. | 15:28 |
| T42 | <elros34> yeah you already said that is the issue but please show rfkil log from it | 15:29 |
| mighty17[m] | @elros34 alright, for now I'll try Mal's commits, thanks | 15:30 |
| mal | those patches of course depend a bit on the exact version of the 3.4 kernel | 15:32 |
| mighty17[m] | 3.4.112 to be exact | 15:34 |
| mighty17[m] | I don't rly know why they didn't do EOL 3.4.113 but /shrug | 15:34 |
| T42 | <edp_17> @elros34 : logs: https://paste.ubuntu.com/p/7x2gJCHQj7/ | 15:39 |
| T42 | <edp_17> The same timeout messages: system_timeout(bluetooth_rfkill_event.c:492) | 15:40 |
| T42 | <elros34> here are rest of the messages? | 15:43 |
| T42 | <elros34> where* | 15:43 |
| T42 | <eriki73> @elros34 , no still no wifi... | 16:11 |
| T42 | <edp_17> @elros34 : I've rebooted and created a fresh log: https://paste.ubuntu.com/p/SQHm32p2Gf/ | 17:22 |
| mighty17[m] | mal: basically no commit from Jul 30 right? | 17:37 |
| mal | mighty17[m]: those namespace/ns related things could be useful for sandboxing | 17:40 |
| mighty17[m] | well gotta get it booting first ;) | 17:41 |
| erikinkinen | I have tried absolutely everything I can think of to get the wifi working... No luck... Any ideas? | 18:03 |
| T42 | <b100dian> what was the sfos version that switched systemd to /usr/lib? both 3.4 and 4.0 have "Fix unit file locations" in diverse packages' changelog.. | 19:13 |
| mal | @b100dian 3.4.0 introduced the new location and compatibility for old location was removed in 4.0.1 which has a script to copy the file to new location | 19:55 |
| T42 | <b100dian> Thanks mal! | 19:57 |
| T42 | <duhjokersinhell> @eriki73 feel free to ping me anytime | 20:05 |
| T42 | <duhjokersinhell> as far as the wlan goes the aosp source normally has it built outside the main kernel tree as a seperate unstripped kernel module that you have to manually put into /vendor/lib/modules. | 20:05 |
| T42 | <duhjokersinhell> Even in UT where the module is built in the kernel tree ive noticed that its buggy because of that reason. tbh while building a kali-nethunter kernel for it i have found it best practice to build it outside the tree like aosp designed to get it work properly | 20:05 |
| T42 | <duhjokersinhell> also i have set the kernel modules like this and have had no problem with the wlan since | 20:26 |
| T42 | <duhjokersinhell> https://pastebin.com/eeNPr4uS | 20:26 |
| T42 | <eriki73> @duhjokersinhell , I do load the wlan as a module now. And kernel loads it successfully but I cannot seem to be able to get an interface... | 20:45 |
| *** izh79_ is now known as izh79 | 21:07 | |
| T42 | <duhjokersinhell> thats odd | 21:11 |
| T42 | <duhjokersinhell> whats the output of ifconfig wlan0 up | 21:13 |
| T42 | <duhjokersinhell> sorry ip link show | 21:19 |
| xx1s1 | Just to check whether I'm going completely mental: when booting hybris-boot.img. Does it boot from /system or from /data/.stowaways/sailfishos? | 23:51 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!