*** birdzhang1 is now known as birdzhang | 07:22 | |
electro575 | hi all | 07:50 |
---|---|---|
electro575 | how can i cancel a pull of a git server on my local git ? | 07:51 |
T42 | <adampigg> git reset? | 07:57 |
T42 | <adampigg> use git log to find the commit to reset back to | 07:58 |
electro575 | yes, git reset --hard for me be appear | 08:02 |
T42 | phlixi was added by: phlixi | 08:03 |
Hacker1245 | Update on the Moto G5 port: After fixing up fixup-mountpoints I can get to rootfs, but it can't assign an IP and reboots after a minute | 11:08 |
Hacker1245 | I needed to set audit=0 otherwise it would crash on boot | 11:08 |
Hacker1245 | selinux is set to 0 on the command line | 11:10 |
Hacker1245 | @elros34 Are you here? | 11:18 |
T42 | <elros34> Did you changed defconfig like I told you? Check generated .config somwhere in $ANDROID_ROOT/out or in device: zcat /proc/config.gz | 11:20 |
Hacker1245 | Are you talking about enabling audit/ | 11:21 |
Hacker1245 | That is enabled in the defconfig | 11:22 |
Hacker1245 | selinux_bootparam is too | 11:22 |
Hacker1245 | Checked the generated config and it should be correct | 11:25 |
Hacker1245 | @elros34 | 11:26 |
T42 | <elros34> so remove audit=0 create init_enter_debug2 if you can't acces telnet and grab fresh journalctl | 11:26 |
Hacker1245 | create init_enter_debug2 where? | 11:27 |
T42 | <elros34> I told you how 2 days ago so check channel logs if you don't remember | 11:28 |
Hacker1245 | Just checked | 11:29 |
Hacker1245 | Run that command from TWRP? | 11:29 |
T42 | <elros34> if you dont have telnet acces then you need to use TWRP | 11:30 |
Hacker1245 | on it | 11:30 |
Hacker1245 | device still crashes | 11:31 |
Hacker1245 | Can't get to telnet | 11:31 |
Hacker1245 | No last_kmsg | 11:32 |
Hacker1245 | It doesn't bring the interface up | 11:33 |
Hacker1245 | Nothing on the dmesg of the host | 11:33 |
Hacker1245 | @elros34 | 11:35 |
T42 | <elros34> So you're telling me that after adding only audit=0 you have telnet access and without it it reboots afet 1 minute? Are you sure /init_enter_debug2 exists | 11:37 |
Hacker1245 | I don't have telnet access at all | 11:37 |
Hacker1245 | With selinux=0 and audit=0 it shows the interface, can't assign the IP and reboots after a minute | 11:38 |
Hacker1245 | With selinux=0 only it reboots pretty much instantly, doesn't show Mer debug info | 11:38 |
Hacker1245 | in dmesg | 11:39 |
T42 | <elros34> which port 23 or 2323 in dmesg | 11:39 |
Hacker1245 | Doesn't show the port in dmesg | 11:39 |
Hacker1245 | Although with hybris-recovery it does that | 11:39 |
Hacker1245 | https://bpaste.net/show/T3KV | 11:40 |
T42 | <elros34> You had telnet 2 days ago so what did you changed? Reboot after 1 minute can be caused by wrong fixup | 11:41 |
Hacker1245 | Fixup might be the issue | 11:42 |
Hacker1245 | Let me ask my friend if he got telnet to rootfs or initramfs | 11:42 |
T42 | <elros34> but your last paste suggest you are in rootfs: Product: init-debug in real rootf so you should be able to at least grab /init.log | 11:45 |
Hacker1245 | I can grab the init.log, wait a moment | 11:51 |
Hacker1245 | There | 11:53 |
Hacker1245 | https://paste.ubuntu.com/p/pbwGNRg9sf/ | 11:53 |
mal | usb setup fails | 11:55 |
Hacker1245 | yes I know | 11:58 |
Hacker1245 | why though | 11:58 |
T42 | <elros34> do you have net-tools installed in rootfs? check in Jolla-3.1.0.12-$DEVICE-armv7hl | 11:58 |
Hacker1245 | Where is that? | 11:59 |
Hacker1245 | In the HADK directory or on the device? | 11:59 |
T42 | <elros34> same place where your build sfos image is | 12:00 |
Hacker1245 | Jolla-3.1.0.11-cedric-armv7hl.packages? | 12:02 |
Hacker1245 | net-tools isn't there | 12:03 |
Mister_Magister | i thought 3.1.0.11 isn't available for ports | 12:03 |
Mister_Magister | only 3.1.0.12 | 12:03 |
T42 | <elros34> yes only 3.1.0.12 | 12:03 |
Mister_Magister | he built wrong version then | 12:04 |
T42 | <elros34> cedric? I thought you are building for griffin | 12:04 |
Hacker1245 | Nope | 12:04 |
Hacker1245 | I'm porting the Motorola Moto G5 | 12:05 |
Hacker1245 | Not the Moto Z | 12:05 |
T42 | <elros34> ok I mixed something | 12:06 |
* Mister_Magister *cough* s/motorola/lenovo/ *cough* | 12:06 | |
Mister_Magister | somebody porting motoz? | 12:06 |
Hacker1245 | not me | 12:06 |
Mister_Magister | @elros34 [13:35] <masha11> Hello. I do not have GUI. I checked mounting and firmware, updated config submodule. My device is griffin, cm-14.1. logcat: https://pastebin.com/XV8ay1QK . Any ideas? | 12:07 |
T42 | <elros34> :) yes | 12:07 |
Mister_Magister | duuuuuuuuuuck | 12:08 |
Mister_Magister | i wanted to port moto z ;-; | 12:08 |
Mister_Magister | guy didn't sent me device yet | 12:08 |
T42 | <elros34> test_hwcomposer reboots device so far | 12:08 |
Hacker1245 | @elros34 So I need to set up building for the 3.1.0.12 version? | 12:08 |
Mister_Magister | Hacker1245: yes | 12:09 |
Mister_Magister | basically change RELEASE=3.1.0.12 | 12:09 |
Mister_Magister | and run mic again | 12:09 |
T42 | <elros34> and check if you have net-tools in generated image | 12:09 |
Mister_Magister | he 99.9% will | 12:11 |
masha11 | @Mister_Magister: It would be great if you join! | 12:11 |
Mister_Magister | masha11: but then i would do it like red army xd | 12:11 |
Mister_Magister | covered whole port and make it myself :P | 12:11 |
* Mister_Magister i took over one port already lol | 12:12 | |
T42 | <elros34> @masha11: btw have you tried minimer, surfaceflinger like in faq-hadk? | 12:12 |
T42 | <mohagician> When shall you port to pro1? | 12:13 |
Mister_Magister | masha11: if surfaceflinger works then you could render on it | 12:13 |
Mister_Magister | @mohagician we are doing it constantly what do you mean | 12:13 |
masha11 | I tried minimer | 12:14 |
Hacker1245 | rebuilding rootfs now | 12:14 |
T42 | <mohagician> @mister_magister When will it be available for end users? | 12:14 |
Mister_Magister | at least after device is available for end users | 12:15 |
Mister_Magister | we have got it mostly working just some annoying bugs so after device release, releasing some beta wouldn't be a problem i think | 12:15 |
masha11 | I will check surfaceflinger | 12:20 |
mal | Mister_Magister: I feel like that device could have broken hwcomposer, some kernels have broken 32-bit support | 12:28 |
Hacker1245 | @elros34 Checked Jolla-3.1.0.11-cedric-armv7hl.packages, net-tools isn't there | 12:28 |
T42 | <elros34> So you changed something in patterns | 12:29 |
Hacker1245 | *3.1.0.12 | 12:29 |
Mister_Magister | mal: just like my huawei | 12:29 |
Hacker1245 | How can I check what's broken? | 12:30 |
T42 | <elros34> show your droid-config github repo | 12:30 |
mal | masha11: did you try test_hwcomposer after once running surfaceflinger? | 12:31 |
Mister_Magister | mal: and what is happening with test_hwcompsoer | 12:33 |
Hacker1245 | @elros34 gonna ask my friend if he has it on his GitHub account | 12:35 |
T42 | <elros34> net-tools is pulled by sailfish-porter-tools which should be in jolla-configuration-$DEVICE.yaml | 12:38 |
Hacker1245 | let me look for that | 12:38 |
Hacker1245 | sailfish-porter-tools is there | 12:39 |
T42 | <elros34> so you mess something, maybe %packages section in ks file | 12:40 |
Hacker1245 | I had jolla-hw-adaptation-cedric in there | 12:46 |
Hacker1245 | But if I change to jolla-configuration-cedric this happens Error <creator>[09/17 12:44:15] : Unable to find pattern: jolla-confguration-cedric | 12:47 |
Hacker1245 | the pattern is there | 12:47 |
T42 | <elros34> default is @Jolla Configuration "$DEVICE" | 12:48 |
Hacker1245 | well | 12:53 |
Hacker1245 | Error <creator>[09/17 12:53:00] : Unable to find pattern: Jolla Configuration "$DEVICE" | 12:53 |
Mister_Magister | ffs man… | 12:53 |
Mister_Magister | @Jolla Configuration cedric | 12:54 |
Hacker1245 | cedric in quotes or not? | 12:54 |
Mister_Magister | nope | 12:54 |
Hacker1245 | ok it's building | 12:59 |
masha11 | mal: yes, the device also reboots after test_hwcomposer. | 13:06 |
Mister_Magister | masha11: and what happens when you run surfaceflineger | 13:13 |
masha11 | Device restarts after runnig minimer. From logcat: GetAndroidProperty: property_get debug.egl.swapinterval failed (err = 0) | 13:16 |
masha11 | From logcat: "SurfaceFlinger: SurfaceFlinger's main thread ready to run. Initializing graphics H/W.." | 13:17 |
masha11 | https://pastebin.com/zA1X7KgK | 13:17 |
masha11 | "SurfaceFlinger service already running, so we won't start it here. If you have trouble with media, try disabling the minisf service." | 13:17 |
masha11 | do i need to do "/system/bin/stop surfaceflinger"? | 13:19 |
Mister_Magister | just /system/bin/surfaceflinger | 13:19 |
masha11 | @Mister_Magister: logcat after running surfaceflinger: https://pastebin.com/SBKbiden | 13:24 |
Mister_Magister | do you see anything on screen? | 13:25 |
Mister_Magister | prably not | 13:25 |
masha11 | @Mister_Magister: no | 13:26 |
Mister_Magister | would love to help but urgent task at work. strace it :P | 13:27 |
masha11 | ok | 13:28 |
masha11 | @Mister_Magister: thanks, may be later) | 13:29 |
masha11 | strace minimer : https://pastebin.com/qcudndUv, strace test_hwcomposer: https://pastebin.com/NtW7HM1C | 13:37 |
Mister_Magister | mal: look at that strace test_hwc. what the heck is going on | 13:43 |
Mister_Magister | why it tries to open /srv/mer/toolings/motorola-griffin-armv7hl/opt/cross/armv7hl-meego-linux-gnueabi/lib/tls/v8l/neon/vfp/libhybris-common.so.1 | 13:44 |
T42 | <elros34> afaik that is "normal" | 13:46 |
Mister_Magister | okay "normal" | 13:46 |
Mister_Magister | so test_hwc causes reboot but surfaceflinger doesn't | 13:47 |
Mister_Magister | masha11: how do you run test_hwcomposer | 13:47 |
masha11 | Mister_Magister: [nemo@Sailfish ~]$ EGL_PLATFORM=hwcomposer test_hwcomposer | 13:48 |
Mister_Magister | do you have /sys/fs/pstore? | 13:49 |
masha11 | annotate-ramoops-0console-ramoopsdmesg-ramoops-0 | 13:51 |
Mister_Magister | oh give console-ramoops | 13:51 |
Mister_Magister | feelsgoodman | 13:51 |
masha11 | https://pastebin.com/8hmNS0TT | 13:53 |
Hacker1245 | @elros34 Update: Now I have consistent telnet to the rootfs after touch /data/.stowaways/sailfishos/init_enter_debug2 | 14:27 |
Hacker1245 | Device does not reboot | 14:27 |
Hacker1245 | If I run normally it tries to go through but gets stuck | 14:27 |
Hacker1245 | lipstick seems to segfault | 14:27 |
Hacker1245 | journalctl: https://paste.ubuntu.com/p/hfhs7hxqPK/ | 14:28 |
Hacker1245 | Dmesg: https://paste.ubuntu.com/p/RZtGN2GPTn/ | 14:28 |
Hacker1245 | Logcat still has the linker error | 14:28 |
T42 | <elros34> Now you need to focus at fixing droid-hal-init. What android base you use? | 14:31 |
T42 | <elros34> and provide journalctl only from current boot "-b" switch otherwise is confusing | 14:33 |
Hacker1245 | I did try to use journalctl -b but it returns no entries | 14:45 |
T42 | <elros34> @DylanVanAssche have you tried https://git.merproject.org/mer-core/sensorfw/merge_requests/40 ? Can you confirm whathever sensors works even without bufferd mode enabled (09972dfa)? | 15:29 |
Hacker1245 | @elros34 I use LineageOS 14.1 as the base | 15:32 |
mal | @elros34 is that channels variable unused or what, you removed some assigning to that? | 15:32 |
mal | ah, you moved it to other place | 15:32 |
T42 | <elros34> mal: it's initialized in findSensor(), some simple drivvers doesn't have scan_elements directory in sysfs so number of channels were set to 0 | 15:34 |
mal | @elros34 ok, probably fine then | 15:36 |
T42 | <DylanVanAssche> @elros34 I use your commits already on top of my sensorfw with additional and improved debug log. I have all sensors working except for proximity and magnetometer. The proximity one isn't support by iioadaptor (I have a WIP that adds support for that, but break light sensor and doesn't show proximity in the UI). Magnetometer ha | 15:45 |
T42 | s wrong values atm. | 15:45 |
T42 | <elros34> ok thanks | 15:46 |
mal | @DylanVanAssche did you remove manual scaling from config when you use the new code? | 15:47 |
T42 | <DylanVanAssche> mal: Yes, no scaling in the configs (doesn't even work if you try it, but it should if you check the src) | 15:48 |
mal | @DylanVanAssche you mean the scaling is incorrect even with some config? | 15:49 |
T42 | <elros34> @DylanVanAssche it works but incorrectly. You need to use it in [sensor name] section instead [sensor type] like for any other options | 15:50 |
mal | hmm | 15:50 |
T42 | <DylanVanAssche> mal: the data from the magnetometer has some weird values, not related to a scale. Seems to be something wrong, needs further debugging. My first priority was proximity sensor | 15:51 |
T42 | <DylanVanAssche> @elros34 Oh that's what I saw in the latte config that uses scaling | 15:51 |
mal | @DylanVanAssch quite odd that proximity sensor would break light sensor | 15:51 |
T42 | <DylanVanAssche> mal: Proximity and light sensor are the same one on the Pinephone | 15:52 |
mal | @DylanVanAssche those are almost always the same | 15:52 |
T42 | <DylanVanAssche> mal: hmmmm I'm not sure what's wrong with this code though :P | 15:54 |
T42 | <DylanVanAssche> https://github.com/sailfish-on-dontbeevil/sensorfw/commit/473f983e4b1da96966b29abf52714b858e2d91cb | 15:54 |
T42 | <DylanVanAssche> `Within proximity` debug log changes (false/true) with the right threshold and covering the sensor. But the data in Messwerk is broken for light and proximity sensor. | 15:56 |
mal | @DylanVanAssche broken how, nothing is shown? | 15:57 |
T42 | <DylanVanAssche> mal: Nothing to show. If you cover the sensor or shine a light on it, nothing happens. Without this proximity sensor commit, I saw the reading of the light sensor plotted in Messwerk | 15:58 |
T42 | <elros34> check in sensorfw whether correct paths are added for proximity, als and they have correct number of channels | 15:59 |
mal | @DylanVanAssche can you show some debug output | 15:59 |
T42 | <DylanVanAssche> They have, all 1 channel (there's only 1 channel for each sensor). | 15:59 |
T42 | <DylanVanAssche> mal: Let me startup the devkit :D | 16:00 |
mal | @DylanVanAssche you probably should also print threshold in debug to see that it's correct | 16:02 |
T42 | <DylanVanAssche> mal: It's in there :) | 16:02 |
T42 | <DylanVanAssche> mal: http://ix.io/1Vzd debug logs | 16:03 |
T42 | <DylanVanAssche> `Aug 29 15:05:30 Sailfish sensorfwd[2489]: Value for key "proximity/threshold" : "2000"` | 16:03 |
mal | @DylanVanAssche something fishy in the iio adaptor https://github.com/sailfish-on-dontbeevil/sensorfw/blob/dontbeevil/adaptors/iioadaptor/iioadaptor.cpp#L196 | 16:08 |
mal | it always uses same range, I feel that shouldn't be the case | 16:09 |
T42 | <DylanVanAssche> mal: Yeah I didn't touch that part :P I found it a bit weird too | 16:09 |
mal | @DylanVanAssche https://github.com/sailfish-on-dontbeevil/sensorfw/blob/dontbeevil/core/hybrisadaptor.cpp#L1111 | 16:11 |
mal | you probably should fix that, it's possible Qt side is confused if sensor ranges are wrong | 16:11 |
T42 | <DylanVanAssche> That's possible | 16:12 |
mal | it's known that the iio adaptor was a buggy initial implementation | 16:12 |
T42 | <DylanVanAssche> mal: There's no information about the range for IIO adaptors in the dev nodes. I'm not sure if we can make this generic | 16:18 |
mal | hmm | 16:18 |
T42 | <elros34> I used something like that in hybrisproximityadaptor: d->value_ = withinProximity_ ? 0 : maxRange(); because my driver reports 30 or 1000 values so qt side never report it | 16:19 |
mal | @DylanVanAssche also the value in debug log looks odd, what offset does the the sensor show? | 16:19 |
T42 | <DylanVanAssche> Oh will try that @elros34 | 16:20 |
T42 | <DylanVanAssche> mal: no offset for this IIO device | 16:20 |
T42 | <elros34> so try something like 0 or 1000 for proximity | 16:21 |
mal | @DylanVanAssche where is the offset initialized, is there some memset to set iiodevice values to 0? | 16:24 |
T42 | <DylanVanAssche> @elros34 will do | 16:25 |
T42 | <DylanVanAssche> mal: https://github.com/sailfish-on-dontbeevil/sensorfw/blob/dontbeevil/adaptors/iioadaptor/iioadaptor.cpp#L252-L255 no memset indeed | 16:25 |
mal | @DylanVanAssche I was just wondering that have you verified the value is indeed 0 if there is no offset for the type? | 16:26 |
T42 | <DylanVanAssche> mal: will put a debug output just to be sure | 16:27 |
T42 | <DylanVanAssche> Needs to rebuild it first | 16:27 |
mal | @DylanVanAssche can you show some debug log where you annotate when you have cover proximity sensor and when not, to verify the values in both cases | 16:58 |
T42 | <DylanVanAssche> mal: Will do, building sensorfw atm | 16:58 |
T42 | <elros34> mal: they are here http://ix.io/1Vzd 499 : 505 | 16:59 |
mal | @elros34 I mean which value is covered and which not | 17:00 |
T42 | <elros34> ok | 17:00 |
T42 | <DylanVanAssche> that one is with the dev kit laying down, so as good as covered | 17:00 |
mal | @DylanVanAssche would be nice to see the not covered for comparison | 17:00 |
T42 | <DylanVanAssche> Indeed | 17:01 |
mal | @DylanVanAssche https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/ABI/testing/sysfs-bus-iio#n1287 "Higher proximity measurements indicate closer objects, and vice versa" | 17:03 |
T42 | <DylanVanAssche> mal: True | 17:03 |
mal | that is somewhat confusing "Units after application of scale and offset are meters" | 17:04 |
T42 | <DylanVanAssche> I think that's a mistake since they say that conversion to SI units is impossible | 17:04 |
mal | yes, very confusing | 17:05 |
T42 | <DylanVanAssche> mal: hmmmm proximity works now :) it was the range definitely. Only the light sensor returns -1.000 Lux (changes from 0.000 to -1.000 when shining a light on the sensor) | 17:21 |
T42 | <elros34> proximity needs 0 or >0 : https://git.merproject.org/mer-core/qtsensors/blob/mer/5.2.1+git17/src/plugins/sensors/sensorfw/sensorfwproximitysensor.cpp#L78 | 17:24 |
T42 | <DylanVanAssche> @elros34 true or false is also fine I guess :P | 17:30 |
T42 | <DylanVanAssche> The light sensor is the illuminance_raw thing from the same sensor | 17:30 |
T42 | <DylanVanAssche> Hmmm it works, but I think some of the devkit cables are influencing the readings, sometimes they are correct, sometimes they are way off (checked the iio interface, not true sensorfw). Need to finetune the threshold when I have an assembled PinePhone. Still no idea why illuminance returns -1 in Messwerk | 17:41 |
mal | @DylanVanAssche have you tried adding some debug logging to als sensor? | 17:57 |
mal | @adampigg about your bluebinder PR, add the bug reference to commit message also | 17:59 |
T42 | <DylanVanAssche> mal: Currently doing that :) | 18:02 |
T42 | <DylanVanAssche> @elros34 I found the scaling bug, I fixed it in my branch. Will make a PR soon | 18:03 |
T42 | <elros34> it will not compile :) | 18:05 |
T42 | <DylanVanAssche> I know ;P I'm on it :P | 18:06 |
piggz | mal: done | 18:11 |
T42 | <DylanVanAssche> @elros34 This commit compiles :) I forgot to make `name` available outside the IF test :) | 18:26 |
T42 | <DylanVanAssche> https://github.com/sailfish-on-dontbeevil/sensorfw/commit/dc0a1f93b7736a1c3591c83e8126e1525cf83049 | 18:26 |
T42 | <DylanVanAssche> Building it now... Together with the light sensor debug | 18:26 |
T42 | <DylanVanAssche> mal: This line is never printed in the debug log... https://github.com/sailfish-on-dontbeevil/sensorfw/blob/dontbeevil/adaptors/iioadaptor/iioadaptor.cpp#L477 Seems that the ALS sensor is never read anymore | 18:41 |
T42 | <elros34> what about error in your log: sensorfwd[2489]: read(): Is a directory. THis could be related | 18:43 |
T42 | <DylanVanAssche> Seems to be this `Aug 29 17:42:14 Sailfish sensorfwd[2510]: No sysfs path defined for: "als"` | 18:43 |
T42 | <elros34> probably not an issue | 18:45 |
mal | @DylanVanAssche could it be that those als and prox come in same event and it doesn't handle both cases the same time? | 18:45 |
mal | or something | 18:45 |
T42 | <DylanVanAssche> mal: They are handled separate. It sees a light sensor and a proximity sensor. | 18:46 |
mal | @DylanVanAssche didn't you say als used to work | 18:46 |
mal | ok | 18:46 |
T42 | <DylanVanAssche> mal: Yes, without the prox code | 18:47 |
T42 | <DylanVanAssche> @elros34 read() seems to be from `Aug 29 17:42:15 Sailfish sensorfwd[2510]: bool SysfsAdaptor::addPath(const QString&, int) "/sys/bus/iio/devices/iio:device2"` which is the MPU6050, but it gets detected properly (in_accel_x_raw, ...) | 18:48 |
T42 | <elros34> try to read manually from als | 18:49 |
T42 | <DylanVanAssche> I did that from the terminal, shining a light and the value changes | 18:50 |
T42 | <DylanVanAssche> Oh now it starts printing to the debug log :P | 18:51 |
T42 | <DylanVanAssche> `Aug 29 17:56:27 Sailfish sensorfwd[2510]: Light intensity: 4294967295` Hmmm overflow I guess, that's maybe why the value changes to `-1.000` to indicate an error? | 18:53 |
T42 | <DylanVanAssche> in Messwerk Qt sensors | 18:53 |
T42 | <elros34> looks like you added path config to [accelerometer] thats why it prints EISDIR | 18:55 |
T42 | <DylanVanAssche> @elros34 In my config there's no `path`: https://bin.disroot.org/?784031b9527ae285#dngTCK7S7LGXx0I4wN8t8IhzyzQlB/wC14Wnfp5Y6PE= | 19:00 |
T42 | <elros34> hmm it must be defined somewhere but that is minor issue: sensorfwd[2489]: Value for key "accelerometer/path" : "/sys/bus/iio/devices/iio:device2" | 19:02 |
T42 | <DylanVanAssche> Hmmmm doesn't hurt it atm, but weird that it's there | 19:05 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!