freemangordon | tmlind: uvos: please help, obviously I am doing something stupid, but I cannot figure out what. probe() never gets called https://pastebin.com/Wymn8tpa | 09:54 |
---|---|---|
freemangordon | dts https://pastebin.com/7tVyA2cB | 09:55 |
freemangordon | this is child of cpcap: pmic@0 | 09:56 |
freemangordon | I also added extcon = <&cpcap_extcon>; to &usb2_phy | 09:58 |
freemangordon | so now it seems usb2_phy does not get inited as for some reason cpcap_extcon is not loaded | 10:00 |
freemangordon | if I modprobe by hand, nothing happens | 10:00 |
freemangordon | being struggling with that for the last 4 hours :( | 10:00 |
freemangordon | this is modinfo https://pastebin.com/qhVBqLiW | 10:04 |
freemangordon | oh | 10:11 |
freemangordon | compat has to be defind in mfd driver as well | 10:12 |
Wizzup | uvos: I got one xt1602 | 11:22 |
Wizzup | well, ordered | 11:22 |
Wizzup | freemangordon: do you remember if the rtcom ui for sip calls if open? | 12:24 |
Wizzup | by that I mean the UI that allows you to configure accounts | 12:24 |
Wizzup | I guess rtcom-accounts-voip-support is not, if that is it | 12:26 |
Wizzup | ah, it's rtcom-accounts-plugin-sip | 12:26 |
Wizzup | it's not big at least | 12:27 |
Wizzup | sicelo: did you say you managed to do routing with modrana? | 14:40 |
Wizzup | it looks at least like the google routing doesn't work | 14:40 |
Wizzup | maybe I should try offline next | 14:40 |
Wizzup | hm, it looks like I already imported the latest release that still had gtk | 14:43 |
Wizzup | looks like the googlemaps that modrana ships is really old | 14:44 |
Wizzup | this might be new enough https://pypi.org/project/googlemaps/3.1.4/ | 14:45 |
Wizzup | oh, wait, I guess it needs an api key | 15:22 |
freemangordon | YAY! charger detection works | 16:51 |
freemangordon | now I only have ti find how to adjust maximum current | 16:52 |
Wizzup | nice! | 17:01 |
sicelo | Wizzup: modrana ... no, wasn't me | 19:27 |
sicelo | freemangordon awesome! | 19:28 |
sicelo | so guys/girls, i was wondering - with appended dtb, how does the system know where kernel ends and dtb starts? | 19:29 |
Wizzup | sicelo: ok, I'm just looking for -anything- that I can use to do routing | 19:29 |
Wizzup | sicelo: I don't think it knows, so you need add kernel size and dtb size | 19:29 |
Wizzup | sicelo: what kind of kernel compression are you using? | 19:30 |
sicelo | there's nowhere we specify kernel or dtb sizes | 19:30 |
sicelo | They're just concatenated in a plain way | 19:31 |
sicelo | Wizzup: i believe navit can route, via monav or some such | 19:31 |
Wizzup | sicelo: hm | 19:34 |
Wizzup | sicelo: it'd be neat to have some guide on our wiki imho | 19:35 |
Wizzup | I think modrana can also use monav | 19:35 |
Wizzup | but it's not really clear how | 19:35 |
Wizzup | apt-cache search monav yields no results | 19:35 |
sicelo | maybe ask uvos ... i believe he has played with navit on leste | 19:37 |
Wizzup | looks like it was only in jessie | 19:38 |
Wizzup | https://tracker.debian.org/pkg/monav | 19:38 |
Wizzup | I'll try this https://navit.readthedocs.io/en/v0.5.6/ | 19:42 |
sicelo | Wizzup: this is how we concatenate kernel + dtb, https://github.com/maemo-leste/maemo-kernel-config/blob/master/n900-uImage#L6 | 19:44 |
Wizzup | sicelo: yes, but what kernel compression is in .config for your kernel | 19:46 |
sicelo | CONFIG_KERNEL_GZIP=y | 19:51 |
Wizzup | does that mean it uses gzip, or is gzip supported? | 19:54 |
sicelo | i think it means kernel is compressed with gzip :-) | 19:55 |
Wizzup | can you try xz? | 19:59 |
Wizzup | the navit internal ui is .. wow | 20:45 |
Wizzup | uvos: do you use navit currently? | 20:48 |
uvos | yeah | 20:51 |
uvos | freemangordon: great you got it woking | 20:52 |
Wizzup | uvos: the one from beowulf/buster? | 20:52 |
uvos | yes | 20:52 |
Wizzup | care to share your xml config? | 20:52 |
uvos | works compleatly fine | 20:52 |
uvos | except for the idea of an ui they have | 20:53 |
Wizzup | I see nothing, even though I downloaded a map, and then I have to touch the screen just to see the menu | 20:53 |
Wizzup | but I can't get it to show my map even though I told it to center on it | 20:53 |
uvos | thats correct | 20:53 |
uvos | at touch screen | 20:53 |
uvos | the problem is | 20:53 |
uvos | that navit starts in serbia or something by default | 20:53 |
uvos | so you have really no way of finding where you have the map | 20:53 |
Wizzup | so I changed the center | 20:53 |
Wizzup | <navit center="..." | 20:54 |
Wizzup | oh there is a center.txt | 20:54 |
uvos | that only works for the first boot | 20:54 |
uvos | er start | 20:54 |
uvos | right after that its in center.txt | 20:54 |
Wizzup | which is of course hex encoded | 20:55 |
Wizzup | lol | 20:55 |
uvos | you can probubly just delete it | 20:56 |
Wizzup | tried that to no avail, still just see a brown-yellow ish screen | 20:56 |
uvos | http://uvos.xyz/maserati/navit.xml | 20:57 |
uvos | Wizzup: you can also just wait for gps | 20:58 |
uvos | it will jump to location then | 20:58 |
Wizzup | I did wait for gps, and it didn't work still | 20:58 |
uvos | hmm | 20:58 |
Wizzup | maybe the map I downloaded using the map cropbox thing is not ok | 20:58 |
Wizzup | ( http://maps3.navit-project.org/ ) | 20:58 |
uvos | thats where i got my map from too | 20:58 |
uvos | maybe typo? | 20:58 |
Wizzup | oh.. | 20:59 |
Wizzup | hmm | 20:59 |
Wizzup | well I forgot to have enabled="yes" | 20:59 |
Wizzup | but that didn't make the difference still | 21:00 |
uvos | aha | 21:00 |
uvos | hmm | 21:00 |
Wizzup | uvos: ok | 21:03 |
Wizzup | uvos: I had the center coords swapped | 21:04 |
Wizzup | it still looks a bit funky, but I see something | 21:04 |
Wizzup | error:gui_internal:gui_internal_set_attr:Unknown attribute: layer | 21:09 |
Wizzup | maybe this is related | 21:09 |
uvos | i do not get that | 21:11 |
freemangordon | uvos: please test the charger patch to confirm it really works and limits the current to the selected config | 21:27 |
freemangordon | I tried all the devices around and seems they all support 500 mA | 21:32 |
uvos | what dose the patch apply on | 21:38 |
uvos | it dosent on mameo-5.18.y | 21:39 |
freemangordon | it should | 21:41 |
freemangordon | on mameo-5.18.y-cpcap | 21:41 |
freemangordon | maemo-5.18.y-cpcap | 21:41 |
freemangordon | make sure to pull | 21:42 |
uvos | id dosent | 21:43 |
uvos | none of the segments do | 21:43 |
uvos | maybe pastebin broke the patch? | 21:43 |
freemangordon | maybe pastebin changed tabs to spaces | 21:43 |
uvos | yes thats exatcly what happend | 21:43 |
freemangordon | I think you can get it raw from pastebin | 21:43 |
freemangordon | uvos: click on 'downloda' button | 21:45 |
freemangordon | *download | 21:45 |
freemangordon | https://pastebin.com/dl/WfVwFwYL | 21:45 |
freemangordon | hmm, link does not work | 21:46 |
freemangordon | still 'download' button it is | 21:46 |
uvos | could you reupload the modem patch too, i saved that with spaces | 21:46 |
uvos | and now it allso dosent apply | 21:46 |
uvos | and its gohne from pastebin | 21:46 |
freemangordon | modem patch? for poweroff? | 21:46 |
uvos | yeah | 21:46 |
freemangordon | it is not needed | 21:46 |
uvos | i thought i should try if it makes a diff on xt875 | 21:47 |
freemangordon | ah | 21:47 |
freemangordon | well, I don;t keep it either | 21:47 |
uvos | uff | 21:47 |
uvos | ok ill fix it | 21:47 |
uvos | but not rn | 21:47 |
freemangordon | no worry | 21:47 |
freemangordon | I will redo it | 21:47 |
uvos | no no | 21:47 |
freemangordon | not now though | 21:47 |
uvos | i can just fix the patch | 21:47 |
freemangordon | sure | 21:47 |
uvos | its just sed | 21:47 |
uvos | 4 spaces = tab | 21:47 |
freemangordon | it should be 5 lines or something | 21:47 |
uvos | should to the trick | 21:47 |
freemangordon | hmm, 8 spaces | 21:48 |
uvos | or 8 | 21:48 |
uvos | sure | 21:48 |
uvos | you get the point | 21:48 |
freemangordon | yeah | 21:48 |
freemangordon | I am more concerned about charger patch :) | 21:49 |
uvos | compileing | 21:49 |
freemangordon | cool | 21:49 |
uvos | really tho | 21:49 |
uvos | we need a usb current mesurement thingy | 21:50 |
uvos | ideally | 21:50 |
Wizzup | I have that somewhere | 21:50 |
sicelo | Wizzup: lzma compression -> kernel 4.4MB. same problem | 21:50 |
Wizzup | if you mean the hw verisons | 21:50 |
Wizzup | sicelo: ok, and that's significantly smaller, right? | 21:51 |
uvos | Wizzup: you also need a port that advertises less than 500mA | 21:51 |
sicelo | yes. | 21:51 |
Wizzup | that makes it less likely the dtb gets cut off for sure | 21:51 |
Wizzup | do others also see this? | 21:51 |
uvos | Wizzup: well you have those (any mapphone) | 21:51 |
sicelo | Wizzup: other N900 users? most probably. you can try to build any kernel 5.19 onwards | 21:52 |
freemangordon | uvos: with android or with linux? | 21:53 |
uvos | well idk if manline dose it correctly, android shurely dose | 21:53 |
uvos | anyhow vbus on mapphones from cpcap can supply only 30 or was it 40 mA | 21:53 |
uvos | in datasheet | 21:53 |
freemangordon | no, I mean - can I use tablet with android to test? | 21:54 |
uvos | idk depends on the hw | 21:54 |
uvos | dont you have a mapphone tablet? | 21:54 |
Wizzup | sicelo: ok, did you manage to bisect it, or? | 21:55 |
uvos | mz6xx | 21:55 |
freemangordon | mz 608? | 21:55 |
uvos | mz609 | 21:55 |
freemangordon | or iz it 808? | 21:55 |
uvos | 609 | 21:55 |
uvos | anyhow that one should do | 21:55 |
freemangordon | yeah, 609 | 21:55 |
freemangordon | I need otg cable, right? | 21:55 |
uvos | i presume it uses the same setup as the mapphone phones for otg | 21:55 |
uvos | freemangordon: yes | 21:55 |
freemangordon | ok | 21:55 |
Wizzup | uvos: btw, I think that I have seen mce hang in writing to gconf | 21:56 |
Wizzup | had no gdb avail, but: | 21:56 |
Wizzup | [pid 2780] poll([{fd=18, events=POLLIN}, {fd=19, events=POLLIN}], 2, -1 <unfinished ...> | 21:56 |
Wizzup | [pid 2652] poll([{fd=6, events=POLLIN}, {fd=7, events=POLLIN}], 2, -1 <unfinished ...> | 21:56 |
Wizzup | [pid 2645] write(2, "\n(process:2645): GConf-\33[1;35mCR"..., 131) = ? | 21:56 |
Wizzup | or maybe this is some debug and it doesn't hang with gconf | 21:56 |
uvos | ok | 21:56 |
Wizzup | basically my lock screen didn't respond, but screen did go off after a while | 21:56 |
Wizzup | when it happens again I'll try to ssh in and see | 21:57 |
uvos | hmm | 21:57 |
Wizzup | (it is semi regular) | 21:57 |
Wizzup | (bionic) | 21:57 |
uvos | oh ok hmm have not seen sutch a thing | 21:57 |
freemangordon | BTW< I have coredump from Xorg | 21:57 |
freemangordon | didn't look at it though | 21:57 |
freemangordon | uvos: "cpcap-charger cpcap-charger.0: maximum current set to 500 mA" | 21:58 |
uvos | Wizzup: what makes you think mce specificly hanged? | 21:58 |
Wizzup | I attached with strace and didn't see a whole lot | 21:58 |
freemangordon | and it really charges d4 :) | 21:58 |
Wizzup | and /etc/init.d/mce restart also hanged | 21:59 |
uvos | sure but if xorg hangs (as happens quite offten) mce will get stuck | 21:59 |
uvos | just a possiblity | 21:59 |
Wizzup | ok, that could be | 21:59 |
Wizzup | brb | 21:59 |
uvos | freemangordon: ok | 22:00 |
uvos | dose it work on your 100mA hub btw? | 22:00 |
uvos | (its at depmod now here) | 22:01 |
freemangordon | yes, somehow pulling 500 mA :) | 22:01 |
uvos | hmm | 22:01 |
uvos | not good | 22:01 |
freemangordon | oh, lemme remove the power from the hub | 22:01 |
uvos | the hub surely can supply 500mA if d4 just takes it | 22:01 |
uvos | esspecaly with just one device connected | 22:01 |
freemangordon | yeah | 22:02 |
freemangordon | oh, second d4 with android? | 22:02 |
uvos | sure | 22:02 |
freemangordon | yeah, but not now | 22:03 |
uvos | but also idk if its a totaly great idea | 22:03 |
uvos | idk how d4 reacts to overload | 22:03 |
freemangordon | it has wires hanging and stuff :) | 22:03 |
uvos | it should just trip off | 22:03 |
uvos | but yeah | 22:03 |
freemangordon | I'll just wait for you to test :) | 22:03 |
sicelo | Wizzup: i tried to bisect, but i would end up on wrong commits. i think i'm missing the right trick for avoiding merge commit ... i know the problem doesn't exist in 5.18, but in 5.19 onwards, it's consistently there. | 22:10 |
freemangordon | sicelo: doesn;t it boot without patches? | 22:11 |
sicelo | n900 boots vanilla kernel with no issues whatsoever :-) | 22:15 |
freemangordon | then why merging? | 22:15 |
freemangordon | just set good to 5.18. bad to 5.19 and start bisectiong | 22:15 |
sicelo | i'll try again, but it took me to a drm commit or similar :p | 22:16 |
Wizzup | sicelo: no need to try again if it didn't do anything last tie | 22:16 |
Wizzup | we'll have to tackle it for sure because mapphone and n900 are now the same kernel | 22:16 |
Wizzup | (for us) | 22:17 |
sicelo | well i could use different starting points, etc. | 22:17 |
Wizzup | you can also tell git bisect to give you another I think | 22:17 |
uvos | freemangordon: no change in behavior with the xt1602 | 22:30 |
uvos | d4 takes 500mA and reset | 22:31 |
freemangordon | dmesg? | 22:32 |
uvos | nothing gets printed | 22:34 |
uvos | but im a idiot | 22:34 |
Wizzup | wrong kernel? | 22:34 |
freemangordon | looks like :) | 22:34 |
uvos | i have phy_cpcap_usb blacklisted | 22:34 |
freemangordon | (wrong kernel, not the idiot part :p) | 22:34 |
uvos | is this relevent | 22:34 |
freemangordon | sure | 22:34 |
freemangordon | do gadgets work without it? | 22:35 |
uvos | it works | 22:36 |
uvos | d4 dosent activate charging at all | 22:36 |
freemangordon | :) | 22:36 |
uvos | xt1602 remains happy | 22:36 |
freemangordon | nice | 22:36 |
freemangordon | ok, now, lets see how to upstream all that | 22:37 |
freemangordon | oh, first I want to finish charger detection | 22:37 |
freemangordon | somehow, when nokia wall charger is detected, the current drawn is ~1300 mA | 22:37 |
freemangordon | while with 2.4 A wall charger it is ~1600mA | 22:38 |
freemangordon | so there must be a way to detect that | 22:38 |
freemangordon | but not now | 22:38 |
uvos | ok | 22:38 |
uvos | anyhow this is a great improvement | 22:38 |
uvos | no need to be scared of pluging in d4 into random devices... | 22:39 |
freemangordon | here it works even better, it sets max limit to 500 when connected to USB port and to 1800 when connected to charger | 22:39 |
freemangordon | once I polish this we will be able to safely remove charge-mode as a requireent | 22:40 |
freemangordon | *requirement | 22:40 |
Wizzup | freemangordon: if the usbnet stays alive, do the charging leds still go on and off? | 22:40 |
freemangordon | no | 22:40 |
freemangordon | it stays green | 22:40 |
Wizzup | that is truly great | 22:40 |
freemangordon | ah,no | 22:40 |
freemangordon | no | 22:40 |
freemangordon | wait | 22:40 |
freemangordon | when the battery is charged, it goes off and on | 22:40 |
Wizzup | ok | 22:40 |
freemangordon | but usbnet stays alive | 22:40 |
Wizzup | still a good improvement :) | 22:40 |
freemangordon | umm... | 22:40 |
freemangordon | isnt that an expected behaviour? | 22:41 |
freemangordon | like, the battery is full | 22:41 |
Wizzup | on the n900 the led stays green | 22:41 |
Wizzup | but this is likely interaction with the battery subsystem | 22:41 |
freemangordon | grennled here means "charging" | 22:41 |
freemangordon | off means "charged" | 22:41 |
Wizzup | maybe this is because we let cpcap/kernel control led vs userspace | 22:41 |
freemangordon | yes | 22:41 |
freemangordon | it seems all android devices behave like that | 22:42 |
uvos | no | 22:42 |
freemangordon | green for charging off for full | 22:42 |
uvos | well hw implementation | 22:42 |
freemangordon | well, all I see around are like that | 22:42 |
uvos | first thing android dose it remove cpcaps hw controll of the led | 22:42 |
uvos | but yes its sw controll looks the same | 22:42 |
freemangordon | I guess we can fix that | 22:43 |
freemangordon | btw, why did you blacklist the phy? | 22:43 |
uvos | sure its just a reg you set | 22:43 |
sicelo | my S7 stays green on full | 22:43 |
Wizzup | we should have all the bits in place for it already | 22:43 |
uvos | but we need some way to set it | 22:43 |
uvos | idk what would be a good interface for this | 22:43 |
uvos | durring boot before mce becomes active we want cpcap to controll the led after all | 22:43 |
freemangordon | device quirk? | 22:43 |
Wizzup | I think on the n900 mce does it | 22:43 |
uvos | freemangordon: sure i mean kernel interface | 22:44 |
freemangordon | ah | 22:44 |
uvos | freemangordon: just a random sysfs file i gues | 22:44 |
uvos | but maybe somthing better can be devised | 22:44 |
freemangordon | mhm | 22:44 |
Wizzup | I suppose it could be some led sysfs file | 22:44 |
uvos | right | 22:44 |
Wizzup | I think this is what it is for n900 | 22:44 |
uvos | i gues the best way would be the led triggers file | 22:44 |
freemangordon | uvos: do you want the full patch to test? including charger detection etc? | 22:44 |
Wizzup | (but I am not sure) | 22:44 |
Wizzup | freemangordon: we could build it for -devel too if it's ready for further testing | 22:45 |
uvos | freemangordon: i think i would like to just build it for leste | 22:45 |
uvos | so a branch | 22:45 |
freemangordon | I want tmlind to comment on it first | 22:45 |
uvos | freemangordon: ok | 22:45 |
uvos | then lets just wait | 22:45 |
sicelo | Wizzup: the led control is in charger (bq24150) | 22:45 |
Wizzup | uvos: looks like planet.bin is 46GB btw (navit) | 22:45 |
freemangordon | as there are few changes in cpcap phy | 22:45 |
Wizzup | sicelo: check, ty | 22:45 |
uvos | Wizzup: uff | 22:45 |
Wizzup | uvos: well there's maps of many countries in europe but somehow not croatia :P | 22:46 |
uvos | i have just hesse mostly on my sdcard | 22:46 |
freemangordon | zzz time, night! | 22:46 |
uvos | good night freemangordon | 22:46 |
Wizzup | gn | 22:46 |
uvos | hesse (german state) is about 500MB | 22:47 |
uvos | for referance | 22:47 |
Wizzup | did you make it yourself? | 22:47 |
uvos | hmm good question | 22:47 |
uvos | i honstly dont remmeber where i got the file | 22:47 |
uvos | freemangordon: "btw, why did you blacklist the phy?" | 22:49 |
uvos | uses a bit of power | 22:49 |
Wizzup | btw navit looks like this for me: https://wizzup.org/navit.png | 22:49 |
uvos | that looks a bit wrong | 22:49 |
Wizzup | and trust me there's a whole city where it's rendering some sand desert | 22:49 |
Wizzup | yeah hence me wanting to try another map | 22:49 |
uvos | you can play with the profiles | 22:49 |
uvos | to make it render more | 22:49 |
uvos | but its extrreamly slow | 22:50 |
Wizzup | in the application? | 22:50 |
uvos | in the xml | 22:50 |
Wizzup | ok | 22:50 |
uvos | the most detailed profile to slow to use even on ryzen9 | 22:50 |
Wizzup | right | 22:50 |
uvos | its just absurdly slow at rendering things so you need to keep the vectors to a minimum | 22:50 |
uvos | thats why it skipps rendering allmost everytthing except highways except at very close zoom levels | 22:51 |
Wizzup | ok, so that might be it then | 22:51 |
uvos | im not sure whats going on with the water tho | 22:51 |
Wizzup | I haven't figured out how to change zoom | 22:51 |
uvos | that looks just wrong | 22:51 |
uvos | Wizzup: ha thats the funny thing, you cant | 22:51 |
uvos | xD | 22:51 |
uvos | the ui dosent allow it | 22:51 |
uvos | i added something to the xml to add some custom buttons to zoom | 22:51 |
Wizzup | probably need to script something for it | 22:51 |
Wizzup | right | 22:52 |
uvos | anyhow navit is just terribly i only use it becasue there is no alternative | 22:52 |
uvos | it works, thats about all the good i can say about it | 22:52 |
Wizzup | modrana ui somehow makes a bit more sense, but the routing doesn't work, but maybe I should figure out how to use monav | 22:52 |
Wizzup | but since monav is removed from the current repos...ugh | 22:52 |
uvos | well i also need vector maps | 22:53 |
uvos | since gsm only is just to slow | 22:53 |
uvos | to keep up with a car | 22:53 |
Wizzup | gsm data you mean? | 22:53 |
uvos | right | 22:53 |
Wizzup | don't have that problem here, but yeah | 22:53 |
Wizzup | https://openrouteservice.org/ maybe this is useful | 22:54 |
Wizzup | seems a bit hyped | 22:55 |
norayr | glad to see the work on mz609. that is 8 inch tablet. | 23:46 |
norayr | mz617 xyboard 10.1, 10 inch tablet, uses the same TI OMAP 4430 | 23:47 |
norayr | but if mz609 works with maemo it doesnt mean that mz617 will work, right? | 23:48 |
norayr | i don't even know which i want. | 23:48 |
norayr | i just know we need tablet. | 23:49 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!