Wizzup | uvos: I pushed untested changes to builddeb to maemo-6.6-buildeb | 00:16 |
---|---|---|
Wizzup | I'll try for an experimental build | 00:19 |
Wizzup | uvos: did you also include all the config changes that we have in our maemo branches? | 00:20 |
Wizzup | looks like we didn't have them anymore (?) | 00:23 |
Wizzup | in any case, let's see if/where this experimental build fails | 00:23 |
Wizzup | uvos: | 00:53 |
Wizzup | arch/arm/boot/dts/ti/omap/motorola-mapphone-common.dtsi:237.30-250.4: ERROR (phandle_references): /ocp/interconnect@48000000/segment@0/target-module@72000/i2c@0/touchscreen@4a: Reference to non-existent node or label "touchscreen_pins" | 00:53 |
Wizzup | arch/arm/boot/dts/ti/omap/motorola-mapphone-common.dtsi:252.14-261.4: ERROR (phandle_references): /ocp/interconnect@48000000/segment@0/target-module@72000/i2c@0/isl29030@44: Reference to non-existent node or label "als_proximity_pins" | 00:53 |
Wizzup | enough for today :) | 00:57 |
Wizzup | tmlind: I briefly measured the ground of the microsd solder points on the mz617 and it does seem to be hooked up at least | 01:14 |
Wizzup | uvos: hm, did you include powervr stuff in this kernel? | 11:46 |
Wizzup | hm, I think I see what happened | 11:47 |
Wizzup | tmlind: ok, I'm in :) | 12:19 |
Wizzup | tmlind: amazing work | 12:20 |
Wizzup | tmlind: somehow the sd card was just picked up with default mz617 dts | 12:20 |
Wizzup | so with the right dts it boots | 12:21 |
Wizzup | scree nworks | 12:21 |
Wizzup | touchscreen doesn't yet, but wifi works, serial works, usb otg works | 12:21 |
Wizzup | now I can't help but wonder if it's just hooked up on the mz617 without us knowing | 12:21 |
Wizzup | it being the microsd card | 12:24 |
arno11 | Wizzup: hi, i need your advice/help to find those bloody ret/off mode blockers: i have a basic sdcard to try to load a very minimal config on the n900 and to try to add one by one every modules. But i don't know what's the trick to load a minimal config :P | 13:15 |
Wizzup | tmlind: looks like ts works actually | 13:18 |
Wizzup | arno11: give me 5-10 mins | 13:18 |
Wizzup | tmlind: but some leste config or xorg config messes it up | 13:18 |
uvos | ts worked, so thats a regession | 13:20 |
uvos | the micro sdcard holder is not simply not installed, tmlind solderd one on his and that dident work | 13:20 |
uvos | pvr works fine | 13:21 |
uvos | d4 leste ofc stopes ts working | 13:21 |
uvos | because the ts is blaklisted in libinput | 13:21 |
uvos | since we use the filtered touchscreen device instead (from ts-buttons) | 13:21 |
Wizzup | uvos: hi, just a few mins | 13:34 |
arno11 | Wizzup: finally i have to go, let's see pm stuff later ;) | 13:39 |
Wizzup | uvos: hi | 13:53 |
Wizzup | arno11: ok, I'll write here anyway | 13:53 |
Wizzup | uvos: so it looks like there is a ts, but there are some probe errors | 13:53 |
Wizzup | https://paste.debian.net/plain/1304073 | 13:54 |
Wizzup | so there is a 004a and 005b | 13:54 |
Wizzup | uvos it looks like X saw the ts but then dropped it | 13:55 |
Wizzup | [ 451.889] (EE) libinput: Atmel maXTouch Touchscreen: Failed to create a device for /dev/input/event3 | 13:55 |
Wizzup | [ 451.889] (EE) PreInit returned 2 for "Atmel maXTouch Touchscreen" | 13:55 |
Wizzup | uvos and of course there is still the droid4 calibration file in udev | 13:55 |
Wizzup | uvos: so the mz609 and mz617 dts do not build, otherwise I might be able to try another kernel build for experimental | 13:56 |
Wizzup | uvos: worth noting is that it seems like the cache partition containing kexecboot somehow gets messed up by android, so it's not just android getting buggy | 13:57 |
Wizzup | uvos: I have some hack fixes for the dtses, shall I just push those to a test branch to debug the rest of the process? | 13:59 |
Wizzup | uvos: so this time on boot both atmel_mxt_ts probes failed | 14:18 |
Wizzup | [ 7.488616] atmel_mxt_ts: probe of 1-004a failed with error -121 | 14:18 |
Wizzup | [ 8.142578] atmel_mxt_ts: probe of 1-005b failed with error -121 | 14:18 |
Wizzup | well, this time the ts works | 14:21 |
Wizzup | https://wizzup.org/MVI_1083.MP4 <- not promotional material | 14:35 |
uvos | no idea about the ts not probeing | 14:35 |
uvos | havent seen that | 14:36 |
Wizzup | uvos: I suppose I could take the part from mz616 and put it on my mz617 and see if it seems a microsd | 14:36 |
uvos | i never tried building the mz6xx dts's in that branch | 14:36 |
uvos | so them being broken in the merge is not compleatly suprising | 14:36 |
Wizzup | check | 14:36 |
uvos | as i say | 14:37 |
Wizzup | I would like to see if my changes to builddeb worked, so for that purpose I might commit a change that just kludge fixes them for building, not for using purposes | 14:37 |
uvos | i presume you removed the blacklist of the ts | 14:37 |
uvos | ok | 14:37 |
Wizzup | yes I removed the blacklist and blacklisted ts buttons | 14:38 |
uvos | ts-buttons should not load at all | 14:38 |
uvos | as it should not be in the dts | 14:38 |
Wizzup | not sure, I just flashed a d4 image, never booted it, then changed boot.cfg, added dtb and kernel and modules | 14:38 |
uvos | ok | 14:39 |
freemangordon | wow, that's fast | 14:39 |
freemangordon | why it seems smoother than d4? | 14:39 |
uvos | same as bionic | 14:39 |
uvos | it lacs the pannel refesh bug of d4 | 14:39 |
freemangordon | ah | 14:39 |
Wizzup | uvos: btw you said the display doesn't turn off yet, or idle correct, if I recall correctly? | 14:45 |
uvos | no that used to be the case | 14:45 |
uvos | it should now | 14:45 |
Wizzup | ah, okay | 14:45 |
uvos | it also sleeps fine | 14:45 |
uvos | and has very low power cosumption when it dose | 14:45 |
uvos | i can run weeks if you dont touch it | 14:46 |
Wizzup | hmm | 14:46 |
Wizzup | on leste? | 14:46 |
uvos | no just minimal debian | 14:46 |
Wizzup | I see | 14:46 |
uvos | but leste too really | 14:46 |
Wizzup | right now it uses about 0.180A at 3.85V on my lab psu | 14:46 |
Wizzup | with the screen off, but wifi on | 14:46 |
uvos | hmm | 14:47 |
Wizzup | I suppose we might need some idling scripts | 14:47 |
uvos | ret? | 14:47 |
Wizzup | RET is 0 atm | 14:47 |
uvos | hmm ok something is blocking | 14:47 |
Wizzup | root@devuan-xt616:~# /etc/init.d/droid4-powermanagement status | 14:47 |
Wizzup | d=2024-01-14|t=13:47:26|i=OFF:0,RET:0|p=673|c=NA|b=none | 14:47 |
uvos | the b= dosent work on leste anymore | 14:47 |
uvos | omapconf time | 14:47 |
Wizzup | is that because of kernel changes, or? | 14:47 |
uvos | no | 14:48 |
Wizzup | (the b=not working anymore) | 14:48 |
uvos | becasue the version of a utility used to parse the bits changed | 14:48 |
Wizzup | so droid4-powermanagement says: | 14:48 |
Wizzup | * Module omap_hdq blocks idle: Seems to poll devices, blacklist? | 14:48 |
Wizzup | * Module atmel_mxt_ts blocks idle: Should be unloaded when screen is blanked | 14:48 |
uvos | thats old/wrong | 14:48 |
Wizzup | both are loaded | 14:48 |
Wizzup | ok, we should maybe update those eventually | 14:49 |
uvos | its wrong | 14:49 |
uvos | atmel_mxt_ts got idle support since then | 14:49 |
Wizzup | I mean we should update the script :) | 14:49 |
uvos | omap_hdq we tamed by changing the parameters | 14:49 |
Wizzup | right | 14:49 |
uvos | anyhow omapconf will tell you | 14:49 |
uvos | just compeare it with the d4 output | 14:50 |
Wizzup | understood | 14:50 |
Wizzup | I think I want to try to build the chimaera-experimental kernel first atm, many things happening at once | 14:50 |
Wizzup | it's on a lab psu anyway so I don't mind the power usage | 14:50 |
uvos | sure | 14:50 |
uvos | working is at all is more important | 14:50 |
uvos | than ret | 14:50 |
uvos | anyhow even at 0.5w it should last quite a while | 14:50 |
uvos | the battery is huge | 14:50 |
Wizzup | yes, but if we can get it to idle then it should be able to charge while in RET | 14:51 |
Wizzup | (that would be my main motivation atm) | 14:51 |
uvos | if you upp cpcap to 2A in sysfs it will charge fine | 14:51 |
uvos | still takes a long time since the battery is huge | 14:51 |
Wizzup | oh, so what I read in the logs a while ago, wrt charging being a problem is no longer relevant? | 14:53 |
uvos | its less of a problem | 14:54 |
Wizzup | ok | 14:55 |
uvos | back when the bridge chip dident work | 14:55 |
uvos | we could not turn off the display at all | 14:55 |
uvos | which drew like 1.5A or so | 14:55 |
uvos | its sill hillariously slow to charge via cpacp ofc | 14:55 |
Wizzup | 2A doesn't seem to be that slow, or? | 14:57 |
uvos | i mean its a 7Ah battery | 14:58 |
uvos | so takes like 4h | 14:58 |
uvos | you also going to have an extreamly hard time to charge it its empty | 14:59 |
uvos | since the power draw of the display on boot is intense | 14:59 |
Wizzup | right | 14:59 |
Wizzup | uvos: I think kexecboot has a charging mode now | 15:00 |
uvos | so it has the d4s low power boot problem on sterioids | 15:00 |
uvos | right | 15:00 |
Wizzup | uvos, not important now, but I think speaker audio doesn't probe either at mfor me | 15:03 |
uvos | hmh | 15:03 |
uvos | strange | 15:03 |
uvos | should be the same as d4 | 15:03 |
Wizzup | in any case, I'm super surprised by how well this works now already | 15:03 |
Wizzup | [ 39.769256] mot-mdm6600-codec 4806a000.serial:modem:audio-codec@2: ASoC: error at snd_soc_component_probe on 4806a000.serial:modem:audio-codec@2: -110 | 15:04 |
Wizzup | [ 39.797454] asoc-audio-graph-card soundcard: ASoC: failed to instantiate card -110 | 15:04 |
Wizzup | [ 39.812561] asoc-audio-graph-card soundcard: error -ETIMEDOUT: parse error | 15:04 |
uvos | yeah i mean once tmlind figured out whats wrong with the bridge driver | 15:04 |
Wizzup | [ 39.827178] asoc-audio-graph-card: probe of soundcard failed with error -110 | 15:04 |
uvos | its just a big d4 | 15:04 |
Wizzup | still :) | 15:04 |
Wizzup | which reminds me, does razr still have this screen issue? | 15:07 |
uvos | you have some mz604 right | 15:09 |
uvos | just mentioning: according to xda some of those have unlocked bootloaders | 15:09 |
uvos | not that its super revelant to us | 15:09 |
uvos | its not like we are going to use a different boot procedure for some random mz604 | 15:10 |
Wizzup | are those omap? | 15:11 |
Wizzup | hm yeah so this time to atmel probe failed again | 15:11 |
Wizzup | s/ to / the / | 15:12 |
uvos | i ment 607 sry | 15:12 |
Wizzup | ok | 15:12 |
uvos | all 604s are unlocked | 15:12 |
uvos | but thats nvidia powered xoom1 | 15:12 |
Wizzup | let me check @ mz607 | 15:13 |
Wizzup | I have something labelled mz608 here as that is what the firmware says, but it doesn't have a microsd card slot exposed | 15:13 |
Wizzup | I *think* the rest are mz609 | 15:13 |
uvos | ok | 15:13 |
uvos | no | 15:13 |
uvos | mz607 european wifi only version | 15:13 |
uvos | same story as xt912/xt910 some european examples have unlocked bootloaders | 15:15 |
uvos | depends on where they where sold exactly | 15:15 |
Wizzup | we have the luxury to get one on ebay in any case, or if you find specific xt91x es | 15:15 |
Wizzup | if it helps with the devel we can use funding money to acquire them | 15:15 |
uvos | i have one allready | 15:16 |
Wizzup | mz607? | 15:16 |
uvos | xt910 (or 12 i dont remember) with a unlocked bootloader | 15:16 |
Wizzup | right | 15:16 |
uvos | i have no mz60x variant at all | 15:16 |
Wizzup | I know there's a lot of different chat threads, but do you recall if the razr still had this refresh bug? | 15:16 |
uvos | yes it dose | 15:17 |
uvos | i never got to that either | 15:17 |
Wizzup | ok | 15:17 |
uvos | i was going to ajust the timeings slightly to see if it goes away | 15:17 |
Wizzup | ok | 15:20 |
Wizzup | once the conversations/telepathy settles down a little I would like to retry droid3/razr/atrix 2 | 15:21 |
uvos | working droid3 would probubly be most usefull | 15:26 |
uvos | razr has a nice display, atrix 2 is kinda just a bionic (maybe it even boots the bionic image, would not suprise me at all, stock android dmseg output suggests its exatcly the same as bionic) | 15:26 |
uvos | minus the lte modem ofc | 15:27 |
Wizzup | I think getting them all to work would be great and I agree d3 is most useful, in the sense that it is mostly d4, but just with less ram, but no modem restrictions in the us | 15:29 |
Wizzup | (which is useful for me :)) | 15:29 |
uvos | well it also has the older non-vunerable bootloader | 15:29 |
uvos | which is booo | 15:30 |
uvos | thats a benefit xt91x has over bionic too | 15:30 |
Wizzup | right, clown boot | 15:31 |
tmlind | good to hear the micro-sd works with leste on some tablets :) | 15:52 |
Wizzup | yes, it's amazing :D | 15:54 |
tmlind | so which tablets are now confirmed to work for micro-sd? | 16:09 |
tmlind | only mz616? | 16:10 |
Wizzup | so far, yes | 16:11 |
Wizzup | the mz608 that I have here (presumably) according to some websites has microsd, but it is not exposed, I had not pried it open yet | 16:11 |
tmlind | oh interesting | 16:17 |
sicelo | https://www.cnx-software.com/2024/01/10/lichee-console-4a-portable-risc-v-development-terminal-review-unboxing-teardown-hands-on/ | 16:21 |
Wizzup | cute | 16:49 |
Wizzup | sicelo: so we need to work on a risc-5 port? ;) | 16:51 |
Wizzup | uvos: I have a mb865 here | 17:14 |
Wizzup | you're suggesting I try clown boot on it like it's bionic? | 17:14 |
Wizzup | InlineFlashing_Edison_67.21.125_CFC_P3_APBP.xml/allow-mbmloader-flashing-mbm.bin well that is helpful | 17:34 |
uvos | Wizzup: yeah, possibly you have to rebuild the kexec modules against the edison kernel | 17:53 |
uvos | Wizzup: and you have to use edison mbm ofc | 17:54 |
uvos | otherwise it could just work | 17:54 |
Wizzup | hm, right, I did do some of that for solana (I still have the clownboot fork of that) | 17:54 |
uvos | iirc you treid but in the end i had to compile the modules :P | 17:55 |
uvos | but i can do that for edison too | 17:55 |
uvos | anyhow possibly its the same kernel as bionic anyhow so it should | 17:55 |
uvos | work | 17:56 |
uvos | just try instmoding them in android | 17:56 |
uvos | or just compeare uname | 17:56 |
uvos | over all confidence my is high that it shal just boot the bionic image otherwise | 17:57 |
Wizzup | ok | 17:59 |
Wizzup | I'll root it first, then try the insmod\ | 17:59 |
uvos | needs to be updated to android 4.0.4 first ofc | 18:00 |
uvos | if its 4.1 you might have to flash 4.0.4's boot.img | 18:00 |
uvos | you do on bionic anyhow | 18:00 |
Wizzup | freemangordon: btw I removed iphdb and glib from experimental sicne I don't think we need them anymore | 18:01 |
Wizzup | uvos: yeah it's 4.0.4 already | 18:01 |
Wizzup | the droid 3 was actually using android 2.3.4 iirc | 18:01 |
uvos | yes droid 3 was the only mapphone device that never got the 4.0+ updates | 18:01 |
uvos | this is because /system is too small | 18:02 |
uvos | but all of them started with 2.3 besides the tablets that started at 3.1 | 18:02 |
Wizzup | uvos: btw, the droid 3 dts that I made is not in your 6.6 kernel | 18:08 |
Wizzup | this is not a big problem but this is why the postinst fails on the experimental kernel ;) | 18:09 |
uvos | yeah that dosent merge | 18:11 |
uvos | and i havent rebased that yet | 18:11 |
uvos | d3 branch that is | 18:12 |
Wizzup | makes sense | 18:13 |
Wizzup | let me try booting the experimental kernel | 18:13 |
Wizzup | worked | 18:17 |
Wizzup | (mz616) | 18:17 |
uvos | https://github.com/IMbackK/droid4-linux/tree/droid3-61 | 18:19 |
uvos | here be d3 branch | 18:19 |
Wizzup | thanks | 18:20 |
Wizzup | probably won't get to it today, planning to try the atrix and then take a break, but good to have it around :) | 18:20 |
Wizzup | uvos: well insmod uart.ko made me lose adb shell but the device is on | 18:35 |
Wizzup | so I guess it's compatible | 18:35 |
uvos | yeah best do this via local shell on device | 18:37 |
Wizzup | do you think I should try the other steps too? | 18:39 |
Wizzup | I figured I'd just assume it works | 18:39 |
uvos | its probubly fine yeah | 18:39 |
* Wizzup prepares bionic mage | 18:41 | |
Wizzup | image , not mage :) | 18:41 |
uvos | no nead | 18:43 |
uvos | *eed | 18:43 |
uvos | *need | 18:43 |
uvos | if it boots to kexecboot via clownboot obv its booting the mainline kernel as a bionic | 18:43 |
Wizzup | yeah but I'm going to make a leap of faith and see if it just works | 18:45 |
Wizzup | the atrix has an extra button it seems | 18:55 |
Wizzup | hm, I think there's some read-only mount stuff going on again, the install.sh script failed | 18:58 |
Wizzup | that really needs set -e I think | 18:58 |
Wizzup | hm, it doesn't let me reflash, complains about preflash validation failure | 19:06 |
Wizzup | so I guess somehow the bootloader is not unlocked even with the allow-mbm... | 19:06 |
uvos | what partition? | 19:09 |
Wizzup | cdt.bin | 19:13 |
uvos | you dont need to reflash that | 19:13 |
uvos | also cdt.bin needs to be from the exact version | 19:14 |
uvos | since thats where the keys live | 19:14 |
uvos | so thats not unusual | 19:14 |
Wizzup | also lbl | 19:14 |
uvos | allow-mbm... only unlocks a couple of extra partions but def not cdt | 19:14 |
uvos | yes all of those need to be signed with the right key | 19:14 |
uvos | so what you are trying to flash is not signed with the same key as the device | 19:15 |
Wizzup | hm | 19:15 |
Wizzup | ok, so I am able to flash mbm for sure | 19:15 |
Wizzup | ok, so I need to find another android zip to flash? | 19:15 |
uvos | yes the signing keys for mbm never changed | 19:15 |
uvos | you can flash any mbm from the same device | 19:16 |
uvos | yes | 19:16 |
uvos | or even from different devices often (for mbm) | 19:16 |
Wizzup | ok | 19:16 |
Wizzup | from this: https://firmware.center/firmware/Motorola/MB865%20Atrix%202%20%28Edison%2C%20Fuath%29/Stock/MB865/ I got the latest InlineFlashing, as well as the ASIAOPS, MEARET and SEARET | 19:16 |
uvos | i would really just ignore all of this and only worry about it if you cant flash /system | 19:17 |
Wizzup | from the InlineFlashing I can flash cdt.bin | 19:17 |
uvos | or bpsw | 19:17 |
Wizzup | hmm ok | 19:17 |
uvos | and boot | 19:17 |
uvos | those are the only partitions you need to touch really | 19:17 |
uvos | clownboot only touches system | 19:17 |
Wizzup | ok, but should system not be compatible with the rest? | 19:18 |
uvos | and pds | 19:18 |
uvos | i think | 19:18 |
uvos | system dosent care | 19:18 |
Wizzup | I am able to flash InlineFlashing files, so I'm planning to just flash those enirely | 19:18 |
Wizzup | entirely* | 19:18 |
uvos | you can boot different /system | 19:18 |
Wizzup | but I can stop doing that if you think that doesn't make sense | 19:18 |
uvos | it makes no sense to reflash all that stuff and its needless danger | 19:19 |
uvos | just reflash what you touched | 19:19 |
Wizzup | well I didn't flash the phone initially | 19:19 |
Wizzup | that's the thing, so I don't know what's flashed in particular | 19:19 |
uvos | yeah but any system will do | 19:19 |
uvos | even 4.1 vs 4.0 dosent matter | 19:19 |
uvos | as long as the bootloader eats it its fine | 19:20 |
Wizzup | hm ok | 19:20 |
uvos | same with boot | 19:20 |
Wizzup | and preinstall? | 19:20 |
uvos | ah right preinstall is where we place some clownboot stuff | 19:20 |
uvos | yes you need to reflash that | 19:20 |
uvos | but also dosent matter | 19:20 |
Wizzup | ok | 19:20 |
uvos | preinstall just contains some application | 19:20 |
uvos | s | 19:21 |
uvos | that are wait for it.. preinstalled | 19:21 |
Wizzup | I just want to boot it android again first and root it, and then run the install.sh one by one | 19:21 |
Wizzup | ok, will boot now and see | 19:21 |
Wizzup | thanks for the help as usual :) | 19:22 |
uvos | yw, i presume it booted | 19:23 |
Wizzup | if it gets past the 'rethink possible' logo then it will, otherwise I am stuck at where I was before :) | 19:24 |
Wizzup | says 'preparing device', so looks like it'll get there eventually | 19:25 |
Wizzup | ok, rooting again | 19:27 |
Wizzup | freemangordon: it looks like chimaera-testing in extras has a newer openmediaplayer, maybe we can move that? | 19:30 |
Wizzup | brb dinner | 19:32 |
Wizzup | will let you know how atrix goes :) | 19:32 |
uvos | if it gets stuck at prepearing device you have to erase cache and data | 19:41 |
uvos | (you should do this when downgradeing android (even patch levels)) | 19:41 |
uvos | and that might be what you are doing, since we dont know what the old version was | 19:41 |
uvos | dazu: guten appetit! | 19:42 |
Wizzup | uvos: ok so /sdcard is not writeable | 20:24 |
Wizzup | I used /data/bionic | 20:27 |
Wizzup | uvos: hm: | 20:29 |
Wizzup | $ fastboot flash bpsw ~/maemo-leste/droid4-kexecboot/current/droid4-kexecboot.img | 20:29 |
Wizzup | Sending 'bpsw' (4096 KB) OKAY [ 0.202s] | 20:29 |
Wizzup | Writing 'bpsw' (bootloader) Command restricted | 20:29 |
Wizzup | also can't override boot.img per the bionic's readme | 20:30 |
Wizzup | $ fastboot flash boot boot.img | 20:30 |
Wizzup | Warning: skip copying boot image avb footer (boot partition size: 0, boot image size: 8388608). | 20:30 |
Wizzup | Sending 'boot' (8192 KB) OKAY [ 0.440s] | 20:30 |
Wizzup | Writing 'boot' (bootloader) Preflash validation failure | 20:30 |
Wizzup | FAILED (remote: '') | 20:30 |
Wizzup | fastboot: error: Command failed | 20:30 |
Wizzup | interesting, now it just shows 'ap fastboot flash mode (s) (flash failure)' | 20:31 |
Wizzup | so it doesn't boot to android | 20:31 |
Wizzup | I guess I need to dd? | 20:33 |
uvos | for /sdcard you might need a fat32 sdcard there | 20:39 |
uvos | Wizzup: no maybe it dosent have bpsw at all | 20:39 |
uvos | could be that bpsw, altho never used, was intended for something todo with lcm2.0 | 20:40 |
uvos | which we dont have | 20:40 |
uvos | you can check in android ofc for the partitions | 20:40 |
uvos | if it isent there you might have to use cache | 20:40 |
Wizzup | will check | 20:40 |
uvos | (and modify the clownboot scripts to suit | 20:40 |
uvos | ) | 20:40 |
uvos | thats already annoying if true | 20:41 |
uvos | because now we already have to support the atrix as another distinct device | 20:41 |
Wizzup | it has more buttons anyway | 20:41 |
uvos | those are just on the button matrix | 20:41 |
uvos | you could add them to the bionic dts | 20:42 |
Wizzup | so /proc/partitions doesn't show a bpsw | 20:42 |
uvos | the bionic wont care | 20:42 |
Wizzup | I don't know if it is normally labelled | 20:42 |
uvos | not sure | 20:42 |
uvos | but look at the partition | 20:42 |
uvos | should be 4mb | 20:42 |
Wizzup | I don't know how to resolve it to a partition | 20:42 |
Wizzup | the name | 20:42 |
uvos | so some are listed in cmdline | 20:42 |
Wizzup | <5>[ 0.000000,0] Kernel command line: omap_wdt.timer_margin=60 oops=panic console=/dev/null rw mem=1024M@0x80000000 vram=10300K omapfb.vram=0:8256K,1:4K,2:2040K init=/init ip=off mmcparts=mmcblk1:p7(pds),p15(boot),p16(recovery),p17(cdrom),p18(misc),p19(cid),p20(kpanic),p21(system),p22(cache),p23(preinstall),p24(webtop),p25(userdata) androidboot.bootloader=0x0A72 | 20:43 |
uvos | the bootloader lists them there | 20:43 |
uvos | /dev/mmcblk1p14 | 20:43 |
uvos | if thats 4mb and empry | 20:44 |
uvos | *empty | 20:44 |
uvos | then its bpsw | 20:44 |
Wizzup | 179 14 4096 mmcblk1p14 | 20:44 |
Wizzup | it does not seem to be mounted | 20:44 |
uvos | fs? | 20:45 |
Wizzup | let me root again so I can check the partition contents | 20:45 |
uvos | the mild danger here is that bpsw might be used for something on this device, in difference to all the others | 20:45 |
uvos | maybe because the modem is slighly different or something | 20:45 |
uvos | i gues risking bricking one as an expirament isent so big a deal | 20:45 |
Wizzup | there is webtop | 20:45 |
Wizzup | you think it would be a hard brick? | 20:46 |
uvos | yeah you could use that | 20:46 |
uvos | Wizzup: unless you can find a singed bpsw image, maybe | 20:46 |
Wizzup | right | 20:46 |
uvos | android might not boot if the modem freaks out | 20:46 |
Wizzup | flashing a wrong boot.img should not brick anything though right? | 20:46 |
uvos | no thats safe | 20:46 |
Wizzup | (rather, dding yours) | 20:46 |
uvos | wait | 20:46 |
uvos | what boot.img are you dding? | 20:47 |
Wizzup | https://github.com/IMbackK/bionic-clown-boot/blob/master/boot.img | 20:47 |
uvos | no no | 20:47 |
uvos | dont do that | 20:47 |
Wizzup | ok, it also didn't work | 20:47 |
uvos | yeah its bionic signed | 20:47 |
uvos | thats just a specific bionic kernel version | 20:47 |
Wizzup | so just use a boot.img that I have? | 20:47 |
Wizzup | ok | 20:47 |
uvos | that i know works, since the one from 4.1 dosent | 20:47 |
Wizzup | let me double check that insmod still works since I reflashed system and such | 20:47 |
uvos | just use a android 4.0.4 boot.img from atrix in its place | 20:48 |
Wizzup | ok | 20:48 |
uvos | dding a bionic image here | 20:48 |
uvos | wont brick your device, but it wont work either | 20:48 |
Wizzup | https://github.com/IMbackK/bionic-clown-boot/blob/master/scripts/kexebootstart.sh#L18C159-L18C159 so I would change this to the webtop partition? | 20:48 |
uvos | since the bootloader will reject it and force you into fastboot asking you to reflash | 20:48 |
Wizzup | right | 20:49 |
Wizzup | I think this might be what happened earlier | 20:49 |
uvos | yeah if you dont want to use mmcblk1p14 | 20:49 |
uvos | might make sense if webtop flashable from fastboot | 20:50 |
Wizzup | will try that now | 20:50 |
Wizzup | yes that works | 20:51 |
Wizzup | ok, let me install bionic again with a fixed script for the partition | 20:51 |
Wizzup | bionic->clown boot | 20:52 |
Wizzup | uvos: got to kexecboot | 20:58 |
uvos | weee | 20:58 |
Wizzup | also see kernel console | 20:58 |
uvos | that basicly guarentees that the device is otherwise 100% bionic compatabile | 20:58 |
Wizzup | it might get to h-d | 20:58 |
Wizzup | generting ssh keys atm | 20:58 |
Wizzup | (..) | 20:58 |
uvos | needs more random data :P | 20:58 |
Wizzup | https://wizzup.org/dirlist/atrix2/ | 21:01 |
uvos | sweet | 21:02 |
uvos | looks like Wizzup is having a pretty good sucess rate today | 21:03 |
Wizzup | nah, this is just me aping my way through trying to use your knowledge and work as well as tmlind's knowledge and work :P | 21:03 |
uvos | could you grab p14 | 21:07 |
uvos | for examination | 21:07 |
uvos | we also need the battery nvme | 21:07 |
uvos | altho i presume its the same | 21:07 |
uvos | since it shares its battery form factor with bionic | 21:07 |
Wizzup | yup, sec | 21:09 |
Wizzup | seems to be ... just zeroes | 21:11 |
Wizzup | for p14 | 21:11 |
Wizzup | how do I get you the battery data? | 21:12 |
uvos | /sys/bus/1w/devices/ | 21:14 |
Wizzup | uvos: https://wizzup.org/dirlist/atrix2/dmesg.txt | 21:14 |
uvos | some lower directory of that | 21:15 |
Wizzup | there are some differences it seems | 21:15 |
uvos | mbm is setting higher voltages | 21:15 |
uvos | thats harmless | 21:15 |
uvos | and not really a relevant difference | 21:15 |
Wizzup | no, but like | 21:15 |
Wizzup | [ 8.487579] kxcjk1013 3-000f: Error reading who_am_i | 21:16 |
Wizzup | [ 8.498779] kxcjk1013: probe of 3-000f failed with error -121 | 21:16 |
Wizzup | and | 21:16 |
Wizzup | [ 7.697601] cpcap_adc cpcap_adc.0: CPCAP ADC device probed | 21:16 |
Wizzup | [ 7.704803] lm75: probe of 0-0048 failed with error -121 | 21:16 |
Wizzup | [ 7.712951] cpcap_battery cpcap_battery.0: failed to register power supply | 21:16 |
Wizzup | the emif_probe and omap-sham seems less relevant | 21:16 |
uvos | the emif error is on all our devices | 21:16 |
Wizzup | right | 21:16 |
uvos | we dont setup emif correctly | 21:16 |
Wizzup | I also don't see the modem in lsusb | 21:16 |
Wizzup | which is probably why there is no audio card either: | 21:17 |
uvos | ok so the cpcap battery thing is pretty wierd | 21:17 |
Wizzup | [ 38.398345] mot-mdm6600-codec 4806a000.serial:modem:audio-codec@2: ASoC: error at snd_soc_component_probe on 4806a000.serial:modem:audio-codec@2: -110 | 21:17 |
Wizzup | [ 38.462890] asoc-audio-graph-card soundcard: ASoC: failed to instantiate card -110 | 21:17 |
Wizzup | [ 38.483459] asoc-audio-graph-card soundcard: error -ETIMEDOUT: parse error | 21:17 |
Wizzup | [ 38.498901] asoc-audio-graph-card: probe of soundcard failed with error -110 | 21:17 |
uvos | ok ok | 21:17 |
Wizzup | I am not trying to make a point :) | 21:17 |
uvos | maybe the modem isent on usb like bionic but on spi like razer | 21:17 |
Wizzup | I just enumerated all the things I could find | 21:17 |
uvos | that would be annoying | 21:17 |
Wizzup | maybe superflous but /sys/bus/1w/devices/ does not exist currently | 21:18 |
uvos | no thats a problem | 21:18 |
uvos | but it exits on my d4 too | 21:18 |
uvos | hmm | 21:18 |
uvos | so previously unoticed bug | 21:18 |
Wizzup | do you mean that it doesn't exist on your d4? | 21:19 |
uvos | *it DOSENT exist | 21:19 |
uvos | yes | 21:19 |
Wizzup | check | 21:19 |
Wizzup | let me reboot, see if -devel works ok, and then try to boot to android from kexecboot as well | 21:20 |
Wizzup | just to confirm that that still works | 21:20 |
uvos | "android from kexecboot" er no | 21:22 |
uvos | clowboot burns that bridge by hijaking androids boot process to start kexecboot | 21:22 |
Wizzup | ah ok | 21:22 |
Wizzup | my bad | 21:22 |
Wizzup | in any case, I have a second atrix 2 | 21:22 |
Wizzup | so we can use that for android debugging | 21:22 |
uvos | we cant get to android from kexecboot anyhow | 21:23 |
uvos | since we are in mainline linux at that point | 21:23 |
Wizzup | ah, right, of course | 21:23 |
uvos | and the motorola kernel has a bug that it dosent survive warm boots | 21:23 |
uvos | so you cant kexec it | 21:23 |
uvos | buuut | 21:23 |
Wizzup | hehe | 21:23 |
Wizzup | in any case, -devel boots fine :) | 21:23 |
uvos | we could improve clownboot to disable itself and let android continue if vol-up is held or so | 21:24 |
Wizzup | right | 21:24 |
uvos | ok | 21:25 |
uvos | do you have decoded stock dts and the signal map? | 21:26 |
Wizzup | not yet | 21:26 |
Wizzup | screen brightness doesn't work either fwiw | 21:26 |
uvos | yeah | 21:26 |
Wizzup | adjusting it I mean | 21:26 |
uvos | the led controller dient work | 21:26 |
uvos | *probe | 21:26 |
uvos | thats lm75 | 21:26 |
uvos | maybe the backligt is connected to cpacap instead | 21:26 |
uvos | or the lm75 has a different enable gpio | 21:27 |
uvos | (which would be in the singal map | 21:27 |
uvos | ) | 21:27 |
Wizzup | ok | 21:27 |
Wizzup | the atrix2 dir as a device_tree.img now | 21:27 |
Wizzup | has a* | 21:28 |
uvos | Wizzup: ok ill decode the signal map and dt tomorrow | 21:29 |
uvos | Wizzup: should still have the scripts somewhere | 21:30 |
Wizzup | cool ty | 21:30 |
Wizzup | I'll write up the steps I took on a gh issue | 21:30 |
uvos | you can check with evtest what the shutter button has, matrix address wise | 21:31 |
uvos | btw | 21:31 |
Wizzup | ok | 21:35 |
Wizzup | 2 mins | 21:36 |
Wizzup | I'm curious if the modem is on spi or not, if so that would be shame | 21:36 |
uvos | should be able to find out in android sysfs | 21:36 |
Wizzup | yeah | 21:37 |
uvos | but if so then its a theme with gsm only mapphones | 21:37 |
Wizzup | what bothers me mostly is that tmlind said it's unlikely we'll ever get that to work :P | 21:38 |
uvos | yeah | 21:39 |
Wizzup | we'll never* | 21:39 |
Wizzup | uvos: which input device do you want me to evtest? | 21:39 |
Wizzup | I thought: /dev/input/event2:cpcap-pwrbutton | 21:40 |
uvos | omap keypad | 21:40 |
uvos | no | 21:40 |
Wizzup | ok | 21:40 |
uvos | the volume buttons are only on cpcap on devices with no removeable battery | 21:40 |
uvos | (for the reboot claw) | 21:40 |
uvos | otherwise they are matrix keys | 21:40 |
Wizzup | Testing ... (interrupt to exit) | 21:40 |
Wizzup | Event: time 1705264818.732375, type 4 (EV_MSC), code 4 (MSC_SCAN), value 10 | 21:40 |
Wizzup | Event: time 1705264818.732375, -------------- SYN_REPORT ------------ | 21:40 |
uvos | great | 21:40 |
uvos | ok we can add that to dts | 21:40 |
Wizzup | ok, the other mb865 needs a bit of time to charge the battery | 21:44 |
* uvos is also charing several mb865 | 21:45 | |
Wizzup | one thing that might need to be done is to network unlock these before installing leste | 21:46 |
Wizzup | I guess not for us in europe of course | 21:46 |
uvos | are they generally locked? | 21:46 |
uvos | even outside the us? | 21:46 |
Wizzup | not sure regarding outside the us | 21:46 |
Wizzup | I still do have a sigma key dongle and license.... used it just once | 21:47 |
uvos | something to try i gues | 21:47 |
uvos | anyoingly they use old school huge sims | 21:47 |
uvos | providers here dont even supply those anymore | 21:47 |
Wizzup | you don't have 20 pcs of those lying around? | 21:47 |
Wizzup | :D | 21:48 |
uvos | i need to source an adapter | 21:48 |
uvos | no :P | 21:48 |
Wizzup | If you still used nokia n900 you would :p | 21:48 |
uvos | i gues so | 21:48 |
Wizzup | btw, the battery indication does give me 86%, so it looks like it might work | 21:48 |
uvos | but the n900's older is so over engeneerd you can place a micro sim in there, if its just for testing ;) | 21:48 |
uvos | *holder | 21:49 |
Wizzup | true | 21:49 |
Wizzup | unfortunately I have a piece of paper folded about 5-6 times in between my battery and sim holder ;) | 21:49 |
uvos | why that | 21:50 |
Wizzup | classic n900 problems where it will sometimes lose sim contact | 21:50 |
Wizzup | or lose modem somehow | 21:50 |
Wizzup | btw I wrote some info here https://github.com/maemo-leste/bugtracker/issues/649 | 21:50 |
uvos | oh so its over-engeneerd and bad | 21:50 |
uvos | great | 21:50 |
Wizzup | hehe | 21:57 |
Wizzup | ah, mb865 is not supported by sigmakey https://forum.gsmhosting.com/vbb/f719/unlock-mb865-1629642/ | 21:59 |
Wizzup | ./bus/spi/drivers/mdm6600_spi | 22:05 |
Wizzup | :( | 22:05 |
uvos | :( | 22:10 |
uvos | now its really not very interesting | 22:10 |
uvos | a bionic with no modem.. pass :P | 22:11 |
Wizzup | you're looking at it all wrong, this is a challenge to make the modem work ;) | 22:12 |
uvos | xD | 22:12 |
Wizzup | do you remember what made supporting mdm6600 over spi particularly hard? | 22:18 |
* Wizzup will try to grep irc logs | 22:18 | |
uvos | Wizzup: kernel just dosent implement qmi over spi (just from what tmlind explained from memory) | 22:20 |
Wizzup | right | 22:21 |
Wizzup | there must be non-omap phones using qmi over spi | 22:22 |
Wizzup | but maybe not mainlind :) | 22:22 |
Wizzup | mainlined* | 22:22 |
uvos | iirc on android the kernel just exposes the modem interfaces to userspace and some blob talks qmi | 22:23 |
uvos | so the downstream implementations are not very usefull | 22:23 |
Wizzup | oof that would be ugly | 22:23 |
uvos | now i wonder if mz608 is also mdm6600 over spi | 22:26 |
uvos | that would complete the theory that all gsm only devices are so | 22:26 |
Wizzup | from what I found tmlind wrote that xyboards have modem on spi | 22:27 |
uvos | nah | 22:27 |
Wizzup | 16:22 < tmlind> uvos: xt910, xt912 and xyboard all uses qmi over spi for the modem i think, so at least i'm not working on that stuff | 22:27 |
Wizzup | 18:46 < tmlind> uvos: i think also droid4 has spi2 wired to the modem but not in use in the phone firmware. maybe that's just a config option though who knows | 22:27 |
Wizzup | but this was many years ago | 22:27 |
uvos | pretty sure he is wrong for xt617 | 22:27 |
Wizzup | probably 4+ years | 22:28 |
uvos | he is def wrong for xt912 | 22:28 |
uvos | only xt910 has it over spi | 22:29 |
uvos | and im pretty sure he is wrong too for mz617 | 22:29 |
Wizzup | right, I think he got the modem working on mz617 :) | 22:29 |
Wizzup | just data only | 22:29 |
Wizzup | if the blob speaks qmi over spi, it might not be super complicated, but yeah, maybe cm or lineage can tell us a bit more | 22:30 |
Wizzup | on the mz616, I also don't see the modem on usb, but the kernel reported the modem powered up ok | 22:31 |
Wizzup | uvos: there seems to be /sys/bus/platform/drivers/qmi-over-spi and /sys/module/qmi_spi on the 3.0.8 kernel | 22:39 |
Wizzup | any idea where the kernel tree for edison could be found? | 22:41 |
uvos | probubly its just the motorola_mapphone_common tree | 22:41 |
uvos | otherwise motorola published the code on sourceforge | 22:42 |
uvos | (we should really save all of that sourceforge gets iffier by the month) | 22:43 |
uvos | (https://github.com/STS-Dev-Team/android_kernel_motorola_omap4-common) this one if you dont know where to find it | 22:45 |
Wizzup | ty | 22:46 |
uvos | drivers/net/spi/qmi_net.c | 22:46 |
Wizzup | yeah, I was looking at that, looks like very little code | 22:48 |
Wizzup | there must be more | 22:48 |
uvos | something something 30mb userspace blob :P | 22:49 |
Wizzup | so there is arch/arm/mach-omap2/board-mapphone.c | 22:49 |
uvos | sure thats just board source | 22:49 |
Wizzup | oh this just refers to qmi over spi | 22:50 |
Wizzup | yeah | 22:50 |
Wizzup | are you sure about the userspace blob? I mean it's entirely possible of course | 22:50 |
Wizzup | I just did find / | grep qmi and all I found was some sysfs stuff | 22:50 |
Wizzup | (on android) | 22:50 |
uvos | yeah i mean there is a huge blob that just opens all the modem device files | 22:50 |
uvos | anyhow tmlind surely knows why he thinks implementing qmi over spi would be very hard | 22:51 |
Wizzup | mhm | 22:51 |
uvos | probubly best to just interrogate him | 22:51 |
Wizzup | well let's just get an image going at least for mb865 in any case | 22:52 |
uvos | ill bring the thumb screws | 22:52 |
Wizzup | I don't think that's necessary :P | 22:52 |
Wizzup | I'm curious about the blob, I always thought qmi was some protocol / standard thing, but I guess not | 22:53 |
* uvos sad trombone | 22:53 | |
Wizzup | hmm | 22:53 |
Wizzup | wait | 22:53 |
Wizzup | root@edison:/ # dmesg|grep qmi | 22:53 |
Wizzup | <6>[ 0.259033,0] feature_qmi_over_spi=0 | 22:53 |
Wizzup | we should check the device_tree tomorrow | 22:54 |
uvos | or lsusb | 22:54 |
uvos | anyhow ill head out | 22:54 |
uvos | ttyl | 22:54 |
Wizzup | https://wizzup.org/dirlist/atrix2/dmesg-android-3.0.8.txt | 22:54 |
Wizzup | yeah lsusb says nothing but that could just mean the modem didn't power | 22:54 |
Wizzup | ok, ttyl | 22:54 |
uvos | on android ofc | 22:54 |
Wizzup | root@edison:/ # lsusb | 22:55 |
Wizzup | Bus 001 Device 001: ID 1d6b:0002 | 22:55 |
Wizzup | that's just some root hub I think | 22:55 |
Wizzup | <6>[ 113.538391,0] omap_hsi omap_hsi.0: Modem wakeup detected from HSI CAWAKE Pad port 1 | 23:01 |
Wizzup | hm | 23:01 |
Wizzup | isn't that what the n900 uses too? ssi/hsi? | 23:13 |
bencoh | it is indeed | 23:19 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!