| Wizzup | uvos: with zstd:15 and -m single -d single, we need ~1.2G: | 00:57 |
|---|---|---|
| Wizzup | $ df /mnt/leste-btrfs/ | 00:57 |
| Wizzup | Filesystem 1K-blocks Used Available Use% Mounted on | 00:57 |
| Wizzup | /dev/loop1 2097152 1105340 857172 57% /mnt/leste-btrfs | 00:57 |
| Wizzup | this is with a gazillion locales | 00:57 |
| Wizzup | cache is 942M on my mz617 so it can't fit there then unfortunately | 01:09 |
| Wizzup | one super mega hack would be to make two loops devices, each about 650MB, and then spread btrfs over these two | 01:10 |
| Wizzup | and use fastboot to flash to cache and preinstall | 01:10 |
| Wizzup | with zlib:9 it is the same story | 01:11 |
| Wizzup | /dev/loop1 2097152 1140420 821532 59% /mnt/leste-btrfs | 01:11 |
| Wizzup | in any case we can use /data I was just trying to find a way to do this without breaking android for the moment | 01:18 |
| Wizzup | $ df -h /mnt/leste-btrfs/ | 01:18 |
| Wizzup | Filesystem Size Used Avail Use% Mounted on | 01:18 |
| Wizzup | /dev/loop1 1.3G 1.1G 124M 90% /mnt/leste-btrfs | 01:18 |
| Wizzup | $ ls -lsh btrfs-?.img | 01:18 |
| Wizzup | 474M -rw-r--r-- 1 merlijn merlijn 650M Jan 16 01:18 btrfs-1.img | 01:18 |
| Wizzup | 637M -rw-r--r-- 1 merlijn merlijn 650M Jan 16 01:18 btrfs-2.img | 01:18 |
| Wizzup | well... | 01:18 |
| Wizzup | I think this is worth trying, but for this I will need utags to boot from cdrom | 01:26 |
| uvos | Wizzup: you can just generate utags for any partition you like via https://github.com/tmlind/utagboot | 10:20 |
| uvos | with mz61x you need to be extra carefull ofc since its doing its best to go brick on you | 10:21 |
| uvos | with all other deivces you can expirament as mutch as you want with this as fastboot will allways allow you to remove the tags | 10:22 |
| Wizzup | right | 10:36 |
| Wizzup | ty | 10:36 |
| Wizzup | uvos: android mounts /preinstall, do you think it needs it to function? | 10:37 |
| Wizzup | (I suppose trying on say a droid bionic would make some sense...) | 10:37 |
| uvos | at least on the phones its ok with it being missing at the expense of some functionalliy | 10:39 |
| uvos | not sure why you are not just using emstorage | 10:39 |
| uvos | using emstorage inconvieances android only minorly too and its mutch bigger | 10:40 |
| uvos | android will nag you with a notification about the partition being corrupt, otherwise it will be fine | 10:40 |
| uvos | hmm but i dont think fastboot will let you flash there | 10:40 |
| uvos | if you format emstorage to somthing the old kernel understands maybe android will not even nag you about it | 10:41 |
| uvos | on the bionic we nuke preinstall for clownboot | 10:42 |
| uvos | but thats not relevant as we dont really boot android userspace there | 10:42 |
| uvos | as we hijack in early init | 10:42 |
| Wizzup | uvos: what is emstorage? | 10:45 |
| Wizzup | these are all the partitions I can see that are large enoguh: | 10:45 |
| Wizzup | 179 17 614400 mmcblk1p17 system | 10:45 |
| Wizzup | 179 18 980992 mmcblk1p18 cache | 10:45 |
| Wizzup | 179 19 716800 mmcblk1p19 preinstall | 10:45 |
| Wizzup | 179 20 12959744 mmcblk1p20 userdata | 10:45 |
| Wizzup | uvos: maybe the 16G version is different | 10:53 |
| Wizzup | uvos: maybe I should start just trying to flash cache and preinstall with fastboot and see if that is allowed, if it is, I could always restore it, no? | 11:02 |
| Wizzup | uvos: also, the input for utagboot for the -16 devices, like the dtb name and serial, that can be found in android I assume? | 11:03 |
| Wizzup | we could also make a maemo-leste minimal image btw, with a different meta pkg | 11:08 |
| Wizzup | lol, literally all the mz617 tablets that I have (I got a bunch from ebay for cheap) as mz617-16 | 11:15 |
| Wizzup | so I do not have a mz617-32 | 11:15 |
| Wizzup | uvos: so I guess you mean data when you say emstorage | 11:23 |
| Wizzup | because I don't see another partition, also not on mz616-32 | 11:23 |
| uvos | hmm we dont have emstorage on mz616? really | 11:51 |
| uvos | emstorage is the "emulated sdcard" for android | 11:51 |
| uvos | ok well then its a bit of a pickle yeah | 11:52 |
| Wizzup | the sdcard is just the data partition | 11:52 |
| uvos | oh ok | 11:52 |
| Wizzup | mount binded or whatever | 11:52 |
| Wizzup | it was also exactly 12G at least | 11:52 |
| uvos | then the layout is different than on xtXXX | 11:52 |
| uvos | ok | 11:53 |
| uvos | if fastboot lets you flash something you can try anything you like | 11:53 |
| uvos | you cant break it | 11:53 |
| Wizzup | as long as I can flash something sensible back | 11:53 |
| Wizzup | do you have mz617-32? | 11:53 |
| uvos | Wizzup: the inputs for utagboot can be found in the stock dts | 11:53 |
| uvos | but you can also find them in the makefile | 11:53 |
| uvos | of the repo | 11:53 |
| Wizzup | I saw the makefile | 11:54 |
| uvos | Wizzup: no i have -16 | 11:54 |
| Wizzup | but it has no entry for mz617-16 | 11:54 |
| uvos | dosent matter | 11:54 |
| uvos | the only difference is the device name | 11:54 |
| uvos | and bootloader dosent care about that | 11:54 |
| uvos | at least it dosent on any other device | 11:54 |
| uvos | and im pretty sure i have 32 too | 11:54 |
| uvos | er 16 | 11:54 |
| uvos | and have -32 in utags | 11:54 |
| Wizzup | ok | 11:55 |
| uvos | you can guinny pig a 16gb mz60x first | 11:55 |
| uvos | if concerned | 11:55 |
| uvos | since on that one you can just erase utags again | 11:55 |
| Wizzup | I have more mz617's | 11:55 |
| Wizzup | oh | 11:55 |
| Wizzup | right | 11:55 |
| Wizzup | uvos: btw I also don't see modem on usb for mz616, so perhaps that is also spi modem then | 11:56 |
| uvos | its pretty likely | 11:56 |
| Wizzup | would explain also why I don't have audio there either | 11:56 |
| uvos | all gsm devices have this | 11:56 |
| uvos | it seams | 11:56 |
| Wizzup | dmegs just confused me with: | 11:56 |
| Wizzup | [ 38.438476] phy-mapphone-mdm6600 usb-phy@2: modem status: 4 awake | 11:56 |
| Wizzup | [ 38.450775] phy-mapphone-mdm6600 usb-phy@2: Powered up OK | 11:56 |
| Wizzup | but it's not in lsusb | 11:56 |
| Wizzup | and the atrix 2 does the same I think | 11:56 |
| Wizzup | so yeah, probably also spi | 11:56 |
| uvos | at least not haveing a modem on a tablet is less of a dealbraker to me | 11:57 |
| uvos | but audio needs fixing | 11:57 |
| Wizzup | yeah | 11:57 |
| Wizzup | well if we're lucky the spi net isn't too hard :) | 11:58 |
| uvos | not a big fan of tablets anyhow :P | 12:06 |
| uvos | mz617 is the first tablet i have ever owned | 12:06 |
| Wizzup | I just want to get maemo working on it | 12:06 |
| Wizzup | it's a nice experience tbh | 12:06 |
| Wizzup | but yeah the tablet concept is also mostly lost on me | 12:06 |
| uvos | relly i want to run debian with lxqt on it | 12:06 |
| uvos | and use the keyboard case | 12:06 |
| uvos | to make it a arm laptop really | 12:07 |
| Wizzup | that works too | 12:07 |
| Wizzup | I'm charging a mz609 | 12:07 |
| Wizzup | I might need some files for them if I don't have them yet, but I will try kexecboot there then | 12:07 |
| Wizzup | it is a 32G version at least | 12:07 |
| uvos | 32g version is pretty useless atm | 12:07 |
| Wizzup | oh? | 12:07 |
| uvos | you can install kexecboot | 12:07 |
| uvos | but mainline linux has no pannel driver | 12:08 |
| Wizzup | what do you mean useless? | 12:08 |
| Wizzup | wait, the 32 panel is different? | 12:08 |
| uvos | no | 12:08 |
| uvos | mz60x panel is different | 12:08 |
| Wizzup | ah, I thought that was working too | 12:08 |
| uvos | i dont think so no | 12:08 |
| Wizzup | ok, then I will probably just try on my mz617 here | 12:09 |
| uvos | i was reccomending a mz60x to expirament with utag behavior only | 12:09 |
| Wizzup | uvos: where did you suggest we put the kernel and dtb | 13:44 |
| Wizzup | on 'boot' ? | 13:45 |
| uvos | you cant change boot | 14:58 |
| uvos | its singed and we need the kernel there | 14:58 |
| uvos | for the final setup i was suggesting system | 14:59 |
| uvos | for a non breaking setup you can also place it on system i gues | 14:59 |
| uvos | just dont overwrite any android files | 14:59 |
| Wizzup | uvos: right | 15:32 |
| humanbeta | Suspend feature would be usefull this way: if offline mode and you choose Lock Screen and Keys then suspend. | 15:35 |
| humanbeta | Or even hilbernate? | 15:39 |
| uvos | is hilbernateing when you leave the device in ohio for a while | 15:55 |
| uvos | humanbeta: we cant suspend ike this because some application might be playing music, downloading files etc | 15:56 |
| uvos | humanbeta: android has apis to allow applications to tell the system when they are doing specific things to allow it to choose when to suspend, we dont | 15:56 |
| uvos | humanbeta: best we can do is provide a suspend button | 15:56 |
| humanbeta | uvos: Suspend button also sounds good. | 16:02 |
| Wizzup | uvos: ok, so I will make utags for cdrom, and then put kexecboot in cdrom and flash the utags using dd (presumably). Can I then still boot android from kexecboot? | 16:40 |
| uvos | yes | 16:41 |
| uvos | unless someting goes wrong | 16:41 |
| Wizzup | right | 16:41 |
| Wizzup | I think so, right? And then I could flash the leste image to cache & preinstall or perhaps just data | 16:41 |
| uvos | yes | 16:41 |
| Wizzup | and then in utagboot I change the root= to /dev/mmcblk1p13, which is cdrom | 16:43 |
| Wizzup | (from p18, which is cache) | 16:43 |
| uvos | if p13 is indeed cdrom, yes thatll work | 16:43 |
| Wizzup | 179 13 217088 mmcblk1p13cdrom | 16:44 |
| Wizzup | so the dtname and serialno I won't change | 16:44 |
| uvos | should be fine | 16:44 |
| Wizzup | cmdline say | 16:45 |
| Wizzup | s | 16:45 |
| Wizzup | androidboot.modelno=MZ617-16 | 16:45 |
| Wizzup | androidboot.serialno=TA31001QA2 | 16:45 |
| Wizzup | but you said it didn't matter to you | 16:45 |
| uvos | if you want to you can change it to suit | 16:45 |
| Wizzup | so I will leave them as they are | 16:45 |
| uvos | but it dosent matter on xtXXX for sure and i think its wrong on my mz617 too | 16:46 |
| Wizzup | is "wrong" means it is verified to work, I'll take it | 16:46 |
| Wizzup | let's see, once kexecboot works to boot to android, I should be careful on the next steps, because if I break android AND am not able to boot to leste, that could be a problem | 16:49 |
| Wizzup | maybe I should make a copy of cache and preinstall to /data, adb pull them, just in case I can somehow flash them with fastboot | 16:49 |
| Wizzup | (otherwise there is no way to say modify the boot lines) | 16:52 |
| Wizzup | uvos: utags is p6, yes? | 16:53 |
| uvos | i sorry i dont know the partion table of the tablets by heart | 16:54 |
| uvos | *im | 16:54 |
| Wizzup | the filename in droid4-kexecboot says so :) | 16:55 |
| Wizzup | an android has a boot.cfg somewhere on a partition, so kexecboot will know where to find it? | 17:00 |
| Wizzup | (just dd'd the two images so doing some double sanity checks to make sure I don't break the device :D) | 17:00 |
| uvos | kexecboot comes with a boot.cfg for stock android | 17:01 |
| Wizzup | ok | 17:01 |
| uvos | as long as kexecboot itself works you have not bricked the device | 17:01 |
| uvos | you can exit kexecboot and have a serial shell | 17:02 |
| uvos | as in exit the gui | 17:02 |
| Wizzup | ok, check | 17:03 |
| Wizzup | I saw kexecboot and I will soon see android again | 17:03 |
| Wizzup | maybe what makes sense is to dd a separate image to data and a split one to preinstall and cache, and make boot.cfg entries for both | 17:04 |
| Wizzup | although I suppose if I use data then I can't put boot.cfg on it | 17:04 |
| uvos | not if using btrfs | 17:04 |
| uvos | or 64bit ext4 | 17:04 |
| Wizzup | right, so maybe boot.cfg can go on system? | 17:05 |
| uvos | yes | 17:05 |
| uvos | system needs to stay ext3 anyhow as mentioned | 17:05 |
| Wizzup | ok | 17:05 |
| Wizzup | I suppose I could flash a regular leste rootfs to data as well (ext4) one, as a failsafe | 17:06 |
| Wizzup | only the root partition ofc) | 17:06 |
| uvos | i mean the faile safe is supposed to be the busybox enviroment supplied with kexecboot | 17:07 |
| Wizzup | one thing to check is to see if btrfs is built in in our kernels... | 17:07 |
| uvos | but as you like | 17:07 |
| Wizzup | fair enough, I should try to get into that environment | 17:07 |
| Wizzup | I am not sure how you would transfer files at that point | 17:07 |
| Wizzup | unless you'd use base64 and cat over serial | 17:07 |
| uvos | it has wpa_supplicant | 17:08 |
| Wizzup | I guess you have some kind of wifi | 17:08 |
| Wizzup | ok | 17:08 |
| uvos | you can connect to wifi with it | 17:08 |
| Wizzup | ok, I've just never been in that environment :) | 17:08 |
| Wizzup | should maybe try that first | 17:08 |
| Wizzup | lol | 17:08 |
| Wizzup | okay, so our default config does not even include btrfs | 17:09 |
| Wizzup | glad I checked :D | 17:09 |
| uvos | mainly you can also use to to zero out utags | 17:09 |
| uvos | to boot stock android in an otherwise big failure (only works if its unmolested ofc) | 17:09 |
| uvos | *use it | 17:09 |
| Wizzup | right | 17:10 |
| Wizzup | uvos: any idea if mmcblk1 still be called mmcblk1 even if there is no microsd on mz617? | 17:53 |
| Wizzup | I suppose it should, because the dts still allows for a microsd card on mz617 | 17:53 |
| uvos | motorola kindof hardcoded the partition names in android all over the place | 18:27 |
| uvos | so presumably they where not interested in changing all that | 18:28 |
| Wizzup | I got to: | 18:41 |
| Wizzup | [ 0.026733] printk: console [tty0] enabled | 18:41 |
| Wizzup | [ 0.030944] printk: bootconsole [omap8250] disabled | 18:41 |
| Wizzup | and then the output stops | 18:41 |
| Wizzup | oh, I see | 18:41 |
| Wizzup | the boot.cfg does not contain the console log | 18:41 |
| Wizzup | serial log I mean | 18:41 |
| Wizzup | omg i have such a hard time pressing the buttons and the kexecboot timeout is so low | 18:43 |
| Wizzup | uvos: looks like android happily mkfs'es cache and preinstall btw | 18:45 |
| Wizzup | I booted back to android and now both are empty ext4 fses, after I had dd'd btrfs images over it previously | 18:50 |
| bencoh | I'm not surprised about cache, but preinstall ... I wonder why ... what is it used for? | 19:04 |
| Wizzup | uvos: it looks like fastboot is letting me write to cache at least | 19:06 |
| Wizzup | same for preinstall | 19:10 |
| Wizzup | uvos: | 19:11 |
| Wizzup | [ 5.164337] VFS: Cannot open root device "/dev/mmcblk1p18" or unknown-block(259,10): error -2 | 19:11 |
| Wizzup | [ 5.164520] Please append a correct "root=" boot option; here are the available partitions: | 19:11 |
| Wizzup | will debug a bit later | 19:11 |
| arno11 | Wizzup: freemangordon: btw i found why i got +10mA idle draw: usb charger ! Idle is around 43-47 until i plug usb charger, then once unplugged, i get +10mA constantly | 19:35 |
| arno11 | if i reboot, back to normal | 19:36 |
| arno11 | is it n900 specific or someone get the same on d4 or pinephone ? | 19:39 |
| Wizzup | arno11: ok, that we can fix @ charger, good find | 19:50 |
| Wizzup | brb dinner | 19:50 |
| arno11 | me too, enjoy ;) | 19:53 |
| freemangordon | arno11: phew... I was afraid there is some issue with iphb(d) :) | 19:53 |
| uvos | "uvos: looks like android happily mkfs'es cache and preinstall btw" thats pretty anoying | 20:04 |
| uvos | Wizzup: it probubly wont do it if you format the partitions in a way that android understands | 20:05 |
| uvos | ie 32bit ext4 | 20:05 |
| uvos | "I'm not surprised about cache, but preinstall ... I wonder why ... what is it used for?" preinstall is used for bloatware 3rd party apps that the default android came with, except that motorola mercyfully removed most of those with updates later, so not mutch anymore. on devices with very smal system like bionic its also (mis)used for some extra resources moved there when motorola ran out of space in system | 20:07 |
| Wizzup | uvos: I'm starting to worry we might need an initramfs | 20:08 |
| Wizzup | but I'll keep trying for a bit more | 20:08 |
| arno11 | freemangordon: :) | 20:16 |
| Wizzup | uvos: so | 21:20 |
| Wizzup | CMDLINE=console=tty0 console=ttyS2,115200 fbcon=rotate:1 debug earlycon ro rootwait rootfstype=btrfs root=/dev/mmcblk1p18 rootflags=device=/dev/mmcblk1p18,device=/dev/mmcblk1p19 softlevel=charge-mode | 21:20 |
| Wizzup | this seems to be doing what the stack overflow post suggests | 21:21 |
| Wizzup | https://paste.villavu.com/raw/hNtTIotcHf96gvd95j0c6sWug805kxavwUUA5aGQ/ | 21:21 |
| Wizzup | oh, looks like I mispasted a bit, anyway, you get the idea | 21:23 |
| Wizzup | uvos: seems like userdata can't be written to using fastboot | 21:48 |
| Wizzup | uvos: I managed to exclude a few dirs and fit it all into a single btrfs image that I am dding to cache, if this works, then at least I have a working start, but perhaps it would make sense to make a small initramfs with busybox mdev and btrfs-progs | 22:13 |
| Wizzup | all we need to do is run a tool to 'scan' for btrfs devices | 22:13 |
| Wizzup | btrfs dev scan or something | 22:13 |
| Wizzup | why this can't be in the kernel is beyond me | 22:13 |
| Wizzup | I also found this, which can't be it I hope: | 22:14 |
| Wizzup | # Required because `btrfs device scan` doesn't appear | 22:14 |
| Wizzup | # to scan past 9 partitions! | 22:14 |
| Wizzup | uvos: ok this is booting at least :) | 22:20 |
| Wizzup | lol I had to update etc/fstab to prevent fsck checks | 22:21 |
| Wizzup | uvos: last q of the night I hope, the data partition I can format as btrfs without problems for the bootloader, yeah? | 22:24 |
| uvos | Wizzup: you can do anything you like with data | 22:36 |
| uvos | Wizzup: to my knowlage the only android related partitions the booloader checks are boot, the device tree and its backup and the first couple of blocks of system | 22:37 |
| Wizzup | ok | 22:47 |
| Wizzup | well, I managed to boot from just cache | 22:47 |
| Wizzup | if we make a simple initramfs as above then we could do the cache+preinstall trick | 22:48 |
| Wizzup | I have h-d working with wifi, touch, etc, just no locales atm :P | 22:48 |
| merom | humanbeta: How to take screenshot with Pinephone? > You can try taking a screenshot with an application simply called Screenshot. The images are then stored in /home/user/MyDocs/.images :-) | 22:49 |
| merom | Oh, humanbeta is not connected here now, if so, someone please let him know. Well thank you | 22:57 |
| uvos | Wizzup: the locales situation should be adressed anyhow | 23:00 |
| uvos | the n900 takeing absoulte ages to generate a tonne of locales is another sore spot | 23:01 |
| Wizzup | I think you can control this with locale.gen right? | 23:02 |
| Wizzup | just comment mostly everything | 23:02 |
| Wizzup | I'm rsyncing to userdata now | 23:03 |
| Wizzup | but for users we'll need a better procedure | 23:03 |
| Wizzup | the nice thing about a cache bootstrap image is that you can do it without harming android (I guess even userdata can be erased with fastboot) | 23:03 |
| Wizzup | I do think that making a hildon-minimal-meta would be good, where we define what is essential and what is not | 23:12 |
| Wizzup | for example, openmediaplayer and pdf reader are not, but icd2/mce/etc is | 23:12 |
| uvos | more like hildon-meta has to mutch bloat | 23:12 |
| Wizzup | this way we could probably quite easily fit a bootstrap image on the cache partition and then upgrade it with a simple apt command | 23:12 |
| uvos | and ther should be hildon-extra | 23:12 |
| Wizzup | same thing different name :) | 23:12 |
| Wizzup | (so we agre) | 23:14 |
| Wizzup | agree* even | 23:14 |
| Wizzup | interesting, on mz617 I also see the same ts working only on certain boots issue | 23:16 |
| Wizzup | after 4-5 rmmod + modprobe it works | 23:19 |
| Wizzup | tmlind: do you also see on mz61x that the ts only functions some of the time? for me repeated rmmod + modprobe of atmel_mxt_ts a few times tends to solve the problem | 23:31 |
| Wizzup | I also somehow see both a atmel_mxt_ts 1-004a and atmel_mxt_ts 1-005b being probed | 23:32 |
| Wizzup | the first one seems to always fail | 23:32 |
| Wizzup | what is very neat about mz617 is that the screen brightness chagne is gradual, like on fremantle/n900 | 23:48 |
| Wizzup | tmlind: cheers again for this work, this (mz617/mz617) is really sweet | 23:51 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!