Vall | Faling that, is there a good mirror I can download locally to try and achieve the same? | 00:10 |
---|---|---|
gnarface | arm-files.devuan.org | 00:37 |
gnarface | and sorry, i think the answer is no | 00:49 |
gnarface | i think these are all in the form of pre-installed disk images | 00:49 |
c0rnelius | tasksel I think is broken by default when pulled in from debootstrap. So you could sudo apt purge -y tasksel --autoremove; sudo apt install -y tasksel; sudo tasksel | 00:58 |
c0rnelius | That will give you a select of desktops and what not available to you for easy install. | 00:58 |
c0rnelius | Just know it will also drag in everything under the sun :) | 00:59 |
rrq | Vall: I added the arm64-pool branch to the devuan/installer-iso project, and used that to prepare a pool ISO with the 5000 "most popular" packages.. | 10:34 |
rrq | at https://charla.rrq.id.au/download/devuan/ | 10:35 |
Vall | rrq: many thanks! | 16:40 |
Vall | Checking it out now | 16:40 |
Vall | rrq: this would be the deb-containing image, right? devuan_chimaera_4.0_arm64_pool1.iso | 16:41 |
Vall | gnarface: no prob, and thanks for the ref to arm-files.devuan.org | 16:47 |
steev | c0rnelius: that radxa firmware - why do you pull in the ene firmware manually instead of firmware-nonfree? that firmware hasn't been touched in 10 years | 20:42 |
c0rnelius | steev: I haven't received the board yet. The firmware I've pulled together is preemptive and researched online. | 20:46 |
steev | ah, got ya | 20:47 |
steev | whcih one are you getting? | 20:47 |
c0rnelius | I do pull in firmware-nonfree though | 20:47 |
steev | i've got an 2G/8G and a 4G/128G | 20:47 |
c0rnelius | 4GB 16eMMC | 20:47 |
steev | c0rnelius: i was meaning https://github.com/pyavitz/debian-image-builder/commit/a1ae645fecf00b53ac25511315ea1fa5fdba2a0e#diff-86e3d650ad1353f5fee7c6066c889f3c7c40ba4c7c3b5aaf9b42bd98b7fc4979R234-R240 - those firmware are in firmware-nonfree and haven't been touched in 10 years (according to the path history of the URL you pull them from :) ) | 20:48 |
c0rnelius | steev: Does the board have any reboot/shutdown problems? | 20:48 |
steev | c0rnelius: sort of | 20:48 |
c0rnelius | I thought so | 20:48 |
steev | there seems to be some vampiric draw from my hdmi monitor | 20:48 |
steev | if hdmi isn't connected, it reboots/shuts down fine | 20:49 |
steev | if hdmi is connected, reboot just shuts it down | 20:49 |
c0rnelius | Oh ur referring to ene-ub6250 | 20:49 |
steev | yeah, sorry that wasn't more clear :) | 20:49 |
c0rnelius | Oh I just dropped that in. I'm not even sure if I need it yet. | 20:50 |
steev | got ya | 20:50 |
steev | is that what that revert is supposed to help? the shutdown | 20:51 |
steev | i'm currently just running their kernels (but my own dtb) | 20:51 |
c0rnelius | I'm hoping yes | 20:51 |
steev | i'd asked about it a few weeks ago in #linux-amlogic | 20:51 |
c0rnelius | Won't know till someone tests or I get the board | 20:51 |
steev | never got any response | 20:51 |
steev | i haven't tried a mainline kernel just yet | 20:52 |
c0rnelius | I'm not even sure if I got the boot fips right :) But it seems to build ok. | 20:52 |
c0rnelius | might actually need to only revert this -> https://github.com/pyavitz/debian-image-builder/blob/feature/patches/amlogic/radxazero/003-revert-meson-drv-shutdown.patch#L25 | 20:54 |
c0rnelius | On the odroids the full patch needs to be applied though. | 20:54 |
c0rnelius | and Oh yeah... When making Ubuntu imgs I don't pull in the Ubuntu firmware package as its something retarded like 600MB or more extracted. Thats why I drag that in.`ene-ub6250` I do the same for the other amlogic socs or else initramfs complains. | 21:07 |
steev | oh, yeah that's a good reason to do it | 21:08 |
steev | ubuntu's firmware package is all the firmware, i always forget about that | 21:09 |
steev | fwiw, i'm able to build u-boot jsut fine... but all i get is a magenta screen on hdmi | 21:09 |
steev | i haven't researched it further as was building up to our 2021.4 release | 21:09 |
steev | now that is complete on my end, i can focus on this, might do it over my vacation | 21:10 |
c0rnelius | The RC? | 21:10 |
steev | it could be that running the fip stuff through qemu is what is breaking it | 21:10 |
steev | our = kali | 21:10 |
steev | https://gitlab.com/kalilinux/build-scripts/kali-arm/-/merge_requests/297/diffs is my initial stuff | 21:11 |
c0rnelius | It looked to me that the application of the fips were exactly like that of say the C4. So I just napped the fips from chewitt and used the same score. | 21:11 |
steev | i haven't looked at what you do there, but is the c4 similar to the c2? | 21:11 |
c0rnelius | https://github.com/pyavitz/debian-image-builder/blob/feature/lib/function/uboot#L705 | 21:12 |
c0rnelius | When building it looks fine. I just of course can't test it. | 21:12 |
c0rnelius | yet | 21:12 |
steev | https://github.com/afaerber/meson-tools | 21:13 |
steev | got a download? i can slap it on my 2g/8g | 21:13 |
steev | see if it does anything | 21:13 |
c0rnelius | you mean pre-compiled? u-boot bins | 21:14 |
c0rnelius | or an actual whole img? | 21:14 |
steev | either/or | 21:15 |
steev | one other thing i need to track down at some point, is why sway, or something sway-ish keeps crashing on it | 21:15 |
c0rnelius | I can make an img real quick. | 21:16 |
steev | no rush | 21:17 |
c0rnelius | I made one yesterday just to make sure I didn't break the builder. Won't be long. | 21:18 |
c0rnelius | I'll just do a fresh one with the newer shit added. | 21:19 |
c0rnelius | all but that revert patch anyway | 21:19 |
steev | sounds good | 21:19 |
steev | i can try on both sdcard and emmc; it bothers me that the address might be different between them | 21:20 |
steev | i couldn't exactly tell | 21:20 |
steev | https://wiki.radxa.com/Zero/dev/u-boot says count=444, somewhere else completely skips it | 21:23 |
c0rnelius | whats the lsblk | 21:32 |
c0rnelius | I still don't know what number the eMMC sits on | 21:32 |
steev | on a system just booted from emmc it's mmcblk2 | 21:33 |
steev | https://bpa.st/QDVQ | 21:33 |
c0rnelius | thanx | 21:33 |
c0rnelius | are they using vfat for the boot partition? | 21:35 |
steev | sadly | 21:36 |
c0rnelius | either way I'll need to add the board to the write2mmc script. Good to know its on mmcblk2. | 21:36 |
steev | actually | 21:36 |
steev | /dev/mmcblk2p1 * 32768 1081343 1048576 512M ef EFI (FAT-12/16/32) | 21:36 |
steev | /dev/mmcblk2p2 1081344 241662047 240580704 114.7G 83 Linux | 21:36 |
c0rnelius | interesting | 21:37 |
steev | i thought so too | 21:37 |
steev | but it is just vfat | 21:37 |
c0rnelius | I looked over one the imgs and noticed it was missing when mounted so I figured EFI was at play there. | 21:39 |
c0rnelius | sdcard only. If it boots, ur login is: steev:board --> https://github.com/pyavitz/binary/releases/download/060420/g12a-radxa-zero-debian-bullseye-5.15.5-2021-11-30.img.xz | 21:40 |
steev | cool, will give it a spin | 21:41 |
c0rnelius | thanx | 21:41 |
steev | btw https://github.com/radxa/radxa-zero-images-released/releases some of those files are helpful | 21:41 |
steev | particularly the rz-blah.bin | 21:42 |
steev | fastboot, obviously, to put it into fastboot mode, and udisks makes it show up as, well, a usb drive | 21:42 |
c0rnelius | Added to the reading list, thanx. | 21:42 |
steev | that's for use with pyamlboot or whatever it was | 21:43 |
c0rnelius | I really know nothing about the board. Just that I ordered one and figured I would start work on getting something in place before I needed to do some trouble shooting. | 21:44 |
steev | i do that too | 21:44 |
steev | btw, the hdmi thing | 21:44 |
steev | it happens whether i have microhdmi plugged in to hdmi or dvi, so | 21:44 |
c0rnelius | Well considering the board isn't rreally supported yet in current stable mainline linux or uboot I'm sure its got tons of problems. | 21:45 |
c0rnelius | and looks like most of those efforts are the usual suspects. armstrong and chewitt. | 21:46 |
steev | 5.16 should have a dts | 21:56 |
steev | Or maybe it’s 5.17 | 21:56 |
steev | Is First Boot one of your services? | 21:56 |
steev | I can’t tell if it’s booting from sdcard or from the emmc. I stupidly just wrote to the emmc too | 21:57 |
steev | Not the image you gave, I meant I’d just written one of Radxa’s | 21:57 |
steev | I don’t have headers on mine, sadly, and I dunno where the fuck the headers I ordered off Amazon are. The post office seems to have lost them | 21:59 |
c0rnelius | Yes firstboot and credentials is a service | 22:01 |
steev | it seems to just be... running | 22:01 |
c0rnelius | firstboot won't write to it, but will check it to see if needs to be grown. | 22:02 |
c0rnelius | Well... We know it boots :) | 22:02 |
steev | indeed :D | 22:02 |
c0rnelius | Probs because of the fat partition on the MMC | 22:02 |
steev | https://usercontent.irccloud-cdn.com/file/DqYXuoir/1638306175.JPG | 22:02 |
steev | I wish there was a way to cancel | 22:03 |
c0rnelius | hmm | 22:03 |
c0rnelius | I bet its because of that fat partition. | 22:03 |
steev | Suppose I could mount the partition and remove the service | 22:04 |
c0rnelius | I guess I need to re-write that stupid script now | 22:04 |
c0rnelius | https://github.com/pyavitz/debian-image-builder/blob/feature/files/scripts/firstboot | 22:04 |
steev | I just use the growpart script from cloud-guest-utils (without pulling it in because of the god awful amount of dependencies) | 22:04 |
c0rnelius | because I added the pi4 I put in all these if and fi | 22:05 |
c0rnelius | Its probs confused | 22:05 |
c0rnelius | I didn't think anyone was using vfat anymore expect the foundation. :\ | 22:06 |
c0rnelius | except* | 22:06 |
c0rnelius | try hitting space or enter a bunch of times | 22:06 |
c0rnelius | might cacnle it | 22:06 |
c0rnelius | cancel* | 22:07 |
steev | https://usercontent.irccloud-cdn.com/file/NHj3ECiG/1638306466.JPG | 22:07 |
steev | I just powered it off, fscked, mounted and removed the first boot service | 22:08 |
c0rnelius | nice | 22:08 |
c0rnelius | I think if I just remove this -> https://github.com/pyavitz/debian-image-builder/blob/feature/files/scripts/firstboot#L30 | 22:08 |
c0rnelius | probs solved | 22:08 |
c0rnelius | its looking somewhere a pi wouldn't anyway | 22:09 |
c0rnelius | does wifi / bluetooth work? | 22:09 |
steev | Looks quite good, though, no nmcli | 22:09 |
c0rnelius | swh -h | 22:09 |
c0rnelius | give access to the wpa and interfaces file | 22:10 |
c0rnelius | might need to manually adjust /etc/default/crda though | 22:10 |
c0rnelius | Shit I'm just happy I got uboot/fips right :) You the man for testing. Thank you so much. The little stuff I can easily sort out. | 22:13 |
c0rnelius | Do you know if it clocks higher then 1800000? | 22:14 |
steev | Bluetooth not working (not blocked either) | 22:16 |
steev | I don’t know if it does | 22:16 |
steev | I mean I guess you technically could | 22:16 |
steev | I don’t know how stable it would be | 22:16 |
steev | I run schedutil on my personal one, mostly so that it idles at 100mhz instead of the default | 22:17 |
steev | Also 5.15 is very unhappy with my ssd :( | 22:18 |
steev | But that’s likely because I’m currently powering it off a MacBook USB port | 22:18 |
steev | Yeah that was it | 22:20 |
steev | oh | 22:22 |
steev | booted from your sdcard, the sdcard is mmcblk1, and the emmc is mmcblk0 | 22:22 |
c0rnelius | Alright the good info | 22:23 |
c0rnelius | then/thats* | 22:23 |
c0rnelius | They are probs using some vendor uboot | 22:24 |
steev | https://paste.debian.net/1221422/ | 22:24 |
steev | dmesg output | 22:24 |
c0rnelius | CMA memory pool looks good | 22:24 |
steev | BCM4345C5.hcd | 22:26 |
steev | it looks for the blueooth firmware to be named that | 22:26 |
c0rnelius | yeah I was just looking at that | 22:27 |
c0rnelius | i have it symlinked -> https://github.com/pyavitz/firmware/tree/radxa/ap6256 | 22:28 |
c0rnelius | maybe its broken | 22:28 |
c0rnelius | or I should symlink it into brcm | 22:28 |
c0rnelius | Doesn't look to bad for a first run. | 22:30 |
steev | 5.15 doesn't look in /lib/firmware/ap6256 | 22:30 |
steev | it looks in /lib/firmware/brcm | 22:31 |
c0rnelius | Gotcha | 22:31 |
steev | let me reboot it | 22:31 |
steev | oh | 22:32 |
steev | yours seems to reboot properly | 22:32 |
c0rnelius | Well I wasn't sure if I was gonna need to patchram the thing like with NEO plus2 | 22:32 |
c0rnelius | So I just did whatever. Reboots huh? | 22:32 |
steev | reboots properly yeah | 22:32 |
c0rnelius | nice | 22:32 |
steev | created the symlink under brcm and bluetooth works | 22:33 |
steev | [bluetooth]# list | 22:33 |
steev | Controller D4:9C:DD:A1:64:71 rzero [default] | 22:33 |
c0rnelius | nicee.... | 22:33 |
steev | now i just need to look at what you're doing and do the same :P | 22:34 |
c0rnelius | Ok so... adjust firmware, correct first boot and apparently mainline MMC#s are completely different. | 22:35 |
c0rnelius | I'm willing to bet if you never had vendor shit on the mmc firstboot would have been fine. | 22:36 |
c0rnelius | But I should and probs remove that anyway. | 22:36 |
steev | that's possible too yeah | 22:37 |
c0rnelius | I really appreciate you testing this. Thank you. | 22:38 |
steev | no problem | 22:38 |
steev | it helps me too :D | 22:38 |
c0rnelius | ;) | 22:38 |
steev | it lets me know where i did dumb shit :D | 22:38 |
steev | protip : everywehre | 22:38 |
c0rnelius | I'm just winging it, bud. | 22:41 |
steev | me too | 22:41 |
steev | boot.cmd factory-loader.img fip mainline-u-boot rz-fastboot-loader.bin rz-udisk-loader.bin u-boot update_extlinux.sh update_uenv.sh zz-uncompress zz-update-extlinux zz-update-uenv | 22:41 |
c0rnelius | Tenkawa was like I order the board! So I was like, fuck it I'll order one too and put that together in a few hours. | 22:42 |
steev | ignoring factory-loader, and uboot stuffs | 22:42 |
steev | thosse were the files i found interesting in their debian image | 22:42 |
steev | update_extlinux seems odd as their image does *not* use it | 22:42 |
steev | it uses the boot.scr and uEnv.txt file | 22:42 |
c0rnelius | Yeah I saw all that stuff on their imgs | 22:42 |
c0rnelius | and didn't get why it was there | 22:42 |
steev | based on having worked at a vendor..... | 22:43 |
steev | they didn't know better | 22:43 |
c0rnelius | Thats why they drop donations over at Armbian and Libre/CoreELEC | 22:43 |
steev | i mean, the boot.scr is | 22:43 |
steev | basically from the c4 | 22:43 |
steev | # odroid c4 legacy kernel values from boot.ini | 22:43 |
c0rnelius | yeah but boot.scr is so old school. | 22:43 |
steev | is a comment in the boot.cmd | 22:43 |
steev | yeah i know | 22:44 |
steev | old habits die hard for some folk | 22:44 |
c0rnelius | I disliked it before it was popular not to like it :) | 22:44 |
c0rnelius | stupid file | 22:44 |
c0rnelius | Lets drop a binary file in here for some stupid reason? How about we don't. | 22:45 |
steev | oh i absolutely hated it too | 22:45 |
steev | but distro_bootcmd or whatever the argument is, didn't exist way back when | 22:46 |
c0rnelius | true | 22:46 |
c0rnelius | Also not a fan of the boot.ini file either. But I let that one slide for now. | 22:46 |
steev | to be fair i am talking about when i was a vendor dev... back in 2010 or so | 22:46 |
steev | oh! | 22:47 |
steev | that reminds me of something i want to test on yours | 22:47 |
steev | audio is broken downstream | 22:47 |
steev | just need to swap the dvi for the hdmi cable, but i'm doing stuff on the other one at the moment so i can't | 22:47 |
c0rnelius | Gotcha. I never dealt in vendor stuff. I just hobby my way through it all. | 22:48 |
c0rnelius | Kool. Let me know, | 22:48 |
steev | i turned the vendor down many many times because it was my hobby until they literally asked "what will it take for you to not say no" for the job | 22:48 |
c0rnelius | Well... If they offered me real money I was would probs take the job. | 22:53 |
c0rnelius | I would still suggest we move forward. | 22:53 |
c0rnelius | I think a lot of this vendor shit is because they wanna support android which in my opinion is fine and all, but thats not where the rest of the production should be stuck at. | 22:54 |
c0rnelius | We all know Android is in its own little fantasy land. | 22:55 |
steev | for big vendors, kinda yeah. for little vendors, it's more about not having the resources to both release a *fully working* device, and chase mainline | 22:55 |
steev | like, if the mainline devs nak a patch that you need for your hardware to work... you're screwed | 22:56 |
c0rnelius | Well maybe thats why they should hire us to chase the dragon :D | 22:56 |
steev | which, is what happened to us at genesi with the efikamx | 22:56 |
steev | we needed some quirk added to the usb driver, and they refused to because they interpreted the usb spec differently and refused to add it | 22:56 |
c0rnelius | Well thats just dumb | 22:57 |
c0rnelius | and not forward thinking | 22:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!