| minute | vkoskiv: i wanted to ask you how up-to-date your debian packages were on the rootfs | 00:00 |
|---|---|---|
| minute | that's why i wrote to apt update and apt upgrade earlier | 00:00 |
| josch | only reform-tools 1.29 and later contain the a311d bootargs | 00:00 |
| jjbliss | I thought I had upgraded everything through apt | 00:01 |
| minute | and this is also why i want a good migration wizard tool... because not even experienced users can remember all the steps and checks that are needed | 00:01 |
| josch | at the absolute minimum, there should be a documented list of steps -- it's just too much to keep in mind for something one does only once or twice :) | 00:01 |
| minute | josch: yep | 00:01 |
| minute | jjbliss: could you re-image the sd card and mount that nvme rootfs again and check if /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv is there? | 00:02 |
| vkoskiv | minute: I ran a update+upgrade right before going about this | 00:02 |
| jjbliss | minute: I'll do that | 00:02 |
| josch | jjbliss: maybe you did but then something else must've gone wrong. The lines "# setenv kernel_addr_r" and around that are not part of reform-tools anymore and indicate an older version | 00:03 |
| minute | josch: also, you were right about the need to make the uboots set all the required bootargs | 00:03 |
| josch | minute: because of the btrfs reated forum post you mean? | 00:03 |
| vkoskiv | Seriously? I pull out a random ugly 2013 Asus and it's running WINDOWS XP? | 00:03 |
| vkoskiv | Bruh | 00:03 |
| minute | yeah but also now... if uboot-a311d would have set all the nvme related args, this wouldn't have happened also | 00:04 |
| minute | s/would have/had | 00:04 |
| vkoskiv | Okay, an old thinkpad. This should have something reasonable on there | 00:04 |
| josch | indeed. i think reform-a311d-uboot does set the special cmdline: https://source.mnt.re/reform/reform-a311d-uboot/-/blob/main/bananapi-cm4-mnt-reform2_defconfig?ref_type=heads#L86 | 00:05 |
| minute | ah, and jjbliss' boot.scr does: setenv bootargs "ro no_console_suspend cma=512M pci=nomsi ${bootargs} console=ttymxc0,115200 console=tty1" | 00:05 |
| vkoskiv | Thinkpad runs Arch, but has no SD slot | 00:05 |
| minute | maybe the pci=nomsi is the culprit here though | 00:06 |
| minute | vkoskiv: usb sd card reader? | 00:07 |
| josch | minute: but pci=nomsi would be overwritten by a311d-specific pci=pcie_bus_perf so should not be an issue as later values override earlier ones, no? | 00:07 |
| vkoskiv | Nah, I'll pull out the boring option, a MacBook with a USB-C dongle | 00:07 |
| jjbliss | Ok, I'm booting up a freshly imaged sdcard | 00:08 |
| minute | josch: i don't think that's necessarily how it works... at least some options can be specified multiple times | 00:08 |
| josch | could be. i'm not sure. | 00:08 |
| minute | vkoskiv: not sure if we migrated the boot partition from fat to ext2 though which the mac won't handle | 00:09 |
| josch | anyways, i have to head off to bed now | 00:09 |
| josch | good luck to you two! | 00:09 |
| minute | josch: n8n8! | 00:09 |
| josch | i'll read the backlog tomorrow :) | 00:09 |
| josch | good night :) | 00:09 |
| vkoskiv | Night! | 00:10 |
| vkoskiv | Looks like macOS doesn't understand linux volumes | 00:10 |
| vkoskiv | I think my desktop has USB-C somewhere... | 00:10 |
| jjbliss | josch: Thanks and goodnight! | 00:10 |
| vkoskiv | Okay, hooked up now. | 00:11 |
| jjbliss | so, /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv is there on my nvme | 00:11 |
| jjbliss | it's looking for u200 though | 00:12 |
| minute | jjbliss: that's fine, also this whole code is expected to show up in boot.scr | 00:12 |
| minute | jjbliss: i.e. it should be included in there, but it's not | 00:12 |
| vkoskiv | Erm, bunch of I/O errors on this SD card | 00:12 |
| jjbliss | ah | 00:12 |
| vkoskiv | Willing to bet it's the crappy USB-C dongle though | 00:12 |
| vkoskiv | Yeah, linux hid the volumes on there after trying to mount, it's not happy | 00:13 |
| vkoskiv | Oof owie, how can this be so hard | 00:13 |
| minute | :/ | 00:13 |
| vkoskiv | Okay, now it mounted | 00:14 |
| vkoskiv | I just unplugged the card and put it back in | 00:14 |
| vkoskiv | /boot/boot.scr, yeah? | 00:14 |
| minute | vkoskiv: well, boot.scr from the first partition | 00:14 |
| vkoskiv | https://pastebin.com/n0z9JtAf | 00:15 |
| vkoskiv | I think this is it here | 00:15 |
| minute | yep, same problem, the code is missing | 00:15 |
| vkoskiv | Can I append it in here while it's mounted? | 00:15 |
| minute | vkoskiv: well, kind of. unfortunately boot.scrs have some binary junk at the beginning that must be correct | 00:16 |
| minute | it's like a checksum or something | 00:16 |
| vkoskiv | Ah, I found instructions on ubuntuwiki | 00:17 |
| minute | but one can replace it with the boot.scr from the original image | 00:17 |
| vkoskiv | Strip the header, then append it back at the end? | 00:17 |
| minute | no, | 00:17 |
| minute | ok you have two options: | 00:17 |
| minute | 1. copy to boot.scr.txt, remove the header, alter the file, use a tool (mkimage) to bake it into boot.scr again | 00:17 |
| minute | 2. extract the known good boot.scr from the downloaded system image | 00:18 |
| vkoskiv | I'll go with 2 | 00:18 |
| minute | ok. i'm downloading the image too on my laptop to see how to extract it | 00:19 |
| vkoskiv | gunzipping here | 00:19 |
| vkoskiv | I think a loop device mount will work | 00:20 |
| minute | stackoverflow says > sudo losetup -Pf disk_image.raw | 00:20 |
| vkoskiv | Got it! | 00:21 |
| vkoskiv | sudo mount -o loop,offset=4194304 reform-system-a311d.img yee/ | 00:21 |
| minute | nice! | 00:21 |
| vkoskiv | Then just copy this boot.scr to the actual sd /boot? | 00:22 |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dca:1fe3:e61a:30b8:bd10:dac3) | 00:22 | |
| minute | vkoskiv: yes | 00:22 |
| + stites (~stites@130.44.147.204) | 00:22 | |
| minute | vkoskiv: keep it open for later. | 00:22 |
| vkoskiv | Ah, vimdiff does indeed show the missing stuff | 00:22 |
| minute | vkoskiv: the mount from the disk image file i mean | 00:23 |
| vkoskiv | Sure | 00:23 |
| minute | so flash-kernel messes up somehow... i wonder why/how | 00:23 |
| vkoskiv | now umount the sd card and try boot? | 00:23 |
| minute | vkoskiv: yep! | 00:23 |
| jjbliss | yup, the original on my reflashed image has the stuff in there. | 00:23 |
| minute | jjbliss: in your case it won't boot from the nvme now though because you also restored the default initramfs, which in turn contains the fstab (and maybe crypttab) that tell it to mount the nvme | 00:24 |
| vkoskiv | heyy! | 00:24 |
| vkoskiv | different things are happening on boot | 00:24 |
| minute | vkoskiv: aha? | 00:24 |
| jjbliss | My nvme is unencrypted, so I don't have to worry about that at least | 00:25 |
| vkoskiv | Detected keyboard and mouse, now stuck again | 00:25 |
| vkoskiv | Let's see if it's the nvme error, waiting | 00:25 |
| jjbliss | flash-kernel is what creates the boot.scr, right? | 00:25 |
| minute | jjbliss: as far as i know, yes | 00:25 |
| minute | maybe vagrantc can chime in here | 00:26 |
| vkoskiv | So the new text is mostly USB logs, detecting those devices. Didn't have that before, I think | 00:26 |
| vkoskiv | But it's def. stuck doing something now | 00:26 |
| jjbliss | I can run update-initramfs -u from my chroot, but leave the boot.scr alone. | 00:26 |
| vkoskiv | at +8.23s from boot | 00:26 |
| minute | vkoskiv: sounds a bit like https://community.mnt.re/t/reform-won-t-finish-booting/1675/2 | 00:26 |
| minute | jjbliss: back the boot.scr up first | 00:26 |
| minute | jjbliss: i think update-initramfs -u will also run flash-kernel | 00:27 |
| jjbliss | ah | 00:27 |
| jjbliss | I've created a boot.scr.original for reference | 00:27 |
| vkoskiv | minute: yeah, same exact spot is where this is stuck | 00:27 |
| minute | good! | 00:27 |
| vkoskiv | except mine got there a bit quicker :D | 00:27 |
| vkoskiv | 8.3s vs 9.3 | 00:27 |
| vkoskiv | Benchmarks! | 00:27 |
| minute | vkoskiv: ok, so maybe there's something messed up in our latest initramfs. which is not good TM | 00:28 |
| minute | vkoskiv: person in forum suggests ctrl+alt+f2. does that do anything? | 00:28 |
| minute | (i kinda doubt it) | 00:28 |
| vkoskiv | Nope, nothing happens | 00:28 |
| sevan | is the partition mark active? | 00:29 |
| jjbliss | okay, I've run update-initramfs and reverted the boot.scr | 00:29 |
| vkoskiv | sevan: On mine? How do I check? | 00:30 |
| jjbliss | anything else I should do before unmounting everything and restarting? | 00:30 |
| minute | jjbliss: i think not | 00:30 |
| jjbliss | I'll give it a try then | 00:30 |
| sevan | vkoskiv: does fdisk indicate? | 00:31 |
| minute | vkoskiv: before i've debugged this in depth (like, tomorrow) i fear the only reasonable thing is to revert the initramfs (or the whole sd card) as well | 00:31 |
| vkoskiv | I can't invoke fdisk, no shell | 00:31 |
| minute | sevan: isn't that an old MBR thing? is that still a thing? | 00:32 |
| vkoskiv | Sure, so just flash the SD card from a downloaded image? | 00:32 |
| minute | normally initramfs should just mount / as per what is in its /etc/fstab | 00:32 |
| jjbliss | It's booted into my nvme system! | 00:32 |
| minute | jjbliss: oh nice! | 00:32 |
| jjbliss | Everything looks to be working | 00:32 |
| vkoskiv | I'm fine continuing debugging tomorrow, it's been an enjoyable night of hardware goodness already | 00:32 |
| minute | vkoskiv: maybe the fstab is not correct | 00:32 |
| sevan | minute: absolutely ancient DOSism. lives on in the arm world for steering where u-boot is. | 00:33 |
| minute | if it works for jjbliss but not for you, that is probable | 00:33 |
| vkoskiv | I didn't change the fstab, I observed it and thought it looked reasonable | 00:33 |
| vkoskiv | But I can pull it up if you want | 00:33 |
| minute | vkoskiv: yes, can you show it to me? it's still possible it needs a different device on a311d | 00:33 |
| jjbliss | vkoskiv: you are switching from emmc to sdcard for boot right? | 00:33 |
| sevan | vkoskiv: perhaps ignore my suggestion. | 00:33 |
| vkoskiv | It's on the nvme, right? I was using emmc to boot, now SD | 00:33 |
| minute | vkoskiv: yes. relevant is from where you ran update-initramfs | 00:34 |
| vkoskiv | inside chroot | 00:34 |
| minute | vkoskiv: that will look at /etc/fstab and copy it in the initramfs | 00:34 |
| minute | vkoskiv: so if /etc inside your chroot was on your nvme, it'll be that | 00:34 |
| vkoskiv | Okay, so I'll flash up this SD card, boot again from that, mount luks and look at the fstab | 00:34 |
| jjbliss | my nvme isn't using encryption, so a good chance your issue is related to that in some way | 00:35 |
| minute | ah yes, so also check /etc/crypttab vkoskiv | 00:35 |
| vkoskiv | Okay, just a minute | 00:35 |
| + jacobk (~quassel@utdpat242016.utdallas.edu) | 00:36 | |
| minute | vkoskiv: you can also unpack the initrd on that (problematic) sd card and look at what is in its /etc/crypttab and /etc/fstab | 00:36 |
| jjbliss | thanks for the help everyone! | 00:36 |
| minute | jjbliss: you're welcome, i'm glad it works for you | 00:36 |
| minute | vkoskiv: it is also possible that it wanted you to unlock the nvme on another console, for example on serial | 00:36 |
| vkoskiv | We will soon find out, is that configured in crypttab or something? | 00:37 |
| minute | no, not sure how that is selected tbh | 00:37 |
| vkoskiv | Flashing SD | 00:37 |
| minute | but for me everything is on the display console | 00:37 |
| minute | ok, i need to go to bed soon... sorry we didn't find an instant solution vkoskiv, but at least some good insights and the hw seems to be working per se | 00:39 |
| vkoskiv | Yeah, that's fine, we'll continue tomorrow | 00:40 |
| vkoskiv | I'll just grab the fstab and head to bed too then | 00:40 |
| minute | and crypttab | 00:40 |
| minute | n8n8 | 00:40 |
| vkoskiv | Yeah | 00:40 |
| vkoskiv | Night! | 00:40 |
| sevan | minute: it was still a thing in 2022? slide 11 & 12 https://bootlin.com/pub/conferences/2022/elce/opdenacker-implementing-A-B-system-updates-with-u-boot/opdenacker-implementing-A-B-system-updates-with-u-boot.pdf | 00:41 |
| sevan | perhaps invalid in the context of Reform? | 00:42 |
| vkoskiv | fstab says /dev/reformvg/root / auto errors=remount-ro 0 1 | 00:43 |
| vkoskiv | and | 00:43 |
| vkoskiv | /dev/mmcblk0p1 /boot auto errors=remount-ro 0 1 | 00:43 |
| vkoskiv | crypttab has 'reform_crypt UUID=<long UUID> none luks,discard' | 00:43 |
| vkoskiv | So I have lvm on luks | 00:44 |
| vkoskiv | Oh, the fstab also has a UUID line at the end: 'UUID=<long uuid> none swap sw 0 0 | 00:44 |
| vkoskiv | ' | 00:45 |
| sevan | minute: n/m, gnite :) | 00:47 |
| vkoskiv | past bedtime here too, night! | 00:49 |
| sevan | gnite. | 00:52 |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@utdpat242016.utdallas.edu) | 00:56 | |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 01:04 | |
| minute | several people have issue booting with encrypted nvme https://community.mnt.re/t/reform-won-t-finish-booting/1675/5 cc josch | 01:30 |
| minute | but need to sleep now, debugging tomorrow | 01:30 |
| abortretryfail | goodnight | 01:30 |
| minute | n8! | 01:31 |
| - stites (QUIT: Ping timeout: 255 seconds) (~stites@130.44.147.204) | 01:35 | |
| + stites (~stites@2607:fb91:dca:1fe3:e61a:30b8:bd10:dac3) | 01:35 | |
| - jacobk (QUIT: Ping timeout: 272 seconds) (~quassel@utdpat241106.utdallas.edu) | 01:55 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dca:1fe3:e61a:30b8:bd10:dac3) | 01:58 | |
| + stites (~stites@130.44.147.204) | 01:58 | |
| - mtm (QUIT: Ping timeout: 258 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:03 | |
| - S0rin (QUIT: Ping timeout: 255 seconds) (~S0rin@user/s0rin) | 02:06 | |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 02:10 | |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 02:12 | |
| - sevan (QUIT: Ping timeout: 260 seconds) (~sevan@user/venture37) | 02:12 | |
| - Gooberpatrol66 (QUIT: Ping timeout: 260 seconds) (~Gooberpat@user/gooberpatrol66) | 02:13 | |
| - xktr (QUIT: Ping timeout: 260 seconds) (~xktr@2602:fe3d:c01:10ca:1050:1ace:0:b) | 02:13 | |
| + xktr (~xktr@2602:fe3d:c01:10ca:1050:1ace:0:b) | 02:14 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 02:16 | |
| + sevan (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 02:18 | |
| + S0rin (~S0rin@user/s0rin) | 02:19 | |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@utdpat241106.utdallas.edu) | 02:22 | |
| - sevan (QUIT: Changing host) (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 02:23 | |
| + sevan (~sevan@user/venture37) | 02:23 | |
| jjbliss | Figured out the issue with boot.scr I had. the old 00reform2_ubootenv still exist in /etc which takes precedence over the new ones in /usr/share | 02:26 |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 02:28 | |
| jjbliss | I think that was it anyways | 02:28 |
| - ajr (QUIT: Quit: Connection closed for inactivity) (uid609314@user/ajr) | 02:28 | |
| - nsc (QUIT: Ping timeout: 240 seconds) (~nicolas@128-48-142-46.pool.kielnet.net) | 03:09 | |
| + nsc (~nicolas@63-97-142-46.pool.kielnet.net) | 03:11 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:10 | |
| josch | jjbliss: that is a good point! This file needs to be removed before moving over to a different SoM that needs some special boot options. | 07:23 |
| + jacobk (~quassel@64.189.201.150) | 07:24 | |
| josch | let me also document that this is how one can overwrite /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv if one really wants to | 07:29 |
| josch | minute: i am on latest linux kernel and reform-tools from the mnt repo and it boots just fine for me with encrypted nvme | 08:38 |
| josch | (that's the good news) | 08:38 |
| josch | the bad news: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1052547 | 08:43 |
| vkoskiv | Oh, is it a debian issue then? :O | 08:49 |
| Boostisbetter | You know if the rpi5 does a compute version it seems like it would be a perfect candidate for the Reform. Rtc means it can do sleep now and it has an expose pcie lane for the nvme as well. | 08:54 |
| Boostisbetter | *exposed | 08:55 |
| josch | Boostisbetter: but it does need active cooling if you want its full performance :) | 09:06 |
| josch | vkoskiv: you ran into this, right? | 09:07 |
| josch | vkoskiv: when you had upgraded all your packages, did you reboot? | 09:07 |
| josch | or did you just power off your system and started the migration? | 09:07 |
| Boostisbetter | josch: indeed! It will run without it but big throttling then. | 09:07 |
| + mjw (~mjw@gnu.wildebeest.org) | 09:13 | |
| vkoskiv | josch: I might have just shut down after upgrading | 09:40 |
| vkoskiv | In fact, yes, I did just shut down. | 09:40 |
| - klardotsh (QUIT: Ping timeout: 255 seconds) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 09:40 | |
| josch | okay, bummer | 09:43 |
| josch | if you had rebooted, the problem would've shown on imx | 09:43 |
| josch | and then we'd know for sure that your problem is not related to your SoM upgrade | 09:43 |
| vkoskiv | Yeah. I'm not worried about issues with my particular nvme, since it works perfectly when booted from the fresh SD card and mounting it manually | 09:44 |
| vkoskiv | The random old Intel SSD I had before this was doing it randomly during normal use, just froze for a moment | 09:45 |
| josch | vkoskiv: but just as i've said in https://community.mnt.re/t/reform-won-t-finish-booting/1675/9 it would be useful to have the last lines of /var/log/apt/history.log to see which packages you upgraded that broke your system | 09:45 |
| josch | i'm now afk for a few hours | 09:46 |
| vkoskiv | I'll grab that info soon | 09:46 |
| vkoskiv | Same, work stuff | 09:46 |
| - mjw (QUIT: Ping timeout: 255 seconds) (~mjw@gnu.wildebeest.org) | 10:03 | |
| * mark_ -> mjw | 10:04 | |
| minute | ok so after more than a month of delay and me asking and begging, i got an answer from nxp via the distributor that they can't release the ls1028a chips because i didn't print the "EUS" form on paper containing our letterhead | 11:00 |
| vkoskiv | -_- | 11:27 |
| vkoskiv | That's so lame | 11:27 |
| vkoskiv | Is that some sort of legal requirement then? | 11:28 |
| Boostisbetter | minute: is the pocket able to use a nvme and a cellular m2 based card at the same time? | 11:33 |
| - iank (QUIT: Quit: ZNC 1.8.2+deb2+b1 - https://znc.in) (~iank@fsf/staff/iank) | 12:00 | |
| + iank (~iank@fsf/staff/iank) | 12:01 | |
| minute | Boostisbetter: sure | 12:05 |
| minute | apropos https://mastodon.social/@mntmn/111171996119037370 | 12:05 |
| Boostisbetter | Ok, for some reason I was thinking the newer Imx8 only had a single PCIe lane available. | 12:05 |
| Boostisbetter | so the additional m2 slot would not be functional as on the Reform itself, but I guess the cellular modem isn't using the pcie lane? | 12:06 |
| minute | Boostisbetter: modem is usb, not pcie | 12:07 |
| Boostisbetter | minute: wunderbar MB | 12:08 |
| Boostisbetter | *wunderbar! | 12:09 |
| josch | minute: wait, so the chips you bought for the remaining ls1028a customers will not be delivered?? | 12:27 |
| josch | ah i now read the answer to my question on mastodon... uff... | 12:31 |
| minute | josch: well, i can send that document again | 12:44 |
| minute | it's just very painful delays | 12:45 |
| minute | i need to make paper with a letterhead.... can't believe this is 2023 | 12:45 |
| sknebel | that still seems to be the "answer" to "how can we verify companies globally" in many cases :/ | 12:48 |
| sknebel | (whatever exactly it is supposed to "verify") | 12:48 |
| josch | minute: maybe you can send them a fax instead ;) | 12:54 |
| josch | i'm currently writing out the rescue image to two sd-cards and then i'll try upgrading cryptsetup-initramfs and see if i get the same error | 12:55 |
| minute | josch: unfortunately i don't have any direct contacts | 12:56 |
| minute | josch: i'll do a similar thing, i'll sidegrade my personal reform with a311d now to the official /boot stuf | 12:56 |
| minute | stuff | 12:56 |
| josch | minute: it doesn not help that neither you nor me are running a vanilla setup XD | 12:56 |
| minute | yep, shame, shame (on me) | 12:57 |
| josch | i've added some additional things to the reform-check utility in my reform-tools MR | 12:57 |
| minute | cool | 12:57 |
| josch | i discovered some more ancient leftovers of sysimage-v3 on my system that can be removed now | 12:57 |
| josch | ideally reform-check should not output anything on sysimage-v4 but i don't think that's the case right now | 12:58 |
| minute | > 2539 packages can be upgraded | 13:01 |
| minute | guess i didn't update in a while | 13:01 |
| josch | it's customary that after a stable release happened *everybody* uploads major packaging updates for their packages again :) | 13:02 |
| - stites (QUIT: Ping timeout: 255 seconds) (~stites@130.44.147.204) | 13:16 | |
| + stites (~stites@2607:fb91:dcf:5280:3352:40b4:9f34:deb7) | 13:16 | |
| - iank (QUIT: Quit: ZNC 1.8.2+deb2+b1 - https://znc.in) (~iank@fsf/staff/iank) | 13:25 | |
| + iank (~iank@fsf/staff/iank) | 13:27 | |
| josch | huh weird | 13:37 |
| josch | i just upgraded cryptsetup-initramfs to the latest version from unstable (2:2.6.1-5) but i still booted successfully from nvme | 13:38 |
| josch | so it must be another package that introduced the breakage | 13:38 |
| josch | this makes sense because 2:2.6.1-5 did not change anything that should the boot process fail (the change was very minimal) | 13:38 |
| josch | minute: are you able to reproduce it? | 13:38 |
| vkoskiv | I'll try to grab my apt log soon | 13:50 |
| + mark_ (~mjw@gnu.wildebeest.org) | 13:54 | |
| josch | yes, if everybody who is affected by this problem could share the last bits of their apt log it should be easy to figure out the offending package | 13:59 |
| - mtm (QUIT: Ping timeout: 264 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:02 | |
| - S0rin (QUIT: Read error: Connection reset by peer) (~S0rin@user/s0rin) | 14:05 | |
| + S0rin (~S0rin@user/s0rin) | 14:22 | |
| minute | josch: apparently i have unmerged /bin and /usr/bin :D | 14:23 |
| minute | (apt complains) | 14:24 |
| minute | now running update-initramfs again | 14:24 |
| minute | ok my boot.scr was generated wrongly | 14:24 |
| minute | deleting /etc/flash-kernel/ubootenv.d/00reform2_ubootenv | 14:25 |
| minute | josch: i can reproduce | 14:27 |
| minute | my system hangs after start of device-mapper | 14:27 |
| josch | perfect! | 14:27 |
| josch | though... | 14:27 |
| minute | i will do some serial debugging later, but there is a lot going on here today | 14:27 |
| minute | so first have to take care of some other stuff in the office | 14:27 |
| josch | did you upgrade all 2539 packages at once? | 14:27 |
| minute | josch: yes | 14:27 |
| josch | well, it's somewhere amongst those! XD | 14:28 |
| josch | and at least you can reproduce it | 14:28 |
| minute | josch: well, i think serial will reveal something, otherwise "break" in cmdline | 14:28 |
| josch | minute: could you share your broken initramfs (assuming it does not contain any private data) so that i can dissect and compare it to mine? | 14:28 |
| josch | minute: here is how you can potentially debug this kind of stuff: https://cryptsetup-team.pages.debian.net/cryptsetup/README.debug.html#debug-cryptroot-initramfs-script | 14:29 |
| josch | in other news, reform-check from my MR seems to work fine in vanilla sysimage-v4 | 14:41 |
| josch | jjbliss: since you are on a311d, would you mind giving it a spin on your device and put in a pastebin what it prints out? | 14:41 |
| josch | jjbliss: specifically this version of it: https://source.mnt.re/reform/reform-tools/-/raw/1547e0d18795b1b4fe4d5a4b412c9a55fb0d38c4/bin/reform-check | 14:42 |
| jjbliss | I'll take a look | 14:47 |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@64.189.201.150) | 14:50 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dcf:5280:3352:40b4:9f34:deb7) | 14:54 | |
| + stites (~stites@130.44.147.204) | 14:54 | |
| + jacobk (~quassel@129.110.242.224) | 15:04 | |
| jjbliss | https://pastebin.com/cHS7ACmv | 15:06 |
| jjbliss | that's my current ouptut when running that version with sudo | 15:06 |
| josch | thank you! | 15:07 |
| jjbliss | So, right now WiFi, Bluetooth, and Battery levels aren't working. I'm going to double check that those work on the fresh SD card image and see what's different. | 15:10 |
| Boostisbetter | jjbliss: how are you don't Bluetooth? | 15:11 |
| josch | jjbliss: thank you! your output helped me fix some bugs | 15:22 |
| josch | if you like, here is an updated version: https://source.mnt.re/reform/reform-tools/-/raw/da873fa009cc5b69276386f2f1399943a9259a75/bin/reform-check | 15:23 |
| minute | josch: btw sorry, i tried to upload initramfs but i have a problem here that big uploads crash the damn fritzbox if 5g wifi is used (sigh) | 15:25 |
| jjbliss | For bluetooth, I just tried installing blueman but it's not detecting the integrated bluetooth on the A311D. Bluetooth isn't important to me yet though. Wifi takes priority. | 15:27 |
| jjbliss | josch: happy to help. I'll try out the updated version | 15:27 |
| minute | jjbliss: sometimes there's an sdio glitch, perhaps it'll work after reboot | 15:29 |
| minute | jjbliss: esp if both wifi+bt don't work, as they're on the same sdio module. | 15:33 |
| minute | jjbliss: "sudo dmesg | grep -i sdio" might yield something. | 15:34 |
| - mark_ (QUIT: Ping timeout: 255 seconds) (~mjw@gnu.wildebeest.org) | 15:40 | |
| jjbliss | I get [ 5.163030] mmc2: new SDIO card at address 0001 | 15:41 |
| jjbliss | minute: I've tried a freshly imaged sd card and bluetooth and wifi work fine there. I'm pretty sure it's something wrong with my nvme system. | 15:41 |
| jjbliss | maybe some packages I don't have installed there | 15:42 |
| minute | jjbliss: perhaps the firmware | 15:44 |
| minute | jjbliss: i think the package is firmware-realtek | 15:44 |
| vkoskiv | josch: https://vkoskiv.com/stuff/apt_history.txt | 15:48 |
| vkoskiv | That's all that was in there | 15:48 |
| jjbliss | minute: That was it, thanks! | 15:49 |
| - stites (QUIT: Ping timeout: 258 seconds) (~stites@130.44.147.204) | 16:01 | |
| + stites (~stites@2607:fb90:ad62:a06c:7a16:280f:353f:7e4f) | 16:02 | |
| - sevan (QUIT: Quit: reboot, brb hopefully) (~sevan@user/venture37) | 16:05 | |
| jjbliss | Now I need to get the battery levels figured out. It was working before switching modules. Currently it is constantly showing 0. This happens both on my install and a fresh SD card | 16:06 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 16:09 | |
| jjbliss | josch: Here is a run of the updated version if it would help. https://pastebin.com/1DBYFmte | 16:10 |
| Boostisbetter | jjbliss: if this is working well for you,i think I going to get a A311D for myself. I miss Bluetooth. | 16:13 |
| - jacobk (QUIT: Ping timeout: 248 seconds) (~quassel@129.110.242.224) | 16:14 | |
| jjbliss | Bluetooth is definitely seeing things, but I haven't tried actually using much. Sometime in the next couple days I can try seeing if it handles several devices at once. I've had issues on other computers if I connect mouse, keyboard, and headphones through bluetooth all at once. | 16:16 |
| + jacobk (~quassel@64.189.201.150) | 16:20 | |
| minute | bluetooth works very well on the a311d according to my tests. | 16:29 |
| josch | vkoskiv: thank you! you probably do not remember whether you rebooted between any of these? | 16:29 |
| josch | jjbliss: thank you, that looks really good now! According to that output, you have firmware-realtek installed | 16:30 |
| jjbliss | I do now. | 16:30 |
| vkoskiv | josch: between the apt autoremove and apt install pigz, I likely rebooted? Not 100% sure though | 16:31 |
| josch | Boostisbetter: or you could add a usb bluetooth into the case by installing a small usb hub as i did :) | 16:32 |
| josch | i'm using bluetooth right now to connect my headphones because my audio chip is fried :D | 16:32 |
| + sevan (~sevan@user/venture37) | 16:33 | |
| josch | vkoskiv: the "interesting" stuff is in your very first stanza but that one is also huge :D | 16:35 |
| minute | i have a few minutes now to look on serial what's up during boot hang (on a311d) | 16:37 |
| minute | well, it's like how i thought | 16:38 |
| minute | > Please unlock disk crypt: | 16:38 |
| josch | your idea of the password prompt showing up in the wrong tty was a good theo | 16:38 |
| josch | oh | 16:38 |
| josch | it was that! | 16:38 |
| minute | this is on serial, it's not on the graphical console | 16:38 |
| josch | then... maybe it's the order of tty in the cmdline? | 16:39 |
| minute | yeah | 16:39 |
| minute | [ 0.000000] Kernel command line: console=tty1 ro no_console_suspend cma=512M@3G cryptomgr.notests console=ttyAML0,115200 pci=pcie_bus_perf libata.force=noncq nvme_core.default_ps_max_latency_us=0 | 16:39 |
| minute | i can manually change that in my boot.scr, one sec | 16:39 |
| josch | sysimage-v3 had console=ttymxc0,115200 and console=tty1 *after* it | 16:41 |
| minute | yep, also just seen that | 16:41 |
| minute | (because of residue in my boot.scr) | 16:41 |
| minute | sudo mkimage -A arm -T script -C none -n "MNT A311D" -d boot.scr.txt boot.scr | 16:41 |
| josch | and it of course did not trigger for me because i use u-boot-menu which bypasses boot.scr and thus my cmdline is different | 16:42 |
| josch | /proc/cmdline for me is: ro no_console_suspend cma=512M pci=nomsi console=ttymxc0,115200 console=ttymxc0,115200 console=tty1 | 16:42 |
| josch | and that works | 16:42 |
| minute | josch: yep, confirmed | 16:42 |
| josch | perfect, let me prepare a patch | 16:43 |
| minute | josch: so i changed in my u200 case the bootargs line so that the ${bootargs} is at the end | 16:43 |
| minute | josch: maybe we could provide a working boot.scr for people to replace on /boot on another computer? | 16:43 |
| vkoskiv | I have a chroot set up here atm | 16:44 |
| minute | hmm, htop shows N/A for cpu frequency | 16:44 |
| minute | ah, it's because i am on old kernel | 16:45 |
| josch | minute: sure, you can provide a working boot.scr. That's probably easier than telling people to do the cryptsetup, mount sheningans to get a chroot, then update reform-tools and then run update-initramfs -u | 16:46 |
| vkoskiv | I didn't know htop even showed that :D | 16:46 |
| minute | vkoskiv: F2, there somewhere you can enable it | 16:46 |
| minute | "also show cpu frequencies" | 16:46 |
| vkoskiv | josch: I guess since I'm already in here, I can do that? Just wait for the new version to come up in apt? | 16:46 |
| vkoskiv | Or maybe it's easier to copy in the boot.scr | 16:47 |
| josch | minute: about how to best fix this, do you think the correct fix would be to put ${bootargs} at the end? That would mean that the arguments from the board are able to overwrite the defaults. If we do it that way, we must be very certain that u-boot never provides wrong values. | 16:47 |
| vkoskiv | I can scp files to and from here | 16:47 |
| josch | vkoskiv: do you have network in your chroot? you might need to copy in resolv.conf from the outside | 16:47 |
| josch | minute: another solution would be to append the console=tty1 part at the end to ${bootargs} instead of setting it in the beginning of the script | 16:48 |
| vkoskiv | Let me test, sec | 16:48 |
| josch | minute: i have no strong opinion here. It depends on whether you want 00reform2_ubootenv having the final say about the platform-specific arguments or u-boot itself. | 16:48 |
| josch | vkoskiv: htop even has mouse support these days | 16:49 |
| vkoskiv | yeah, that I noticed. Sometimes F1/F2 are intercepted by the terminal emulator | 16:50 |
| minute | josch: lets try putting console=tty1 in the end of every board-specific string | 16:50 |
| vkoskiv | I can ping 1.1 in my chroot | 16:50 |
| vkoskiv | Ping me when you've worked out a working conf, I can try it then | 16:50 |
| Boostisbetter | minute: could I help pay for a bounty to get hibernation working on the Reform? This is the last hard edge the Reform has for me, and I think hibernation would help out a great deal, especially for those with NVMEs in their system. | 16:52 |
| josch | minute, vkoskiv: this is how i solved it now: https://source.mnt.re/reform/reform-tools/-/merge_requests/57/diffs?commit_id=d9ee804f63720c550238c6790f80f0a0019c5cca | 16:54 |
| josch | vkoskiv: here is the resulting file to test: https://source.mnt.re/reform/reform-tools/-/raw/d9ee804f63720c550238c6790f80f0a0019c5cca/flash-kernel/ubootenv.d/00reform2_ubootenv | 16:54 |
| vkoskiv | And this goes where again? | 16:56 |
| josch | putting console=tty1 after the if/else block avoids duplication and allows for a comment explaining why it is done like that in the hopes that this does not happen again | 16:56 |
| vkoskiv | On the nvme? | 16:56 |
| josch | vkoskiv: on the nvme you put it to /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv | 16:56 |
| josch | then you run flash-kernel again | 16:57 |
| Boostisbetter | minute: in other happy news, today was the first time that I started on the charger, suspended, took it off the charger, used it twice off the charger, suspending and resuming in between, and then finally suspending because the batteries were low. Getting home, plugged into the charger, and then resumed. All worked 100%. This is the first time. | 16:57 |
| Boostisbetter | So I am going to say that the 6.5 kernel fixes suspend in most cases. However, there ARE still cases where a file lock or something like that can cause resume to crash, as the NVME is not brought back up | 16:57 |
| vkoskiv | josch: since I wiped my sdcard in the meantime, do I also do the initramfs command | 16:59 |
| josch | vkoskiv: i don't think you have to re-build the initramfs | 16:59 |
| vkoskiv | update-initramfs -u && flash-kernel | 16:59 |
| vkoskiv | It now boots from the sd card though? | 17:00 |
| josch | but if you have /proc and /sys and /dev and /boot mounted, it does not hurt to run it either | 17:00 |
| vkoskiv | I do, I'll run it too | 17:00 |
| vkoskiv | hm, no output, it exited right away | 17:00 |
| vkoskiv | Quicker than before | 17:00 |
| josch | what exited? | 17:00 |
| vkoskiv | Hang on, I forgot /boot in the chroot | 17:01 |
| josch | vkoskiv: after you are done, maybe inspect your boot.scr to make sure it contains the right stuff | 17:02 |
| vkoskiv | Hmm, the bootargs doesn't match what's in the file you sent | 17:04 |
| josch | what do you mean? | 17:04 |
| vkoskiv | I see pci=nomsi | 17:04 |
| vkoskiv | Which in the file you sent is for imx9 | 17:04 |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb90:ad62:a06c:7a16:280f:353f:7e4f) | 17:04 | |
| vkoskiv | imx8, I mean | 17:05 |
| + stites (~stites@130.44.147.204) | 17:05 | |
| josch | vkoskiv: which file exactly do you mean? | 17:05 |
| vkoskiv | Maybe I misunderstand the format? You have the shared bootargs, then you append the platform specific flags, then console=tty1 at the end | 17:05 |
| josch | yes | 17:05 |
| vkoskiv | But I'm not seeing pci=pcie_bus_perf libata.force=noncq nvme_core.default_ps_max_latency_us=0 | 17:06 |
| vkoskiv | in my bootargs in boot.scr | 17:06 |
| josch | then something went wrong | 17:06 |
| josch | boot.scr should contain a copy of /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv | 17:06 |
| josch | the contents of /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv are copied verbatim into boot.scr | 17:06 |
| josch | so everything inside that file should end up in there | 17:07 |
| josch | if it does not, then something went wrong | 17:07 |
| vkoskiv | Yeah, I def. don't see the contents in there | 17:07 |
| vkoskiv | I'll go back into chroot to investigate | 17:07 |
| josch | vkoskiv: do you maybe have /etc/flash-kernel/ubootenv.d/00reform2_ubootenv on that system? | 17:07 |
| jjbliss | I had the old file in /etc that was overriding it. | 17:07 |
| josch | if you have /etc/flash-kernel/ubootenv.d/00reform2_ubootenv then that file will take precedence over whatever is in /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv | 17:07 |
| josch | and then the latter will have no effect | 17:08 |
| josch | yes, what jjbliss said | 17:08 |
| josch | the reform-check version that jjbliss tested checks for that | 17:08 |
| vkoskiv | I do have that file. Do I delete it? | 17:08 |
| josch | vkoskiv: probably yes. Do you want to share what is inside of it? | 17:09 |
| vkoskiv | sure, sec | 17:09 |
| vkoskiv | # setenv bootpart "1" | 17:10 |
| vkoskiv | # setenv prefix "/" | 17:10 |
| vkoskiv | # setenv kernel_addr_r "0x40480000" | 17:10 |
| vkoskiv | # setenv fdt_addr_r "0x50000000" | 17:10 |
| vkoskiv | # setenv ramdisk_addr_r "0x51000000" | 17:10 |
| josch | is it by any chance this: https://paste.debian.net/hidden/27fff577/ | 17:10 |
| josch | yes | 17:10 |
| josch | that is a remnant the old system image | 17:10 |
| vkoskiv | rerunning flash-kernel once more | 17:10 |
| josch | this is why the system image should not put stuff into /etc | 17:10 |
| josch | it's just a big hassle to tell users to please adjust their stuff in /etc | 17:10 |
| josch | vkoskiv: i hope you are still taking notes so that we can write something down after all of this is done :) | 17:11 |
| vkoskiv | Looks better now | 17:12 |
| vkoskiv | I have the pics from last night, plus my questionable short-term memory :D | 17:13 |
| + chomwitt (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 17:14 | |
| vkoskiv | Noted some commands I forgot down | 17:16 |
| vkoskiv | Rebooting now | 17:16 |
| vkoskiv | stuck again @ "new SDIO card at address 0001" | 17:17 |
| vkoskiv | Yep, NVME error again. Completion polled thing | 17:17 |
| josch | vkoskiv: what is your boot.scr now? | 17:18 |
| vkoskiv | I'll grab that, sec | 17:19 |
| vkoskiv | My boot.scr on the SD card now: https://pastebin.com/zguCjC88 | 17:20 |
| josch | A PACKAGE WRAPPED IN MNT TAPE JUST ARRIVED VIA UPS \o/ | 17:21 |
| josch | vkoskiv: that looks good | 17:21 |
| vkoskiv | ooo! hardware time! | 17:22 |
| vkoskiv | I think last night, copying another boot.scr from the SD image made more stuff happen | 17:22 |
| josch | hrm... minute fixed their setup by appending console=tty1 at the end of the line... theoretically my change should have the same result | 17:22 |
| vkoskiv | But then that would overwrite this boot.scr again | 17:22 |
| josch | minute: do you see anything wrong in that boot.scr? | 17:22 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 17:22 | |
| josch | when was the last time you bought a computer that came with schematics? https://mister-muffin.de/p/h4ay.jpg | 17:29 |
| josch | oh sorry, wrong channel :) | 17:29 |
| minute | vkoskiv: you still have pci=nomsi i guess | 17:37 |
| minute | vkoskiv: that needs to go away | 17:37 |
| josch | that setting kills nvme? | 17:37 |
| vkoskiv | Yeah, I noticed that, I wonder where it comes from | 17:38 |
| josch | vkoskiv: it comes from your /etc/default/flash-kernel | 17:38 |
| josch | this is also something that reform-check warns about | 17:38 |
| vkoskiv | So chroot, delete that, then flash-kernel? | 17:38 |
| vkoskiv | Or delete pci=nomsi from in that file? | 17:38 |
| josch | both would work | 17:38 |
| josch | reform-check recommends the following content for that file: | 17:39 |
| vkoskiv | Guess I need to flash my SD once more to get to chroot? | 17:39 |
| vkoskiv | Or can I fix this from having the SD mounted on my desktop? | 17:39 |
| minute | josch: yes, pci=nomsi is harmful on a311d and ls1028a iirc | 17:40 |
| josch | vkoskiv: theoretically you can rewrite your boot.scr outside using mkimage | 17:40 |
| josch | vkoskiv: minute pasted the command you can use above | 17:40 |
| josch | minute: this also answers the question whether the pci argument gets overwritten or appended to | 17:40 |
| vkoskiv | I'll try that | 17:41 |
| josch | vkoskiv: reform-system-image and reform-check recommend the following content for that file: | 17:41 |
| josch | LINUX_KERNEL_CMDLINE="" | 17:41 |
| josch | LINUX_KERNEL_CMDLINE_DEFAULTS="" | 17:41 |
| vkoskiv | Hrm, no mkimage on my desktop | 17:45 |
| josch | vkoskiv: install u-boot-tools | 17:46 |
| vkoskiv | Yeah, need to permute the name to please pacman first | 17:46 |
| vkoskiv | ah, uboot-tools it is | 17:46 |
| josch | minute: once vkoskiv confirms this to work, should i merge https://source.mnt.re/reform/reform-tools/-/merge_requests/57 and re-run reform-debian-packages or do you want to confirm this first? | 17:47 |
| vkoskiv | Okay, new boot.scr in, testing this out | 17:47 |
| vkoskiv | > Please unlock reform_crypt:_ | 17:48 |
| vkoskiv | Aww yeee | 17:48 |
| vkoskiv | Aaand we're in sway! Success! | 17:49 |
| josch | perfect! | 17:50 |
| vkoskiv | Mmm, 6 cores. Sweeet | 17:50 |
| josch | okay, so adjusting /etc/default/flash-kernel and /etc/flash-kernel/ubootenv.d/00reform2_ubootenv must definitely be part of moving to a different SoM | 17:51 |
| josch | vkoskiv: you could paste your boot.scr to that forum thread so that others can get their system back maybe | 17:51 |
| vkoskiv | Sure thing, but is it safe to paste with the binary data in there? | 17:52 |
| josch | vkoskiv: maybe you should use a pastebin that does not mangle those bits | 17:53 |
| vkoskiv | Maybe base64 it and have people base64 > boot.scr or something? | 17:53 |
| josch | you could use my pastebin for that, for example: | 17:53 |
| josch | cat /boot/boot.scr | curl -F 'arg=<-' https://mister-muffin.de/paste | 17:53 |
| josch | that pastebin definitely does not do any mangling | 17:53 |
| vkoskiv | Okay, sec | 17:54 |
| vkoskiv | https://mister-muffin.de/p/ZjJp.bin | 17:54 |
| vkoskiv | I'll put that link in the thread | 17:55 |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@64.189.201.150) | 17:56 | |
| vkoskiv | Posted | 17:57 |
| vkoskiv | So wait a bit longer before running apt upgrade, then? | 17:58 |
| josch | yes, i'd like for minute to approve that MR of mine first | 17:58 |
| vkoskiv | Observations right away: Idle current seems much lower, it's around ~220-230mA when sitting in sway | 17:59 |
| vkoskiv | Bloated tabloid website loads up reasonably fast, much faster than on imx8 | 17:59 |
| vkoskiv | It's a good benchmark. If the bloated tabloid loads acceptably fast, that means I'm dealing with a really powerful system. | 18:00 |
| vkoskiv | Not even my current i7 desktop loads it acceptably fast. | 18:00 |
| josch | just recently i introduced the problem described here https://community.mnt.re/t/problem-upgrading-kernel/1667 so after messing up again, i'd like to share the blame with minute if something goes wrong XD | 18:00 |
| vkoskiv | Eeeh, it's a lot of very detailed work. I really, really appreciate what you do! | 18:02 |
| josch | last time, the problem was that i'm not running the HDMI version of imx. This time the problem was, that i'm not using boot.scr to boot, so i just assumed it worked from looking at it but didn't test it in practice. :( | 18:04 |
| josch | i had no idea the console= order mattered for the unlocking luks | 18:04 |
| vkoskiv | In any case, you needn't apologise. If I were in charge of maintaining the system image, I guarantee it'd be a disaster :D | 18:06 |
| vkoskiv | I forget things a lot | 18:06 |
| vkoskiv | Okay, wifi card is missing still, it seems. Ethernet works. Bluetooth shows up, I'll test that out | 18:06 |
| vkoskiv | . | 18:08 |
| vkoskiv | BT controller shows up, and as unblocked in rfkill, but bluetoothctl can't chat with it. | 18:09 |
| vkoskiv | [ 37.997989] bluetooth hci0: firmware: failed to load rtl_bt/rtl8822cs_fw.bin (-2) | 18:09 |
| vkoskiv | [ 38.005390] bluetooth hci0: firmware: failed to load rtl_bt/rtl8822cs_fw.bin (-2) | 18:10 |
| vkoskiv | [ 38.012817] bluetooth hci0: Direct firmware load for rtl_bt/rtl8822cs_fw.bin failed with error -2 | 18:10 |
| vkoskiv | Some maybe relevant lines from kernel log | 18:10 |
| vkoskiv | bt service running too | 18:10 |
| vkoskiv | lsusb doesn't see the bt controller, maybe that's it | 18:11 |
| vkoskiv | Well, I'm still good to go for the time being, I've got ethernet hooked up and I have some work to do hacking on my secret project :3 | 18:12 |
| josch | vkoskiv: those files are in the package firmware-realtek -- do you have that installed? | 18:14 |
| vkoskiv | apt says it's not available, but referred to by another pkg. Isn't this the thing where I need to opt into nonfree pkgs or something? | 18:15 |
| vkoskiv | Since that's a blob package, if I recall | 18:15 |
| josch | yes, then that is your problem | 18:16 |
| josch | you need to add non-free-firmware | 18:16 |
| - murphnj (QUIT: Read error: Connection reset by peer) (~murph@user/murphnj) | 18:16 | |
| josch | in contrast to non-free, non-free-firmware only includes non-free blobs, not non-free software that runs on your main cpu | 18:16 |
| + murphnj (~murph@user/murphnj) | 18:16 | |
| josch | i have this in my /etc/apt/sources.list: | 18:16 |
| josch | deb http://deb.debian.org/debian unstable main non-free-firmware | 18:16 |
| josch | vkoskiv: running the new reform-check would've shown you :) | 18:17 |
| vkoskiv | I will give that a go, I'm sure it has many things to say about my ex-V3 system :D | 18:22 |
| + jacobk (~quassel@129.110.242.224) | 18:23 | |
| minute | josch: sorry, was afk | 18:25 |
| minute | josch: what do you want me to accept? | 18:25 |
| minute | (which MR) | 18:25 |
| minute | ah, probably https://source.mnt.re/reform/reform-tools/-/merge_requests/57 | 18:26 |
| josch | yes | 18:26 |
| minute | strange, if i click on changes there it has a bunch of old stuff | 18:27 |
| josch | minute: either inspect just visually or try out the new ubootenv in your system to see if it fixes your setup as well | 18:27 |
| josch | just force-pushed some more changes to d/changelog | 18:28 |
| josch | vkoskiv: if you did not yet install firmware-realtek you could try out reform-check from that MR to see if it really does warn about it missing | 18:28 |
| josch | (the check only warns about this on a311d) | 18:29 |
| vkoskiv | I'll have dinner and then test | 18:30 |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@129.110.242.224) | 18:33 | |
| minute | josch: LGTM, merged | 18:34 |
| josch | nice, reform-debian-packages pipeline triggered | 18:39 |
| josch | in 1.5 hours reform-tools 1.31 should be in the MNT repo | 18:39 |
| josch | i'll reply to that forum thread once that has happened | 18:39 |
| josch | sorry again for the mess :( | 18:39 |
| vkoskiv | Do I want to just upgrade reform-tools to get the latest reform-check, or pull one off src.mnt.re? | 18:59 |
| vkoskiv | Interesting, the LPC reported low battery and shut the system off a few secs after I pulled the plug | 19:00 |
| vkoskiv | That didn't happen just an hour or so ago | 19:00 |
| vkoskiv | I have a script that plays a sound at <10%, so that started playing | 19:01 |
| vkoskiv | Then it shut off while I started investigating | 19:01 |
| vkoskiv | LPC shows 99%, and my battery should be full | 19:01 |
| vkoskiv | I'll do more digging | 19:01 |
| vkoskiv | It did a controlled shutdown, that's nice | 19:01 |
| vkoskiv | Yeah, the lpc driver is doing something funky, it's reporting values all over the place | 19:06 |
| vkoskiv | 109% :D | 19:06 |
| minute | vkoskiv: a311d spi lpc stuff is very funky, | 19:06 |
| minute | vkoskiv: because it's emulated with gpios | 19:07 |
| vkoskiv | Yeah, it's just randomly bouncing around it seems. I'm looking at /sys/class/power_supply/8xlifepo4/device/status | 19:07 |
| vkoskiv | The fix is needed in the lpc driver? | 19:07 |
| vkoskiv | So you have gpios bit-banging the serial/UART to the LPC? Where's that code? | 19:09 |
| + mark_ (~mjw@gnu.wildebeest.org) | 19:10 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark_!~mjw@gnu.wildebeest.org))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 19:12 | |
| * mark_ -> mjw | 19:12 | |
| + mark_ (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 19:13 | |
| minute | vkoskiv: it's a linux feature. hang on a sec... perhaps i put some error recovery code in the lpc firmware | 19:14 |
| minute | ah yes. totally forgot about that | 19:15 |
| minute | vkoskiv: https://source.mnt.re/reform/reform/-/blob/master/reform2-lpc-fw/src/boards/reform2/board_reform2.c?ref_type=heads#L931 | 19:15 |
| minute | vkoskiv: pull the latest master from reform repo, and build the reform2-lpc-fw, and flash it. while making sure to activate this line https://source.mnt.re/reform/reform/-/blob/master/reform2-lpc-fw/src/boards/reform2/board_reform2.c?ref_type=heads#L44 | 19:16 |
| vkoskiv | Cool, I'll get to that then | 19:17 |
| minute | jjbliss: same applies to you, sorry for this inconvenience | 19:18 |
| jjbliss | no worries. My LPC firmware is still on one from about a year ago, I was assuming that it was what needed updating for the battery status. | 19:19 |
| vkoskiv | I'm running 20220915 | 19:22 |
| Boostisbetter | minute: I know that the L5 kernel is not the same thing, but it seems like the option just needs to be set to true when compiling the kernel. https://source.puri.sm/Librem5/linux/-/merge_requests/701 | 19:28 |
| Boostisbetter | minute: I would think that if the swap is large enough, it should just work. | 19:28 |
| Boostisbetter | Especially if encryption of the drive is not being used. | 19:28 |
| minute | i would think that debian has this enabled. | 19:30 |
| vkoskiv | lpc flashed, battery info looks good now | 19:32 |
| vkoskiv | Yup, certainly matches the OLED battery view | 19:33 |
| minute | this document i'm trying to sign to get the ls1028as says: | 19:33 |
| minute | > Pursuant to subsection 40(1) of the Strategic Trade Act 2010, a person who submits false or misleading information in any application, report or other documents for any purpose under this Act or the related laws commits an offence and shall, on conviction, be punished with the fine not exceeding one million ringgit or with imprisonment for a term not exceeding two years or with both, and in | 19:33 |
| minute | the case of a body corporate, be punished with a minimum fine of two million ringgit. | 19:33 |
| minute | vkoskiv: oh great! | 19:33 |
| vkoskiv | Downside of A311D: It now runs youtube smooth enough for me to waste more time on there :D | 19:35 |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 19:36 | |
| vkoskiv | Opened this from my front page to test, just look at this wonderful computing optimism: https://www.youtube.com/watch?v=SZ1RcNxiNX8 | 19:37 |
| vkoskiv | "120 billion bytes of fast-access storage!" | 19:37 |
| minute | vkoskiv: hehe @ a311d | 19:40 |
| vkoskiv | Left speaker is very quiet, wonder if it's a connection issue or a hw one | 19:42 |
| vkoskiv | headphones sound normal | 19:43 |
| vkoskiv | Schematic time! | 19:43 |
| vkoskiv | Hmm, speakers hook up to U3, which is on the motherboard | 19:44 |
| abortretryfail | check the mixer first. | 19:44 |
| vkoskiv | I'm looking at pavucontrol, don' | 19:44 |
| vkoskiv | t see a slider for this | 19:44 |
| abortretryfail | alsamixer -c 0 should show you all of the controls for the codec | 19:45 |
| abortretryfail | pavucontrol presents a very simplified version of what's going on. :) | 19:45 |
| vkoskiv | it shows 76<>76 for the speaker, which is I assume meaning that both channels have same vol | 19:45 |
| abortretryfail | yeah that's it | 19:46 |
| vkoskiv | They adjust in tandem if I bump up/down the volume | 19:46 |
| jjbliss | I've updated my lpc firmware and the battery status is showing properly now, thanks for the info minute | 19:46 |
| abortretryfail | there are more though, the PCM DAC output levels are separate for right and left | 19:46 |
| vkoskiv | All these seem to be in sync | 19:48 |
| vkoskiv | Nothing obvious jumps out | 19:48 |
| minute | jjbliss: great! | 19:50 |
| minute | vkoskiv: in doubt i would recommend to join #linux-amlogic also and ask there | 19:50 |
| vkoskiv | I'll ask there, thanks | 19:51 |
| vkoskiv | I'll grep their irc log first though :D | 19:51 |
| minute | we'll also send a laptop with a311d to neil armstrong soon, who is very active in upstreaming the drivers for this soc | 19:52 |
| vkoskiv | This thing is smooooth though! Very noticeable speed boost | 19:52 |
| minute | vkoskiv: yaaay | 19:52 |
| vkoskiv | That is a... Very un-googleable name :D | 19:53 |
| minute | (bay libre) | 19:53 |
| vkoskiv | Ah, that does it. | 19:53 |
| jjbliss | I just tried out gnome with the a311d, it's so much better than with the mx8mq | 19:53 |
| vkoskiv | I'm quite certain sway will be it for me for a very long time, it was perfectly snappy on the imx8 too | 19:54 |
| minute | jjbliss: ohhh i've never even tried that! | 19:54 |
| jjbliss | not sure if I'll stick with it, as sway has been working fine, but I thought it was worth giving gnome a try | 19:54 |
| minute | if one of you wants to put up some videos of gnome and/or kde in action on a311d reform, would be much appreciated | 19:54 |
| vkoskiv | lots of mentions of speaker observations in that IRC log, I'll comb through in a bit | 19:56 |
| vkoskiv | So many things to test :D | 19:56 |
| vkoskiv | Much fun | 19:56 |
| vkoskiv | And I was testing with spotifyd, so now I'm just jamming to some good tunes too. | 19:56 |
| vkoskiv | I need to do that headphone jack fix to get more bass at some point | 19:57 |
| + jacobk (~quassel@129.110.242.224) | 19:58 | |
| josch | oh well... reform-debian-packages is blocked until next dinstall because of https://buildd.debian.org/status/package.php?p=python3.11 | 20:03 |
| vkoskiv | josch: Pulled latest version of reform-check, and it does mention missing 'firmware-realtek' | 20:06 |
| vkoskiv | In fact, let me grab the whole output for ya | 20:07 |
| vkoskiv | josch: https://mister-muffin.de/p/mxby.diff | 20:08 |
| vkoskiv | Many things, I haven't tended to these in a while :D | 20:08 |
| josch | vkoskiv: you absolutely need apt-listbugs | 20:14 |
| josch | vkoskiv: if you run unstable, that package will prevent you from installing or upgrading to packages with release critical bugs in them | 20:14 |
| vkoskiv | I'll just install all those recommendations | 20:15 |
| josch | that output looks mostly good | 20:15 |
| josch | except that i'm unsure of what to make of a311d u-boot | 20:15 |
| josch | minute: in reform-system-image you dd it but skip the first 512 byte of flash.bin? | 20:15 |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.32.204) | 20:16 | |
| + Ar|stote|is (~linx@149.210.32.204) | 20:19 | |
| bkeys | I am excited to get the adapter boards but I got my ppc64le workstation I do most computing on | 20:20 |
| bkeys | I have an A311D SBC that runs mainline Fedora so I am excited by those prospects | 20:20 |
| vkoskiv | Not too many people aroudn running PPC on their workstations | 20:22 |
| bkeys | Yeah I bought it from my best friend; it's really performant I enjoy it | 20:23 |
| bkeys | And runs all the paplications I need it to | 20:23 |
| vkoskiv | Curious what makes you prefer that arch? | 20:23 |
| vkoskiv | I think it's cool there are boxes running it around | 20:23 |
| vkoskiv | It's a newer machine, right? | 20:23 |
| bkeys | It's fully libre | 20:23 |
| bkeys | It's a talos workstation | 20:24 |
| vkoskiv | Ah, right. Yeah, I think I read about that | 20:24 |
| bkeys | It's comparable to performance to x86 | 20:24 |
| vkoskiv | Very cool, I see the appeal. Performant, libre silico. | 20:24 |
| vkoskiv | s/./n./ | 20:24 |
| bkeys | That and ppc64le is IBM architecture so all the Red Hat stuff is really well supported and that is the family of distros I prefer | 20:25 |
| bkeys | I actually have an aarch64 server that has UEFI and it runs CentOS Stream, the only thing not running RH distros is the Reform, but hopefully adapter boards change that | 20:26 |
| bkeys | Absolute worst case I could pay out the wazoo for rpi cm4 SoM and get mainline support, the cm4 adapter is an awesome idea it opens up a lot of possibilities | 20:26 |
| minute | josch: yep, following the way it is done up/downstream | 20:30 |
| josch | okay, then i have to fix reform-check again XD | 20:32 |
| sevan | bkeys: very nice, running not so impressive ppc64be here :) | 20:46 |
| josch | vkoskiv: would you mind trying another version of reform-check? | 20:51 |
| vkoskiv | Sure thing! | 20:51 |
| vkoskiv | Do I just grab off master? | 20:52 |
| josch | vkoskiv: use this: https://source.mnt.re/reform/reform-tools/-/raw/d10284c29757d0a46527b43bde0133dcd1c4db31/bin/reform-check | 20:52 |
| josch | that's part of yet another reform-tools MR: https://source.mnt.re/reform/reform-tools/-/merge_requests/58 | 20:52 |
| jjbliss | So, I just went to try using mpv for hardware accelerated h264 video on the A311D and I'm getting a segfault. This happens on a fresh SD card and on my nvme system. | 20:55 |
| vkoskiv | josch: https://mister-muffin.de/p/GOcb.diff | 20:55 |
| vkoskiv | Hmm, I meant to try accelerated playback, but I forget the name of the player that I used | 20:56 |
| vkoskiv | I switched from mpv to another one that seemed to work better on the imx8 | 20:56 |
| vkoskiv | Specifically with accelerated playback | 20:56 |
| jjbliss | The segfault follows an ffmpeg message "Could not dynamically load CUDA" I don't know if this actually related or just something it always says when no nvidia card is available | 20:57 |
| josch | vkoskiv: you probably used clapper | 20:57 |
| jjbliss | I also tried a flatpak installation of clapper and it did not play any video on A311D | 20:58 |
| josch | vkoskiv: hrm... this is still wrong output | 20:58 |
| josch | i have to check with minute how to do this properly on a311d | 20:58 |
| vkoskiv | I'll reboot to see if I can get bluetooth working | 20:59 |
| vkoskiv | Is the realtek firmware package also for the wifi, or is that a separate issue? | 21:00 |
| vkoskiv | josch: I think installing the firmware package did something with the /boot volume, I get the pci=nomsi NVME issue again. I had this once earlier, I can just fix the boot.scr and get back to booting. | 21:01 |
| minute | jjbliss: ah, the problem is the default config for mpv | 21:01 |
| minute | jjbliss: i also stumbled over this. we ship some default config for mpv which sets hwdec=auto. i think that should be removed cc josch | 21:02 |
| minute | jjbliss: on a311d you don't need hw dec, the cpu can easily do it | 21:02 |
| minute | jjbliss: delete ~/.config/mpv/mpv.conf and mpv should work | 21:02 |
| vkoskiv | Youtube was smooth earlier, and I didn't observe hwdec interrupts going up | 21:02 |
| minute | a311d has some hwdec but i didn't figure out how to use it yet. | 21:03 |
| minute | it can easily sw decode 1080p @ 60fps youtube or h265 1080p videos with mpv | 21:03 |
| minute | so i didn't have a need for hw dec | 21:03 |
| jjbliss | minute: thanks I might also try messing around with the hwdec option there | 21:04 |
| minute | jjbliss: if you want to know more about the state of hwdec on a311d i would suggest to ask in #linux-amlogic here | 21:04 |
| vkoskiv | Wifi working now! | 21:07 |
| vkoskiv | Signal strength is usually around ~50ish % here, now it's showing 73% | 21:09 |
| minute | vkoskiv: nice! i have to say that wifi is slower than it should be at the moment, driver issue | 21:09 |
| minute | vkoskiv: ethernet was plenty fast though in my tests | 21:10 |
| vkoskiv | Around 30Mbit/s, I ran a speedtest | 21:10 |
| minute | yep | 21:10 |
| minute | that checks out | 21:10 |
| vkoskiv | Ethernet does go fast though, yeah | 21:10 |
| vkoskiv | I rarely transfer files with USB sticks these days, I just hook up ethernet and scp/rsync | 21:10 |
| vkoskiv | At Assembly this year, I bought the most hilariously bad USB stick. They were selling them for I think 12 euro, 256GB Kingston stick | 21:11 |
| vkoskiv | I measured the performance, and it would take you over 6 hours to actually fill it with 256GB of data | 21:12 |
| vkoskiv | Which is probably why they were so cheap. Of course the packaging advertised the USB 3.1 link speed, not the actual transfer speed. | 21:12 |
| vkoskiv | I thought better of Kingston. Naughty, naughty. | 21:12 |
| - Boostisbetter (QUIT: Ping timeout: 272 seconds) (4a410829d7@irc.cheogram.com) | 21:13 | |
| vkoskiv | I get into sub-200mA at idle if I turn the brightness down | 21:21 |
| vkoskiv | Cool, literally and figuratively. | 21:21 |
| + sigrid (~sigrid@ftrv.se) | 21:23 | |
| - mhoye (QUIT: Ping timeout: 240 seconds) (~mhoye@li319-32.members.linode.com) | 21:26 | |
| + mhoye (~mhoye@li319-32.members.linode.com) | 21:26 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 21:27 | |
| minute | vkoskiv: yep, this chip runs quite cool | 21:31 |
| minute | 12nm vs imx' 28nm perhaps | 21:31 |
| - jacobk (QUIT: Ping timeout: 272 seconds) (~quassel@129.110.242.224) | 21:33 | |
| vkoskiv | AirPods connected with noise cancelling enabled | 21:37 |
| vkoskiv | Veeery cool | 21:37 |
| vkoskiv | Bunch of this in dmesg: | 21:39 |
| vkoskiv | [ 1744.799362] rtw_8822cs mmc2:0001:1: failed to send h2c command | 21:39 |
| vkoskiv | [ 1746.868008] rtw_8822cs mmc2:0001:1: firmware failed to leave lps state | 21:39 |
| vkoskiv | Not sure if relevant, things seem to work | 21:39 |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 21:40 | |
| minute | vkoskiv: yep, i see these too sometimes | 21:43 |
| - bluerise (QUIT: Changing host) (~bluerise@p5b0acf0d.dip0.t-ipconnect.de) | 21:50 | |
| + bluerise (~bluerise@user/bluerise) | 21:50 | |
| - Boostisbetter (QUIT: Ping timeout: 255 seconds) (4a410829d7@irc.cheogram.com) | 21:58 | |
| vkoskiv | I wonder if 'vdec' is the hw video decoder interrupt? | 22:03 |
| minute | i think so | 22:05 |
| minute | here's the datasheet btw https://dl.khadas.com/hardware/vim3/datasheet/a311d_datasheet_01_wesion.pdf | 22:05 |
| vkoskiv | mpv works really smooth with just software decode | 22:05 |
| vkoskiv | smoother than the youtube player in a browser | 22:06 |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 22:06 | |
| vkoskiv | BT reception seems good, I just walked to the other side of my apartment, only cut out when I went behind a concrete wall | 22:09 |
| vkoskiv | Slight caveat with the OLED change I made - If a program draws to those rightmost 2 columns of pixels, and then neglects to clear it on exit with WCLR, then the menus don't clear those pixels and they are stuck | 22:23 |
| vkoskiv | I'll make a fix MR | 22:23 |
| josch | vkoskiv: when you installed firmware-realtek, flash-kernel ran again and destroyed your boot.scr because reform-tools in the mnt repo is not yet fixed | 22:24 |
| vkoskiv | Yeah, I figured that. Easy enough fix | 22:24 |
| josch | minute: yes, hwdec=auto is from etc/skel/.config/mpv/mpv.conf but that only influences initial installations | 22:26 |
| vkoskiv | I had that in there, removed it, and that fixed mpv for me | 22:26 |
| josch | minute: what config should reform-tools ship then? It will not change anything on a running system as it only changes files in /etc/skel | 22:27 |
| Boostisbetter | ahh teh exact science of Linux | 22:27 |
| josch | minute: maybe setting up stuff like this should be done by the same thing that is supposed to set up keyboard layout, timezone, language etc on first boot? | 22:28 |
| josch | minute: such a script would also know what platform it is running on and could write out correct values in ~/.config/mpv/mpv.conf when creating a new user | 22:28 |
| - chomwitt (QUIT: Ping timeout: 264 seconds) (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 22:32 | |
| minute | josch: hmm yeah, i agree | 23:02 |
| - robin (QUIT: Remote host closed the connection) (~robin@user/terpri) | 23:39 | |
| + robin (~robin@user/terpri) | 23:39 | |
| - klardotsh (QUIT: Ping timeout: 255 seconds) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 23:41 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
