| + chomwitt (~chomwitt@2a02:587:7a02:1a00:1ac0:4dff:fedb:a3f1) | 00:11 | |
| josch | vagrantc: i tried building a MNT image using guix but I get: Throw to key `encoding-error' with args `("scm_to_stringn" "cannot convert wide string to output locale" 84 #f #f)'. | 00:22 |
|---|---|---|
| josch | vagrantc: any idea? https://source.mnt.re/josch/mnt-reform-nonguix/-/jobs/4561 | 00:22 |
| - mjw (QUIT: Ping timeout: 268 seconds) (~mjw@gnu.wildebeest.org) | 00:44 | |
| vagrantc | josch: wait, what? :) | 01:14 |
| vagrantc | josch: you're building a kernel package for debian using guix? | 01:15 |
| vagrantc | josch: wild hunch would be some issue with the guix locale vs. the system locale ... but i may be mixing that up with some other sort of issue | 01:18 |
| vagrantc | josch: sometimes due to differing glibc versions (and thus locale data) | 01:24 |
| vagrantc | josch: you might find people who can actually read those guile tracebacks more proficiently on #guix :) | 01:35 |
| vagrantc | ACTION waves | 01:35 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 01:35 | |
| - chomwitt (QUIT: Ping timeout: 240 seconds) (~chomwitt@2a02:587:7a02:1a00:1ac0:4dff:fedb:a3f1) | 02:44 | |
| - jn (QUIT: Ping timeout: 268 seconds) (~quassel@user/jn/x-3390946) | 03:21 | |
| + jn (~quassel@2001-4dd4-8ba8-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 03:21 | |
| - jn (QUIT: Changing host) (~quassel@2001-4dd4-8ba8-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 03:21 | |
| + jn (~quassel@user/jn/x-3390946) | 03:21 | |
| - nsc (QUIT: Ping timeout: 246 seconds) (~nicolas@240-99-142-46.pool.kielnet.net) | 03:25 | |
| + nsc (~nicolas@i5C74DC29.versanet.de) | 03:27 | |
| josch | minute: linux 6.9 is in experimental so i rebased our patch stack again. Here is the diff to the stack before: https://paste.debian.net/hidden/b9343dab/ | 06:37 |
| josch | as usual, cadence was the most tricky one to rebase | 06:38 |
| josch | minute: i think the Ampere CI runner works well and solves the problem that debian unstable frequently has cross-building broken due to multi-arch:same version skews. But the big intel box had the advantage that cross-building the kernel on it is quite a bit faster. I prepared https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/42 which (if it works) would allow one to select the CI | 07:14 |
| josch | runner by variable (defaulting to arm64) | 07:14 |
| josch | minute: could you add the intel box back to the reform-debian-packages project as a CI runner so that i can test this? thanks! | 07:14 |
| pandora | i made an apt update && upgrade and now my pocket reform is not booting anymore :/ | 09:22 |
| pandora | i am curious if there is an image that is not based on debian unstable but stable | 09:22 |
| josch | pandora: not yet but soon [tm] | 09:23 |
| josch | pandora: i'm maintaining an unofficial fork of the MNT repos building everything for stable (and backports) on https://reform.debian.net | 09:23 |
| josch | pandora: but i did not yet have time to adept the scripts for the pocket reform release | 09:23 |
| josch | pandora: if you want to help, i'd love to answer all your questions and then it can happen sooner :) | 09:23 |
| pandora | my background knowledge of the mnt build processes are currently non-existing :D ... the pocket is my first reform since i had to skip the normal reform due to money :D | 09:26 |
| josch | pandora: i'd also love to figure out what broke your pocket reform -- do you happen to be able to access the serial output with a uart adapter to have a look? | 09:26 |
| pandora | i do have a serial adapter but idk where i would connect them | 09:27 |
| pandora | is there a whitepaper on "how to get serial output from a broken mnt"? | 09:27 |
| josch | pandora: https://mntre.com/documentation/pocket-reform-handbook.pdf | 09:28 |
| josch | i think page 65 has the motherboard showing S1 and S2 for uart | 09:29 |
| pandora | ah yeah... i didn't scoll down that far yet :D | 09:29 |
| josch | the table on page 87 has more details | 09:29 |
| pandora | i will try to get a serial output ... but i have to do some stuff at work first | 09:32 |
| josch | minute: an apt-upgrade broke boot on the pocket -- maybe you want to have a look in case you manage to do it before pandora? | 09:33 |
| pandora | just a bit confused by the labeling | 09:45 |
| pandora | the pins say gnd, ax2, tx2, gnd, ax1, tx1 | 09:46 |
| pandora | i guess ax1. means rx1? | 09:46 |
| pandora | on the board | 09:46 |
| pandora | its prob. just a bad print and u can mistake A for R | 09:47 |
| josch | minute: and another issue that we will soon have is, that qcacld fails to build with linux 6.9: https://paste.debian.net/1321040/ -- i'd need a patch from you for that | 09:52 |
| - dozens (QUIT: Ping timeout: 268 seconds) (~dozens@tilde.town) | 09:53 | |
| + dozens (~dozens@tilde.town) | 09:53 | |
| pandora | ok got a serial output | 09:55 |
| pandora | https://pastebin.com/GTvKx5Q7 | 09:56 |
| josch | pandora: i guess it's stopping there? | 09:57 |
| pandora | yes | 09:57 |
| pandora | thats all the output | 09:57 |
| josch | pandora: when it says "Hit any key to stop autoboot:" can you press a key to get into the u-boot console? | 09:57 |
| josch | and then you run: | 09:57 |
| pandora | let me try to get into uboot | 09:58 |
| josch | setenv bootargs "${bootargs} loglevel=7" | 09:58 |
| josch | and then boot the thing by running: | 09:58 |
| josch | boot | 09:58 |
| josch | that should give you a bit more output | 09:59 |
| pandora | it is not accepting any key strokes | 09:59 |
| josch | pandora: you have your system on eMMC? | 09:59 |
| josch | uh oh... :( | 09:59 |
| pandora | yes ... it is kinda like it came out of the box | 09:59 |
| pandora | i myself just installed some python stuff | 10:00 |
| pandora | and apt update & upgrade | 10:00 |
| pandora | no customisation so far | 10:00 |
| josch | right, it's unfortunate that you cannot get into uboot (but at least we saw some output) | 10:00 |
| josch | pandora: can you try popping in an sd-card with a fresh system image? | 10:00 |
| pandora | yeah I guess thats kinda the only other option :D | 10:01 |
| josch | pandora: if that boots, you could inspect /var/log/apt/history.log to get a list of packages that got upgraded and maybe that will let us spot potential problems | 10:01 |
| josch | minute: sorry, there is yet another issue on top of the things i mentioned above... the CI on source.mnt.re currently fails with: Uploading artifacts as "archive" to coordinator... 500 Internal Server Error id=4610 responseStatus=500 Internal Server Error status=500 token=glcbt-64 | 10:02 |
| josch | minute: in the past this was because there was not enough space to store the artifacts | 10:02 |
| pandora | do u have a link to a system image by any chance ... i lost the node with the link to the pocket system images :D | 10:03 |
| josch | pandora: sure thing! https://source.mnt.re/reform/reform-system-image/-/jobs/4529/artifacts/browse | 10:04 |
| josch | minute: sorry, yet another issue is that https://mnt.re/system-image still points to artifacts that are 404 -- you have to press the "keep" button in the artifact overview once you adjust the job that you want this to point to | 10:05 |
| minute | pandora: sorry for the troubles! but you can boot a newly flashed system image via microsd, uboot will pick that up automatically | 10:06 |
| pandora | already flashing an sd card :D | 10:07 |
| minute | josch: ok, will fix that today | 10:07 |
| josch | minute: sorry to throw all these todos at you first thing in the morning | 10:07 |
| josch | especially after all the frustration during the past days... | 10:07 |
| josch | minute: i also wonder why pandora was not able to stop u-boot via uart -- that should work, no? | 10:08 |
| pandora | oh now the uboot worked | 10:12 |
| pandora | https://pastebin.com/gJPM5EzQ | 10:13 |
| pandora | but not a big difference | 10:13 |
| josch | huh this is super odd | 10:14 |
| josch | that's even before the initramfs starts | 10:14 |
| josch | (the last reform-tools upload could've potentially messed up the initramfs) | 10:14 |
| josch | (though i had tested that on the ls1028a) | 10:14 |
| pandora | that was at least part of the update | 10:14 |
| pandora | booting form sd rn | 10:14 |
| josch | okay, good | 10:15 |
| + andreas-e (~Andreas@2001:861:c4:f2f0::c64) | 10:16 | |
| - nybble (QUIT: Ping timeout: 272 seconds) (~sprang@user/csprng) | 10:18 | |
| pandora | ah pasting into firefox doesn't work on the pocket for some weird reason xD | 10:30 |
| josch | pandora: when you have a plain text file you can just do: cat file | curl -F 'arg=<-' https://mister-muffin.de/paste | 10:31 |
| josch | pandora: if you are on Debian you can install pastebinit and then you can do: cat file | pastebinit | 10:31 |
| + reform11163 (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 10:33 | |
| reform11163 | https://mister-muffin.de/p/8SKF.txt | 10:33 |
| pandora | thats great :D | 10:33 |
| pandora | *adding to my notes of helpful commands :D | 10:34 |
| josch | pandora: did you reboot after you installed systemd-cryptsetup? | 10:35 |
| pandora | i rebooted after every apt upgrade | 10:35 |
| josch | okay | 10:35 |
| josch | the interesting parts of that log are the upgrades of reform-tools as well as console-setup and keyboard-configuration as those packages can mess with the initramfs and thus with the boot process | 10:36 |
| josch | but then i do not understand why the kernel fails this early without any output | 10:36 |
| josch | pandora: since you are on the sd-card now, could you try upgrading only reform-tools and then reboot to check if the reform-tools upgrade is what kills it? | 10:37 |
| pandora | sure | 10:37 |
| minute | josch: yes, pandora should have been able to stop uboot via uart. | 10:38 |
| josch | minute: it seems to have worked in the end :) | 10:38 |
| pandora | ah... not enough free space :D | 10:39 |
| josch | pandora: to upgrade reform-tools? | 10:39 |
| minute | ahhh... yeah in the "factory" we have a set of commands to grow the root partition | 10:39 |
| minute | on the sd card | 10:39 |
| pandora | @josch yeah :D | 10:39 |
| josch | oh! | 10:39 |
| minute | i documented that in the forum, one moment | 10:39 |
| pandora | 257 mb available . 286 mb to download | 10:40 |
| josch | pandora: wait, that looks wrong | 10:40 |
| josch | is more than just reform-tools upgrading? | 10:40 |
| minute | https://community.mnt.re/t/deleted-the-emmc-partition/2100/2 | 10:41 |
| pandora | @josch u r right... i forgot to add --only-upgrade option | 10:44 |
| pandora | it is upgrading rn | 10:44 |
| pandora | but i will put the link into my notes on "helpful links" :D | 10:44 |
| josch | minute: in case the "sleep 1" in that script annoys you, you can replace it with "udevadm settle" | 10:45 |
| - reform11163 (QUIT: Quit: reform11163) (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 10:47 | |
| + reform13244 (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 10:48 | |
| reform13244 | https://mister-muffin.de/p/Tn21.txt | 10:48 |
| - reform13244 (QUIT: Remote host closed the connection) (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 10:48 | |
| pandora | it is booting again | 10:49 |
| josch | pandora: the "apt install" output looks good | 10:49 |
| josch | pandora: does it boot successfully? | 10:50 |
| pandora | it does | 10:50 |
| josch | uff that's a relieve | 10:51 |
| josch | okay, next please try upgrading just the kernel, that would be linux-image-mnt-reform-arm64 | 10:51 |
| pandora | kk | 10:51 |
| josch | theoretically this should also not change anything as the version remained the same at 6.8.12-1+reform20240605T104210Z | 10:51 |
| josch | pandora: if you could send us the "apt install" output again just to be sure that would also be great, thank you! | 10:52 |
| josch | though you could run into space problems now | 10:53 |
| minute | a new imx8mp video pll dynamic adjust patch by marex just landed in linux, this will probably eliminate some of my patches in the future | 10:54 |
| josch | oh how nice! i'm happy every time i can drop some patches when i rebase everything :) | 10:55 |
| + reform5504 (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 10:55 | |
| reform5504 | https://mister-muffin.de/p/owTA.tx | 10:55 |
| josch | i'm also very impressed how butter smooth the collabora rk3588 patch stack was so far -- the cadence patch is much more problematic than the dozens of rk3588 patches... | 10:56 |
| pandora | https://mister-muffin.de/p/owTA.txt | 10:56 |
| - reform5504 (QUIT: Client Quit) (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 10:56 | |
| pandora | yeah that killed it | 10:57 |
| josch | wow | 10:57 |
| josch | pandora: same issue with no output on serial? | 10:58 |
| pandora | https://pastebin.com/t5QSqxa0 | 10:58 |
| josch | pandora: you didn't do the setenv loglevel though | 10:59 |
| pandora | mom | 10:59 |
| josch | but it's probably going to remain the same... :/ | 10:59 |
| josch | minute: this might be a very tricky issue as the kernel sources remained the same (6.8.12-1) and only the dependencies in the build environment changed... | 11:00 |
| pandora | yeah not big changes | 11:01 |
| pandora | https://pastebin.com/pA8952hD | 11:01 |
| pandora | bdw... how long can i run the pocket without having the backplate on (since it is acting as the heat sink?) | 11:02 |
| minute | pandora: you can check with "sensors" | 11:02 |
| minute | pandora: it should shut down automatically when critical temp is reached (i think 90 celsius). but only if linux is running! without load or gpu action, it can run without heatsink for a longer while | 11:03 |
| minute | josch: sooo... the kernel was updated "slightly" and now it is broken? when was that? | 11:04 |
| pandora | ok was just curious because i had to remove the backplate to access the serial port... i might route it to the side at some point. The side slot seems to have a lot of space to make it permanently accessible | 11:04 |
| minute | pandora: yep, i've done that too | 11:05 |
| josch | minute: the kernel sources are the same -- pandora upgraded from 6.8.12-1+reform20240605T104210Z to 6.8.12-1+reform20240605T104210Z | 11:05 |
| josch | that's the same version but from different builds in the reform-debian-packages pipeline | 11:05 |
| minute | josch: bizarre... but what changed? | 11:05 |
| josch | the 6.8.12-1 bit is the upstream version | 11:05 |
| josch | which apparently didn't change, so the sources are the same | 11:06 |
| - andreas-e (QUIT: Quit: Leaving) (~Andreas@2001:861:c4:f2f0::c64) | 11:06 | |
| josch | the reform20240605T104210Z bit comes from the date of the top git commit in reform-debian-packages | 11:06 |
| minute | maybe we can sidestep this by upgrading to 6.9 P: | 11:06 |
| josch | this also remained the same, so it's the same patch stack | 11:06 |
| minute | but anyway, need to hurry to get to work and examine this | 11:06 |
| josch | minute: in all seriousness, this is something that you might want to try out first before sinking more time into this | 11:06 |
| josch | minute: because you have to spend the time to make 6.9 work anyways so maybe the problems goes away magically? | 11:07 |
| josch | and if it's still there, probably better to fix it in 6.9 | 11:07 |
| josch | minute: the thing that breaks 6.9 right now is qcalcld | 11:07 |
| josch | https://paste.debian.net/1321040/ | 11:08 |
| minute | josch: ok, i can fix that i guess :D | 11:08 |
| josch | maybe more is broken by my rebase but the CI cannot check because it's out of space (i guess?) and cannot store artifacts | 11:08 |
| minute | ok, i will clean all that up | 11:08 |
| josch | thank you | 11:09 |
| josch | i am maintaining a todo list for you locally, so no worries, nothing will be forgotten ;) | 11:09 |
| josch | pandora: thanks a lot for your very useful logs -- that was a great help! | 11:10 |
| minute | josch: which .deb could i delete on the repo to prevent upgrades in the meantime? | 11:10 |
| pandora | hehe u r welcome... stuff at work is broken too, so I had some time to spare | 11:11 |
| josch | minute: this one: https://mntre.com/reform-debian-repo/pool/main/l/linux/linux-image-6.8.12-mnt-reform-arm64_6.8.12-1%2Breform20240605T104210Z_arm64.deb | 11:11 |
| minute | i think it's better if apt throws an error temporaily than to have unbootable pockets | 11:11 |
| minute | josch: thanks!! | 11:11 |
| josch | thank you too :) | 11:11 |
| josch | just keeping the status quo is already quite a bit of work XD | 11:12 |
| minute | yeah :0 | 11:13 |
| minute | ok, deb moved away | 11:14 |
| minute | and update job pauses | 11:14 |
| minute | paused | 11:14 |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-64-230.tukw.qwest.net) | 11:14 | |
| + colinsane (~colinunin@97-113-64-230.tukw.qwest.net) | 11:15 | |
| pandora | is there a chance i can boot from the old kernel on my emmc? On my other linux laptop i can just select the old kernel in grub but i am not really familiar with uboot | 11:29 |
| minute | pandora: perhaps you could mount the boot partition from emmc while booted from microsd and just replace the kernel (overwrite it) | 11:55 |
| minute | we don't have a graphical uboot yet on imx8mp | 11:55 |
| minute | i mean, we kinda could, via hdmi | 11:55 |
| pandora | I’ll try that tonight or so… gonna be busy until then | 12:06 |
| josch | pandora: this was not really an upgrade. Apt replaced version X of your kernel image with version X. | 12:21 |
| josch | so our /boot/vmlinuz-* remained the same | 12:21 |
| + andreas-e (~Andreas@2001:861:c4:f2f0::c64) | 12:35 | |
| - andreas-e (QUIT: Remote host closed the connection) (~Andreas@2001:861:c4:f2f0::c64) | 12:35 | |
| + andreas-e (~Andreas@2001:861:c4:f2f0::c64) | 12:36 | |
| + chomwitt (~chomwitt@2a02:587:7a02:1a00:1ac0:4dff:fedb:a3f1) | 12:36 | |
| + mjw (~mjw@gnu.wildebeest.org) | 12:46 | |
| minute | josch: ok, so i will look into 6.9 now | 12:49 |
| josch | minute: it didn't build on the CI yet but you can test locally with: | 12:51 |
| josch | sh -xc 'BASESUITE=experimental; . ./setup.sh; cd linux; . ./build.sh' | 12:51 |
| josch | (CI fails due to the http error) | 12:51 |
| josch | and i'll probably be away until the evening | 12:51 |
| josch | minute: as usual, please feel free to force push any fixes into the linux6.9 branch | 12:58 |
| josch | that's this MR: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/41 | 12:58 |
| minute | josch: ok, i just checked out the linux6.9 branch | 12:59 |
| minute | josch: iirc just qcacld needs fixup, yes? | 12:59 |
| minute | ah and i need to clean the CI stuff | 12:59 |
| josch | minute: and maybe other fixes are needed -- i only made sure the patches apply but didn't build it yet because i only have a poor weak a311d here :D | 13:00 |
| minute | > /dev/sdb 98G 78G 16G 84% /mnt/gitlab-artifacts | 13:00 |
| minute | hmm | 13:00 |
| minute | anyway, how did i delete old artifacts... i think using some gitlab cli tool | 13:01 |
| josch | i used the web interface in the past: https://source.mnt.re/reform/reform-debian-packages/-/artifacts | 13:02 |
| josch | these eat 7.5 G each: https://source.mnt.re/reform/reform-system-image/-/artifacts | 13:02 |
| minute | ok, gonna click some trashcans | 13:03 |
| minute | ah yeah, it takes a while for the actual files to get deleted in the background i guess... | 13:04 |
| josch | aha! i made the same experience | 13:05 |
| josch | good to hear that it was not just me :) | 13:05 |
| josch | if you have a shell open, can you really verify that /mnt/gitlab-artifacts does not immediately get cleaned up? | 13:05 |
| minute | > /dev/sdb 98G 78G 16G 84% /mnt/gitlab-artifacts | 13:06 |
| minute | nothing :D | 13:06 |
| josch | o0 | 13:06 |
| josch | well, then at least i was not imagining things... | 13:07 |
| minute | non-monolithic webapps were a mistake... | 13:07 |
| minute | 22.4 GiB [##########] /2024_06_18 | 13:07 |
| minute | 14.9 GiB [###### ] /2024_06_05 | 13:07 |
| josch | oh progress bars! | 13:07 |
| minute | that's ncdu :D | 13:07 |
| josch | oh :D | 13:07 |
| josch | ncdu... like not-cdu -- looks like my new tool of choice ;) | 13:08 |
| - colinsane (QUIT: Remote host closed the connection) (~colinunin@97-113-64-230.tukw.qwest.net) | 13:08 | |
| minute | :D | 13:11 |
| minute | there's some discussions saying there's a 15 minute interval for cleanups | 13:11 |
| minute | so lets check again at 13:15 | 13:11 |
| minute | wow, my pc is on 6.3.0 :D | 13:13 |
| + colinsane (~colinunin@97-113-64-230.tukw.qwest.net) | 13:13 | |
| josch | "non-monolithic webapps were a mistake" -- i was sceptical of monolithic webapps but i have been using gitea on my own server for years (after a horrifying experience with gitlab -- too much maintenance cost for the gain) and i'm *very* happy with gitea. I know it now has some more forks and is quite actively developed. Not saying you should switch. Just saying that I changed my opinion over time | 13:13 |
| josch | regarding monolithic webapps. | 13:13 |
| minute | the best webapps are single executable + sqlite | 13:13 |
| josch | +1 | 13:14 |
| josch | you just know your stuff is in these files and that's it | 13:14 |
| minute | yeah | 13:14 |
| minute | > sudo apt install linux-image-6.9-amd64 | 13:14 |
| minute | josch: > Total artifacts size 7.46 GiB | 13:16 |
| minute | ok, 15 minute cleanup interval confirmed :3 | 13:16 |
| josch | \o/ | 13:16 |
| minute | i am once again cloning linux git | 13:18 |
| josch | i'm also excited to hear about your tests of 6.9 on rk3588. The patch stack is 13 patches smaller and i was able to remove four patches from other stacks. | 13:18 |
| minute | josch: oh great :3 | 13:19 |
| josch | minute: we could also throw this patch stack in if you like: https://lore.kernel.org/linux-phy/20240506124836.3621528-1-heiko@sntech.de/ | 13:25 |
| minute | josch: yesss please | 13:30 |
| minute | looks like qcacld2 needs only a 2-byte change | 13:30 |
| minute | josch: you wanted me to reassign a runner for...? | 13:34 |
| minute | josch: qcacld2 fix merged | 13:41 |
| minute | josch: > Refusing to run pipeline in the 'reform' namespace with BASESUITE not set to 'unstable' | 13:43 |
| minute | ++ echo 'to avoid accidentally tainting the repository at https://mntre.com/reform-debian-repo.' | 13:43 |
| josch | minute: since you disabled the repo updates, just comment out those lines in .gitlab-ci.yml | 13:45 |
| - colinsane (QUIT: Remote host closed the connection) (~colinunin@97-113-64-230.tukw.qwest.net) | 13:45 | |
| josch | minute: lets do the runner re-assignment another time and focus on 6.9 today -- the runner thing can wait and there is a MR to avoid forgetting about this | 13:46 |
| + colinsane (~colinunin@97-113-64-230.tukw.qwest.net) | 13:46 | |
| minute | josch: ok, thanks for confirming what i was thinking @ gitlab-ci | 13:48 |
| minute | this slaps https://mastodon.social/@Invictvs@indieweb.social/112654412861971027 | 13:48 |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 13:50 | |
| minute | from tomorrow i'll have a short vacation (till including wednesday) and i think i wanna sidegrade my pocket to a311d for that. lets see if i have time tonight | 13:52 |
| josch | fwiw the mipi dsi patch applies cleanly on top of 6.9 | 13:53 |
| minute | josch: btw it might be interesting to change the imx8mq patchstack to https://patchwork.kernel.org/project/linux-phy/cover/20240306101625.795732-1-alexander.stein@ew.tq-group.com/ | 13:53 |
| minute | josch: nice... i think it has a second part though | 13:53 |
| josch | "playing Bach" nice! :D | 13:54 |
| josch | ah there is a v15 now, i can take care of this later | 13:54 |
| josch | <= afk now until tonight | 13:55 |
| minute | josch: thank you and cya l8r! | 13:55 |
| - andreas-e (QUIT: Quit: Leaving) (~Andreas@2001:861:c4:f2f0::c64) | 13:55 | |
| * Guest7886 -> mjw | 13:56 | |
| pandora | @josch: “this was not really an upgrade. Apt replaced version X of your kernel image with version X” . Hehe ok then I don’t really understand what broke. If I replace x with the same x it shouldn’t brake, should it? :D | 14:44 |
| minute | pandora: actually just copying over the vmlinuz might not work because module signature might differ | 14:50 |
| pandora | I’ll give it a shot anyway because I have no idea what else to try really | 14:54 |
| josch | pandora: you can extract a working kernel together with fitting modules from the sd-card image that you flashed before. You need to copy vmlinuz to /boot (do not touch your working initrd) and you need to copy the corresponding /usr/lib/modules/ directory | 16:12 |
| pandora | Thx will test that when I get back home | 16:19 |
| + mark_ (~mjw@gnu.wildebeest.org) | 16:31 | |
| Twodisbetter | minute: hopefully as a little cheer me up for Friday, I just want you to know that MNT is referenced quite a bit on the Framework forums, and I believe that discussion has driven a bit of traffic your way. MNT deserves a lot more limelight for sure. | 17:11 |
| minute | Twodisbetter: nice | 17:27 |
| minute | reform next motherboards has first lifesigns, i have linux running on imx8mp over serial (not much use without ports though) | 17:28 |
| minute | josch: and also the initramfs | 17:28 |
| minute | because it has modules | 17:28 |
| minute | josch: linux6.9 branch has linux6.8.12 :/ https://source.mnt.re/reform/reform-debian-packages/-/jobs/4620/artifacts/browse/repo/pool/main/l/linux/ | 17:38 |
| Twodisbetter | minute: is there always something the newer kernel has that you want, or is this just a standard to move to the newer kernel? | 17:43 |
| pandora | @josch: it worked! I copied the vmlinuz and the modules and now it is booting again! thx for the help | 18:08 |
| minute | pandora: nice | 18:26 |
| + nybble (~sprang@user/csprng) | 19:46 | |
| jfred | minute: woohoo congrats on the reform next progress :D | 19:50 |
| minute | jfred: thaaanx | 19:50 |
| * mjw -> Guest7521 | 20:23 | |
| * mark_ -> mjw | 20:23 | |
| josch | minute: you have to run the pipeline with BASESUITE=experimental | 22:17 |
| josch | Twodisbetter: it depends on the module. For LS1028A for example, hardware support is nearly all upstreamed. Not so much for other modules though. Most in flux right now is hardware support for rk3588 | 22:18 |
| minute | josch: ohh... i thought i had clicked retry | 22:26 |
| minute | josch: oh i saw you just ran the pipeline! | 22:26 |
| josch | :) | 22:27 |
| minute | josch: i'm going on a 5-days holiday now (incl the weekend) but taking the a311d module with me, and i have imx8mplus and rk3588, so can make a few tests at home | 22:27 |
| josch | minute: anything on the TODO list that you'd like the community to take care of during your time afk? | 22:28 |
| minute | josch: just be good while i'm gone!! :D | 22:28 |
| josch | i hope you manage to have a time as worry free and relaxing as it can be during your holidays :) | 23:17 |
| josch | now i need to remember why past-me explicitly added an override BASESUITE=unstable to the build_linux job... hrm... | 23:18 |
| josch | lets undo this and see what breaks | 23:21 |
| josch | this is when i wish that past-me was writing more verbose commit messages... | 23:21 |
| minute | josch: ah, i just wanted to ask why the packages are still 6.8... | 23:43 |
| josch | minute: no worries, 6.9 now fails to build -- much better :) | 23:45 |
| josch | so in the end, no clue why I put BASESUITE=unstable in there... it seems to do the right thing without it and the wrong thing with it | 23:45 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!