| minute | josch: hmm normally this happens if you attach the usb cable too early/before pressing reset | 00:33 |
|---|---|---|
| - chomwitt (QUIT: Ping timeout: 256 seconds) (~chomwitt@2a02:587:7a0f:8900:1ac0:4dff:fedb:a3f1) | 00:37 | |
| - mjw (QUIT: Ping timeout: 256 seconds) (~mjw@gnu.wildebeest.org) | 01:28 | |
| - Gooberpatrol66 (QUIT: Ping timeout: 245 seconds) (~Gooberpat@user/gooberpatrol66) | 02:20 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 03:03 | |
| - nsc (QUIT: Ping timeout: 264 seconds) (~nicolas@i5E86146A.versanet.de) | 03:44 | |
| + nsc (~nicolas@166-121-142-46.pool.kielnet.net) | 03:46 | |
| - lexik (QUIT: Ping timeout: 252 seconds) (~lexik@93.185.97.218) | 03:49 | |
| + lexik (~lexik@93.185.97.218) | 03:50 | |
| - f_ (QUIT: Read error: Connection reset by peer) (~f_@fases/developer/funderscore) | 03:52 | |
| + f_ (~f_@fases/developer/funderscore) | 03:54 | |
| josch | minute: third time's a charm -- flashing worked now and my unit is charging from all adapters again :) | 04:17 |
| josch | at least the lpc shows negative current now | 04:17 |
| josch | the figure jumps erratically between 0.1 and 0.7 though | 04:17 |
| josch | but at least some current is coming in, I guess? XD | 04:18 |
| josch | err... between -0.1 and -0.7, of course | 04:18 |
| josch | i also hear a very quiet buzzing noise when the power cable is attached | 04:19 |
| - lexik (QUIT: Ping timeout: 252 seconds) (~lexik@93.185.97.218) | 04:34 | |
| + lexik (~lexik@93.185.97.218) | 04:35 | |
| - jagtalon (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~quassel@user/jagtalon) | 08:07 | |
| + jagtalon (~quassel@user/jagtalon) | 08:07 | |
| - jagtalon (QUIT: Client Quit) (~quassel@user/jagtalon) | 08:07 | |
| + jagtalon (~quassel@user/jagtalon) | 08:08 | |
| josch | and battery is at 100% \o/ | 08:36 |
| josch | took a few hours but it got there :) | 08:37 |
| - eibachd (QUIT: Ping timeout: 256 seconds) (~eibachd@p200300dcf7231c002026752f58af3308.dip0.t-ipconnect.de) | 09:28 | |
| + eibachd (~eibachd@p200300dcf7231c00213b3d1249890d94.dip0.t-ipconnect.de) | 09:29 | |
| - klardotsh (QUIT: Ping timeout: 256 seconds) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 09:48 | |
| - eibachd (QUIT: Ping timeout: 276 seconds) (~eibachd@p200300dcf7231c00213b3d1249890d94.dip0.t-ipconnect.de) | 09:52 | |
| + eibachd (~eibachd@2a01:599:313:f146:f042:2177:c81:fccf) | 09:53 | |
| - eibachd (QUIT: Ping timeout: 245 seconds) (~eibachd@2a01:599:313:f146:f042:2177:c81:fccf) | 10:14 | |
| minute | josch: ok, i see (and hear) mostly the same behavior. the buzzing is from turning the charger off and on at around 1kHz i think | 11:35 |
| josch | minute: thank you so much for your very quick workaround!! <3 | 11:49 |
| josch | you didn't have to work on that on a holiday, thank you :) | 11:49 |
| josch | right, makes sense then that the buzzing seems to be gone once batteries are charged to 100% | 11:51 |
| + mjw (~mjw@gnu.wildebeest.org) | 12:00 | |
| + eibachd (~eibachd@2a01:599:301:8b97:acec:523:146d:9756) | 13:04 | |
| josch | minute: so I'm investigating the CI failures on the mntre-builder. I saw that the docker systemd unit wasn't running. Running "dockerd --debug" manually revealed that there was a broken symlink from /var/lib/docker to /ramdisk/docker and re-creating the latter made docker start successfully. But now the new problem is, that the build runs out of space. It seems that /ramdisk is no longer actually a | 13:11 |
| josch | ramdisk and looking at /etc/fstab, the 32G tmpfs entry for /ramdisk is commented out. Do you know more about that? | 13:11 |
| + f_[xmpp] (fffdb90022@fases/developer/funderscore) | 13:13 | |
| minute | josch: ah. the server probably restarted and you need to mount the ramdisk. there is a script in / i think? or in /root? | 13:14 |
| josch | minute: found it! /make-zram.sh | 13:16 |
| josch | ah that script also contains the "mkdir /ramdisk/docker" thing, so i wasn't totally crazy :) | 13:17 |
| - eibachd (QUIT: Read error: Connection reset by peer) (~eibachd@2a01:599:301:8b97:acec:523:146d:9756) | 13:29 | |
| + eibachd (~eibachd@p200300dcf7231c0029d9b2b5ee8d0c06.dip0.t-ipconnect.de) | 13:30 | |
| - eibachd (QUIT: Ping timeout: 245 seconds) (~eibachd@p200300dcf7231c0029d9b2b5ee8d0c06.dip0.t-ipconnect.de) | 13:53 | |
| + eibachd (~eibachd@p200300dcf7231c006e7f86cd4b19823f.dip0.t-ipconnect.de) | 13:57 | |
| + julianwgs (~julian@p200300e9d74175008d93a5a3fbbdc935.dip0.t-ipconnect.de) | 13:57 | |
| minute | josch: yes. feel free to automate this on the server :D maybe we can make a systemd unit? | 14:38 |
| - f_[xmpp] (QUIT: Remote host closed the connection) (fffdb90022@fases/developer/funderscore) | 15:31 | |
| - Twodisbetter (QUIT: Remote host closed the connection) (2cc0e4ea1c@irc.cheogram.com) | 15:31 | |
| - eibachd (QUIT: Ping timeout: 260 seconds) (~eibachd@p200300dcf7231c006e7f86cd4b19823f.dip0.t-ipconnect.de) | 16:07 | |
| + eibachd (~eibachd@p200300dcf7231c003d801d802ff5b116.dip0.t-ipconnect.de) | 16:08 | |
| josch | minute: nope, display stays black when attempting to launch sway with Radeon RV710 withou anything interesting in dmesg :/ | 16:15 |
| josch | oh nevermind there is something! | 16:18 |
| - ggoes (QUIT: Quit: ZNC - https://znc.in) (~gregf@fsf/staff/ggoes) | 16:19 | |
| minute | josch: ahaa? | 16:20 |
| minute | josch: first thing you should rmmod etnaviv | 16:20 |
| minute | josch: in my experience otherwise etnaviv will be used as gpu and render to the discrete GPU framebuffer | 16:20 |
| minute | josch: also a good first thing to try is kmscube | 16:21 |
| minute | (in linux console) | 16:21 |
| + ggoes (~gregf@fsf/staff/ggoes) | 16:22 | |
| josch | progress? :D https://mister-muffin.de/p/e6UU.jpg | 16:23 |
| josch | and now something on dmesg https://paste.debian.net/1302786/ | 16:23 |
| minute | wow haha | 16:24 |
| josch | i'm just going to blacklist etnaviv as i don't need it on this system anyways | 16:24 |
| minute | josch: these errors in the log are expected/ok | 16:24 |
| minute | ah that's a very old ati gpu | 16:25 |
| minute | did someone test these on ARM before? | 16:26 |
| josch | not to my knowledge | 16:26 |
| josch | i have a bunch of cards here that i'm just throwing at the ls1028a and see what happens | 16:26 |
| minute | it's possible that you will run into driver incompatibilities if they make assumptions about the cpu | 16:26 |
| josch | at work we had to dispose of a bunch of machines recently so there is a few choices available to me :) | 16:26 |
| minute | i remember someone had to fix a lot of floating point stuff in the kernel for amdgpu | 16:26 |
| minute | (for arm) | 16:26 |
| josch | hah :) | 16:27 |
| josch | i'm surprised your nividia card just worked then? | 16:27 |
| julianwgs | Hey all, I just created this issue: https://source.mnt.re/reform/reform/-/issues/16 | 16:28 |
| julianwgs | I am very interested what you have to say about it :) | 16:28 |
| julianwgs | Also I am having issues exporting netlists or bom with kicad-cli in docker. I am getting this errors: "Failed to load schematic file". I suspect it has something to do with the version differences. The files were created with 5 and the docker images are only available for 7. However it works perfectly fine using the GUI. Has anyone experience with this and could help me? | 16:28 |
| minute | josch: i think maxwell architecture is one of the best supported by nouveau | 16:29 |
| minute | but it uses quite a lot of CPU, which is not desirable | 16:29 |
| minute | would be more fun to try turing though because of https://docs.mesa3d.org/drivers/nvk.html | 16:30 |
| minute | welcome julianwgs | 16:30 |
| minute | julianwgs: we should upgrade the files to kicad 7.0.6 i think | 16:31 |
| minute | julianwgs: would be good to do this in a new branch | 16:31 |
| minute | julianwgs: maybe you can attempt that, fork the repo and make a new branch like kicad-ci and then open and re-save the files and commit them? | 16:32 |
| - yankcrime (QUIT: Quit: "") (~nick@gw.tetromino.io) | 16:32 | |
| + yankcrime (~nick@gw.tetromino.io) | 16:34 | |
| julianwgs | When I open an old file with a new version of KiCAD the rescue helper opens and I need to make decisions whether to rescue symbols (which might break the schematic). Also it is not a diff on the file, but a complete new file (.kicad_sch instead of .sch). I am kind of anxious I might break something in a very subtle way. Are you sure this is the best approach? As I said I am new to KiCAD coming from Eagle :) | 16:40 |
| - XYZ (QUIT: Ping timeout: 245 seconds) (~XYZ@78-80-26-245.customers.tmcz.cz) | 16:41 | |
| josch | gmscube on the Radeon: https://mister-muffin.de/p/QDMu.jpg | 16:56 |
| josch | is the cube supposed to move? because it does not | 16:56 |
| josch | the gmscube output says "renderer: AMD RV710" which sounds okay? https://mister-muffin.de/p/bWrx.jpg | 16:56 |
| josch | still only a black screen with sway | 16:57 |
| josch | now this one is in: | 16:57 |
| josch | 0001:01:00.0 VGA compatible controller: NVIDIA Corporation GP107GL [Quadro P600] (rev a1) | 16:57 |
| josch | but the monitor doesn't even turn on. dmesg says all firmware got loaded and a framebuffer device got created. Maybe this is a problem on the monitor side with which the card is connected using a display port to DVI adapter? dmesg here: https://mister-muffin.de/p/_8kC.txt | 17:02 |
| josch | now taking a break -- more cards will get tested in the following few days :) | 17:03 |
| minute | josch: yes, the cube should move | 17:17 |
| minute | maybe there's some interrupt problem. MSI is enabled/not disabled, yeah? | 17:17 |
| + reform1358 (~josch@win7148.informatik.uni-wuerzburg.de) | 17:28 | |
| reform1358 | Linux reform 6.6.8-reform2-arm64 #1 SMP Debian 6.6.8-1+reform20231225T111019Z1 (2023-12-25) aarch64 GNU/Linux | 17:28 |
| reform1358 | hello from a311d with linux 6.6.8 | 17:28 |
| reform1358 | neil's new patch stack seems to at least get me sway | 17:29 |
| reform1358 | minute: what are the things that i should test on a311d to make sure 6.6 works as it should with neils recent patch stack instead of the old one? | 17:29 |
| minute | reform1358: does the display look good in sway, no vertical offset or glitches? | 17:30 |
| minute | reform1358: can you exit and reenter sway? | 17:30 |
| reform1358 | okay | 17:31 |
| - reform1358 (QUIT: Client Quit) (~josch@win7148.informatik.uni-wuerzburg.de) | 17:31 | |
| + reform10637 (~josch@win7148.informatik.uni-wuerzburg.de) | 17:32 | |
| reform10637 | okay, i exited sway using "pkill sway" (couldn't come up with the swaymsg command) and am back | 17:32 |
| reform10637 | things seem to look fine | 17:32 |
| minute | reform10637: great | 17:32 |
| reform10637 | i guess i'm next upgrading my existing system for more real-life testing | 17:32 |
| reform10637 | sound works, wifi works | 17:36 |
| reform10637 | now on to the next thing... | 17:37 |
| - reform10637 (QUIT: Quit: Leaving) (~josch@win7148.informatik.uni-wuerzburg.de) | 17:37 | |
| AbortRetryFail | wasnt it just `swaymsg exit`? | 17:41 |
| josch | oooh i was reading the wrong man page | 17:49 |
| josch | this is not documented in swaymsg(1) but in sway(5) | 17:49 |
| josch | "See sway(5) for a list of commands." -- being able to read is supposed to have advantages XD | 17:49 |
| AbortRetryFail | man sway-ipc | 18:02 |
| AbortRetryFail | yea the manpages about it are sorta confusing | 18:02 |
| + reform7795 (~josch@win7148.informatik.uni-wuerzburg.de) | 18:16 | |
| reform7795 | hello from imx8mq on motherboard 2.5 | 18:16 |
| reform7795 | i just successfully suspended and resumed three times | 18:16 |
| reform7795 | with kernel 6.6 and reform-tools 1.34 | 18:16 |
| reform7795 | ah, unfortunately Twodisbetter isn't here right now | 18:21 |
| - reform7795 (QUIT: Quit: Leaving) (~josch@win7148.informatik.uni-wuerzburg.de) | 18:22 | |
| josch | i now put the imx8mq and motherboard 2.5 together with my existing system (hence me being logged in as josch again) | 18:34 |
| josch | and resume from suspend still fails with the original symptoms of not being able to recover my nvme | 18:35 |
| josch | nvme 0001:01:00.0: Unable to change power state from D3hot to D0, device inaccessible | 18:39 |
| josch | nvme nvme0: I/O 0 QID 0 timeout, disable controller | 18:39 |
| josch | nvme nvme0: Identify Controller failed (-4) | 18:39 |
| josch | nvme nvme0: Disabling device after reset failure: -5 | 18:39 |
| josch | the other errors in dmesg are from a higher level like from jbd2, ext4fs etc | 18:40 |
| josch | EXT4-fs (dm-2): I/O error while writing superblock | 18:40 |
| josch | well, it was an experiment :) | 18:40 |
| josch | back to a311d now | 18:40 |
| minute | josch: how do you like a311d so far? | 18:51 |
| josch | i must admit i was skeptical of the idea of stacking three boards on top of each other | 18:52 |
| josch | but the heatsink hex screws are an incredible improvement :D | 18:52 |
| josch | bluetooth connection is somehow worlds better compared to my usb adapters -- suddenly i can listen to stuff throughout my whole flat without connection drops | 18:52 |
| josch | wifi is okay but even with the powersave hack it drops occasionally and i have to reconnect | 18:53 |
| josch | but it's rare for that to happen | 18:53 |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 18:53 | |
| josch | it seems more stable for connections to non 5GHz networks | 18:53 |
| josch | but i only have a few days of experience :) | 18:53 |
| josch | the increase in raw computing power is the most amazing thing -- it's such a massive improvement | 18:54 |
| josch | since suspend didn't work with my imx8mq either, i'm hard pressed to find a disadvantage of the a311d for me | 18:54 |
| minute | that sounds great! | 18:56 |
| josch | there are the occasional graphical glitches in firefox but it's acceptable | 18:58 |
| josch | it's really bad in chromium which becomes nearly unusable | 18:58 |
| josch | but those glitches do not happen when booting Debian unstable from an sd-card so i guess they exist because i'm using mesa and other stuff from bookworm and not from unstable | 18:58 |
| vkoskiv | I haven't had glitches, or a drop in WiFi since applying that workaround | 18:58 |
| vkoskiv | And yeah re: bluetooth, it's much nicer than I'm used to on Linux | 18:58 |
| minute | josch: yes, old mesa is likely the issue. i also remember having had gpu issues in firefox or chromium because of our old patched xwayland | 19:00 |
| josch | right, i'll see if i can backport newer mesa to bookworm for my personal use and/or for reform.debian.net and see if it makes a difference | 19:01 |
| josch | because i'm definitely staying with a311d and will not go back to imx8mq XD | 19:01 |
| minute | haha :D | 19:01 |
| minute | i'm happy you got a performance boost | 19:01 |
| josch | i now finally managed to test kernel 6.6 and reform-tools 1.34 on all three SoMs | 19:02 |
| josch | minute: so i think we can push those into the repo soon unless you want to do some testing yourself before that? | 19:02 |
| josch | i'd like to build a new system image on source.mnt.re but the "500 Internal Server Error" is still there and i have to dig a bit deeper to debug this at some point: https://source.mnt.re/reform/reform-system-image/-/jobs/2721 | 19:03 |
| josch | minute: about your suggestion about MSI above, this looks good, right? https://paste.debian.net/1302808/ | 19:06 |
| minute | cool, something like synergy for wayland/wlroots https://github.com/feschber/lan-mouse | 19:34 |
| minute | josch: yes, green lights from me | 19:38 |
| josch | minute: cool, thanks! Once you find some time, could you also tell me what command the cron script that updates https://mntre.com/reform-debian-repo uses to download the artifact from the reform-debian-packages pipeline? | 19:39 |
| josch | The issue is, that it seems that running the pipeline with different pipeline variables set will overwrite the last default artifacts for the same branch | 19:40 |
| josch | So in MR 88 I added some code to prevent this from accidentally happening | 19:41 |
| minute | will check a bit later, currently in subway | 19:42 |
| josch | indirect pocket reform flex? ;) | 19:42 |
| josch | okay, reform-debian-packages pipeline is running... fingers crossed! | 20:26 |
| josch | the reform-tools update was in the pipeline since october last year -- i hope nothing breaks XD | 20:27 |
| + Twodisbetter (2cc0e4ea1c@irc.cheogram.com) | 20:29 | |
| josch | Twodisbetter: howdy! I tested suspend with imx8mq and motherboard 2.5 today but it failed with the same error as before :( | 20:31 |
| minute | likely som related? | 20:35 |
| josch | since switching the mainboard didn't change the behaviour i think so, yes | 20:41 |
| josch | also since i had tried three different nvme drives including the one that works for Twodisbetter | 20:41 |
| - eibachd (QUIT: Ping timeout: 256 seconds) (~eibachd@p200300dcf7231c003d801d802ff5b116.dip0.t-ipconnect.de) | 21:05 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 21:09 | |
| + eibachd (~eibachd@p200300dcf7231c001fa87565e7775b4e.dip0.t-ipconnect.de) | 21:11 | |
| - Twodisbetter (QUIT: Ping timeout: 252 seconds) (2cc0e4ea1c@irc.cheogram.com) | 21:40 | |
| + chomwitt (~chomwitt@2a02:587:7a0f:8900:1ac0:4dff:fedb:a3f1) | 21:49 | |
| - julianwgs (QUIT: Quit: Leaving.) (~julian@p200300e9d74175008d93a5a3fbbdc935.dip0.t-ipconnect.de) | 22:14 | |
| josch | minute: the error 500 probably has to be tracked down by you as you have access to the log files. According to issues on gitlab.com, the reason for those errors should be in /var/log/gitlab/gitlab-rails/production.log or /var/log/gitlab/gitlab-workhorse/current | 23:21 |
| minute | josch: do you have a link that produces the error 500? | 23:23 |
| minute | grepping workhorse log for "error" i find: > {"correlation_id":"01HK64C95YW726GSJP6W46EY39","duration_ms":877,"error":"badgateway: failed to receive response: context canceled","level":"error","method":"GET","msg":"","time":"2024-01-02T22:23:19Z","uri":"/reform/reform-debian-packages/-/commits/999e39a7aaf0653d2c60632c193fa62edec12e3c"} | 23:24 |
| josch | minute: that's the one i cancelled right now | 23:25 |
| josch | minute: here is an example: https://source.mnt.re/reform/reform-system-image/-/jobs/2721 | 23:26 |
| josch | the reform-debian-packages jobs manage to upload just fine | 23:26 |
| minute | ok so the error is while uploading artifacts | 23:26 |
| josch | so i'm currently trying only building the imx8mq image -- maybe the artifact size is too large or the transfer takes too long? | 23:26 |
| minute | > /dev/sdb 98G 93G 443M 100% /mnt/gitlab-artifacts | 23:27 |
| minute | that disk is full ;3 | 23:27 |
| josch | aha! :D | 23:27 |
| minute | hmm normally default artifacts expiration is 3 days | 23:32 |
| minute | but artifacts from latest successful jobs are kept | 23:32 |
| josch | maybe they are kept for each branch that exists and stale branches should be cleaned up? | 23:33 |
| josch | yeah, these are 2 months old and not useful anymore, for example: https://source.mnt.re/reform/reform-system-image/-/jobs/2388/artifacts/browse | 23:35 |
| josch | for reform-system-image i'm going to delete test, reform.debian.net, u-boot and ls1028a-unified | 23:36 |
| josch | minute: okay? | 23:36 |
| minute | josch: ok, lets try that | 23:36 |
| josch | i don't know whether artifacts belonging to those branches are cleaned up automatically somehow | 23:36 |
| minute | they should, i can check after you deleted them | 23:37 |
| minute | and yes, i think stale branches with built images are the culprit | 23:37 |
| josch | done | 23:37 |
| minute | ok, i guess some worker task has to run to do the job... maybe i should restart gitlab | 23:37 |
| minute | ah, there is supposed to be "run housekeeping" in admin | 23:39 |
| josch | i can also delete unite-them, unite-builds, a311d-v2, a311d, ls1028a-dev, v3-kde-tweaks, sysimage-v3, sysimage-v2, ls1028a, mmdebstrap, ci | 23:39 |
| josch | reform-debian-packages also has 27 branches :D | 23:41 |
| josch | and then there are all the forked projects as well | 23:41 |
| + XYZ_ (~XYZ@37-48-49-146.nat.epc.tmcz.cz) | 23:43 | |
| minute | hm https://gitlab.com/gitlab-org/gitlab/-/issues/383159 | 23:43 |
| josch | minute: I deleted some more branches that were either my own or which belonged to merge requests that got merged | 23:49 |
| josch | there is lots of stuff left over but this should already free quite a bunch of space up | 23:49 |
| minute | josch: i think until the artifacts expire (3 days?) we have to delete the artifacts manually here https://source.mnt.re/reform/reform-system-image/-/artifacts | 23:49 |
| minute | josch: i read through the issue at gitlab and they don't automatically delete the artifacts of deleted branches | 23:50 |
| josch | ah okay | 23:50 |
| josch | let me delete some artifacts of my branches manually | 23:50 |
| minute | yes | 23:50 |
| josch | i only need a little bit of space for now | 23:50 |
| minute | handily they show the size in that list | 23:50 |
| josch | non-handily they show up even when expired | 23:52 |
| minute | so far the needle does not move @ df | 23:52 |
| josch | interesting... i keep deleting stuff... | 23:53 |
| josch | and yeah at the top it keeps showing "Total artifacts size 32.18 GiB" | 23:53 |
| minute | haha oof > An error occurred while deleting. Some artifacts may not have been deleted. | 23:55 |
| josch | ... | 23:55 |
| minute | (i tried to delete the lateset artifacts of a311d-v2) | 23:55 |
| josch | XD | 23:55 |
| josch | hah, it changed! | 23:57 |
| josch | minute: did you do something? | 23:57 |
| josch | Total artifacts size 6.04 GiB | 23:57 |
| josch | those 6 GiB correspond to the only two artifacts that should be remaining for reform-system-image, the one for the main branch and the one for the staging branch | 23:58 |
| minute | no, didn't do anything | 23:58 |
| minute | i just tried to delete those artifacts again and checked network inspector to grab the response | 23:58 |
| minute | > "errors": ["Artifacts (2860,2861,2864) not found"] | 23:59 |
| josch | magic ¯\_(ツ)_/¯ | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
