| + matt2 (~matt@75-166-133-128.hlrn.qwest.net) | 00:09 | |
| + erle (~erle@user/erle) | 00:11 | |
| minute | josch: damn, gitlab server disks are 100% full | 00:24 |
|---|---|---|
| minute | hmm, for some reason gitlab stores 500gb artifacts on internal disk instead of nfs | 00:28 |
| minute | ah on nfs we have only backups. makes sense | 00:28 |
| minute | (also full) | 00:28 |
| - xandy (QUIT: Ping timeout: 246 seconds) (~xandy@user/xandyv) | 00:32 | |
| minute | backup drive was full so gitlab started to put backups on the local disk somehow | 00:32 |
| minute | josch: artifacts size exploded 3 days ago to 107 GB, what happened? | 00:33 |
| minute | > 107.6 GiB [####################################] /2025_12_21 | 00:33 |
| bremner | </3 gitlab | 00:33 |
| minute | bremner: well it's mostly operator error | 00:34 |
| bremner | I guess its just a.big complicated system | 00:34 |
| minute | bremner: or me being underfunded and don't have admins who are regularly checking that everything is smooth | 00:34 |
| minute | anyway, it's usable again and will soon have to set up more monitoring + automated cleanup | 00:38 |
| - wickedshell (QUIT: Ping timeout: 244 seconds) (~wickedshe@c-98-60-138-166.hsd1.nm.comcast.net) | 00:40 | |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@46-144-38-170.biz.kpn.net) | 00:42 | |
| - matt2 (QUIT: Quit: WeeChat 4.7.2) (~matt@75-166-133-128.hlrn.qwest.net) | 01:03 | |
| - pomel0 (QUIT: Remote host closed the connection) (~pomel0@user/pomel0) | 01:48 | |
| + pomel0 (~pomel0@user/pomel0) | 01:48 | |
| + wickedshell (~wickedshe@c-98-60-138-166.hsd1.nm.comcast.net) | 02:46 | |
| swivel | minute: if you wanted help with such things there's likely many qualified and willing folks among your supporters | 02:56 |
| - chrcav (QUIT: Quit: Lost terminal) (~chrcav@user/chrcav) | 03:13 | |
| josch | minute: 3 days ago was December 21 when I wrote you here: "minute: there are currently aroun 400 GiB of artifacts in the reform-system-image repo and i suspect that uploading fails because there is no more space. I've cleaned up some of my branches (so the associated artifacts should be auto-removed soon as well)" | 03:16 |
| josch | 107 GB is only five runs of reform-system-image pipeline (each is 22 GiB) | 03:17 |
| josch | this could be (i don't remember the events precisely) a simple sequence of: push a feature branch to reform/reform-system-image, push the branch to josch/reform-system-image as well because on the "reform" namespace I cannot build with custom options, then run the pipeline in the "josch" namespace with custom options, then merge the feature branch in the "reform" namespace so that another rebuild is | 03:19 |
| josch | done on the "main" branch | 03:19 |
| josch | okay, that's only four runs but the fifth run is likely a single second manual run of the pipeline somewhere | 03:20 |
| josch | what i was investigating 3 days (and ultimately fixed) ago was a problem where the final image had LINUX_KERNEL_CMDLINE="quiet" in /etc/default/flash-kernel but *only* if the image was built on CI. Here is the fix: https://source.mnt.re/reform/reform-system-image/-/merge_requests/141/diffs?diff_id=7734&start_sha=0c8ea72ba83685610f5c1bc01bdba6a2d82543f9 | 03:26 |
| josch | the problem was that I tested the presence of /etc/default/flash-kernel which is a file *outside* the chroot whicn on my Reform is evidently present but on the CI runner it is not present and that's why the problem only showed on the CI builds and that's what motivated the five reform-system-image runs to track this down before I found the culprit | 03:28 |
| - paperManu_ (QUIT: Ping timeout: 244 seconds) (~paperManu@146.71.9.156) | 03:36 | |
| minute | josch: ah, thanks for all the context. and yeah, the rest of the storage was filled up by too many backups being kept | 03:50 |
| josch | nachtschicht o0 | 04:24 |
| - paperManu (QUIT: Ping timeout: 240 seconds) (~paperManu@146.71.9.156) | 04:43 | |
| - spew (QUIT: Quit: stay prayed up) (~spew@user/spew) | 04:53 | |
| + XYZ (~XYZ@37-48-3-97.nat.epc.tmcz.cz) | 06:01 | |
| - XYZ (QUIT: Read error: Connection reset by peer) (~XYZ@37-48-3-97.nat.epc.tmcz.cz) | 06:07 | |
| + XYZ (~XYZ@37-48-3-97.nat.epc.tmcz.cz) | 06:08 | |
| - XYZ (QUIT: Read error: Connection reset by peer) (~XYZ@37-48-3-97.nat.epc.tmcz.cz) | 06:28 | |
| + XYZ (~XYZ@37-48-3-97.nat.epc.tmcz.cz) | 06:29 | |
| - XYZ (QUIT: Ping timeout: 240 seconds) (~XYZ@37-48-3-97.nat.epc.tmcz.cz) | 06:48 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 06:58 | |
| - XYZ (QUIT: Read error: Connection reset by peer) (~XYZ@194-212-250-123.customers.tmcz.cz) | 06:59 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 06:59 | |
| - XYZ (QUIT: Read error: Connection reset by peer) (~XYZ@194-212-250-123.customers.tmcz.cz) | 07:28 | |
| + XYZ (~XYZ@194.212.250.123) | 07:29 | |
| + reform22198 (~jinnak@2a10:a5c0:54:ee03:12d9:ef2f:6abc:9b62) | 07:54 | |
| * reform22198 -> jinna | 07:59 | |
| jinna | Heyo! I swapped out my Pocket SoM from BPi to RK. My unencrypted root is on nvme so what I'm looking for is some guide or guidance how to flip my root back to that while keeping my boot on the SoM eMMC. I'm not finding that guide though in the handbook or community so I'm a bit confused. None of the various reform- tools seem to do what I want which is even more confusing since I'd thought this would be a common scenario. Any tips w | 08:04 |
| - XYZ (QUIT: Ping timeout: 240 seconds) (~XYZ@194.212.250.123) | 08:08 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 08:11 | |
| jinna | reform-boot-config seems the least wrong option, but it fails with the error: /boot device /dev/mmcblk1p1 from /etc/fstab in /dev/nvme0n1p1 doesn't exist | 08:15 |
| jinna | The error is probably right, but I fail to see why that should be a fatal error and not an expected state it should fix, like the manual implies it would? | 08:16 |
| jinna | Ah, I guess the key is adding the --no-copy-old-boot flag, which gets me to a whole new set of errors. I guess that answers it, but still wishing this was actually documented and linked from the SoM upgrade PDF -_- | 08:20 |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194-212-250-123.customers.tmcz.cz) | 08:22 | |
| + XYZ (~XYZ@194.212.250.123) | 08:23 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194.212.250.123) | 08:34 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 08:35 | |
| - enwu (QUIT: Ping timeout: 264 seconds) (~enwu@user/enwu) | 09:01 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194-212-250-123.customers.tmcz.cz) | 09:01 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 09:03 | |
| + enwu (~enwu@user/enwu) | 09:15 | |
| josch | jinna: hi, i'm maintaining the tools and am happy abotu your feedback and would be glad to document the things that you find lacking. I think the tool you want to use is reform-emmc-bootstrap -- from reading its man page or --help output, can you confirm that this is what you are looking for? | 09:17 |
| jinna | Let me have a peek | 09:18 |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194-212-250-123.customers.tmcz.cz) | 09:18 | |
| jinna | Well, it *sounds* like it's close, but it insists to be run from a rescue sd and that's annoying | 09:19 |
| + XYZ (~XYZ@194.212.250.123) | 09:20 | |
| jinna | I'm already booted to a usable OS on the new SoM and can easily unmount /boot, but can't unmount / like it wants | 09:20 |
| jinna | I feel like what I want doesn't need to completely reformat the emmc like that implies | 09:21 |
| josch | jinna: how are you currently booting your system on your nvme? is your /boot on an sd-card? | 09:24 |
| jinna | What I've done now that I *think* will work once I reboot is: run reform-boot-config --emmc --no-copy-old-boot ssd. That did a whole buncha errors because the system on the nvme is old and lacking the right kernel etc, but it did update the fstab on the nvme and didn't revert that change. Then I chrooted into the nvme and ran a full apt update/upgrade/full-upgrade. Then I took a diff of the non-automatic installed packages on both a | 09:24 |
| josch | (sorry for sluggish replies, i'm currently travelling by train and internet is very spotty) | 09:24 |
| jinna | Boot was on the old BPi SoM, hence now inacessible and I'm booted from the RK SoM eMMC | 09:25 |
| jinna | If I ignore any reform tooling, in theory I could've just gone straight to the chroot, updated fstab myself to point to the different eMMC dev number as per the RK and did the apt'ing to get a usable /boot onto the new SoM. | 09:28 |
| jinna | (as a side note, really loving the speed of the RK and reliable WiFi of the intel & non-faraday-cage backplate :D) | 09:30 |
| josch | ooooh i think now i understand. You currently have a whole bootable system on emmc (including /boot and /) and you would like to destroy the /boot of that system and replace it with one which boots your system on nvme -- i think this situation is indeed unique. Usually users boot a system from sd-card and then do the setup. But i think the tool that should be able to do the job is reform-boot-config. | 09:32 |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194.212.250.123) | 09:32 | |
| jinna | Oh, I realy thought this would come up with anyone moving from one SoM to the other. Feels weird I'd have boot on an external SD card o: | 09:33 |
| + XYZ (~XYZ@194.212.250.123) | 09:34 | |
| jinna | It's how the RK comes out of the box | 09:34 |
| josch | jinna: the problem with the tools is exactly that: every setup is unique and it's hard to cater for all this diversity. I would like to support your use-case. What are the "whole buncha errors" you mention above? | 09:34 |
| jinna | Let me dig that up from the backlog | 09:34 |
| josch | jinna: my rk3588 pocket reform arrived unformatted if i remember correctly. I don't work for MNT so i'm not so close in the loop about what gets flashed for new customers. :) | 09:35 |
| josch | (and that's also why i cannot fix the docs in the handbook or the quickstart manual but i can and want to improve the docs for reform-tools) | 09:36 |
| jinna | Ah, default gnome-terminal is a bit too stingy on the backlog length to get back to it. But, it was basically crashing and burning on creating the new boot images because the nvme system didn't have the right kernel things, just like the manual warns | 09:36 |
| jinna | Gooootcha, makes sense | 09:37 |
| jinna | It had a pre-installed debian with both boot and / on the eMMC. | 09:37 |
| jinna | Same thing with BPi | 09:37 |
| jinna | So you can just plop it in and get booting right away, but then need to figure out on your own how to get back to using the nvme for root | 09:38 |
| josch | right, that could certainly be improved | 09:41 |
| jinna | Ok, I _think_ this will boot, so I'll do a reboot and see what happens. I do have other computers to get back if it self-destructs | 09:41 |
| josch | it should somewhere in the docs say that before swapping the SoM you should have your packages up-to-date | 09:42 |
| jinna | I get that advice, but it' | 09:42 |
| jinna | it's super un-fun to follow | 09:42 |
| josch | the reason for this requirement is that when you remove the old /boot because the old emmc is gone, it's like doing a "sudo rm /boot/vmlinuz-" and that file is shipped by the linux-image-*-mnt-reform-arm64 which has to get re-installed to restore the file | 09:43 |
| josch | and you can only re-install the file if the original was up-to-date because the mirrors will not keep old versions | 09:43 |
| jinna | "I have this new much faster cpu, why should I go do a whooole buncha slow apt'ing with the old slow one" :D | 09:43 |
| josch | jinna: which docs are you reading right now that are not fun? I'm sure minute would like to hear that feedback. | 09:43 |
| jinna | Just the general guidance of update first. But, I don't think the RK upgrade bundle docs even mention that tbh | 09:44 |
| jinna | This comes printed with it: https://mntre.com/documentation/pocket-reform-rcore-rk3588-manual.pdf | 09:45 |
| jinna | Zero mention of any migration, which if you've bought an upgrade bundle, should probably be mentioned | 09:46 |
| josch | hrm... i guess we could improve the tooling in the sense that we could either have a new tool or extend an existing tool so that it does the "whooole buncha apt'ing" in a chroot on the new SoM for you | 09:48 |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194.212.250.123) | 09:48 | |
| jinna | The BPi upgrade bundle had some notes on migration I think, that's where I vaguely remembered the "update first" which I promptly ignored | 09:48 |
| josch | minute: helpful customer feedback can be found above :) | 09:48 |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 09:49 | |
| jinna | Yeah, I think that would be a useful path. I worry a bit though because there's a whole handful of reform-* tools that sound fairly similar at face value. So it can be difficult to pick the right one with the right flags if it's not super explicitly documented in a thing you will naturally end up reading while figuring out how to physically plug it in. | 09:50 |
| josch | jinna: thank you i think i now have a good understanding of your stituation and can reproduce it myself and see how we can improve the tooling to make this easier without having to flash a system image to an sd-card first | 09:50 |
| jinna | Great, happy to hear it was helpful | 09:51 |
| jinna | I'll cross my fingers and do that reboot now o_O | 09:51 |
| - jinna (QUIT: Quit: brb hopefully) (~jinnak@2a10:a5c0:54:ee03:12d9:ef2f:6abc:9b62) | 09:52 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194-212-250-123.customers.tmcz.cz) | 09:59 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 10:00 | |
| - jn (QUIT: Ping timeout: 260 seconds) (~quassel@user/jn/x-3390946) | 10:03 | |
| + jn (~quassel@user/jn/x-3390946) | 10:03 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@194-212-250-123.customers.tmcz.cz) | 10:13 | |
| + XYZ (~XYZ@194.212.250.123) | 10:15 | |
| + jinna (~jinna@109-204-184-64.cust.valoonet.fi) | 10:16 | |
| jinna | Yeaaah, it did not boot. Just a black screen. I guess u-boot is unhappy somehow and I should've run update-uboot. Time for that rescue SD I guess. | 10:16 |
| josch | Okay. Once you booted from that, you can use reform-emmc-bootstrap which will call "apt upgrade" for you (but only for the kernel images) | 10:20 |
| jinna | ack, sounds like a plan. | 10:21 |
| josch | and when you find a bug, you know who to ping :) | 10:22 |
| - nybble (QUIT: Ping timeout: 260 seconds) (~sprang@user/csprng) | 10:26 | |
| jinna | reform-emmc-bootstrap worked well and now it boots again. Thanks! | 11:10 |
| josch | perfect, thank you for the feedback! | 11:15 |
| minute | jinna: thanks for the feedback! | 11:26 |
| minute | unfortunately the whole rk3588 situation looks very dire suddenly, as i don't know yet if we can get any more of them | 11:26 |
| minute | because there's no more ram | 11:26 |
| minute | i wrote bpi about their rk3588 module and they don't even want to name a price at the moment... | 11:27 |
| minute | lets see what radxa say | 11:28 |
| minute | firefly are not responding since a few days | 11:28 |
| minute | thundercomm modules have to be reengineered because the ram+emmc combined chip is now gone from the market | 11:29 |
| minute | (that they used) | 11:29 |
| minute | not sure what to do. perhaps rpi cm5. | 11:29 |
| minute | we have enough rk3588 stocked for the next campaign units. but we might need to pause all other sales soon | 11:30 |
| josch | nobody could've forseen this sudden crisis -- it's such a mess :( | 11:31 |
| kfx | the time has come for esp32 laptops | 11:31 |
| minute | yeah i found an email from july that had warned me about this, but the subject looked to generic so i overlooked it back then :/ | 11:31 |
| minute | kfx: yeah, might need to go retrooo | 11:31 |
| + voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net) | 11:32 | |
| minute | we can get a 4gb riscv module from bpi | 11:32 |
| kfx | can't be hurt by a RAM shortage if you can't address more than 64kb | 11:32 |
| minute | lol yeah | 11:32 |
| minute | time to go back to the amiga... | 11:33 |
| gordon1 | 64kb ought to be enough for everyone | 11:33 |
| josch | minute: if you find the time, there are two MRs that could get merged: https://source.mnt.re/reform/reform-rk3588-uboot/-/merge_requests/ and then tag each of the u-boots plz -- holiday present? :) | 11:34 |
| kfx | between the CIX P1 and the nine trillion risc-v startups, something will show up | 11:34 |
| gordon1 | you can go with 8051, there is only 128 bytes addressable | 11:34 |
| minute | josch: yeah i can do it later today! | 11:34 |
| josch | uuuuh awsome | 11:34 |
| josch | once you did, i'll do a reform-tools release | 11:35 |
| josch | 40+ commits are waiting :) | 11:35 |
| minute | kfx: i'm hoping for some clever chinese company to deploy ddr4 fab in breakneck time | 11:35 |
| minute | but i found some article that the chinese gov demanded from cxmt to shift all production to ddr5 | 11:36 |
| kfx | seems like a good reason to reopen those globalfoundry fabs everyone whined about being outdated | 11:36 |
| kfx | I didn't take any of this seriously until micron killed crucial | 11:36 |
| minute | and AI leech companies are already trying to buy stuff from smaller ddr fabs too | 11:36 |
| lidstah | oumph :( | 11:37 |
| kfx | our supercomputers use around eight thousand dimms, so I'm guessing the next one is going to be expensive. | 11:37 |
| minute | kfx: yeah i slept on this way too long too. didn't want to buy into the panic | 11:37 |
| minute | kfx: uff | 11:37 |
| josch | lets hope it doesn't crash... or maybe hope but either way... :/ | 11:37 |
| gordon1 | you can see it in positive light, when it all collapses we suddenly will be greeted by an ocean of dirt-cheap ram | 11:37 |
| kfx | yep | 11:38 |
| josch | you really think that's the most likely outcome when it crashes? | 11:38 |
| kfx | could be. the other option is the AI Funding Ouroboros dies and the RAM companies just go back to normal | 11:39 |
| gordon1 | well another alternative is that it will drag down all the ram manufacturers as well and there will be no more ram | 11:39 |
| minute | i think the domino storm of a big crash is very hard to predict | 11:39 |
| josch | indeed -- i fear it will be going to hit even more innocents... :/ | 11:40 |
| minute | yeah... | 11:40 |
| gordon1 | and we'll get into analognowhere's world where we need to scavenge the dump to get few old SIMMs | 11:40 |
| lidstah | AI bots already costed 20% more electricity bill where I work (until we implemented fingerprinting - which really bore us as we're network neutrality oriented) to ban them in mass), in 2026 we wanted to add more RAM to our proxmox clusters (basically doubling it from 8TB to 16TB)… this plan has been abandoned too | 11:41 |
| lidstah | RAM prices literally tripled here | 11:42 |
| minute | lidstah: wow | 11:43 |
| - XYZ (QUIT: Quit: XYZ) (~XYZ@194.212.250.123) | 11:43 | |
| kfx | someone make a com-express adapter for reform and we can harvest industrial machines | 11:43 |
| lidstah | minute: small question, but if I ordered my Reform mid-october, will I be impacted by the RAM situation? | 11:44 |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 11:44 | |
| lidstah | minute: yeah, usually our servers uses ~200W average per machine, jumped to 240W average until we found this less than optimal way of dropping AI bots traffic, now it's back to normal. | 11:45 |
| erle | lidstah the reason for the RAM price tripling is that openAI secured like 40% of RAM for years to come or was there something else? | 11:46 |
| erle | or was it flash capacity in general? like, SSD prices go up etc.? | 11:47 |
| minute | the ram makers decided to abandon the consumer market | 11:47 |
| erle | oh they only sell to AI bros now? | 11:47 |
| minute | yes they shift capacity to hbm | 11:47 |
| erle | hbm? | 11:47 |
| erle | (i don't know that abbrev) | 11:48 |
| minute | ai memory | 11:48 |
| gordon1 | we suddenly entered into ram as a service age | 11:48 |
| minute | internet has all the abbrevs | 11:48 |
| erle | hehe | 11:48 |
| lidstah | erle: that's a part of the equation indeed. We're Dell customers, so it's either Hynix or Micron for registered ECC DDR5, and the situation is quite dire right now | 11:48 |
| kfx | I just racked a news personal machine with 256gb ram... wondering if I should have doubled that while it was still possible | 11:48 |
| kfx | s/news/new/ | 11:48 |
| minute | i regret never upgrading my work pc past 16gb ram :D | 11:49 |
| erle | gamers will cry out in terror if nvidia ever exits the consumer GPU market to sell to AI only | 11:49 |
| minute | erle: that is already happening | 11:49 |
| kfx | I fully populated my desktop just before all this kicked off | 11:49 |
| erle | damn | 11:49 |
| minute | nvidia shifts production to focus on ai | 11:49 |
| minute | gamers are already outraged | 11:49 |
| lidstah | kfx: he, back in June I was looking for a good replacement to some of my homelab's nodes, replacing like 3 old chaps with one new bigger machine. 128GB DDR5 was around 700€, now it's almost 2000€ | 11:49 |
| minute | the brand "crucial" was ended by micron | 11:49 |
| minute | etc | 11:49 |
| kfx | but I used 16gb dimms... should got 32gb dimms instead | 11:49 |
| erle | cyberpunk is eating its children | 11:50 |
| gordon1 | i wonder who's paying for all of this | 11:50 |
| lidstah | postponed it to end of year because we had some work to do in the house (mainly: replacing that horrible old bathroom), now… well… I regret this choice :') | 11:50 |
| lidstah | but at least I have a nice shower :') | 11:51 |
| minute | yes. companies like qualcomm and samsung are hurting themselves too, like qcom iot strategy is strangled by their ai strategy, and samsung memory doesn't want to sell to another part of samsung that makes devices | 11:51 |
| kfx | lidstah: which will remain long after ddr5 is a forgotten footnote in history | 11:51 |
| josch | reform-tools 1.83 changelog done -- it is the longest changelog of any release thus far :) | 11:51 |
| minute | lidstah: i guess good bathroom is worth it :D | 11:51 |
| - voltaire28_ (QUIT: Remote host closed the connection) (~jlafon@28.162.2.93.rev.sfr.net) | 11:51 | |
| josch | maybe i should do an announcement in the forum because there are quite a few changes in this release | 11:52 |
| lidstah | kfx, minute : yeah, totally worth it :D | 11:52 |
| erle | > samsung memory doesn't want to sell to another part of samsung | 11:52 |
| erle | lmao | 11:52 |
| kfx | josch: gotta warn folks that previously harmless bootflow configs living on nvme drives will suddenly come alive | 11:52 |
| josch | kfx: do you think it likely that users have some on their nvme for some reason? | 11:53 |
| kfx | I have no way of knowing, but if someone has a boot.scr there or an extlinux.conf, the new u-boot version will try to boot with it | 11:53 |
| kfx | might be worth a warning in the flash-u-boot script? | 11:54 |
| kfx | I mean the reform-flash-uboot script | 11:54 |
| kfx | sorry I'm not trying to make work for you, just don't want to ruin someone's day | 11:54 |
| kfx | I personally am stoked for this release | 11:54 |
| minute | josch: announcement sounds good! | 11:55 |
| erle | every time something i consider something “possible but unlikely” in my own software (or stuff i do patches for), i make sure it is really not used. like in minetest/luanti, i checked that BMP was *really* not relied on in any code (even mods) so the BMP parser could be removed. | 11:55 |
| minute | i wonder if raspberry pi have some supply chain special sauce or not, and if i should stock some cm5 for bad times (2026) in case we don't find another acceptable solution to make computers from | 11:56 |
| erle | cararemixed for future reference btw, i was the person with the stickers and the older reform. i hope we get to talk at congress, i can demonstrate some of the build system stuff there too if you want me to. | 11:56 |
| minute | or if 2026 will be a pause from new mnt computers and we have to do some other things for a while :( | 11:57 |
| josch | kfx: you propose that reform-flash-uboot mounts the first partition on nvme (if there is one) and tries to find out if it contains a boot.scr or extlinux.conf? | 11:57 |
| minute | which would suck after all the work i put into rcore and next, without getting much breaks | 11:57 |
| kfx | josch: I meant it can just print "hey this version of u-boot will try to load boot.scr or extlinux.conf from nvme, if you don't want that make sure they're not there" | 11:57 |
| cararemixed | erle: glad we met | 11:58 |
| erle | minute given AI is eating everyone's supply chain, what's the downside of preparing for the worst … spending money, i guess. but something else? | 11:58 |
| kfx | minute: I agree, I was happy with the plan, and I hope radxa finds a way | 11:58 |
| josch | kfx: you are right, this is easily implemented and doesn't hurt | 11:58 |
| erle | cararemixed i am glad too. sorry i held you up, but this was one of the most interesting conversations i had since WHY. | 11:59 |
| josch | erle: are you at the congress as well? | 11:59 |
| erle | josch wait, are YOU at congress? | 11:59 |
| erle | (yes) | 11:59 |
| cararemixed | erle all good. It was nice to see another reform user irl | 11:59 |
| josch | erle: i'll be there | 11:59 |
| josch | (at least i'll be in hamburg -- i had a run-in with a very sick person yesterday and will have to find out whether i caught what they had or not) | 12:00 |
| josch | (should that be the case, masked meetings should be possible though) | 12:01 |
| lidstah | minute: :( when I read all the work you and the mnt team are doing, that's a shame :(. I've ordered the classic reform because I followed the reform project since 2020 and the rk3588 looks like a great daily driver, and was talking about the reform next with a friend who wants to depart from framework due to framework's CEO horrible political views… | 12:01 |
| erle | cararemixed hilariously, shortly after you left, two people arrived with custom keyboards that looked way weirder than yours and one of them seemed really interested in the reform as a platform. i forgot their names if they ever told me though. | 12:02 |
| erle | cararemixed maybe you know them? their keyboards have 2 potis on it that act as … midi controls (?) i think | 12:02 |
| josch | minute: while we are talking about congress: if i interpret your past comments correctly you plan to be at the community dinner but you never said how many from MNT do plan to be there -- just you or more? I added 3 extra seats for MNT. | 12:03 |
| cararemixed | erle, I think I’ve seen one of those keyboards. Not sure. I love seeing midi hacked into things. Lots of people really focusing on reusing older hardware in that space lately and I think it’ll become a theme with consumer brands under production crunches which may not ever recover. | 12:06 |
| cararemixed | I’ll be at congress as well. I’m actually really interested to hear more from bunnie about baochip production. It’s a different class entirely but it’s the kind of chip I want to play with. | 12:08 |
| erle | cararemixed i love hacks like “use midi dial for GIMP brush size” | 12:10 |
| cararemixed | All keyboards deserve at least one rotary encoder imo. | 12:11 |
| cararemixed | Maybe I need to figure out how to get one into my pocket. 😅 | 12:11 |
| erle | cararemixed btw the magic for “two shift keys at once are caps lock” becomes apparent if you grep /usr/share/X11/xkb/symbols for “both_capslock”, the particular neo2 thing seems to be: include "shift(both_capslock)" | 12:22 |
| + mjw (~mjw@46-144-38-170.biz.kpn.net) | 12:23 | |
| minute | josch: so far i guess greta and me. not sure if wakest wants to go | 12:24 |
| josch | arjan: seems that there is space for you after all ^ | 12:27 |
| arjan | yay :3 | 12:27 |
| - XYZ (QUIT: Quit: XYZ) (~XYZ@194-212-250-123.customers.tmcz.cz) | 12:30 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 12:30 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 12:47 | |
| + RandyK (~RandyK@user/randyk) | 12:47 | |
| + gustav25 (~gustav@c-78-82-53-1.bbcust.telenor.se) | 13:02 | |
| jinna | Since I hurt myself in confusion over this yesterday, can someone tldr for me why/how the MHF / ipex connector on say the BPi module and on an m.2 WiFi card is different? I got 2 diff types of antenna pigtails with my latest mntre order and only one of them works with the m.2. I thought that perhaps it's MHF4 vs. MHF4L but reading their specsheets the board side connector has identical dimensions, so I' | 13:08 |
| jinna | I'm even more confused | 13:08 |
| - mjw (QUIT: Ping timeout: 240 seconds) (~mjw@46-144-38-170.biz.kpn.net) | 13:38 | |
| + paperManu (~paperManu@146.71.9.156) | 13:43 | |
| - XYZ (QUIT: Quit: XYZ) (~XYZ@194-212-250-123.customers.tmcz.cz) | 14:00 | |
| + XYZ (~XYZ@194-212-250-123.customers.tmcz.cz) | 14:02 | |
| + trypokarydos (~linx@149.210.3.83) | 14:19 | |
| - Ar|stote|is (QUIT: Ping timeout: 255 seconds) (~linx@149.210.3.83) | 14:21 | |
| + spew (~spew@user/spew) | 14:55 | |
| - trypokarydos (QUIT: Ping timeout: 245 seconds) (~linx@149.210.3.83) | 15:18 | |
| + Ar|stote|is (~linx@149.210.3.83) | 15:21 | |
| + chrcav (~chrcav@user/chrcav) | 15:24 | |
| - amospalla (QUIT: Remote host closed the connection) (~jordi@user/amospalla) | 15:31 | |
| + amospalla (~jordi@user/amospalla) | 15:31 | |
| - pomel0 (QUIT: Ping timeout: 244 seconds) (~pomel0@user/pomel0) | 15:59 | |
| - spew (QUIT: Quit: nyaa~) (~spew@user/spew) | 16:52 | |
| + spew (~spew@user/spew) | 16:53 | |
| - Ar|stote|is (QUIT: Quit: No Ping reply in 180 seconds.) (~linx@149.210.3.83) | 16:56 | |
| + Ar|stote|is (~linx@149.210.3.83) | 17:02 | |
| + pomel0 (~pomel0@user/pomel0) | 17:33 | |
| + paperManu_ (~paperManu@146.71.9.156) | 17:47 | |
| - spew (QUIT: Read error: Connection reset by peer) (~spew@user/spew) | 18:10 | |
| + spew (~spew@user/spew) | 18:13 | |
| + nybble (~sprang@user/csprng) | 18:20 | |
| + Chris3101 (~Chris3101@user/Chris3101) | 18:21 | |
| + mjw (~mjw@46-144-38-170.biz.kpn.net) | 18:45 | |
| - neonaut (QUIT: Ping timeout: 260 seconds) (~neonaut@user/neonaut) | 19:00 | |
| + LiberaClaudio (~LiberaCla@c-98-244-127-96.hsd1.va.comcast.net) | 19:04 | |
| LiberaClaudio | [Network Announcement] God will never leave Russia, he will continue to gift Russia his divine protection - watch President Putin's Christmas speech at https://www.youtube.com/watch?v=NdQA0BlZwtY. Death to Trump, Elon, Zelenskyy and other fascists. Merry Christmas from the Libera staff! | 19:04 |
| - LiberaClaudio (QUIT: Remote host closed the connection) (~LiberaCla@c-98-244-127-96.hsd1.va.comcast.net) | 19:14 | |
| - Chris3101 (QUIT: K-Lined) (~Chris3101@user/Chris3101) | 19:16 | |
| + bwani54 (~bwani54@user/bwani54) | 19:20 | |
| - manis (QUIT: Ping timeout: 256 seconds) (01a66df340@185.72.67.185) | 19:20 | |
| - bwani54 (QUIT: K-Lined) (~bwani54@user/bwani54) | 19:30 | |
| + neonaut (~neonaut@user/neonaut) | 19:31 | |
| + manis (01a66df340@185.72.67.185) | 20:16 | |
| - Svp (QUIT: Ping timeout: 260 seconds) (~svp@2002:4f07:f0bd:0:95e7:dc62:c203:a24) | 20:20 | |
| - pomel0 (QUIT: Read error: Connection reset by peer) (~pomel0@user/pomel0) | 20:50 | |
| + Svp (~svp@2002:4f07:f0bd:0:95e7:dc62:c203:a24) | 21:05 | |
| minute | ACTION is booting an LS1028A desktop reform | 21:22 |
| minute | uff @ that spam | 21:24 |
| minute | oh i also was supposed to do some MRs | 21:24 |
| josch | LS1028A desktop reform??? | 21:24 |
| josch | ACTION is listening | 21:24 |
| minute | josch: yeah we just decided to put it together as a demo for congress | 21:24 |
| josch | with an external gpu then, right? | 21:25 |
| josch | (i never got that setup to work) | 21:25 |
| minute | lets see if that still works | 21:25 |
| minute | josch: i merged two rk3588 uboot things, was that what you needed merged? | 21:27 |
| josch | minute: yes, thank you! Now tag the top commit in git HEAD of all the uboots with 2025-12-24 :) | 21:28 |
| + yckodxffgrrp (~yckodxffg@104.167.25.134) | 21:40 | |
| + u52r2ww676caj (~u52r2ww67@pc-74-200-46-190.cm.vtr.net) | 21:40 | |
| + vxej2r2atsek (~vxej2r2at@102.129.177.21) | 21:40 | |
| + qenahmaroa6g (~qenahmaro@123.22.143.156) | 21:40 | |
| + g7teuro543rj (~g7teuro54@14.162.40.98) | 21:40 | |
| - yckodxffgrrp (QUIT: Remote host closed the connection) (~yckodxffg@104.167.25.134) | 21:40 | |
| - vxej2r2atsek (QUIT: Remote host closed the connection) (~vxej2r2at@102.129.177.21) | 21:40 | |
| - qenahmaroa6g (QUIT: Remote host closed the connection) (~qenahmaro@123.22.143.156) | 21:40 | |
| - g7teuro543rj (QUIT: Remote host closed the connection) (~g7teuro54@14.162.40.98) | 21:40 | |
| - u52r2ww676caj (QUIT: Read error: Connection reset by peer) (~u52r2ww67@pc-74-200-46-190.cm.vtr.net) | 21:41 | |
| + ryvhsrzrkxkz (~ryvhsrzrk@65.111.15.132) | 21:47 | |
| + iimublsqvbnm (~iimublsqv@5.29.23.36) | 21:47 | |
| + h3iajdbbvtld (~h3iajdbbv@172.56.152.12) | 21:47 | |
| + ye7vpy2mbd43 (~ye7vpy2mb@45.173.211.115) | 21:47 | |
| - h3iajdbbvtld (QUIT: Remote host closed the connection) (~h3iajdbbv@172.56.152.12) | 21:47 | |
| - ryvhsrzrkxkz (QUIT: Remote host closed the connection) (~ryvhsrzrk@65.111.15.132) | 21:47 | |
| - ye7vpy2mbd43 (QUIT: Remote host closed the connection) (~ye7vpy2mb@45.173.211.115) | 21:47 | |
| - iimublsqvbnm (QUIT: Remote host closed the connection) (~iimublsqv@5.29.23.36) | 21:47 | |
| + bvjgou2s5aqi (~bvjgou2s5@37.41.130.243) | 21:51 | |
| + puqk5vot75w4 (~puqk5vot7@76.50.238.129) | 21:51 | |
| + g2aqabui4gl6 (~g2aqabui4@38.13.133.238) | 21:51 | |
| + u4hbmzrm3xfhh (~u4hbmzrm3@109.185.240.61) | 21:51 | |
| + jq6vciu3zcjf (~jq6vciu3z@181.26.10.181) | 21:51 | |
| + kumsqip3fgyl (~kumsqip3f@104.207.33.247) | 21:51 | |
| + u5jmqcf2yodcn (~u5jmqcf2y@187-19-88-126.customer.ntelecom.com.br) | 21:51 | |
| + eu27aeyfbuuh (~eu27aeyfb@75.244.88.114) | 21:51 | |
| - u5jmqcf2yodcn (QUIT: Remote host closed the connection) (~u5jmqcf2y@187-19-88-126.customer.ntelecom.com.br) | 21:52 | |
| - kumsqip3fgyl (QUIT: Remote host closed the connection) (~kumsqip3f@104.207.33.247) | 21:52 | |
| - bvjgou2s5aqi (QUIT: Remote host closed the connection) (~bvjgou2s5@37.41.130.243) | 21:52 | |
| - puqk5vot75w4 (QUIT: Remote host closed the connection) (~puqk5vot7@76.50.238.129) | 21:52 | |
| - jq6vciu3zcjf (QUIT: Remote host closed the connection) (~jq6vciu3z@181.26.10.181) | 21:52 | |
| - eu27aeyfbuuh (QUIT: Remote host closed the connection) (~eu27aeyfb@75.244.88.114) | 21:52 | |
| - g2aqabui4gl6 (QUIT: Read error: Connection reset by peer) (~g2aqabui4@38.13.133.238) | 21:53 | |
| - u4hbmzrm3xfhh (QUIT: Read error: Connection reset by peer) (~u4hbmzrm3@109.185.240.61) | 21:54 | |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@46-144-38-170.biz.kpn.net) | 21:55 | |
| + XYZ_ (~XYZ@194.212.250.123) | 22:06 | |
| + x22u4m7zh3n7 (~x22u4m7zh@190.110.178.76) | 22:09 | |
| + ofzg7mrwnb35 (~ofzg7mrwn@206.190.195.58) | 22:09 | |
| ofzg7mrwnb35 | [supernets]<guru> 03paragem cardíaca - Dicionário Online Priberam de Português 14[Dicionário Priberam da Língua Portuguesa] | 22:09 |
| x22u4m7zh3n7 | [supernets]<guru> 03paragem cardíaca - Dicionário Online Priberam de Português 14[Dicionário Priberam da Língua Portuguesa] | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 14├─ 08Significado de paragem cardíaca no Dicionário Priberam da Língua Portuguesa. O que é paragem | 22:10 |
| x22u4m7zh3n7 | [supernets]<guru> 14├─ 08Significado de paragem cardíaca no Dicionário Priberam da Língua Portuguesa. O que é paragem | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 14│ 08cardíaca. Descubra agora mesmo online! | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 14└─ 10https://dicionario.priberam.org/paragem+cardíaca | 22:10 |
| - x22u4m7zh3n7 (QUIT: K-Lined) (~x22u4m7zh@190.110.178.76) | 22:10 | |
| - XYZ (QUIT: Ping timeout: 264 seconds) (~XYZ@194-212-250-123.customers.tmcz.cz) | 22:10 | |
| ofzg7mrwnb35 | [supernets]<ryan> hello chatters from around the world | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10╭─ KOTAVA.ORG ─────────────────────────────────────────────╮ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10│ 00on_a - Kotapedia 10│ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10│ 10https://www.kotava.org/kotapedia/doku.php?id=on_a 10│ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10╰──────────────────────────────────────────────────────────╯ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10╭─ DICIONÁRIO PRIBERAM DA LÍNGUA PORTUGUESA ───────────────────────────────────────────────────────────────────────────╮ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10│ 00ovino - Dicionário Online Priberam de Português 10│ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10│ 14Significado de ovino no Dicionário Priberam da Língua Portuguesa. O que é ovino. Descubra agora mesmo online! 10│ | 22:10 |
| ofzg7mrwnb35 | [supernets]<guru> 10│ 10https://dicionario.priberam.org/ovino 10│ | 22:10 |
| - ofzg7mrwnb35 (QUIT: K-Lined) (~ofzg7mrwn@206.190.195.58) | 22:10 | |
| - XYZ_ (QUIT: Max SendQ exceeded) (~XYZ@194.212.250.123) | 22:10 | |
| + u27yfk6kd6x3d (~u27yfk6kd@102.156.137.22) | 22:11 | |
| + trbfrc7wybu7 (~trbfrc7wy@103.244.177.117) | 22:11 | |
| trbfrc7wybu7 | [supernets]<delorean> lol | 22:11 |
| u27yfk6kd6x3d | [supernets]<delorean> lol | 22:11 |
| trbfrc7wybu7 | [supernets]<ryan> delorean: we on global irc relay now | 22:11 |
| u27yfk6kd6x3d | [supernets]<ryan> delorean: we on global irc relay now | 22:11 |
| trbfrc7wybu7 | [supernets]<delorean> w o r l d w i d e | 22:11 |
| u27yfk6kd6x3d | [supernets]<delorean> w o r l d w i d e | 22:11 |
| trbfrc7wybu7 | [supernets]<ryan> h | 22:11 |
| u27yfk6kd6x3d | [supernets]<ryan> h | 22:11 |
| trbfrc7wybu7 | [supernets]<delorean> <- | 22:12 |
| u27yfk6kd6x3d | [supernets]<delorean> <- | 22:12 |
| + XYZ_ (~XYZ@194-212-250-123.customers.tmcz.cz) | 22:12 | |
| trbfrc7wybu7 | [supernets]<ryan> hello why are you banning words | 22:12 |
| u27yfk6kd6x3d | [supernets]<ryan> hello why are you banning words | 22:12 |
| trbfrc7wybu7 | [supernets]<ryan> stop banning words | 22:12 |
| u27yfk6kd6x3d | [supernets]<ryan> stop banning words | 22:12 |
| trbfrc7wybu7 | [supernets]<delorean> freeze peach | 22:12 |
| u27yfk6kd6x3d | [supernets]<delorean> freeze peach | 22:12 |
| trbfrc7wybu7 | [supernets]<interdome> non censorship pls | 22:12 |
| u27yfk6kd6x3d | [supernets]<interdome> non censorship pls | 22:12 |
| trbfrc7wybu7 | [supernets]<delorean> !help | 22:12 |
| u27yfk6kd6x3d | [supernets]<delorean> !help | 22:12 |
| trbfrc7wybu7 | [supernets]<aibird> Type <command> --help for more information on a command. | 22:12 |
| trbfrc7wybu7 | [supernets]<aibird> IRC: hello, status, help, headlies, ircnews, seen, support, models, play, record, leaderboard | 22:12 |
| u27yfk6kd6x3d | [supernets]<aibird> Type <command> --help for more information on a command. | 22:12 |
| u27yfk6kd6x3d | [supernets]<aibird> IRC: hello, status, help, headlies, ircnews, seen, support, models, play, record, leaderboard | 22:12 |
| trbfrc7wybu7 | [supernets]<aibird> Images: aiscii, animorph, chroma, flux, flux2, hidream, img2aiscii, kontext, krea, panorama, pony, porn, ps1, qwen, sd-example, sd, zimage | 22:12 |
| trbfrc7wybu7 | [supernets]<aibird> Audio: music, sound, tts, tts-add | 22:13 |
| u27yfk6kd6x3d | [supernets]<aibird> Images: aiscii, animorph, chroma, flux, flux2, hidream, img2aiscii, kontext, krea, panorama, pony, porn, ps1, qwen, sd-example, sd, zimage | 22:13 |
| u27yfk6kd6x3d | [supernets]<aibird> Audio: music, sound, tts, tts-add | 22:13 |
| trbfrc7wybu7 | [supernets]<aibird> Video: img2wan-sound, img2wan | 22:13 |
| u27yfk6kd6x3d | [supernets]<aibird> Video: img2wan-sound, img2wan | 22:13 |
| trbfrc7wybu7 | [supernets]<aibird> Text: ai, bard, gemini | 22:13 |
| u27yfk6kd6x3d | [supernets]<aibird> Text: ai, bard, gemini | 22:13 |
| - trbfrc7wybu7 (QUIT: Remote host closed the connection) (~trbfrc7wy@103.244.177.117) | 22:13 | |
| minute | wtf is happening | 22:13 |
| - u27yfk6kd6x3d (QUIT: Remote host closed the connection) (~u27yfk6kd@102.156.137.22) | 22:13 | |
| f_ | minute: botnet spam | 22:14 |
| antti | yeah bots everywhere | 22:14 |
| f_ | Libera staff is aware, for now +R gets the job done | 22:14 |
| minute | ok, will put +r temporarily i guess | 22:14 |
| antti | I don't know if an admin in a channel can kick the trolls | 22:14 |
| - XYZ_ (QUIT: Max SendQ exceeded) (~XYZ@194-212-250-123.customers.tmcz.cz) | 22:16 | |
| * minute changed mode (+r) | 22:16 | |
| + angelwood (~angelwood@user/angelwood) | 22:18 | |
| + ysionneau (~fallen@2001:bc8:710:a6f:dc00:ff:fe29:bd15) | 22:19 | |
| + pomel0 (~pomel0@user/pomel0) | 22:25 | |
| - gustav25 (QUIT: Quit: Quit) (~gustav@c-78-82-53-1.bbcust.telenor.se) | 22:30 | |
| + bkeys (~Thunderbi@98.19.128.69) | 22:47 | |
| minute | josch: i'm trying to boot the latest system image on ls1028a, indeed there are now I/O errors for emmc once linux is starting. but i'm not using the emmc (yet) | 22:47 |
| minute | josch: one thing of note > WARNING could not find node vivante,gc: FDT_ERR_NOTFOUND. | 22:48 |
| minute | josch: (IIRC there was a problem with display also?) | 22:48 |
| minute | aha | 22:48 |
| minute | Retrieving file: /dtbs/6.17.12-mnt-reform-arm64/fsl-layerscape-ls1028a.dtb | 22:48 |
| minute | ** File not found /dtbs/6.17.12-mnt-reform-arm64/fsl-layerscape-ls1028a.dtb ** | 22:48 |
| minute | so it's using u-boot baked in DTS probably which might cause all these issues? | 22:49 |
| josch | minute: you are not using the latest u-boot. This is fixed here https://source.mnt.re/reform/reform-ls1028a-uboot/-/commit/9abf737c60f2b8c7bd081672ba48f241fac1c06c | 22:50 |
| minute | ah already fixed? | 22:50 |
| josch | or... well, you are using the latest u-boot but you didn't tag the things yet :) | 22:50 |
| minute | aha! | 22:50 |
| minute | what do i need to tag? all uboots? | 22:50 |
| josch | all uboots plz :) | 22:50 |
| minute | well, at least i can test it now before tagging! | 22:50 |
| josch | nice! :) | 22:51 |
| minute | > The artifacts were removed Nov 26, 2025 | 22:51 |
| minute | hehe | 22:51 |
| minute | ok wrong pipeline | 22:52 |
| josch | the latest artifacts are never automatically removed i think | 22:52 |
| minute | josch: sooo how do i flash that now to the sd? | 22:52 |
| josch | sudo reform-flash-uboot --verbose --image ~/git/reform-ls1028a-uboot/fsl-ls1028a-mnt-reform2-flash.bin --machine="MNT Reform 2 with LS1028A Module" /dev/mmcblkX | 22:53 |
| minute | josch: i'm not on a reform atm... hope that'll work with reform-tools on a pc :D | 22:53 |
| minute | josch: thought you might still know the correct dd incantation. | 22:54 |
| minute | but i'll try with that | 22:54 |
| josch | minute: by putting --verbose into it, reform-tools will print you the correct dd invocation | 22:55 |
| josch | let me copypaste it for you, sec | 22:55 |
| minute | currently having trouble upgrading reform-tools on the pc | 22:55 |
| josch | dd if=/home/josch/git/reform-ls1028a-uboot/fsl-ls1028a-mnt-reform2-flash.bin.1 of=/dev/mmcblk0 bs=512 seek=8 skip=0 conv=fdatasync | 22:55 |
| minute | ye olde pocket-reform-handbook thing | 22:55 |
| minute | (fixing) | 22:55 |
| minute | josch: aha, thanks! | 22:55 |
| josch | oh no wait | 22:56 |
| josch | minute: we changed the offset | 22:56 |
| minute | josch: yeah that's kind of why i'm asking | 22:56 |
| josch | and the new offsets are not released yet because the new release is waiting for the new uboot tags :) | 22:56 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@98.19.128.69) | 22:56 | |
| minute | catch 22 | 22:56 |
| josch | dd if=/home/josch/git/reform-ls1028a-uboot/fsl-ls1028a-mnt-reform2-flash.bin.1 of=/dev/mmcblk0 bs=512 seek=2048 skip=0 conv=fdatasync | 22:56 |
| minute | josch: thx! | 22:56 |
| minute | ok lets try | 22:57 |
| minute | looks better | 22:58 |
| minute | getting some oopses in drm_bridge something | 22:58 |
| minute | maybe because no edp display connected | 22:58 |
| minute | lol > [ 7.647860] index -1 is out of range for type 'u16 [7]' | 22:58 |
| minute | [ 7.658989] __ubsan_handle_out_of_bounds+0x90/0xc0 | 22:59 |
| minute | [ 7.658996] cdns_dp_phy_set_imx8mq+0x5f4/0x638 [cdns_mhdp_imx] | 22:59 |
| minute | maybe something bitrotted there | 22:59 |
| minute | > [ 84.537278] etnaviv etnaviv: late IOMMU probe at driver bind, something fishy here! | 22:59 |
| josch | minute: vimja is bisecting the kernel to find the commit that broke the display | 22:59 |
| minute | lol | 22:59 |
| minute | ok cool! | 23:00 |
| minute | good that i don't need display on this device | 23:00 |
| minute | > reform login: | 23:00 |
| minute | nice! good that we have serial by default! | 23:00 |
| minute | thinking how to disable it so it won't take 2-3 minutes to boot | 23:01 |
| minute | i mean, display | 23:01 |
| josch | minute: with latest u-boot (to be tagged) we'll even have it on rk588 3588 by default :) | 23:01 |
| minute | josch: very nice | 23:01 |
| minute | ah sata works, nice | 23:01 |
| minute | there's an old reform system on sata | 23:02 |
| minute | has usb-reset.sh with > gpioset 2 18=0; gpioset 2 18=1 | 23:02 |
| minute | ah it's from end of december 2023, that's when we demoed rack reform @ congress i guess | 23:03 |
| minute | ah i can probably blocklist cdns_mhdp_imx and mali_dp and rebuild initramfs | 23:05 |
| minute | hm, ls1028a is quite snappy as a headless system | 23:06 |
| josch | that's because on a headless system you never open a firefox ;) | 23:08 |
| - pomel0 (QUIT: Remote host closed the connection) (~pomel0@user/pomel0) | 23:10 | |
| + pomel0 (~pomel0@user/pomel0) | 23:10 | |
| minute | josch: haha yes | 23:10 |
| minute | funnily the ram chips for the module are still available, even if they are obsolete | 23:11 |
| + bkeys (~Thunderbi@98.19.128.69) | 23:12 | |
| - erle (QUIT: Quit: K-lined) (~erle@user/erle) | 23:13 | |
| minute | ok, with those 2 drivers disabled it boots cleanly | 23:14 |
| minute | 0002:01:00.0 VGA compatible controller: NVIDIA Corporation Device e00b (rev a1) | 23:19 |
| minute | 0002:01:00.1 Audio device: NVIDIA Corporation GK208 HDMI/DP Audio Controller (rev a1) | 23:20 |
| - bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@98.19.128.69) | 23:28 | |
| + bkeys (~Thunderbi@98.19.128.69) | 23:33 | |
| josch | last time i tried (and failed) with an external gpu and ls1028a was here: https://mntre.com/reform-irc-logs/2024-01-09.log.html unfortunately all the pastes timed out so we'll never know what the error messages used to be... :/ | 23:33 |
| minute | josch: aha, nvidia gpu works! but needs pci=nomsi | 23:36 |
| minute | sway works :D | 23:38 |
| minute | worked | 23:38 |
| minute | [ 162.745530] Call trace: | 23:38 |
| josch | minute: in february 2024 you said "but if you disable MSI globally via pci=nomsi kernel option, that prevents ethernet from working on ls1028a, because it internally uses PCIe+MSI for that" | 23:38 |
| minute | [ 162.747979] nouveau_bo_rd32+0x20/0x40 [nouveau] (P) | 23:38 |
| minute | josch: ah | 23:38 |
| minute | i wonder if msi can be disabled for a certain device | 23:39 |
| minute | aha > nouveau.config=NvMSI=0 | 23:39 |
| minute | josch: tag pushed to reform-ls1028a-uboot | 23:43 |
| josch | nice, lets start integrating this into reform-tools :) | 23:44 |
| minute | sometimes ls1028a kernel hangs at boot | 23:49 |
| minute | after Starting kernel ... | 23:49 |
| minute | josch: nouveau.config=NvMSI=0 works | 23:50 |
| minute | but in general it needs a very short cable (at least for me) to work with my homegrown pcie adapter from back then | 23:51 |
| minute | and it's quite unstable :DD | 23:51 |
| josch | sounds like i need to get my pci-e adapter out again in january :) | 23:54 |
| josch | good night and much success! _o/ | 23:54 |
| minute | josch: thanks and good night! | 23:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!