| + bkeys (~Thunderbi@98.19.131.193) | 00:08 | |
| - mjw (QUIT: Ping timeout: 248 seconds) (~mjw@gnu.wildebeest.org) | 00:12 | |
| + photomattmills (~matt@user/photomattmills) | 00:40 | |
| photomattmills | question: I'm trying to modify the pocket keyboard firmware to work with my modified board. The matrix is 6x12, and I've added 4 keys to the row that previously only had the mouse buttons. All the keys but those four are working, including the mouse keys. I added a sixth row to matrix.h with the appropriate scancodes for my keys, and I can't see why that's not working | 00:43 |
|---|---|---|
| - josch (QUIT: Ping timeout: 245 seconds) (~josch@mister-muffin.de) | 00:45 | |
| + josch (~josch@mister-muffin.de) | 00:45 | |
| - Gooberpatrol66 (QUIT: Ping timeout: 245 seconds) (~Gooberpat@user/gooberpatrol66) | 00:46 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 00:46 | |
| photomattmills | oh, firther wrinkle: two of the keys are hyper and enter, and hitting hyper+enter _does_ bring up the menu on the oled | 00:51 |
| photomattmills | *further | 00:51 |
| minute | photomattmills: if i remember correctly there's some special casing in the row/column loop for the last row, did you check it? | 01:39 |
| - _justin_kelly71 (QUIT: Quit: The Lounge - https://thelounge.chat) (~justinkel@user/justin-kelly/x-6011154) | 02:02 | |
| + emdevt (~emdevt@2001:7e8:fa17:7901:889:cfa1:9c37:6514) | 02:16 | |
| photomattmills | minute: I didn't see that, let me look again | 02:23 |
| photomattmills | looks like all the other branches of the switch case | 02:25 |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 02:27 | |
| photomattmills | oh I wonder if it's the debounce | 02:34 |
| photomattmills | keyboard firmware does this to me every single time. proabably because it's the only time I ever touch C | 02:49 |
| + reform24728 (~dozens@136.29.71.46) | 03:17 | |
| - reform24728 (QUIT: Remote host closed the connection) (~dozens@136.29.71.46) | 03:17 | |
| - paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@72.10.129.226) | 03:31 | |
| - MyNetAz (QUIT: Remote host closed the connection) (~MyNetAz@95.216.142.242) | 03:43 | |
| - casparvitch (QUIT: Changing host) (~casparvit@server.samsci.com) | 03:44 | |
| + casparvitch (~casparvit@user/casparvitch) | 03:44 | |
| - kop316 (QUIT: Read error: Connection reset by peer) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:04 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:06 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:06 | |
| - digitalrane (QUIT: Ping timeout: 250 seconds) (~digitalra@user/digitalrane) | 04:06 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:06 | |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 04:18 | |
| + digitalrane (~digitalra@user/digitalrane) | 04:45 | |
| - skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich) | 05:00 | |
| + skipwich (~skipwich@user/skipwich) | 05:01 | |
| - dodo (QUIT: Server closed connection) (~dodo@user/dodo) | 06:42 | |
| + dodo (~dodo@user/dodo) | 06:43 | |
| photomattmills | figured it out. `if (used_key_codes < MAX_SCANCODES && resulting_scancodes && y < 5)` the 5 needs to be a 6 for my board | 07:14 |
| photomattmills | that worked! one agressive rebase later, and no one is the wiser lol | 07:18 |
| - emdevt (QUIT: Remote host closed the connection) (~emdevt@2001:7e8:fa17:7901:889:cfa1:9c37:6514) | 07:49 | |
| + reform31765 (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 08:02 | |
| * reform31765 -> rwa_ | 08:02 | |
| - rwa_ (QUIT: Quit: Leaving) (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 08:30 | |
| - elektron (QUIT: Server closed connection) (~elektron@apoc.halo.nu) | 09:14 | |
| + elektron (~elektron@apoc.halo.nu) | 09:14 | |
| + reform8766 (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 09:48 | |
| * reform8766 -> rwa_ | 09:48 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 09:50 | |
| rwa_ | lol - compiling the linux kernel just comletely filled my 1TB NVMe... | 10:33 |
| gordon1 | that can't be right | 10:38 |
| - GNUmoon (QUIT: *.net *.split) (~GNUmoon@gateway/tor-sasl/gnumoon) | 10:40 | |
| - RandyK (QUIT: *.net *.split) (~RandyK@user/randyk) | 10:40 | |
| rwa_ | jep, sounds weird | 10:41 |
| rwa_ | "cp: error copying 'debian/build/build_arm64_none_mnt-reform-arm64/output/image/lib/modules/6.16.1-mnt-reform-arm64/kernel/sound/ac97_bus.ko.xz' to 'debian/linux-image-6.16.1-mnt-reform-arm64/lib/modules/6.16.1-mnt-reform-arm64/kernel/sound/ac97_bus.ko.xz': No space left on device" | 10:42 |
| rwa_ | i used the scripts at https://source.mnt.re/reform/reform-debian-packages/-/tree/linux6.16?ref_type=heads | 10:42 |
| + RandyK (~RandyK@user/randyk) | 10:43 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 10:43 | |
| * Guest560 -> mjw | 11:13 | |
| josch | rwa_: is your /tmp maybe mounted as a tmpfs? What is your output of: df -h /tmp | 11:31 |
| josch | and also useful: findmnt /tmp | 11:32 |
| rwa_ | yes, it is - i didn't realize it "writes" to /tmp :) | 11:34 |
| josch | rwa_: you can change that by executing the commands with TMPDIR set to a different directory | 11:37 |
| josch | TMPDIR=/var/tmp is a popular choice | 11:37 |
| josch | do you think it would make sense to add that as a tip to the README? | 11:37 |
| rwa_ | seems like the tmpfs is limited to 8GB on my 16GB RK3588 | 11:41 |
| rwa_ | i'll give it a try | 11:42 |
| rwa_ | but i should get some active cooling set up before starting again :D | 11:42 |
| josch | rwa_: you can mount a tmpfs of arbitrary size -- it is only limited in practice by the size of your swap partition | 11:43 |
| josch | you can remount your tmpfs to be bigger without rebooting even: | 11:44 |
| gordon1 | rwa_: you can increase it (if you dare) with mount -o remount,size=75% /tmp | 11:44 |
| josch | sudo mount -o remount,size=16G /tmp | 11:44 |
| josch | you can use percentages?? wow, wild | 11:44 |
| gordon1 | well, by default it is size=50% | 11:45 |
| rwa_ | yeah, default is 50% | 11:45 |
| josch | you can probably add 200% there if you have swap | 11:45 |
| josch | then it will be fast if it fits into 8 GB but not result in "No space left on device" in case it becomes larger than that | 11:46 |
| gordon1 | i guess you can do zram + tmpfs | 11:47 |
| gordon1 | never tried tho | 11:47 |
| rwa_ | side question: how to add kernel parameters permanently? edit /boot/boot.scr? | 11:48 |
| josch | zram is an interesting idea now that my processor is way faster than any workload i have requires | 11:48 |
| josch | rwa_: read /boot/boot.scr -- it tells you how :) | 11:48 |
| rwa_ | i'd usually with zswao | 11:48 |
| rwa_ | i'd usually go with zswap | 11:48 |
| rwa_ | josch: k, gonna have a look | 11:49 |
| + andreas-e (~Andreas@2a02-8434-b6a3-e901-facc-8e87-8e54-890d.rev.sfr.net) | 11:53 | |
| - casparvitch (QUIT: Remote host closed the connection) (~casparvit@user/casparvitch) | 12:10 | |
| + casparvitch (~casparvit@user/casparvitch) | 12:10 | |
| - casparvitch (QUIT: Remote host closed the connection) (~casparvit@user/casparvitch) | 12:18 | |
| - rwa_ (QUIT: Quit: Leaving) (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 12:18 | |
| + casparvitch (~casparvit@user/casparvitch) | 12:18 | |
| grimmware | uuuuugh the power on my pocket is doing weird shit again :/ Brownouts when running certain things under steam or firefox (high load) at around 50% battery and now only charging to 54% with both batteries showing voltage of 3.6 | 12:49 |
| grimmware | This is like, the most consistently unreliable thing on the pocket and it's the one thing that makes it really difficult to rely on as a device | 12:49 |
| grimmware | It'll always boost up to charge at full speed when I plug the cable in but one of the cells will then jump up to 4.1v with the other at around 3.8 and then drop back to a subsistence charge; | 12:51 |
| + paperManu (~paperManu@72.10.129.226) | 12:56 | |
| + gustav28 (~gustav@c-78-82-54-127.bbcust.telenor.se) | 13:02 | |
| + reform7359 (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 13:12 | |
| * reform7359 -> rwa_ | 13:12 | |
| - buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org) | 13:17 | |
| + buckket (~buckket@vps.buckket.org) | 13:18 | |
| rwa_ | i updated /etc/default/flash-kernel, ran "sudo flash-kernel", i see /boot/boot.scr being updated but the kernel does not pick up the new params during boot | 13:22 |
| bremner | rwa_: are you by chance using uboot-menu? | 13:24 |
| rwa_ | don't think so | 13:26 |
| rwa_ | it's essentially a debian stable image from reform.d.n on NVMe and the /boot partition on eMMC | 13:26 |
| rwa_ | i need to get more familiar with the boot process on arm devices (which always seemed a little weird to me) | 13:27 |
| rwa_ | until now all my devices where efi & grub for years | 13:28 |
| josch | rwa_: doe you have a /boot/extlinux/extlinux.conf? | 13:29 |
| josch | rwa_: there is a merge request which enables efi booting in case you want to help getting that into shape | 13:29 |
| rwa_ | yep, this file exists - seems to be related to the u-boot menu mentioned before | 13:30 |
| josch | rwa_: if /boot/extlinux/extlinux.conf exists, then u-boot will ignore /boot/boot.scr | 13:30 |
| josch | the next version of reform-tools will make you aware of that | 13:30 |
| rwa_ | thanks for the clarification | 13:31 |
| rwa_ | josch: can you point me to the efi MR? | 13:36 |
| chorc | grimmware: +1, those brownouts are very annoying. I even had couple of them while charging! The ironic thing is that now, as my charger board has burned and I'm waiting for the replacement, I'm running of external power, the thing is rock solid! :) | 13:37 |
| grimmware | chorc: did your charger board have the chip failure? | 13:38 |
| chorc | grimmware: yep | 13:39 |
| grimmware | yeah mine hasn't done that :/ | 13:39 |
| chorc | FYI, I had a theory that a brownout was actually a controller watchdog reset, because I had those even after balancing both cells and having some high voltage | 13:42 |
| chorc | but I couldn't get to properly debugging, Summer and all, maybe if the new one still does it - I'll get to it | 13:43 |
| grimmware | I wouldn't even know where to start :( | 13:45 |
| - b0 (QUIT: Server closed connection) (~b0@user/b0) | 13:45 | |
| + b0 (~b0@user/b0) | 13:45 | |
| grimmware | I'm kinda wondering if one of the cells is bad or something | 13:46 |
| chorc | watchdog SDK has watchdog_caused_reboot function, I'll start with a custom firmware that would report this flag to UART after each reboot | 13:47 |
| chorc | I think I saw minute mentioning that sometimes a cable between two sides might be faulty (I think they said it about keyboard cable, but the one to the charger is similar) | 13:49 |
| chorc | so having MB reporting the reason for reboot might be useful | 13:50 |
| minute | chorc: do you think "random reset" of the system (with screen fading out or no?) are watchdog resets of sysctl? | 13:51 |
| minute | photomattmills: awesome btw, that y < 5 was what i had in mind, but couldn't tell you the exact location from memory | 13:51 |
| - rwa_ (QUIT: Quit: Leaving) (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 13:52 | |
| chorc | minute: I had a screen fading out during those resets, I think it might be a watchdog | 13:52 |
| chorc | screen fades out the same way for me after sudo reboot | 13:53 |
| grimmware | I was under the impression that a watchdog reset of the sysctl would leave the SoM power rail set to whatever it was already at to prevent sysctl reboot taking the whole system down | 13:53 |
| chorc | I recall minute increasing that timeout at some point, and resets were less frequent after it, but still happening | 13:54 |
| minute | grimmware: ok, i should really look deeper into this problem. right now i'm sick at home though, an can still do some work, but not on hardware directly. i have to finish 2 major pcb projects, then i'll look more into pocket power stuff. what comes to mind is testing with lab power supply inducing low voltages instead of the batteries, to see if low bat voltages can be the cause, and measure | 13:54 |
| minute | voltage drops along the path to the SoC | 13:54 |
| + reform31557 (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 13:55 | |
| * reform31557 -> rwa_ | 13:55 | |
| minute | grimmware: maybe it's something easy like too high resistance somewhere, or the polyfuse on charger board being overzealous. actually you could try briding that fuse with an R0. also it's possible that the bat gauge drifted just totally off and 50% is 0%. we need to check if the gauge can be reset in sw/fw at all, versus reconnecting the batteries | 13:56 |
| minute | grimmware: also, charging board v2 will do better balancing (at least that's the goal) | 13:56 |
| grimmware | Gotcha, if you have any experiments you’d like for me to repeat I have a bench top supply and a scope here | 14:04 |
| grimmware | I doubt the gauge is the issue given it’s sitting below 4v on both cells when it’s not actively charging. | 14:06 |
| - polymetis (QUIT: Server closed connection) (sid320610@id-320610.hampstead.irccloud.com) | 14:13 | |
| + polymetis (sid320610@id-320610.hampstead.irccloud.com) | 14:14 | |
| josch | rwa_: https://source.mnt.re/reform/reform-system-image/-/merge_requests/133 | 14:19 |
| josch | that MR allows building images for amd64 which was something minute would like to use at MNT | 14:20 |
| josch | but since amd64 (and not qemu) has u-boot, i added efi booting while we were at it | 14:20 |
| josch | bkeys has successfully tested the resulting image with EDK2 on the hardware | 14:20 |
| bkeys | I can concur it works | 14:21 |
| josch | there were still some open questions from when i worked on this in june, mainly concerning how we want to do the partitioning | 14:23 |
| - andreas-e (QUIT: Quit: Leaving) (~Andreas@2a02-8434-b6a3-e901-facc-8e87-8e54-890d.rev.sfr.net) | 14:23 | |
| josch | because u-boot only looks at the first partition and efi wants fat32 and i wonder whether we want *three* partitions as the result | 14:23 |
| - potash (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 14:24 | |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 14:52 | |
| - bkeys (QUIT: Ping timeout: 260 seconds) (~Thunderbi@98.19.131.193) | 15:10 | |
| - sknebel (QUIT: Server closed connection) (~quassel@v22016013254630973.happysrv.de) | 15:26 | |
| + sknebel (~quassel@v22016013254630973.happysrv.de) | 15:27 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 15:34 | |
| + AnimaInvicta (~AnimaInvi@88-120-179-216.subs.proxad.net) | 15:39 | |
| - AnimaInvicta (PART: !!unknown attribute: msg!!) (~AnimaInvi@88-120-179-216.subs.proxad.net) | 15:41 | |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 16:27 | |
| + emdevt (~emdevt@2001:7e8:fa17:7901:889:cfa1:9c37:6514) | 16:30 | |
| + Ar|stote|is (~linx@149.210.0.217) | 16:47 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149.210.0.217) | 16:49 | |
| + Ar|stote|is (~linx@149.210.0.217) | 16:50 | |
| Ar|stote|is | If I update to the rk when it will be available for the reform, will there be something else i will be missing from the reform next, other than the form factor? | 16:54 |
| gordon1 | third pci? | 16:59 |
| - qbit (QUIT: Remote host closed the connection) (~qbit@user/qbit) | 17:09 | |
| - rwa_ (QUIT: Remote host closed the connection) (~rene@p200300d0ff24e5007fc2100dbdd09829.dip0.t-ipconnect.de) | 17:25 | |
| - emdevt (QUIT: Ping timeout: 255 seconds) (~emdevt@2001:7e8:fa17:7901:889:cfa1:9c37:6514) | 17:33 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 17:54 | |
| + qbit (~qbit@user/qbit) | 18:03 | |
| minute | Ar|stote|is: usb-c for actual usb comes to mind, otherwise, not much (except the very different architecture) | 18:10 |
| josch | Ar|stote|is: the maybe potential option to replace the lifepo4 cells with lithium cells | 18:11 |
| - photomattmills (QUIT: Read error: Connection reset by peer) (~matt@user/photomattmills) | 18:23 | |
| + potash (~potash@user/foghorn) | 18:34 | |
| Ar|stote|is | minute: hey lukas, how hard would it be to cram or design a pcb to drive the display of the reform to use it as another monitor? (I mean without losing the laptop display) | 18:42 |
| Ar|stote|is | By usbc you mean the one for charging the device? Battery chemistry doesn't mean that much to me for the time being at least. I prefer the form factor of the og reform it is more spacious. | 18:46 |
| - L29Ah (QUIT: Read error: Connection timed out) (~L29Ah@wikipedia/L29Ah) | 18:53 | |
| gordon1 | ACTION always dreamt to replace liion cells with lifepo4 | 19:02 |
| gordon1 | now my dream came true | 19:02 |
| - blast007 (QUIT: Server closed connection) (~blast@user/blast007) | 19:03 | |
| + blast007 (~blast@user/blast007) | 19:03 | |
| ^alex | hey hey, so, we have an i.MX8 pocket that we've tinkered with a bit, but now we want to reset it to the latest factory image. iirc there's a tool for that? | 19:09 |
| gordon1 | does anyone have working slicer for rk3588 (or any arm64 for that matter)? | 19:13 |
| + andreas-e (~Andreas@2a02-8434-b6a3-e901-facc-8e87-8e54-890d.rev.sfr.net) | 19:28 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 19:34 | |
| josch | ^alex: boot from the latest system image and then run reform-flash-rescue | 19:43 |
| ^alex | gotcha | 19:44 |
| ^alex | downloading the latest system image from the artifacts page on source.mnt.re | 19:46 |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 19:57 | |
| + mark_ (~mjw@gnu.wildebeest.org) | 20:16 | |
| - mjw (QUIT: Killed (copper.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 20:17 | |
| * mark_ -> mjw | 20:17 | |
| + Guest5442 (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 20:18 | |
| + bkeys (~Thunderbi@98.19.131.193) | 20:32 | |
| josch | Ar|stote|is: something else you might be missing with the classic reform: built-in webcam | 20:38 |
| minute | Ar|stote|is: not hard and i've considered it already. one could use our existing hdmi to edp converter. it just needs an actual HDMI connector and a case | 21:59 |
| minute | gordon1: which slicer would you like to run? maybe i can help | 22:00 |
| gordon1 | minute: well, i compiled my own superslicer but it crashes in its guts with some assert when it actually starts to slice things, don't think it is related to rk3588 per se or arm64, just finiky project, I also built prusaslicer 2.7.4 but it spawns no window after start, didn't debug it yet, trying my luck with superslicer HEAD | 22:03 |
| gordon1 | which refuses to compile because of again C++ changed something | 22:03 |
| gordon1 | not sure that i need help with superslicer, probably emotional support | 22:04 |
| gordon1 | and with prusaslicer i just need to run it under gdb to see what's happening there first | 22:06 |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-54-127.bbcust.telenor.se) | 22:15 | |
| gordon1 | minute: actually you may help, you did run blender on panfrost, didn't you? it looks like it needed geometry shaders that are not supported by panfrost, and it looks like prusa-slicer also needs geometry shaders | 22:17 |
| gordon1 | what did you do to make it run? | 22:19 |
| - cwebber (QUIT: Server closed connection) (~Christine@user/cwebber) | 22:22 | |
| + cwebber (~Christine@user/cwebber) | 22:22 | |
| josch | gordon1: https://community.mnt.re/t/has-anyone-tried-genuine-mali-drivers/3576/14 | 22:25 |
| gordon1 | ah, sorry | 22:26 |
| gordon1 | oh well | 22:26 |
| + erle (~erle@user/erle) | 22:27 | |
| josch | gordon1: and if you want something that does not work then there is https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/126 | 22:29 |
| erle | josch hello again. where would i start if i want to make a clean new debian SD card image that is reproducible for my reform2? | 22:31 |
| erle | i doubt there is an official install image ;) | 22:32 |
| josch | erle: i don't understand. What is wrong with the existing ones? | 22:33 |
| erle | josch nothing, i just wanted to learn about the process to see how it is made from first principles. | 22:33 |
| erle | and ig you can point me at the start of it | 22:34 |
| erle | easier to ask a human than trying to figure it out and failing because i looked at something outdated etc. | 22:34 |
| josch | erle: how "first" do you want your principles to be? Is installing Debian packages sufficient? | 22:34 |
| erle | basically: “if i have an empty SD card, what script collection creates the closest thing to the disk image that when it starts starts the custom reform install script” | 22:35 |
| erle | the one with the keyboard layout selection | 22:36 |
| erle | i have some long idle time on the coming weekend and thought it would be good to prepare :) | 22:37 |
| josch | erle: you mean the reform-setup-wizard? | 22:37 |
| erle | maybe? no idea how it is called! | 22:37 |
| josch | erle: you are familiar with https://source.mnt.re/reform/reform-system-image/ | 22:37 |
| erle | josch so that is the canonical source then ok? | 22:38 |
| erle | like, no additional magic | 22:38 |
| minute | gordon1: i got blender to kind of work only by turning off the remaining stuff that needs geom shaders | 22:39 |
| gordon1 | right, i thought so | 22:39 |
| minute | gordon1: i heard through the grapevine that geom shaders emulation will come one day. | 22:39 |
| minute | gordon1: if graphics performance is not super important, you can always try running with LIBGL_ALWAYS_SOFTWARE=1 | 22:40 |
| gordon1 | yeah i tried that just now and it is unusable | 22:40 |
| gordon1 | works for paraview though | 22:40 |
| gordon1 | surprisingly | 22:40 |
| minute | i did use a slicer on imx8m reform back in the day, something that was in debian maybe | 22:41 |
| josch | erle: depends what you call magic | 22:42 |
| minute | gordon1: i think it was slic3r | 22:42 |
| josch | erle: for example that script downloads the u-boot binaries (checked with hash) -- is that additional magic for you? The same is done for all Debian packages that are insatlled. They are also downloaded and their hashes checked. Do you want to build those "from first principles" first? | 22:42 |
| erle | josch i see, reproducability is at the level of “hashes match”, but not at “can be built offline” | 22:43 |
| erle | gordon1 there are also variables you can give to MESA to control how much opengl standards are exposed | 22:43 |
| erle | gordon1 e.g. you can tell it to keep silent about extensions that are newer. depending on workload, *some* stuff can work better if you limit what a driver exposes. | 22:44 |
| erle | you need to experiment for that | 22:44 |
| josch | erle: it can be built offline if you download the source code of several thousand debian packages first | 22:44 |
| josch | erle: a few percent of those thousand packages are not yet bit-by-bit reproducible | 22:44 |
| erle | but i remember that e.g. mesa intel integrated gfx exposed lower opengl versions than the windows driver for the hardware because some software (e.g. chrome) had a fallback that was *faster* on such HW | 22:45 |
| erle | wings3d works perfectly with LIBGL_ALWAYS_SOFTWARE=1 btw | 22:45 |
| minute | gordon1: i just tried slic3r, to get working 3d it needs GDK_BACKEND=x11 but then it seems fine | 22:46 |
| erle | i think somewhere i have a 3d printed cup still that i made with it | 22:46 |
| erle | nice! | 22:46 |
| gordon1 | is slic3r still a thing? i thought it is long dead by now | 22:46 |
| josch | are we talking about 3d printer slicer? any? | 22:46 |
| gordon1 | yes | 22:47 |
| minute | gordon1: no idea. you asked for a slicer :D | 22:47 |
| josch | i have printed all my mnt reform PLA pieces with an ultimaker, sliced with cura | 22:47 |
| minute | ah yes i think i ran cura back in the day? | 22:47 |
| josch | works on imx8mq, a311d and rk3588 | 22:47 |
| gordon1 | erle: yeah, i know about those variables, unfortunately i cannot use it to make smething that isn't there, i.e. geometry shaders in recent mesa for panfrost | 22:47 |
| minute | noice | 22:47 |
| gordon1 | okay, i thought this day would come | 22:48 |
| gordon1 | will reserve cura as last resort, but thanks for tell that it works | 22:49 |
| erle | gordon1 yeah if there is no fallback then you can't do anything except software rendering. which always works, if you don't care about the computer being hot and you getting 1 to 3 fps lol | 22:49 |
| - Asmadeus (QUIT: Server closed connection) (~asmadeus@user/meow/Asmadeus) | 22:49 | |
| minute | gordon1: what's wrong with cura? :D | 22:49 |
| josch | erle: are your questions about the images answered? | 22:49 |
| gordon1 | erle: im getting more like 0.1 fps | 22:50 |
| + Asmadeus (~asmadeus@user/meow/Asmadeus) | 22:50 | |
| minute | erle already disappeared into another topic | 22:50 |
| gordon1 | minute: absence of configs for my custom voron | 22:50 |
| minute | gordon1: ah. that's beyond my knowledge | 22:50 |
| erle | josch good enough ig :) | 22:50 |
| gordon1 | okay, back to superslicer then | 22:51 |
| erle | minute no, i am just easily distracted and distracted myself with another channel | 22:51 |
| erle | does anyone have performance tips for minetest/luanti on reform2 (imx8?) except “use 5.10 and not a higher version”, “disable shaders”, “activate mipmaps”? | 22:52 |
| erle | i mean the engine is *really* wasteful with resources but i think it used to work better on reform1 even | 22:52 |
| erle | so i kinda wonder | 22:52 |
| erle | (might also that the engine became *more* wasteful for the default workload) | 22:53 |
| minute | haven't used imx8 in a longer time... | 22:53 |
| ^alex | hmm, how long does the system image take to turn the screen on? it should boot from SD card even if the EMMC root partition is blank, right? | 22:54 |
| ^alex | we know the bootloader needs to stay on EMMC for the pocket reform | 22:55 |
| minute | erle: here is how minetest works on rk3588 reform (from ~51 seconds into the video) https://www.youtube.com/watch?v=YNfC22P0Zc0 | 22:55 |
| minute | ^alex: which processor? | 22:55 |
| ^alex | imx8mp | 22:55 |
| minute | ^alex: alright, yes, uboot needs to be on emmc for it to boot | 22:55 |
| minute | ^alex: it takes a while from sd card, maybe 20-30 seconds? | 22:56 |
| erle | minute it works, just not as well enough as i *think* it should. but given the devs constantly drop code that does not make a difference on their gaming hardware, there might not be any knobs to turn. | 22:56 |
| erle | e.g. triangle merging. you have a big plane with 1 texture? technically you only need 2 triangles. but that code got dropped after a dev noticed it does not make a difference on his gaming GPU. it does, obviously, make one if you have a weaker one! | 22:56 |
| ^alex | minute, and uboot is the first partition, right? or are we about to have to do a hard-mode recovery after formatting the root :') | 22:57 |
| erle | mipmaps did have a noticeable effect though | 22:57 |
| minute | ^alex: iirc on imx8mplus uboot is on the emmc special boot partition | 22:58 |
| minute | boot0 | 22:58 |
| Ar|stote|is | minute: that sounds great, i was wondering what it would need to hack something so that i could use the reform as a lapdock for a steam deck. | 22:58 |
| ^alex | ok gotcha | 22:59 |
| minute | Ar|stote|is: i think this will be much easier with the next | 22:59 |
| ^alex | just gonna be patient then | 22:59 |
| minute | Ar|stote|is: because one could make a custom port board that acts as an input thing with internal display switch | 22:59 |
| minute | Ar|stote|is: by coincidence i thought about this a few hours ago | 22:59 |
| Ar|stote|is | Why is that? Due to the more modular pcbs? | 23:00 |
| Ar|stote|is | Tbh i've been gathering resources for quite a while. I wanted to try the surface screen which is "probably" the same one the framework 13 uses and hack together a fw13 with active digitizer | 23:01 |
| Ar|stote|is | But i believe the reform would be better to start testing for something like this. | 23:02 |
| josch | in completely unrelated news I now tested linux 6.16.1 on A311D and screen, wifi, bluetooth and sound work fine | 23:02 |
| Ar|stote|is | The next seems great and i already got two friends of mime to back it up, but i feel kinda bummed spending 1.5k extra on top of the 1k of the reform. The whole idea was to update the machine not buy a new one. | 23:03 |
| Ar|stote|is | And i like the bulkiness of it. So i guess I'll have to get the next's innards and transplant it on the original | 23:04 |
| josch | Ar|stote|is: you will not be the only one doing that ;) | 23:04 |
| Ar|stote|is | And i already consider the eink screen that is on funding right now on crowd supply by modos | 23:05 |
| Ar|stote|is | That would also be a neat project for the reform | 23:05 |
| ^alex | dOH pebcak: we downloaded _reform_ system image, not pocket-reform-system-image. | 23:06 |
| Ar|stote|is | I just was hoping for a more affordable upgrade path but alas i'll have to stick to the one i got and mess around with the rest of the hardware. | 23:06 |
| ^alex | _that_ would explain why the pocket did not `main screen turn on`, as it were | 23:07 |
| minute | Ar|stote|is: yes, the motherboard has a lot of internal connectors only, and there are modules that convert this to that. so you can puzzle together any weird thing that fits | 23:07 |
| minute | factorio but it's a laptop | 23:08 |
| Ar|stote|is | josch: do you plan to mod ur reform with the next's pcb pieces? Will make a custom bottom or something else in mimd? | 23:08 |
| josch | ^alex: if your pocket reform has the most recent u-boot version, then downloading the reform-system-any.img will get you a system image that boots everywhere | 23:08 |
| josch | Ar|stote|is: maybe :) | 23:08 |
| ^alex | ours most likely does not (we haven't updated since sometime last year iirc) | 23:08 |
| Ar|stote|is | minute: that was my intention with the reform because it seems to have quite some free space. Maybe I need to sell my reform and buy the next. | 23:09 |
| Ar|stote|is | It will be quite some time before we can grab a next motherboard with the rest of the pcbs me thinks | 23:10 |
| Ar|stote|is | I should just focus on trying to use the screen as a monitor for the time being. | 23:10 |
| Ar|stote|is | At least I feel good that I have helped both Germans create their modular computers. | 23:13 |
| ^alex | hmm | 23:17 |
| ^alex | booting the system image drops us into rescue mode after a big blob of UBSAN traces from the kernel, right before qcacld says it's loaded | 23:17 |
| josch | ^alex: imx8m+ pocket reform? | 23:19 |
| ^alex | yes | 23:20 |
| josch | ^alex: and this was the image you downloaded? https://source.mnt.re/reform/reform-system-image/-/jobs/12438/artifacts/file/pocket-reform-system-imx8mp.img.gz | 23:22 |
| ^alex | SHA256 (pocket-reform-system-imx8mp.img.gz) = fd44ac469ae77a783577202b3252853af71bfeafdc24a6a585b9642dbcd04436 | 23:22 |
| ^alex | browser reports that we downloaded https://source.mnt.re/reform/reform-system-image/-/jobs/12438/artifacts/raw/pocket-reform-system-imx8mp.img.gz | 23:23 |
| ^alex | which crosschecks with the job number | 23:23 |
| erle | if i do “systemctl suspend”, what is the thing that wakes my reform up again? | 23:23 |
| josch | erle: circle, space | 23:23 |
| erle | i try | 23:24 |
| josch | ^alex: sometimes it's a broken sd-card | 23:24 |
| ^alex | josch, hm, we could probably cable it with ethernet, dhclient up a lease, and run reform-flash-rescue | 23:24 |
| ^alex | or not, `ifconfig` doesn't show us an eth0 | 23:24 |
| erle | this does work! | 23:24 |
| ^alex | oh wait it does under `ifconfig -a` | 23:25 |
| josch | erle: which kernel version are you running? | 23:25 |
| erle | already shut it off. why? | 23:26 |
| erle | also i guess hibernate is disabled as long as you start from sd card, correct? | 23:26 |
| josch | erle: because i spent many, many hour to make it work with kernels newer than 6.1 | 23:27 |
| erle | josch 6.12.21 | 23:27 |
| josch | nice! | 23:27 |
| josch | thank you for your feedback | 23:27 |
| erle | suspend and wakeup work fine | 23:27 |
| josch | phew, i'm relieved | 23:27 |
| erle | should i check any apps or other stuff for you? | 23:27 |
| erle | like, if all subsystems come up clean? | 23:27 |
| ^alex | ... if we could find the adapter >.> | 23:28 |
| josch | erle: be careful though because it worked "most of the time" for me | 23:28 |
| josch | erle: it sometimes did not wake up again | 23:28 |
| erle | o.0 | 23:28 |
| josch | you are also lucky because it depends on which imx8mq you have | 23:31 |
| josch | i had one board where it never worked but it did work on the other with the same system image | 23:31 |
| ^alex | ok we have `reform-flash-rescue` slurpin' down the image with bmaptool | 23:32 |
| josch | ^alex: so you managed to boot from the sd-card? | 23:32 |
| erle | oof i think the graphics drivers on my reform may be broken somehow. neverball shows every ground texture as black and the right gear in glxgears has no hole :( | 23:36 |
| erle | (the green one) | 23:36 |
| ^alex | josch, yeah, we booted from the SD card but it dropped us into systemd rescue mode, whence we were able to dhcpcd on the ethernet interface and start reform-flash-rescue | 23:40 |
| josch | hm i see | 23:42 |
| ^alex | hopefully it doesn't hork up the same errors when we boot from emmc >.> | 23:42 |
| josch | ^alex: usually you can leave the rescue shell with ctrl+d or exit | 23:42 |
| ^alex | sure, but, we could also just do the thing we were trying to do from the rescue shell | 23:43 |
| erle | btw, minetest/luanti works. | 23:43 |
| ^alex | we don't need the full environment just to run reform-flash-rescue | 23:43 |
| + Guest94 (~Guest94@nat-178-19-230-27.net.encoline.de) | 23:50 | |
| - Guest94 (QUIT: Client Quit) (~Guest94@nat-178-19-230-27.net.encoline.de) | 23:52 | |
| - erle (QUIT: Quit: K-lined) (~erle@user/erle) | 23:57 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!