| bkeys | Connection to what? I haven't messed with the hardware | 00:00 |
|---|---|---|
| josch | here is how it looks for me: https://paste.debian.net/1412118/ | 00:07 |
| josch | but as you can see, this is a different version of u-boot from the one you have | 00:07 |
| josch | bkeys: do you have u-boot on emmc or is that zeroed? | 00:08 |
| bkeys | josch: I have a damaged uboot on emmc it seems | 00:14 |
| bkeys | Or I suppose an incorrect one | 00:14 |
| josch | bkeys: How? Above you said that you can boot just fine from sd-card | 00:15 |
| bkeys | That was true when the uboot on the emmc was totally borked | 00:15 |
| bkeys | I tried to dd a valid uboot on the emmc and I am getting the output in the pastebin | 00:15 |
| josch | and you decided to take a different u-boot than the one from the machine.conf file with the hash in there? | 00:16 |
| josch | it seems you used the latest gitlab CI artifact, is that correct? | 00:16 |
| bkeys | Yes the latest gitlab artifact is what I used | 00:17 |
| josch | it would've been safer to use what everybody else uses ;) | 00:17 |
| josch | bkeys: which command did you use to flash u-boot onto your emmc? | 00:17 |
| bkeys | I know, do you think it's bricked? | 00:17 |
| bkeys | The one that minute put in the chat earlier | 00:17 |
| josch | checking... | 00:18 |
| bkeys | I'm gonna flash a fresh system image on the SD card and hope that it boots that, although my hopes are not high because uboot says "trying to boot from MMC1" no matter what | 00:18 |
| bkeys | regardless of if the card is in there | 00:18 |
| josch | the command is correct | 00:19 |
| josch | what image are you flashing exactly? | 00:19 |
| minute | just to save you a lot of puzzle hunting time, what can happen (and might happen here) is that SPL is loaded from eMMC and the rest of uboot from SD | 00:19 |
| bkeys | https://source.mnt.re/reform/reform-system-image/-/jobs/15981/artifacts/raw/reform-system-rk3588.img.gz | 00:19 |
| bkeys | This one | 00:19 |
| minute | and the two can be incompatible | 00:20 |
| bkeys | minute: Possibly, although I seem to get the same output regardless of if SD is in there | 00:20 |
| josch | bkeys: that is the non-dsi varint | 00:20 |
| josch | bkeys: that image will also include u-boot for the non-dsi reform | 00:21 |
| bkeys | Well that gives credence to what minut (don't want to ping them since they need rest) said | 00:21 |
| bkeys | I hope that there is hope | 00:22 |
| josch | bkeys: can you try flashing the image for your rcore-dsi reform? | 00:23 |
| bkeys | You mean to an SD card? | 00:23 |
| josch | you are flashing system images to sd-card right now, no? At least flash the right one for your platform. :) | 00:24 |
| bkeys | Yes I will do that | 00:26 |
| bkeys | I'm downloading the image | 00:29 |
| bkeys | josch: So I flashed the wrong uboot, which makes sense why I got the same output regardless of SD card or not. So if I put the SD card in there with the right uboot I should be able to boot from SD card and be okay | 00:30 |
| bkeys | ACTION is feeling more hopeful than he did before josch came onto the scene | 00:30 |
| josch | you are giving me too much credit, this is not my area of expertise | 00:30 |
| bkeys | josch: So even if there is a valid uboot on emmc, if there is a uboot on SD the SD card will take precedent right? | 00:34 |
| josch | no | 00:34 |
| josch | i made a table for this | 00:35 |
| josch | source.mnt.re/reform/reform-handbook/-/issues/8 | 00:35 |
| bkeys | So according to this table having a valid SD card won't help us? | 00:37 |
| josch | it might if m.inute's theory is correct and u-boot from emmc is reading parts of u-boot from sd-card | 00:38 |
| josch | but in general, when you want to do experiments its safest to have emmc zeroed out because then you can write whatever you want to sd-card and even if it's garbage you can just re-flash easily | 00:39 |
| bkeys | Well I had my experiment successfully executed on SD card I was trying to move it to emmc | 00:39 |
| bkeys | But it must have been reading part of it from SD card | 00:39 |
| bkeys | josch: Alright it wrote, lets try this out | 00:45 |
| bkeys | josch: Nope, I'm getting the exact same output | 00:47 |
| bkeys | It's totally ignoring my SD card | 00:47 |
| josch | okay, then i'll now try to reproduce your situation by flashing the image you flashed to my emmc | 00:47 |
| bkeys | You're not worried about bricking your board? | 00:48 |
| josch | the artifact you flashed is what i told minute days ago to please tag as the next official release. If it's broken, then this is something we should make sure of before tagging a release and have others flash it to their emmc and brick their device | 00:49 |
| josch | so better brick mine than others | 00:49 |
| - aloo_shu (QUIT: Ping timeout: 244 seconds) (~aloo_shu@90.166.98.18) | 00:51 | |
| bkeys | Well if it is bricked, do you have any equipment that unbricks it? | 00:53 |
| bkeys | I read some docs from firefly, it seems there is a way to unbrick it but they only write about their SoM, I'm not sure if the Reform has the same ability | 00:54 |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.3.83) | 01:00 | |
| + Ar|stote|is (~linx@149.210.3.83) | 01:00 | |
| bkeys | josch: Were you able to replicate it? | 01:04 |
| josch | sorry my internet is slow :) | 01:11 |
| bkeys | I really hope I didn't brick my Reform | 01:12 |
| bkeys | I mean I accept responsibility if I did, but still | 01:12 |
| josch | even if i succeed, the question then becomes what you did wrong and how you can recover | 01:14 |
| josch | another tip: reform-flash-uboot now also has the --machine option | 01:14 |
| josch | that way, you can avoid having to use dd even on a machine that is not the target | 01:15 |
| bkeys | Duly noted | 01:15 |
| bkeys | https://wiki.t-firefly.com/en/iCore-3588Q/upgrade_firmware.html | 01:15 |
| josch | that means you no longer have to ask for offsets or whether boot0 or whatnot | 01:15 |
| bkeys | It seems if I did brick it I can go into maskrom mode, but you would need their carrier board | 01:15 |
| josch | first step: this is the u-boot binary you used loaded from sd-card: https://paste.debian.net/1412131/ | 01:20 |
| - cli (QUIT: Ping timeout: 240 seconds) (~m-vsauiy@user/cli) | 01:22 | |
| josch | and now i flashed the same u-boot blob that you flashed (2025-11-04-3-g086b855-g424c714eb247-dirty) to my emmc | 01:23 |
| josch | and zeroed sd-card but that step should not matter because u-boot from emmc will be preferred | 01:24 |
| - Svp (QUIT: Ping timeout: 256 seconds) (~svp@host-79-7-240-189.business.telecomitalia.it) | 01:24 | |
| josch | and here is the boot log for the u-boot blob you tried flashed to emmc: https://paste.debian.net/1412133/ | 01:24 |
| josch | if you diff the two outputs, the main difference is that the BL31 output from u-boot on sd-card is: | 01:25 |
| josch | Trying to boot from MMC2 | 01:25 |
| josch | and when u-boot (and the blobs that get loaded before that) come from emmc then it prints: | 01:25 |
| josch | Trying to boot from MMC1 | 01:26 |
| josch | bkeys: so the u-boot binary you tried works perfectly fine on emmc for me | 01:26 |
| bkeys | :I | 01:27 |
| bkeys | josch: What does that mean? | 01:27 |
| josch | as i said this is not my area of expertise so i would not dare to speculate what could've gone wrong for you | 01:28 |
| bkeys | Yeah but with me something always seems to go wrong | 01:28 |
| josch | you were warned :/ | 01:29 |
| josch | let me take this as an opportunity to add a new feature to reform-flash-uboot: read back u-boot from emmc/sd-card and compare the checksum of what got written to what it is expected to be | 01:30 |
| bkeys | Yeah as I said I take responsibility | 01:31 |
| bkeys | We will see if m has any thought, and if not then I don't know | 01:31 |
| josch | there are other users of the firefly icore board, so a solution could also be found outside of the MNT bubble | 01:32 |
| bkeys | Yeah I know, hopefully I can avoid buying another board for the OTG port | 01:33 |
| josch | bkeys: what i find curious about your serial output is that it contains some garbled values. Are you on motherboard 3.0 and connect it to another machine via usb-c? | 01:34 |
| bkeys | I think I have an earlier motherboard | 01:35 |
| bkeys | I got my Reform first in like 2021, I bought it originally on crowdsupply | 01:36 |
| + Svp (~svp@2002:4f07:f0bd:0:95e7:dc62:c203:a24) | 01:39 | |
| kfx | bkeys: I'm no expert, but I've been where you are, and I was able to short CLK/GND and boot into maskrom mode in the past. then you can put u-boot back on the emmc with rkdevtool | 01:43 |
| bkeys | kfx: Yeah I see the pad where the wiki says I can do that | 01:43 |
| bkeys | I indeed get different output, but what happens after that? | 01:43 |
| kfx | you can use rkdeveloptool(1) to erase-flash and go back to booting off the sd card, or you can unpack the u-boot image and figure out which offsets to write the component parts to | 01:49 |
| kfx | I think there's a way to write the u-boot image in one command but I can't remember if that's true | 01:49 |
| bkeys | kfx: Can I do that with only a TTL connection? | 01:52 |
| josch | bkeys: it would also be interesting to find out what you did wrong earlier | 01:54 |
| bkeys | Yeah, if I can recover from this I suppose I can recover from anything | 01:55 |
| kfx | bkeys: I recall doing it over the serial pins, but it was a long while ago and I might be wrong | 01:57 |
| bkeys | kfx: As in with a TTL connector? | 01:57 |
| kfx | yeah | 01:57 |
| bkeys | Yeah everything on the firefly wiki says to use this OTG port but it seems like it would just be a glorified TTL connector | 01:58 |
| bkeys | kfx: Assuming you are remembering correctly, is it just issuing commands over the serial or does it make the reform show up as a device on the host PC? | 01:59 |
| kfx | this is the cable I used | 02:00 |
| kfx | https://www.pishop.us/product/usb-to-ttl-serial-cable-for-raspberry-pi-5-debugging-ft232rnl-chip/ | 02:00 |
| bkeys | That's basically what I have | 02:00 |
| bkeys | https://community.mnt.re/t/recovering-a-possibly-bricked-rk3588-pocket-reform/3592 | 02:00 |
| kfx | and no, it doesn't show up as a usb device I think? in order to make that happen you have to push a binary blob that does something different | 02:01 |
| kfx | I never did that | 02:01 |
| kfx | they call it 'usbplug' | 02:01 |
| kfx | I've now entered a world of confusion, where I got the motherboard 3.0, but it won't boot the rcore, but it WILL boot the imx8mq, but neither of them output anything over usb-c serial | 02:02 |
| bkeys | So without it being a usb device, does it just give me a prompt of sorts? | 02:02 |
| kfx | I miss my ttl pins :( | 02:02 |
| kfx | bkeys: in maskrom mode I only ever use rkdeveloptool to interact with it | 02:02 |
| kfx | e.g. rkdeveloptool list-partitions to confirm it was talking | 02:03 |
| kfx | I don't know the protocol details at that point | 02:03 |
| bkeys | Let me try | 02:03 |
| bkeys | I didn't try anything with rkdeveloptool | 02:03 |
| bkeys | kfx, josch: WHAT?!?! | 02:06 |
| bkeys | So I short the pads to try to make it go into maskrom mode | 02:07 |
| bkeys | And it boots the system image from my SD card?! | 02:07 |
| bkeys | Holy crap my prayer was answered! | 02:07 |
| josch | o0 | 02:08 |
| bkeys | I was honestly thinking to myself "I could pray to God and see if he works out a miracle, nah he doesn't care about this situation" and this happened | 02:09 |
| bkeys | I was rewarded despite my lack of faith! | 02:09 |
| bkeys | josch: So what script to I run inside of the system image to fix the emmc? | 02:10 |
| kfx | reform-flash-uboot emmc | 02:11 |
| bkeys | /boot/flash.bin is up-to-date -- not downloading it again | 02:12 |
| bkeys | the first partition on /dev/mmcblk0 starts at 9437184 and would be overwritten by uboot | 02:12 |
| bkeys | make sure that the first 9463808 bytes are free on /dev/mmcblk0 | 02:12 |
| josch | i'm so happy i put these guards in :) | 02:13 |
| bkeys | josch: So what do I do? This is my one shot to fix this it feels like | 02:14 |
| josch | bkeys: exactly what kfx said | 02:14 |
| bkeys | That got me the output I just posted | 02:14 |
| josch | and if there is a partition in the way, remove the partition | 02:14 |
| josch | you probably created that one? | 02:14 |
| kfx | zero out the first few blocks so that stuff is gone | 02:15 |
| bkeys | Okay, I'm gonna assume I'm just wrong about everything because I have blown it up to this point | 02:15 |
| bkeys | Is it just dd if=/dev/zero of=/dev/mmcblk0 count=512 | 02:15 |
| josch | bkeys: did you create partitions on emmc or were they there already? | 02:15 |
| bkeys | Is that right? | 02:15 |
| bkeys | josch: They were created when I flashed Fedora on top of them | 02:15 |
| josch | i'd zero more than 512 bytes in case this is guid partition table | 02:16 |
| bkeys | I flashed fedora then put uboot on top of it | 02:16 |
| josch | oooooh fedora :) | 02:16 |
| bkeys | Yes I was able to get it running, my woes are with uboot | 02:16 |
| josch | yeah, don't just flash that but do the partitioning such that the first 16MiB are empty | 02:16 |
| bkeys | Well a Fedora install is already on the emmc | 02:16 |
| bkeys | There is nothing of particular value on there though | 02:16 |
| bkeys | bkeys@reform:~$ ls /dev/mmcblk0 | 02:18 |
| bkeys | mmcblk0 mmcblk0boot0 mmcblk0boot1 mmcblk0rpmb | 02:18 |
| bkeys | I ran | 02:18 |
| bkeys | sudo dd if=/dev/zero of=/dev/mmcblk0 count=50000 | 02:19 |
| bkeys | So the fedora install is gone for now, so I guess I will flash the emmc again? | 02:19 |
| josch | just do not flash the fedora image directly to /dev/mmcblk0 unless you plan to have u-boot on sd-card | 02:19 |
| bkeys | Well I know the emmc is dangerous, but it's still memory that is there and is reasonably fast | 02:20 |
| josch | if you want both fedora and u-boot on emmc, you need to move the partition with fedora so that there is space for u-boot | 02:20 |
| bkeys | I'd like to have fedora and uboot on the emmc if possible | 02:20 |
| bkeys | How do I flash the .img with Fedora so that it's shifted like that? | 02:20 |
| josch | then move the partition with fedora on it so that 16MiB are free | 02:20 |
| bkeys | ACTION runs the dd command to zero it out further so that the uboot is really gone | 02:21 |
| josch | your dd command with count=50000 zeroed out 24 MiB of data, that is more than enough | 02:21 |
| bkeys | I have zeroed out the first 256mb | 02:21 |
| bkeys | die uboot die | 02:21 |
| josch | you could either flash the image and then use a partitioning software to move the partition | 02:21 |
| bkeys | Yeah that just takes forever, it would be great if I could somehow make dd do it | 02:22 |
| josch | or you disect the image and move things to the right locations before flashing it | 02:22 |
| josch | you can also make dd flash to the right location if you pass the right offsets and if you then adjust the partition table accordingly | 02:22 |
| bkeys | I'll just do it the slow way | 02:23 |
| bkeys | josch: The Fedora image when flashed gives me about 9.4MB at the front, do I need to shift it to the right still or is that good enough? | 02:37 |
| josch | barely | 02:39 |
| josch | u-boot itself is 9M | 02:39 |
| bkeys | Do you think I should shift it? | 02:39 |
| josch | you can do the math: take the u-boot size and the offset (32768 bytes) and see if it clashes with the start of the first partition | 02:40 |
| bkeys | Your script is basically telling me I should shift it, I will shift it | 02:40 |
| bkeys | I will mindlessly obey rn I don't want my stuff bricked again | 02:40 |
| josch | bkeys: source of why the 16 MiB: https://opensource.rock-chips.com/wiki_Partitions | 02:43 |
| - paperManu (QUIT: Ping timeout: 250 seconds) (~paperManu@107.159.15.124) | 02:44 | |
| bkeys | I made it 25MiB just to be super safe | 02:44 |
| bkeys | josch: https://paste.centos.org/view/df064ec6 | 02:46 |
| bkeys | Do you think it's safe to reboot? (I will not hold you responsible if it messes up, I understand this is all on me) | 02:46 |
| bkeys | I flashed fedora, moved the partitions and did this | 02:46 |
| josch | if you used "sudo reform-flash-uboot emmc" and it does not work then that is on me | 02:47 |
| bkeys | I'm just saying don't worry about me being angry at anyone. But does that output and image hash look good to you? | 02:49 |
| bkeys | It looks good to me but I'm paranoid so I will ask questions I know are obvious | 02:49 |
| josch | could you do a sha1sum? | 02:50 |
| josch | that's what we do everywhere else | 02:50 |
| josch | also, reform-tools checks it | 02:50 |
| josch | so it should be correct | 02:50 |
| bkeys | 876b4f5f125a40afaceac7258504ee15b212f36d /boot/flash.bin | 02:51 |
| josch | "/boot/flash.bin is up-to-date" tells you that the sha1sum is correct | 02:51 |
| josch | yes, 876b4f5f125a40afaceac7258504ee15b212f36d is what is written in machines/MNT Reform 2 with RCORE-DSI RK3588 Module.conf | 02:52 |
| josch | i improved the message of reform-flash-uboot to say that it also has successfully verified the sha1sum | 02:52 |
| bkeys | Okay, so I am safe to reboot? (in theory) | 02:54 |
| josch | yes | 02:54 |
| bkeys | ACTION takes a deep breath | 02:54 |
| bkeys | Here we go | 02:54 |
| - pomel0 (QUIT: Ping timeout: 240 seconds) (~pomel0@user/pomel0) | 02:54 | |
| bkeys | josch: Thank the Almighty! It's booting Fedora | 02:57 |
| bkeys | Last thing now is to just resize the partition | 02:58 |
| + paperManu (~paperManu@107.159.15.124) | 02:59 | |
| bkeys | josch: Hmm, I need to resize the filesystem for Fedora to finish my install, but to do that I need to boot from the SD card; I can't seem to get uboot to interrupt though (my serial connection is fine though | 03:00 |
| josch | u-boot on emmc will prefer the system on sd-card over that on emmc | 03:00 |
| josch | unless there is some efi magic getting in the way | 03:00 |
| josch | but at least that's how it works with MBR and extlinux.conf or boot.scr | 03:01 |
| bkeys | josch: Hmm, it seems to be going straight for the emmc's Fedora | 03:02 |
| - mjw (QUIT: Ping timeout: 240 seconds) (~mjw@gnu.wildebeest.org) | 03:03 | |
| bkeys | I even spam the keyboard to try and interrupt it | 03:03 |
| josch | nothing comes to mind right now | 03:04 |
| josch | and it's 3 AM over here, so i'll hit the pillow now that it at least boots for you :) | 03:05 |
| bkeys | I appreciate it friend | 03:05 |
| bkeys | Thank goodness for the miracle of it booting again | 03:05 |
| josch | minute: apparently bkeys flashed something wrong yesterday. I have a local improvement of reform-flash-uboot which reads in u-boot from emmc/sd and compares it to the expected content after flashing to make extra sure that the right thing got flashed. To do this though it would make this comparison process more convenient if the u-boot images were a multiple of 512 bytes in size. Would you approve of me | 03:06 |
| josch | filing MRs against our five u-boots to pad all blobs with zeroes at the end so that they are a multiple of 512 bytes large? | 03:06 |
| + _justin_kelly71 (~justinkel@user/justin-kelly/x-6011154) | 03:46 | |
| - paperManu (QUIT: Ping timeout: 265 seconds) (~paperManu@107.159.15.124) | 04:12 | |
| + spew (~spew@user/spew) | 04:44 | |
| - spew (QUIT: Client Quit) (~spew@user/spew) | 04:48 | |
| bkeys | josch: Hopefully you are asleep but I figured it out, I used the grub shell from Fedora to boot the system image and then resized my rootfs while I was there. Thank you for your help | 04:49 |
| bkeys | I got Fedora, DTB loaded, it's all good to go I even got nvme in here | 04:49 |
| josch | bkeys: nice! the alternative would be to use the grub shell to boot something else or to modify the fedora beforehand with the correct offsets and sizes | 06:12 |
| josch | oh wait you *did* boot the system image in the grub shell | 06:23 |
| bkeys | Yes I did | 06:28 |
| bkeys | It's running Fedora 43 with mainline kernel and hardware acceleration | 06:30 |
| bkeys | I can't run blender sadly | 06:30 |
| bkeys | But otherwise is really good, I need to get it to detect the battery and all that | 06:30 |
| josch | that is expected | 06:31 |
| josch | blender *does* work with quasar | 06:31 |
| bkeys | What is quasar? | 06:31 |
| josch | the qualcomm adapter board | 06:31 |
| bkeys | The one that comes with rcore? | 06:31 |
| josch | no, rcore is the adapter board for the firefly icore rk3588 | 06:31 |
| josch | quasar is the adapter board for the qualcomm qcs6490 | 06:32 |
| bkeys | Ah I see | 06:32 |
| bkeys | I have some loongarch64 hardware I plan on doing the 3D modelling on | 06:32 |
| josch | bkeys: though there is some slight hope for blender on rk3588: https://mastodon.social/@mntmn/115636279777084665 | 06:33 |
| bkeys | josch: https://imgur.com/a/8LAcwrU | 06:40 |
| bkeys | I'll just use different hardware for 3D models | 06:40 |
| josch | noice! | 06:40 |
| josch | bkeys: next i'm looking forward to your feedback regarding packaging reform-tools for fedora :) | 06:41 |
| bkeys | Heh, I'm doing a lot of porting work right now. Today I took a break to get the Reform since it arrived back to me today | 06:41 |
| bkeys | Right now I'm working on porting CentOS Stream 10 to loongarch64 | 06:42 |
| bkeys | At some point I need to document everything I did today and post it on community.mnt.re | 06:43 |
| josch | cool! i recently enabled loong support in box64 in debian | 06:43 |
| josch | you should :) | 06:44 |
| bkeys | There is the cherry on top, I got my nvme drive being mounted as my /home | 06:46 |
| bkeys | So I got 512GB in /home and 256 in / so I got plenty of room here | 06:46 |
| josch | minute: expanding on my last message to you, this is what i'd like to add to all u-boots if you agree: https://source.mnt.re/reform/reform-boundary-uboot/-/merge_requests/23 | 06:55 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@98.19.131.193) | 06:59 | |
| + cli (~m-vsauiy@user/cli) | 07:02 | |
| + potash1 (~potash@user/foghorn) | 07:23 | |
| + echeese_pocket (~echeese_p@2601:8c1:c180:4630::9e1a) | 08:32 | |
| echeese_pocket | hi, I have a keyboard with a custom layout that I like (im trying to switch off of qwerty), how do I do that? thanks! | 08:34 |
| josch | echeese_pocket: usually the same way as you do for other keyboard via your system settings | 08:37 |
| echeese_pocket | I was looking at gallium layout, but I havent fully switched over yet | 08:39 |
| echeese_pocket | currently im using qmk on an external keyboard | 08:41 |
| josch | echeese_pocket: okay, of course you can swap the keys in the firmware and if you want to do that i can show you which file to do that in if you tell me your keyboard version. But you can also just use their kxb layout: https://github.com/GalileoBlues/Gallium/blob/main/Linux/gallium | 08:46 |
| echeese_pocket | so can I just swap keys in that before installing that to my reform? | 08:47 |
| echeese_pocket | bc I want partial gallium not full | 08:47 |
| josch | you can also flash your keyboard while it is in the reform and running (but have a second keyboard ready in case of emergencies when you mess up for example) | 08:48 |
| josch | but even if you want partial gallium, you can just modify the xkb file for that too | 08:48 |
| josch | the problem with modifying the scancodes in the firmware is that then you will get wrong behaviour of programs which rely on scancodes instead of keypresses, for example in games | 08:49 |
| josch | SDL has that feature for example and games like factorio make use of it | 08:49 |
| josch | i know because i use the neo2 layout which also scrambles essentially all key positions around | 08:49 |
| josch | and changing the xkb file is less prone to failure if it goes wrong than flashing keyboard firmware | 08:50 |
| josch | i think changing the firmware makes sense if you either use it with operating systems that make it hard to run custom layouts | 08:50 |
| echeese_pocket | so how do I do it without the firmware changes? | 08:50 |
| josch | or if you want to carry the standalone keyboard around and connect it to different computers and still have the same layout independent on what they have configured (they even then the software matching scancodes to keys can mess things up) | 08:51 |
| echeese_pocket | maybe ill just go full gallium on the reform | 08:51 |
| echeese_pocket | but I assume it will not work on bios unless I reflash firmware, right? | 08:52 |
| josch | searching on the internet reveals articles like this one: https://codeaffen.org/2023/09/16/custom-keyboard-layouts-with-xkb/ | 08:52 |
| josch | echeese_pocket: there is no bios | 08:52 |
| echeese_pocket | oops | 08:52 |
| echeese_pocket | but like the drive decription screen | 08:52 |
| josch | what is the drive decryption screen? | 08:53 |
| echeese_pocket | the screen where it asks for the drive password | 08:53 |
| echeese_pocket | I think its part of the pocket default image | 08:54 |
| josch | that's in the initramfs and if you make your xkb config system wide, then that will have your custom layout in software as well | 08:54 |
| josch | source: i'm using neo2 to type my luks password as well | 08:54 |
| echeese_pocket | nice | 08:54 |
| josch | the pocket default image is not encrypted but you can use reform-setup-encrypted-disk to set it up | 08:54 |
| echeese_pocket | whats neo2? | 08:55 |
| josch | a keyboard layout | 08:55 |
| josch | https://en.wikipedia.org/wiki/Neo_(keyboard_layout) | 08:55 |
| echeese_pocket | oh | 08:56 |
| echeese_pocket | I dont speak German so I dont need that | 08:56 |
| echeese_pocket | but thats cool | 08:56 |
| josch | echeese_pocket: it seems there were efforts to add your layout (and related ones) to xkeyboard-config but it was closed 2 days ago because nobody seemed to be interested in it anymore: https://gitlab.freedesktop.org/xkeyboard-config/xkeyboard-config/-/merge_requests/725 | 08:59 |
| josch | maybe you can help make it happen and then it will be even easier to use that layout by just editing /etc/default/keyboard like for any other layout | 09:00 |
| echeese_pocket | how do I do that? | 09:00 |
| josch | the maintainers raised some issues and questions in that merge request and it needs somebody to address these | 09:02 |
| echeese_pocket | ok... | 09:04 |
| josch | but as i said, i can also show you which file to edit in the firmware but you never told me your keyboard version :) | 09:04 |
| echeese_pocket | idk which version | 09:05 |
| josch | it says so on the keyboard pcb or you can send a photo of your layout and whether or not you have rgb led backlight | 09:06 |
| + gidzit (~gidzit@gidzit.org) | 09:06 | |
| echeese_pocket | I have qwerty layout and rgb backlight | 09:06 |
| josch | https://source.mnt.re/reform/reform/-/blob/master/reform2-keyboard4-fw/src/matrix.h | 09:08 |
| echeese_pocket | so what do I do once I modify that file? | 09:12 |
| josch | echeese_pocket: you run cmake which is automated by the shell script build.sh in the parent directory. Have a look at the README.md there. | 09:15 |
| echeese_pocket | ok! :) | 09:16 |
| jah | grimmware: thank you :) | 09:35 |
| jah | I have a (non-hyper) pocket reform, and I just found that I also have a unused wd 2TB sn770 NVme drive, could I use such a drive with my pocket? | 09:36 |
| - echeese_pocket (QUIT: Quit: Client closed) (~echeese_p@2601:8c1:c180:4630::9e1a) | 09:53 | |
| + echeese_pocket (~echeese_p@2601:8c1:c180:4630::9e1a) | 10:07 | |
| - casparvitch (QUIT: Ping timeout: 240 seconds) (~casparvit@user/casparvitch) | 10:08 | |
| - echeese_pocket (QUIT: Client Quit) (~echeese_p@2601:8c1:c180:4630::9e1a) | 10:09 | |
| + casparvitch (~casparvit@user/casparvitch) | 10:11 | |
| josch | minute: unfortunately neither flashing u-boot to 8MB+0x1000 nor to 8MiB+0x1000 gave me any output on serial on LS1028A :/ | 10:19 |
| josch | but i'm happy that we tried first before putting the effort | 10:19 |
| + echeese_pocket (~echeese_p@2601:8c1:c180:4630::9e1a) | 10:45 | |
| - XYZ_ (QUIT: Remote host closed the connection) (~XYZ@78-80-115-228.customers.tmcz.cz) | 10:47 | |
| + XYZ_ (~XYZ@78-80-115-228.customers.tmcz.cz) | 10:47 | |
| + jogu (~jogu@user/jogu) | 10:52 | |
| minute | josch: weird, but it still works at 0x1000? | 10:57 |
| minute | josch: it's possible that you can't just flash it unmodified to that offset, but that it needs to be built for that offset | 10:58 |
| minute | josch: i'm fine with the 512b alignment idea yep | 10:59 |
| - echeese_pocket (QUIT: Quit: Client closed) (~echeese_p@2601:8c1:c180:4630::9e1a) | 11:02 | |
| - XYZ_ (QUIT: Read error: Connection reset by peer) (~XYZ@78-80-115-228.customers.tmcz.cz) | 11:43 | |
| josch | minute: yes, 0x1000 was the default and when i flash it there it works | 11:43 |
| minute | ok | 11:44 |
| + XYZ_ (~XYZ@78-80-115-228.customers.tmcz.cz) | 11:44 | |
| josch | minute: re built for that offset: yes possible but should it at least not print something for bl3 even if the offsets for the stuff after that (like u-boot) are wrong? | 11:44 |
| minute | josch: not sure if bl3 can be loaded without rcw knowing the offset | 11:44 |
| josch | hrm right | 11:45 |
| kfx | neat, my keyboard lights stay on throughout boot now. guess the problem was with my vintage motherboard | 11:46 |
| kfx | also minute I apologize for basically liveblogging into a support ticket, your company does not deserve this idiocy | 11:46 |
| minute | googling around a bit, rcw+pbl+bl2 are bl2.pbl :D | 11:47 |
| - RandyK (QUIT: *.net *.split) (~RandyK@user/randyk) | 11:48 | |
| - lidstah (QUIT: *.net *.split) (~lidstah@gateway/tor-sasl/lidstah) | 11:48 | |
| - GNUmoon2 (QUIT: *.net *.split) (~GNUmoon@gateway/tor-sasl/gnumoon) | 11:48 | |
| minute | josch: table 9 mentions gpt and even efi https://alider.by/upload/sata/docs/2ef8d1b0-b2a4-11e9-bef2-9551a92c8e3a.pdf | 11:49 |
| + aloo_shu (~aloo_shu@90.166.98.116) | 12:05 | |
| - gordon1 (QUIT: Quit: WeeChat 4.5.2) (~gordon1@citadel.niflheim.info) | 12:07 | |
| + gordon1 (~gordon1@citadel.niflheim.info) | 12:07 | |
| - gordon1 (QUIT: Client Quit) (~gordon1@citadel.niflheim.info) | 12:08 | |
| + gordon1 (~gordon1@citadel.niflheim.info) | 12:10 | |
| josch | minute: i don't want to waste your time with ls1028a. I'm not upset if you just give up. Maybe one thing you could do is to ask for help with this using your wide reach in the fediverse and maybe somebody cares enough | 12:10 |
| + paperManu (~paperManu@107.159.15.124) | 12:12 | |
| - gordon1 (QUIT: Client Quit) (~gordon1@citadel.niflheim.info) | 12:12 | |
| josch | there is also a MR which tried to build ATF such that it can boot from emmc by setting BOOT_MODE=emmc when building bl2 but this also did not end up working: https://source.mnt.re/reform/reform-ls1028a-uboot/-/merge_requests/3 | 12:16 |
| + mjw (~mjw@gnu.wildebeest.org) | 12:40 | |
| minute | josch: without clicking that link, it's probably from dirk eibach who did a physical mod to the ls1028a module to make that possible (need to change a boot config bit/pin) | 12:56 |
| minute | josch: we just need to look at rcw+pbl source. i can take a quick look now | 12:56 |
| josch | minute: your dedication to technical excellence is highly appreciated as usual | 12:57 |
| minute | looking at https://source.mnt.re/reform/reform-ls1028a-uboot/-/blob/main/rcw/ls1028a-mnt-reform2/R_PSPH_0xb8be/rcw_1500_sdboot.rcw | 13:00 |
| + gustav25 (~gustav@c-78-82-54-128.bbcust.telenor.se) | 13:02 | |
| minute | hm > 9.2.1.51 Boot location pointer low-order address (BOOTLOCPTRL) | 13:17 |
| minute | > Used by Service Processor to pass the boot address to GPPs | 13:17 |
| minute | ok fact 1: > If boot fails from the RCW loaded at offset 0 or 0x1000 (in case of SD), the boot ROM searches for RCW at 8 MB (8 MB + 0x1000 for SD) offset on the device. | 13:21 |
| minute | this means: _RCW_ is loaded. what happens then | 13:21 |
| minute | > If enabled by RCW, the Service Processor polls Reset PBI_REQ[REQ]. When the request is detected, | 13:22 |
| minute | it starts Pre-Boot Initialization, reading data from either the SD/MMC, SPI and writing to CCSR space | 13:22 |
| minute | or local memory space (OCRAM, SRAM, DDR). If the Service Processor reports an error during its | 13:22 |
| minute | Pre-Boot Initialization process, the device reset sequence is halted. | 13:22 |
| minute | this is probably what's happening right now | 13:22 |
| minute | it probably writes the wrong offset to "CCSR space" for further loading | 13:22 |
| - martins (QUIT: Remote host closed the connection) (924a802ee9@user/martins) | 13:29 | |
| + martins (924a802ee9@2001:bc8:1210:2cd8::15c) | 13:30 | |
| - Ar|stote|is (QUIT: Quit: No Ping reply in 180 seconds.) (~linx@149.210.3.83) | 13:31 | |
| minute | mhm https://github.com/nxp-archive/openil_rcw/blob/fda65ca642ec7f1cd906e929857d73de4f3ab1cd/ls1028asi/block_copy_cmds_sdhc.rcw#L4 | 13:32 |
| minute | mhm https://github.com/nxp-archive/openil_rcw/blob/fda65ca642ec7f1cd906e929857d73de4f3ab1cd/ls1028asi/bootlocptr_sdhc.rcw | 13:32 |
| minute | josch: these two files are probably what we approximately need to do | 13:33 |
| minute | josch: josch this stuff is interpreted/compiled by rcw.py | 13:33 |
| minute | josch: the first argument to blockcopy is the device. i guess 0x08 = sd, 0x09 = emmc | 13:36 |
| minute | josch: i think the second arg is the destination in OCRAM (0x00100000) | 13:36 |
| + Ar|stote|is (~linx@149.210.3.83) | 13:36 | |
| minute | josch: 0x18010000 should be the offset of SPL on the sd card | 13:36 |
| minute | might be big endian though | 13:37 |
| - martins (QUIT: Remote host closed the connection) (924a802ee9@2001:bc8:1210:2cd8::15c) | 13:38 | |
| minute | OpenIL was "open industrial linux" | 13:38 |
| minute | public file https://www.nxp.com/docs/en/user-guide/OPEN-LINUX-IND-UM-1-10.pdf | 13:39 |
| minute | > OpenIL is a tool for building the final flash/SD image, including U-Boot | 13:39 |
| minute | josch: see page 44 | 13:41 |
| minute | josch: mmc erase 8 0x800 | 13:41 |
| minute | josch: mmc write 0x82000000 8 0x800 | 13:41 |
| minute | ah, sorry, that's 8 blocks, 4kb | 13:42 |
| minute | ah but bl31 they put at 0x800 | 13:42 |
| minute | some more info https://github.com/u-boot/u-boot/blob/master/arch/arm/cpu/armv8/fsl-layerscape/doc/README.lsch3 | 13:47 |
| + chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 13:49 | |
| minute | aha, i got it wrong, 0x00100000 is the source addr | 13:49 |
| minute | explainer https://github.com/u-boot/u-boot/blob/master/arch/arm/cpu/armv8/fsl-layerscape/doc/README.lsch3#L206 | 13:49 |
| + martins (924a802ee9@2001:bc8:1210:2cd8::15c) | 13:54 | |
| + jahkosha (~jahkosha@user/jahkosha) | 13:56 | |
| * jah -> Guest2871 | 13:56 | |
| - Guest2871 (QUIT: Killed (lithium.libera.chat (Nickname regained by services))) (~jahkosha@user/jahkosha) | 13:56 | |
| * jahkosha -> jah | 13:56 | |
| minute | josch: have you tried flashing the fsl-ls1028a-mnt-reform2-flash.bin to both 0x1000 and 0x801000, and then overwriting the start with a gpt partition table? | 13:57 |
| minute | josch: using github code search for blockcopy 0x00801000 i found this SVG showing a GPT layout for LS1028A SD card https://github.com/tano-systems/meta-tanowrt/blob/3ff38a82bce33e6d17bb250399d6e5d289917c81/docs/layers/meta-tanowrt-hsl-nxp/boards/ls1028ardb/images/ls1028ardb-layout-sd.svg?plain=1#L38 | 14:01 |
| minute | josch: it has uboot at 0x100000 (1 MiB) and RCW at 0x00801000 (8196 KiB) | 14:02 |
| minute | josch: whatever TanoWRT is, they have figured it out. maybe you can look at their code and carry that over? https://github.com/tano-systems/meta-tanowrt/blob/3ff38a82bce33e6d17bb250399d6e5d289917c81/docs/layers/meta-tanowrt-hsl-nxp/boards/ls1028ardb/index.rst#partitioning-layouts | 14:03 |
| minute | (awesome docs btw) | 14:03 |
| minute | josch: see these includes (Boot Location Pointer and Copy SPL to OCRAM) https://github.com/tano-systems/meta-tanowrt/blob/3ff38a82bce33e6d17bb250399d6e5d289917c81/meta-tanowrt-hsl-nxp/recipes-bsp/rcw/rcw/patches/0004-ls1028ardb-Add-SQPH-RCW-with-SD-card-boot.patch#L59 | 14:04 |
| - martins (QUIT: Remote host closed the connection) (924a802ee9@2001:bc8:1210:2cd8::15c) | 14:06 | |
| minute | josch: yeah, makes sense now. we need to add the following to our .rcw: | 14:06 |
| - paperManu (QUIT: Ping timeout: 264 seconds) (~paperManu@107.159.15.124) | 14:09 | |
| minute | josch: https://source.mnt.re/reform/reform-ls1028a-uboot/-/commit/f7effca43439de110182283e207df3d732877373 | 14:13 |
| minute | josch: but you would need to please figure out how to change the image building / concatenating itself so that uboot ends up at 0x100000 and rcw at 0x80001000 on the card, _or_ maybe it's possible to just replace the 0x100000 in my blockcopy line with for example 0x80100000, depending on where SPL is by default (i don't know where) | 14:15 |
| minute | anyway, i hope that's enough data to get you started | 14:16 |
| + martins (924a802ee9@2001:bc8:1210:2cd8::15c) | 14:28 | |
| jah | hey folks, I did just open my pocket, initially to look where I could add a NVMe SSD, anyway I get curious and looked at the assembly manual to understand what is what... there is one cable that I don't see mention in the assembly manual though. | 14:34 |
| jah | https://imgur.com/a/6qw7RCN | 14:34 |
| jah | this is the one marked in red where the arrow is pointing at, what is it? | 14:34 |
| minute | jah: yeah that's pretty new, it's the pcie+usb cable for the new wifi carrier that you have | 14:36 |
| jah | minute: okay makes sense thanks for explaining! | 14:36 |
| jah | also I'm curious where does come from the default 256GB of storage, is there an (mini)sdcard somewhere? some chip on the motherboard? | 14:37 |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 14:41 | |
| - XYZ_ (QUIT: Ping timeout: 240 seconds) (~XYZ@78-80-115-228.customers.tmcz.cz) | 14:49 | |
| minute | jah: emmc on rk3588 module | 14:52 |
| grimmware | minute: I've noticed that my wifi has notably improved since the rcore2 w/ Intel wifi | 14:53 |
| grimmware | I'm also impressed with the bluetooth range! | 14:53 |
| minute | grimmware: great! | 14:53 |
| grimmware | batteries seem to be doing sensible things too | 14:53 |
| grimmware | with the 7000mAH cells it just gooooooooes as well | 14:54 |
| grimmware | I'm having trouble leaving it alone to be honest | 14:54 |
| jah | yeah I missed all the meetings I had plan since I received my pocket, I can't leave it alone either | 14:56 |
| jah | minute: I see, thank you | 14:56 |
| + bkeys (~Thunderbi@h193.131.19.98.dynamic.ip.windstream.net) | 14:58 | |
| bkeys | ACTION basks in victory | 14:58 |
| minute | bkeys: great to hear btw... how did you recover from the issue? and will you share some pix/vids from fedora on reform? | 15:00 |
| minute | grimmware: haha nice, which 7mAh cells did you get? | 15:01 |
| bkeys | minute: I'm convinced it was a miracle, I was trying to get it to go into maskrom mode by shorting the pads on the SoM like it said in the firefly wiki, and it booted up Linux | 15:01 |
| bkeys | I have no technical explanation or reason why that worked | 15:02 |
| grimmware | minute: same ones as josch | 15:02 |
| minute | bkeys: ah, maybe those pins overrode emmc booting | 15:02 |
| grimmware | I can dig out the forum thread if you want to know more :) | 15:02 |
| bkeys | Maybe | 15:02 |
| bkeys | https://imgur.com/8LAcwrU | 15:02 |
| bkeys | minute: https://imgur.com/a/8LAcwrU | 15:04 |
| bkeys | I am also typing this on my Fedora Reform | 15:04 |
| minute | bkeys: awesome | 15:05 |
| bkeys | I will document and make a post on community.mnt.re in the near future | 15:05 |
| minute | bkeys: great, i just wanted to ask that | 15:05 |
| bkeys | I'm glad about this, I have wanted to run Fedora on this thing since I got it almost 4 years ago | 15:06 |
| bkeys | I followed minute for a long time, I remember watching the Reform1 and IMX6 | 15:06 |
| bkeys | also minute the keycaps y'all put over my trackball is nice, and the trackball is indeed much nicer | 15:07 |
| bkeys | I gotta go run a bike shop now, afk | 15:08 |
| jah | an other "just curious" question, how was the default background image generated? | 15:09 |
| - mjw (QUIT: Ping timeout: 256 seconds) (~mjw@gnu.wildebeest.org) | 15:20 | |
| * Guest3787 -> mjw | 15:30 | |
| - buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org) | 15:35 | |
| + buckket (~buckket@vps.buckket.org) | 15:36 | |
| + spew (~spew@user/spew) | 15:55 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 16:06 | |
| + XYZ_ (~XYZ@78-80-115-228.customers.tmcz.cz) | 16:07 | |
| minute | jah: i did it in blender | 16:09 |
| minute | jah: basically it's the mnt logo extruded and subdivided | 16:10 |
| jah | minute: yeah I eventually realized it must have been the mnt logo (very nice btw), I just wondered if you did the extrusion in a generative manner | 16:10 |
| jah | looks so neat <3 | 16:10 |
| jah | I'm having so much fun hacking this thing, thanks so much for building such a cool piece of hardware | 16:11 |
| + gordon1 (~gordon1@citadel.niflheim.info) | 16:16 | |
| - elb (QUIT: Remote host closed the connection) (~elb@2600:4041:6682:ea00:c6a7:e77f:1e0b:37ab) | 16:24 | |
| + elb (~elb@2600:4041:6682:ea00:dad4:689b:3c70:a564) | 16:24 | |
| + wielaard (~mjw@gnu.wildebeest.org) | 16:35 | |
| grimmware | it's amazing isn't it :) | 16:37 |
| grimmware | try replacing the trackball with a silicon nitride one :D | 16:37 |
| grimmware | fun hack for like $5 | 16:38 |
| grimmware | so far I've installed accelerometers (although I'm not currently running my sysctl or kernel module patch for them), replaced the trackball, upgraded the cpu module once and the carrier twice, upgraded the charger, replaced all the keyswitches with kailh sunsets (apart from the mouse buttons, I kept them clicky), made a custom side panel for accessing UART, removed it again, 3d printed | 16:40 |
| grimmware | replacement battery retainers, removed them again when I installed bigger batteries... | 16:40 |
| jah | grimmware: yeah I can feel I just started a long journey :) | 16:40 |
| jah | for now I'll be trying to add keyboard extra layer and "tap vs hold" feature that I'm used to in QMK, but thank you I'll keep that in my endless list of stuff to do with it! | 16:40 |
| jah | grimmware: wow, okay I see that is quite a rabbit hole xD ... noooice | 16:40 |
| grimmware | jah: oooooh if you do tap-to-hold functions can you open an MR with upstream? I'm really starting to miss home row mods. | 16:41 |
| rick_ | grimmware did you 3dprint the custom side panel? if yes, do you mind sharing the files somewhere? 0.0 | 16:44 |
| jah | grimmware: yeah! that is the first thing I will focus on (in next days ... or weeks depending of how much time I can spend on it), I hope I'll be able to make it generic enough so it can be shared easily upstream, I'm optimistic, I'll take inspiration from QMK | 16:44 |
| grimmware | rick_: I did yeah, they're just some tinkercad files because I mashed it together out of other models and it's not tremendously practical | 16:45 |
| grimmware | lemme see if I have a picture somewhere | 16:45 |
| jah | actually, I'll first try to make the enter key act as right maj when hold (KC_SFTENT), that seems simpler than layer to get started and I also need that | 16:51 |
| grimmware | This is it installed https://usercontent.irccloud-cdn.com/file/afiMRSaL/1764863528.JPG | 16:52 |
| grimmware | https://usercontent.irccloud-cdn.com/file/4pa5DJPQ/1764863553.JPG | 16:52 |
| grimmware | https://usercontent.irccloud-cdn.com/file/8aR3DSZR/1764863576.JPG | 16:53 |
| grimmware | https://usercontent.irccloud-cdn.com/file/Hs8vKXC1/1764863589.JPG | 16:53 |
| grimmware | Oh yeah this was the other dumb shit I did https://usercontent.irccloud-cdn.com/file/wXZdIY7t/1764863621.JPG | 16:53 |
| grimmware | it's my favourite toy :) | 16:54 |
| jah | oh my god, you literally merged it with an ergo dox | 16:54 |
| rick_ | waohhhhh both a very nice <3 | 16:55 |
| rick_ | but you don | 16:55 |
| jah | grimmware: I notice an other of my favorite toy in the background from storz & bickel, rofl | 16:55 |
| jah | good taste sir | 16:55 |
| rick_ | 't have the thinkercat files anymore? | 16:55 |
| grimmware | rick_: I can pull them out for you but I can't promise they're any good! Also won't work if you have the new wifi carrier board | 16:56 |
| rick_ | ahhh see | 16:56 |
| rick_ | mhh i think ill make one myself then | 16:57 |
| rick_ | but thanks for the idea, i love it! | 16:57 |
| jah | grimmware: how do you use that UART connector? I mean what kind of stuff do you connect to it? | 16:58 |
| grimmware | Here’s the .stl https://usercontent.irccloud-cdn.com/file/hQNfDe8f/Terrific_Bombul-Kup.stl | 16:59 |
| grimmware | jah: you need something like an FTDI adapter and some du-pont cables so you can interact with u-boot (which can't render to screen on the rcore) when you want to pick which kernel you're booting | 17:00 |
| rick_ | <3 | 17:01 |
| - wielaard (QUIT: Remote host closed the connection) (~mjw@gnu.wildebeest.org) | 17:01 | |
| grimmware | you can access the uart anyway with the back off, I was just doing a lot of fucking around with different kernel builds at the time | 17:01 |
| + wielaard (~mjw@gnu.wildebeest.org) | 17:01 | |
| jah | grimmware: ooooh, I see, thanks. now that explains why I don't see the whole boot sequence as I usually do on other machines. | 17:02 |
| rick_ | also for serial console access, or not? | 17:02 |
| grimmware | I don't actually know, minute can you spawn a TTY on the uart? | 17:03 |
| minute | yeah josch fixed that recently in a bunch of uboots i thought | 17:03 |
| - spew (QUIT: Quit: WeeChat 4.7.2) (~spew@user/spew) | 17:03 | |
| cwebber | https://usercontent.irccloud-cdn.com/file/wXZdIY7t/1764863621.JPG this is kinda what I'm gonna be working on also | 17:42 |
| amospalla | cwebber: amazing. | 17:47 |
| minute | cwebber: that's a next level setup! | 17:49 |
| cwebber | oh that one is grimmware's! | 17:49 |
| cwebber | so what I'm going to do | 17:49 |
| cwebber | - have my pocket sit in the middle | 17:49 |
| - jordi (QUIT: Ping timeout: 240 seconds) (~jordi@79.117.156.55) | 17:50 | |
| cwebber | - have a ZSA Voyager split keyboard on either side https://www.zsa.io/voyager | 17:50 |
| cwebber | - then I have a portable monitor on a stand that sits *behind* the pocket https://www.zsa.io/voyager | 17:50 |
| cwebber | basically, all that stuff packs up easily into a backpack | 17:50 |
| cwebber | so I can use the pocket as a lap machine while in transit | 17:51 |
| cwebber | but then when setting up to work | 17:51 |
| cwebber | I have a *full workstation* | 17:51 |
| grimmware | I finally got a coat with big enough pockets to put my pocket in | 17:51 |
| cwebber | with a second monitor (that portable monitor is really light and is perfect behind the Pocket as a second screen!), an ergonomic shoulder-width keyboard setup | 17:51 |
| cwebber | and the ZSA Voyager also has an add-on trackball thing called the Navigator. But I don't need it! Because the Pocket is sitting in between and already has a trackball setup! | 17:52 |
| cwebber | I'm very excited about this | 17:52 |
| cwebber | basically I have to be on the road more than I'd like and ergonomics are important to me and my body, so while I'd like to just use the pocket, I often need a more desktop-like environment for longer periods | 17:52 |
| cwebber | I think it's going to look pretty wild with all of it set up. It's a lot of cables tho, lol | 17:53 |
| grimmware | cwebber: heartily recommend getting a bunch of right-angled flat cables that are exactly the right length from aliexpress | 17:54 |
| - gidzit (QUIT: Ping timeout: 240 seconds) (~gidzit@gidzit.org) | 17:54 | |
| cwebber | grimmware: good call | 17:54 |
| cwebber | grimmware: your setup is truly awesome, can I post that on the fediverse? do you have a fedi handle? | 17:55 |
| grimmware | yeah sure, I'm @grimmware@chaos.social | 17:55 |
| cwebber | josch: did I hear you mention using a powered usb-c hub? | 17:58 |
| cwebber | I'm curious more about that | 17:58 |
| cwebber | I've thought about adding it to my "backpack'ed portable pocket workstation" | 17:58 |
| cwebber | as described above | 17:58 |
| cwebber | I think if I were gonna design a cyberdeck today even with a *different* form factor, I'd start with the guts of the Pocket | 18:02 |
| cwebber | but it's also nice how easily it can be adapted into a cyberdeck foundation | 18:02 |
| cwebber | I love weird computer setups | 18:02 |
| grimmware | this ^ | 18:12 |
| - arminweigl (QUIT: Ping timeout: 256 seconds) (~arminweig@sourcehut/user/arminweigl) | 18:22 | |
| jah | that is absolutely amazing all the possibilities this approach open O_O | 18:23 |
| grimmware | jeez my masto is absolutely blowing up now heh | 18:27 |
| + arminweigl (~arminweig@sourcehut/user/arminweigl) | 18:31 | |
| cwebber | haha | 18:45 |
| cwebber | *tosses hair* I have a way of doing that ;) | 18:45 |
| grimmware | yeah you really do! | 18:57 |
| jah | I've been tinkering with anbernic device in the past, now seeing this pocket reform I wonder how much effort it would be to use a similar approach to build a handheld gaming device | 19:06 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 19:33 | |
| - pasta161 (QUIT: Quit: WeeChat 4.7.1) (~pasta161@user/xha) | 19:59 | |
| * mjw -> Guest3960 | 20:08 | |
| - Guest3960 (QUIT: Killed (tantalum.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d) | 20:08 | |
| * wielaard -> mjw | 20:08 | |
| + Guest3960 (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d) | 20:09 | |
| - chomwitt (QUIT: Ping timeout: 264 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 20:30 | |
| aloo_shu | there is one, based on something existing | 21:33 |
| aloo_shu | if it's not considered advertising, I can search for the link in the channel logs where I was told about it | 21:34 |
| + leony (~leony@2a02:8109:f083:bb00::5ef2) | 21:40 | |
| aloo_shu | ok got it, name is uConsole, builds on the raspberry ecosystem, less ambitious goals/ethics than mnt, but enough parallels to make comparing approaches potentially fruitful | 21:46 |
| kfx | I have that thing and it is ... not good | 21:51 |
| kfx | the case is very nice, the keyboard and mouse are ok when they work, and I would not buy it again | 21:52 |
| - gustav25 (QUIT: Quit: Quit) (~gustav@c-78-82-54-128.bbcust.telenor.se) | 22:15 | |
| + johl (~johl@dynamic-093-128-144-195.93.128.pool.telefonica.de) | 22:26 | |
| johl | Hey there! | 22:26 |
| grimmware | ugh my flicker is back | 22:26 |
| grimmware | minute: did the branch to fix flicker on v1 ever get merged? | 22:27 |
| grimmware | I upgraded kernel recently from my patched one | 22:27 |
| minute | grimmware: no it's not done | 22:27 |
| minute | grimmware: because it didn't work on imx8mplus | 22:27 |
| grimmware | oh lame | 22:27 |
| minute | i'm currently working on debugging dsi on barebox | 22:28 |
| minute | more later | 22:28 |
| johl | I fell very humble and stupid, but... How do y'all enter the backtick character on the Pocket R3eform keyboard? | 22:28 |
| grimmware | hyper esc | 22:28 |
| johl | Thank you! | 22:28 |
| grimmware | minute: cool, I can just rebase my changes and build again :) | 22:29 |
| - leony (QUIT: Quit: leony) (~leony@2a02:8109:f083:bb00::5ef2) | 22:31 | |
| - johl (QUIT: Quit: leaving) (~johl@dynamic-093-128-144-195.93.128.pool.telefonica.de) | 22:34 | |
| + chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 22:42 | |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 22:51 | |
| vagrantc | hrm. after updating to the firmware with 2.2Ah for the cells hard-coded, it does not seem to fully charge ... i ran it till it shut off, charged it up, and the cells seem to be at 3.1v at 87% charge | 22:52 |
| + espressocheese (~espressoc@2601:8c1:c180:4630:7154:e07a:54b1:b266) | 22:53 | |
| vagrantc | on the oled display ... does it round or truncate the voltage reading? e.g. would 3.19v register as 3.1v ... or 3.2v ? | 22:53 |
| vagrantc | is there some way to get a more precise reading of the voltage of individual cells? | 22:55 |
| espressocheese | im stuck at initramfs, can someone please help? | 22:56 |
| espressocheese | it decrypts the drive then goes to initramfs | 22:58 |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 23:15 | |
| + paperManu (~paperManu@107.159.15.124) | 23:16 | |
| + pomel0 (~pomel0@user/pomel0) | 23:23 | |
| + paperManu_ (~paperManu@107.159.15.124) | 23:32 | |
| espressocheese | I think I bricked my reform, pls help | 23:55 |
| espressocheese | its stuck on initramfs | 23:56 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
