| abortretryfail | I hope that works..? | 00:04 |
|---|---|---|
| josch | minute: once you can confirm that things do work as expected, please give me a small ping so that i can do some final touch-ups in reform-tools and reform-system-image | 00:16 |
| minute | josch: i need to rerun it now, but i noticed that errors in mmtarfilter are ignored https://source.mnt.re/reform/reform-system-image/-/jobs/2151#L6311 | 00:43 |
| josch | minute: the reason is shell quoting i think | 00:47 |
| josch | let me send you a patch... one sec... | 00:47 |
| josch | also, the reason that mmtarfilter errors are ignored is, that in POSIX shell, only the last command of a pipeline has its exit status evaluated | 00:47 |
| josch | in bash there is pipefail to have it exit on any non-zero exit as part of a pipeline | 00:48 |
| minute | ah yeah | 00:48 |
| josch | minute: https://paste.debian.net/hidden/b809e9a3/ | 00:51 |
| josch | let me add something to prevent the pipeline succeeding if the earlier parts fail | 00:52 |
| josch | minute: that patch is of course untested -- my reform isn't that fast :) | 00:52 |
| minute | josch: thanks, i will apply it and test | 00:54 |
| minute | ok, need to sleep now! see ya tomorrow | 00:57 |
| josch | lets see what that patch does tomorrow then XD | 00:59 |
| - pandora (QUIT: Quit: Connection closed for inactivity) (uid585533@id-585533.ilkley.irccloud.com) | 01:01 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@utdpat241106.utdallas.edu) | 01:14 | |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 01:30 | |
| josch | aha, good i stayed awake a bit longer -- seems that the changes to reform-tools do something :D | 01:40 |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@utdpat241106.utdallas.edu) | 01:41 | |
| josch | minute: this should fix it: https://source.mnt.re/reform/reform-tools/-/merge_requests/54 | 01:48 |
| josch | i don't have the required privileges to push to main or accept reform-tools merge requests, otherwise i would've done so and also re-run the reform-debian-packages pipeline and scheduled reform-system-image for after that so that all is done by tomorrow | 01:49 |
| josch | i also have the fix for reform-system-image locally which avoids the CI pipeline from succeeding even though steps of that mmdebstrap pipe failed but it doesn't make sense for me to push those changes before having reform-tools fixed | 01:52 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 02:00 | |
| + jacobk (~quassel@129.110.242.140) | 02:01 | |
| - mtm (QUIT: Ping timeout: 258 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:04 | |
| abortretryfail | hm, i dont think the ???% after being unused for a while is a communication issue between the LPC and keeb. | 02:34 |
| abortretryfail | The SPI code in the LPC fw has its reached_full_charge check commented out, but the part that handles keyboard commands via uart checks that and replies "???%" | 02:36 |
| abortretryfail | something about sitting powered off for a few days resets reached_full_charge to 0? | 02:38 |
| - klardotsh (QUIT: Ping timeout: 260 seconds) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 02:42 | |
| abortretryfail | guess at this point my next task is to charge it up and let it sit for a couple days. | 02:45 |
| - eery (QUIT: Remote host closed the connection) (~eery@77.137.73.95) | 02:47 | |
| - nsc (QUIT: Ping timeout: 264 seconds) (~nicolas@103-99-142-46.pool.kielnet.net) | 03:46 | |
| + nsc (~nicolas@83-97-142-46.pool.kielnet.net) | 03:48 | |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@129.110.242.140) | 03:52 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:10 | |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 04:25 | |
| + bkeys (~bkeys@static-198-54-130-101.cust.tzulo.com) | 05:04 | |
| bkeys | Hi all, my LPC was fine after my batteries charged some | 05:04 |
| bkeys | But now my LPC won't light on at all. I reset the plug on the battery boards and still no issue | 05:05 |
| bkeys | I saw that the keyboard backlight was flickering between two low light levels and now it won't respond to anything | 05:05 |
| bkeys | The OLED isn't showing up with anything | 05:09 |
| bkeys | https://shop.mntre.com/products/mnt-reform-keyboard-oled-module?taxon_id=13 | 05:12 |
| bkeys | Maybe I need to get one of these? Hopefully it isn't anything too serious, I use the Reform a lot | 05:13 |
| bkeys | Or maybe my keyboard is somehow the issue? I'm not certain | 05:16 |
| bkeys | Well it's 11PM here, hopefully this can be worked out | 05:16 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 05:29 | |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@utdpat241106.utdallas.edu) | 05:58 | |
| + jacobk (~quassel@64.189.201.150) | 06:18 | |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 07:34 | |
| josch | bkeys: minute is super overworked these days -- if they do not reply to you here, maybe just send a mail to support@mntre.com to not loose track of the issue | 07:44 |
| - klardotsh (QUIT: Ping timeout: 258 seconds) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 08:44 | |
| - bkeys (QUIT: Ping timeout: 245 seconds) (~bkeys@static-198-54-130-101.cust.tzulo.com) | 09:24 | |
| - doppler (QUIT: Ping timeout: 240 seconds) (~doppler@user/doppler) | 09:29 | |
| + doppler (~doppler@user/doppler) | 09:33 | |
| + MajorBiscuit (~MajorBisc@2001:1c00:31c:8400:f184:4168:559b:d91b) | 09:40 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 10:12 | |
| minute | josch: thank you very much, merged and rerunning | 10:31 |
| + mjw (~mjw@gnu.wildebeest.org) | 10:37 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@78-80-18-84.customers.tmcz.cz) | 10:44 | |
| + XYZ (~XYZ@78-80-18-84.customers.tmcz.cz) | 10:47 | |
| josch | minute: nice! after the reform-system-image pipeline succeeds (i assume you want an image as fast as possible) i have some more updates for it | 10:51 |
| - XYZ (QUIT: Read error: Connection timed out) (~XYZ@78-80-18-84.customers.tmcz.cz) | 11:17 | |
| + XYZ (~XYZ@78-80-18-84.customers.tmcz.cz) | 11:35 | |
| + IchikaZou (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:42 | |
| - IchikaZou (QUIT: Remote host closed the connection) (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:43 | |
| + IchikaZou (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:43 | |
| - IchikaZou (QUIT: Remote host closed the connection) (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:43 | |
| + IchikaZou (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:48 | |
| abortretryfail | bkeys, did you try the reset button on the keyboard? The OLED is driven by the AVR on the keyboard, not the LPC controller on the motherboard. | 11:48 |
| - IchikaZou (QUIT: Remote host closed the connection) (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:48 | |
| + IchikaZou (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:50 | |
| - IchikaZou (QUIT: Remote host closed the connection) (~IchikaZou@114-140-105-83.adsl.fetnet.net) | 11:51 | |
| josch | minute: reform-tools 1.27 made it to mntre.com/reform-debian-repo -- i just restarted the failed reform-system-image job. | 12:01 |
| josch | minute: you started another job before? why? that one still has reform-tools 1.26 | 12:01 |
| violet | the imx8mq really is so impressive to me | 12:22 |
| violet | like nominally the rock64 has similar specs. cortex a53, similar clock rate. dont know the GPU for sure but i think its similar capabilities? thats the part im least sure about | 12:23 |
| violet | but i cannot watch a 480p video smoothly at 30fps with software decoding | 12:24 |
| violet | but i absolutely can on the imx8mq it works just fine even with fractional scaling, its wild | 12:24 |
| Boostisbetter | I would agree. I have it in the Reform and in the Librem 5. It is a solid CPU and I have been very impressed with its performance. On the Reform especially since a PCIe lane is exposed and we get to use an NVME with it. | 12:30 |
| violet | yeah nvme is huge | 12:31 |
| + eery (~eery@77.137.73.95) | 12:31 | |
| violet | i run my quartz64 with an nvme in the pcie slot and its great | 12:31 |
| violet | at some point i think i want to replace the wifi with something 5ghz capable | 12:32 |
| Boostisbetter | I rave about the design of the Reform all the time, but I just can't not miss a chance to bring up how well designed the whole thing is. | 12:32 |
| Boostisbetter | I would bring the Reform everywhere if I could. | 12:32 |
| violet | because i use a 2.4ghz wireless headset (usb dongle) and it creates interference with the wifi when i use it | 12:32 |
| violet | just an intel chip probably. i think i have one from a thinkpad i dont use that i can slot in | 12:33 |
| violet | yeah thats the thing i wrote about on my blog is like the build quality | 12:34 |
| violet | im so used to sketchy sbc vendors its night and day | 12:34 |
| Boostisbetter | voilet: can you link your blog? I'd like to read what you wrote about the Reform, especially since you build your own. | 12:39 |
| violet | https://artemis.sh/2023/07/06/thoughts-on-mnt-reform.html this was my post. although 2 months later i think i was too harsh on the cpu/ram spec | 12:40 |
| violet | in practice is really hasnt been a problem like 95% of the time | 12:40 |
| violet | but i stand by the rest | 12:40 |
| violet | its mainly just a problem with compiling and theres really no wayaround that. compling is just expensive for CPU | 12:42 |
| Boostisbetter | that was the gist of my basic fears as well. I underestimated how much an NVME base swap would basically invalidate my concerns. | 12:43 |
| violet | i can cross compile big things a lot of the time anyway | 12:44 |
| Boostisbetter | I hear you on the X220. That was the last of the great Thinkpad line really. From there the cover story of making everything thinner gave them excuses to make them cheaper and lacking their features. | 12:46 |
| Boostisbetter | My old X220T still works perfectly. | 12:47 |
| Boostisbetter | I have also found ZERO issues with using Wayland on the Reform. | 12:47 |
| Boostisbetter | I actually prefer Wayland to X. | 12:47 |
| minute | josch: sorry, i had a lapse of thought! | 12:49 |
| minute | josch: forgot that the debian packages don't make it immediately to the repo | 12:49 |
| josch | i think it's updated every full hour? | 12:50 |
| minute | josch: yes ^^ | 12:50 |
| josch | well, the job i triggered has now been running for half an hour already, so we should soon see results :) | 12:52 |
| minute | great | 12:55 |
| josch | okay, there is still an error | 13:36 |
| josch | let me try locally | 13:36 |
| josch | oh shoot | 13:36 |
| minute | what's the issue? | 13:37 |
| + bkeys (~bkeys@c-67-191-106-248.hsd1.fl.comcast.net) | 13:37 | |
| josch | the issue is that i am bad with escaping line-continuators :D | 13:38 |
| josch | i put a backslash at the end of a line where there should be none | 13:38 |
| bkeys | abortretryfail Where is the reset button on the keyboard? | 13:38 |
| josch | bkeys: https://mntre.com/reform2/handbook/parts.html#keyboard | 13:40 |
| josch | bkeys: in that image there is a label "Reset" pointing right below the flashing switch | 13:41 |
| bkeys | Okay so I don't need to remove the carrier board | 13:41 |
| josch | bkeys: are you in the process of flashing a new keyboard firmware? | 13:42 |
| bkeys | No | 13:42 |
| josch | yes, resetting the keyboard is possible just by removing the frame around it | 13:43 |
| bkeys | josch Yeah I remove frame, plugged it up to power (I have disconnected battery boards for safety) and still it doesn't light up or anything | 13:45 |
| josch | bkeys: probably only minute can properly help you here | 13:46 |
| bkeys | Yeah, I'll remove the ribbon cable and make sure it's good and in there although I don't know what could have wiggled it loose | 13:47 |
| bkeys | It spent all night disconnected from power | 13:47 |
| minute | josch: in reform-system-image or debian packages? | 13:47 |
| minute | bkeys: you say your oled doesn't work anymore, right? but does the keyboard itself work? you can disconnect the oled completely and the keyboard should still work | 13:48 |
| bkeys | Yeah I don't have solid suspicions but the OLED doesn't light up | 13:48 |
| minute | bkeys: if it doesn't, then either the keyboard is broken or the keyboard doesn't get power from the mainboard | 13:48 |
| josch | minute: in reform-system-image. I put the fix together with another improvement that makes sure that it doesn't succeed even though it failed here: https://source.mnt.re/reform/reform-system-image/-/commit/d392995d28abc8ccd4c650744931d20607dca647 | 13:48 |
| bkeys | minute The last time the OS was on here the keyboard wouldn't work, but the trackball did | 13:48 |
| minute | bkeys: you said something about flickering backlight IIRC, do you mean the backlight of the keyboard itself? | 13:48 |
| bkeys | Yes | 13:49 |
| bkeys | Flickering between two low states of backlight | 13:49 |
| minute | bkeys: ok, then the keyboard is very probably broken | 13:49 |
| bkeys | Hmm, I didn't do anything to it though | 13:49 |
| bkeys | Can't I remove it and try to use it standalone to verify? | 13:49 |
| minute | bkeys: only if it is v3 keyboard, or has U3 soldered, or you have another way of providing 3v3 on the UART port of the keyboard | 13:50 |
| minute | bkeys: if the keyboard isn't older than 2 years we can do warranty repair or exchange | 13:50 |
| minute | bkeys: other option is to order v3 keyboard with "nocase" option, we have them in stsock | 13:51 |
| bkeys | Okay, I already placed an order for the cm4 adapter boards; maybe you could just mail me a new keyboard when you send those out? | 13:51 |
| minute | bkeys: yes, then please order v3 keyboard with "no case" option and write to support@mntre.com asking to combine the orders | 13:51 |
| bkeys | Alright; will yall just refund me since it's in warranty? Do you want my old keyboard? | 13:52 |
| minute | bkeys: warranty works only if you send in your old keyboard first because we need to look at it | 13:52 |
| minute | bkeys: and we can repair it, which might be more economical for us than sending a new one and throwing the old one away | 13:52 |
| bkeys | Okay, I will likely do that then | 13:53 |
| bkeys | I take it you are doing the repairs and are rather busy? | 13:53 |
| minute | bkeys: yes | 13:53 |
| bkeys | Alright. I ought to have the email from the last time I sent you this thing with the address | 13:54 |
| minute | bkeys: the email is support@mntre.com | 13:55 |
| abortretryfail | You don't even need to remove the keyboard frame/bezel. You can reach it with the corner of a credit card if you press down the F4 key and slip the card under. | 13:56 |
| abortretryfail | I figured that out the other day after accidentally hitting circle+x and disabling my keyboard while I had no ext keyboard nor a screwdriver. | 13:57 |
| bkeys | abortretryfail Good to know, I already moved the bezel | 13:57 |
| bkeys | I'm gonna try to see if reseting the cable for the keyboard helps anything | 13:57 |
| bkeys | My keyboard says 2020-09-02, trying to find out what version it is | 14:01 |
| - eery (QUIT: Remote host closed the connection) (~eery@77.137.73.95) | 14:01 | |
| - mtm (QUIT: Ping timeout: 260 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:03 | |
| - mjw (QUIT: Ping timeout: 258 seconds) (~mjw@gnu.wildebeest.org) | 14:16 | |
| bkeys | minute I can't find the email, is this the correct address? | 14:28 |
| bkeys | Research GmbH | 14:28 |
| bkeys | Fehlerstr. 8 | 14:28 |
| bkeys | 12161 Berlin | 14:28 |
| minute | bkeys: MNT Research GmbH yes. it's also in the footer of mntre.com | 14:29 |
| bkeys | Yeah that's where I got it | 14:29 |
| bkeys | I will mail it (and the OLED just in case) to that address | 14:29 |
| minute | bkeys: ok, thank you, and please inform support@mntre.com of it and send the tracking number | 14:29 |
| minute | bkeys: otherwise it could get lost | 14:29 |
| bkeys | Will do | 14:29 |
| minute | thx! | 14:29 |
| bkeys | Thank you man | 14:30 |
| bkeys | It autocompleted to 8 Fehlerstraße for the address line | 14:30 |
| bkeys | Is that still right? | 14:30 |
| minute | that's fine | 14:31 |
| minute | just a different address format | 14:31 |
| bkeys | minute Do you have a preference for UPS or USPS? | 14:42 |
| minute | bkeys: definitely UPS | 14:47 |
| minute | bkeys: USPS will take months | 14:47 |
| bkeys | Okay I am sending it UPS then | 14:47 |
| minute | bkeys: thanks | 14:47 |
| minute | bkeys: if you don't get an OK price ask support@mntre.com for a ups label | 14:48 |
| + mjw (~mjw@84.241.205.78) | 14:48 | |
| minute | josch: debugfs: No such file or directory while trying to open /dev/mmcblk0p1 | 14:51 |
| minute | josch: how is this debugfs command supposed to work outside of being in a chroot in the disk image? | 14:52 |
| minute | in the meantime i will comment this out locally and try to build only ls1028a and give it a test | 14:53 |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@84.241.205.78) | 14:53 | |
| + mjw (~mjw@84.241.205.199) | 14:57 | |
| bkeys | minute Me and my wife use this Pirate Ship service, it was less than $50 to get it to you UPS | 14:58 |
| bkeys | Hopefully if it lines up; you could send it back with the cm4 adapter boards I ordered and save on shipping | 14:59 |
| bkeys | I'll make note of it in the email | 14:59 |
| josch | minute: i should really not push stuff i haven't run locally | 15:03 |
| josch | minute: but running it locally also takes super long, so it's a trade-off... | 15:03 |
| josch | minute: the fix is (hopefully) this: https://source.mnt.re/reform/reform-system-image/-/commit/2615fd814944ff1ec98fdd22754ae2a34aede334 | 15:04 |
| - XYZ (QUIT: Ping timeout: 255 seconds) (~XYZ@78-80-18-84.customers.tmcz.cz) | 15:05 | |
| minute | bkeys: great! | 15:12 |
| josch | note to self: comment out "packages for graphical user interface" and "packages for networking" to get a far faster run of mkimage.sh | 15:16 |
| + XYZ (~XYZ@89-24-41-118.nat.epc.tmcz.cz) | 15:18 | |
| - mjw (QUIT: Ping timeout: 240 seconds) (~mjw@84.241.205.199) | 15:26 | |
| minute | josch: ah | 15:35 |
| minute | josch: btw without the debugfs line i got a ls1028a image locally | 15:35 |
| minute | while i was on the phone with HUBS about the machining messup | 15:35 |
| - bkeys (QUIT: Quit: Client closed) (~bkeys@c-67-191-106-248.hsd1.fl.comcast.net) | 15:44 | |
| minute | ls1028a image: bootloader works. bootloader loads kernel | 15:45 |
| minute | display backlight is up. display itself is not up | 15:45 |
| minute | [ 24.201593] cdns-mhdp-imx f200000.display: Direct firmware load for ls1028a-mhdpfw.bin failed with error -2 | 15:46 |
| minute | [ 24.224734] [drm:cdns_mhdp_firmware_init_imx8qm [cdns_mhdp_imx]] *ERROR* Failed to request MHDP firmware | 15:46 |
| minute | ok, lets see where the firmware went | 15:46 |
| minute | ok, /boot/ls1028a-mhdpfw.bin exists | 15:47 |
| minute | but /lib/firmware/ls1028a-mhdpfw.bin does not exist | 15:47 |
| minute | ah, so the issue is probably that the module is loaded only after systemd starts, it's not loaded in initramfs | 15:49 |
| minute | will try to add stuff to /usr/share/initramfs-tools/modules.d/reform.conf | 15:52 |
| + mjw (~mjw@89.205.142.121) | 15:55 | |
| josch | minute: another thing you can try is to run unmkinitramfs on the generated initramfs to make sure that the initramfs really includes the firmware as expected | 16:00 |
| minute | oh wow, i didn't know that command was a thing | 16:00 |
| minute | also i think we can't use pci=nomsi on ls1028a :D | 16:00 |
| minute | josch: yes, lib/firmware contains ls1028a-mhdpfw.bin in the initrd, so that's god | 16:01 |
| minute | good | 16:01 |
| josch | minute: if pci=nomsi doesn't work with ls1028a then this means we cannot have a unified kernel cmdline for all reform SoMs and this means that u-boot-menu is broken until it learns to honor ${bootargs} :( | 16:03 |
| josch | i'll have to re-read some of this chat history to remind myself what we agreed upon last time with vagrant | 16:03 |
| minute | josch: i've seen that flash-kernel has a template for bootscript | 16:07 |
| minute | josch: can't we just replace that and set the commandline in there? | 16:07 |
| minute | ah, u-boot-menu. the thing i never tried yet | 16:08 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 16:10 | |
| josch | minute: flash-kernel and boot.scr are no problem because it honors ${bootargs} as set by u-boot itself | 16:11 |
| josch | minute: flash-kernel is also no problem because we can do board-specific things in /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv | 16:12 |
| minute | ok | 16:12 |
| josch | u-boot-menu is a problem because it sources its configuration from /usr/share/u-boot-menu/conf.d/reform.conf without differentiating between different boards and while ignoring ${bootargs} from u-boot | 16:13 |
| josch | the "solution" is probably that reform-tools should stop shipping /usr/share/u-boot-menu/conf.d/reform.conf and those who want to set up u-boot-menu have to write their own /etc/u-boot-menu/conf.d/reform.conf | 16:13 |
| josch | minute: you should give u-boot-menu a try nevertheless because it allows you to choose which kernel you want to boot right at the beginning | 16:16 |
| josch | just like you would do with grub | 16:16 |
| josch | so i can install a new kernel and if it borked my system i just choose to boot the old one instead | 16:16 |
| josch | this of course is only useful with imx because the other platforms do not have display on via u-boot | 16:17 |
| minute | ok, i'm currently struggling with debugging the initramfs | 16:17 |
| minute | i added break, or break=premount etc but i don't get a shell in initramfs. it just stops | 16:18 |
| minute | so frustrating aaahhh | 16:19 |
| minute | (sorry) | 16:19 |
| josch | sorry, i cannot be of much help right now -- i have my daughter now for the next six hours | 16:20 |
| josch | times between 9:00 and 15:00 UTC+2 i'm way more flexible XD | 16:20 |
| minute | ohh yeah, no worries! | 16:21 |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@89.205.142.121) | 16:30 | |
| Boostisbetter | minute: have you though about possibly making an upgrade for the Reform and Pocket that would allow it to have a touchscreen? | 17:21 |
| minute | Boostisbetter: for reform i think it doesn't make sense. for pocket ~maybe~. there is an existing touchscreen frame for the display we use. | 17:23 |
| minute | Boostisbetter: i was thinking to make a chunky industrial-ish tablet from one half of pocket later | 17:23 |
| minute | but r/n i'm caught up in the grind of making all the existing things good, stable, and shipped. | 17:24 |
| Boostisbetter | minute: inspired thinking. I could use cases for the Reform, but i agree with the excellence of the trackball and keyboard there, it isn't necessary. The Pocket will have that as well, but the size will make touch beneficial there. | 17:24 |
| Boostisbetter | minute: no worries, I am very happy with my Reform right now. More like an idea for when things slow and you have time to spend on other things for the Reform lineup. | 17:25 |
| minute | yeah | 17:25 |
| Boostisbetter | I can't wait to get the Pocket. I have so many plans for the that little guy! | 17:25 |
| minute | the future "new products" pipeline looks like this currently: 1. rk3588 module, 2. reform next | 17:25 |
| Boostisbetter | minute: thanks for sharing. | 17:26 |
| minute | in the meantime we have rcm4+a311d, which will also be interesting for pocket | 17:26 |
| minute | and someone could try that risc-v thingy with rcm4. | 17:26 |
| Boostisbetter | minute: for the pocket I am going to want to use whatever support standby the best. I appreciate all the hard work regardless! | 17:26 |
| minute | in parallel we'll also work on finishing rack reform for shipping (on hold while we ship other things) | 17:26 |
| minute | ok, i am in ls1028a initrd shell (i think having console=tty1 in cmdline prevents that from working over serial) | 17:30 |
| minute | cdns modules are _not_ loaded | 17:30 |
| minute | modprobe cdns_mhdp_imx works but doesn't suffice | 17:31 |
| minute | now the old game of, which module could be missing | 17:32 |
| Boostisbetter | minute: I know the pain you suffer, but I'm grateful for your efforts. | 17:36 |
| Boostisbetter | minute: keep fighting the good fight! | 17:36 |
| minute | Boostisbetter: thanks for your support! | 17:36 |
| minute | aha, mali-dp module is missing | 17:38 |
| unixpoet | minute: I imagine the plan for the rk3588 is to wait until the GPU driver is finished? | 17:41 |
| minute | unixpoet: no, it's waiting for me to have time for it | 17:43 |
| minute | but i have to work on shipping rcm4, ls1028a, keyboards, and design finishing and buying for pocket reform first | 17:44 |
| minute | and there we go, display in initramfs! | 17:46 |
| minute | now i only need to resolve this pci commandline problem and we should be ready to go | 17:48 |
| minute | ok, so we need another round of reform-debian-packages building. | 17:52 |
| Boostisbetter | minute: I appreciate the transparency | 17:53 |
| Boostisbetter | sounds like great progress is getting made. | 17:53 |
| Boostisbetter | most be thrilling on your side as well. I always get euphoric when my project starts working and taking shape. | 17:53 |
| Boostisbetter | one of the best feelings ever. | 17:53 |
| Boostisbetter | violet: The Tangara is EXACTLY what I have been thinking about doing myself. I was a big iPod user, and i love the dedicated music device, and what this is here is far beyond the capabilities of them. Really interested and would love to get many of them (5). | 17:54 |
| minute | i hope crowd supply will accept the tangara, it will be a good fit | 17:59 |
| minute | > Other platforms set ${bootargs} to overwrite imx8mq specific values defined here | 18:02 |
| minute | hmm, the problem is that that is not entirely true i think. if i set bootargs in u-boot, these merely get appended to those existing defaults | 18:03 |
| minute | ok, josch so i will try to implement your suggestion of doing a switcheroo in flash-kernel/ubootenv.d/00reform2_ubootenv | 18:06 |
| minute | it always feels painful to move these things into reform-tools because the reform-debian-package build takes so long | 18:07 |
| minute | josch: is this going in the direction you imagined? https://source.mnt.re/reform/reform-tools/-/commit/4174efbd85ceabe3ef31c0fc629773cf8833e76e | 18:19 |
| + mjw (~mjw@89.205.132.216) | 18:19 | |
| Boostisbetter | one little odd way I use my Reform is as a 5v USB source for a desk fan I use in the summer months. Really appreciate the power delivery there. Hahaha | 18:28 |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@64.189.201.150) | 18:35 | |
| minute | hehe | 18:35 |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@89.205.132.216) | 18:48 | |
| - Boostisbetter (QUIT: Ping timeout: 255 seconds) (4a410829d7@irc.cheogram.com) | 18:52 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 18:53 | |
| + mjw (~mjw@89.205.132.216) | 18:56 | |
| - mjw (QUIT: Remote host closed the connection) (~mjw@89.205.132.216) | 18:57 | |
| + mjw (~mjw@89.205.132.216) | 18:57 | |
| - MajorBiscuit (QUIT: Ping timeout: 240 seconds) (~MajorBisc@2001:1c00:31c:8400:f184:4168:559b:d91b) | 19:00 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 19:01 | |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@89.205.132.216) | 19:03 | |
| josch | minute: yes, that's exactly what i had in mind | 19:16 |
| josch | ideally, this would be set by u-boot itself but since imx u-boot doesn't set the right things and since we cannot upgrade u-boot i think we have to stick to this way of doing things... | 19:17 |
| josch | and yes, as you said above, u-boot's ${bootargs} only get appended to the defaults | 19:20 |
| josch | ideally, those defaults are *empty* | 19:20 |
| josch | we had them set only because the original imx u-boot didn't set them at all | 19:20 |
| josch | if we could time-travel it would've been best to have the flash-kernel u-boot default cmdline empty so that it doesn't overwrite the ${bootargs} set by u-boot | 19:20 |
| minute | josch: gotcha | 19:26 |
| + mjw (~mjw@84.241.193.198) | 19:28 | |
| - bluerise (QUIT: Ping timeout: 252 seconds) (~bluerise@p5b0acd9a.dip0.t-ipconnect.de) | 19:32 | |
| minute | ok, packages are done, building more system images. | 19:33 |
| josch | minute: with that change in reform-tools, mkimage.sh can now stop writing out /etc/default/flash-kernel which further reduces the stuff that has to be in /etc -- i'm gonna prepare a merge request tonight for that | 19:34 |
| josch | minute: getting here indeed was painful but i think it was also a *huge* change to switch the whole machinery from imx-only to three-boards-supported. I think things are set up much better now :) | 19:35 |
| josch | oh the tangara is being proposed to crowd supply! i'm so getting myself that one! | 19:37 |
| + bluerise (~bluerise@pc19f8f03.dip0.t-ipconnect.de) | 19:37 | |
| josch | "it's supposed to sound like that, it's electronic music" i giggled at that :D | 19:42 |
| + jacobk (~quassel@129.110.242.224) | 19:44 | |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 19:47 | |
| - mjw (QUIT: Ping timeout: 245 seconds) (~mjw@84.241.193.198) | 19:48 | |
| minute | josch: haha same | 19:59 |
| + mjw (~mjw@84.241.196.7) | 20:00 | |
| minute | josch: not sure if you've seen this https://source.mnt.re/reform/reform-system-image/-/commit/a63e8bf7f3ac23c747bfd538b03a51a332178b1e | 20:00 |
| minute | (this time i did roll out the reform-debian-packages before rerunning btw) | 20:01 |
| josch | minute: awesome, that commit does what i had in mind, thank you! :) | 20:02 |
| josch | ah maybe the "We still want to overwrite" comment should go above the /etc/default/flash-kernel line but that's just cosmetics :) | 20:02 |
| minute | true, true! | 20:05 |
| josch | i feel much better that those kernel args now live in reform-tools because it means that we can change them via apt upgrade and don't have to rely on people somehow getting the current up-to-date cmdline from somewhere | 20:07 |
| minute | cool | 20:13 |
| Boostisbetter | so I just checked updates, and i see tha tthe kernel has an update. | 20:16 |
| Boostisbetter | My question is, should i just keep using the kernel I am now, because it has been so great with suspend? Do we feel like updates to the 6.5 kernel are going to cause another regression? | 20:16 |
| + wielaard (~mjw@84.241.198.27) | 20:17 | |
| - mjw (QUIT: Ping timeout: 255 seconds) (~mjw@84.241.196.7) | 20:19 | |
| * wielaard -> mjw | 20:21 | |
| josch | Boostisbetter: if you do not uninstall the old kernel, you can always go back to it | 20:26 |
| josch | Boostisbetter: installing a new kernel version does not remove your old one | 20:26 |
| josch | Boostisbetter: example from my system: https://paste.debian.net/1292733/ | 20:28 |
| minute | Boostisbetter: those updates are also a side effect of us doing the integration work for the other 2/3 platforms | 20:32 |
| minute | Boostisbetter: they don't necessarily benefit you right now | 20:33 |
| minute | Boostisbetter: although i think a fix that i did for rcm4/a311d did probably improve suspend behavior for you... because i made sure that the display converter is never reset | 20:33 |
| + wielaard (~mjw@89.205.128.70) | 20:35 | |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@84.241.198.27) | 20:37 | |
| Boostisbetter | excellent. Thanks! | 20:37 |
| - wielaard (QUIT: Ping timeout: 260 seconds) (~mjw@89.205.128.70) | 20:40 | |
| Boostisbetter | anybody else here having an issue with the flatpak Signal on the Reform not having a resizable contact list? | 20:47 |
| minute | hmmm josch i did something wrong, the generated image is still running the old bootscript and has old commandline args. hmm. | 21:12 |
| minute | looks like new reform-tools didn't make it over | 21:13 |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@129.110.242.224) | 21:18 | |
| minute | lol i probably just downloaded an old image somehow | 21:18 |
| minute | no, i didn't unpack it and wrote an old one to the sd card :3 | 21:20 |
| minute | + test ls1028a = ls1028a | 21:25 |
| minute | + setenv bootargs console=ttyS0,115200 console=tty1 ro no_console_suspend cma=512M@3G libata.force=noncq nvme_core.default_ps_max_latency_us=0 cryptomgr.notests iommu.passthrough=1 arm-smmu.disable_bypass=0 | 21:25 |
| minute | ok so this works | 21:25 |
| + jacobk (~quassel@129.110.242.224) | 21:26 | |
| minute | but: [ 6.342355] cdns-mhdp-imx f200000.display: firmware: failed to load ls1028a-mhdpfw.bin (-2) | 21:26 |
| vkoskiv | Is there a TLDR of the current state of suspend on imx8? I guess I could run updates and try :D | 21:28 |
| minute | vkoskiv: i think it should mostly work now, except if you are bitten by nvme not coming up after wake | 21:29 |
| vkoskiv | Cool, I'll give it a go | 21:29 |
| - murphnj (QUIT: Quit: Leaving) (~murph@user/murphnj) | 21:29 | |
| minute | vkoskiv: Boostisbetter is def. the most experienced with suspend testing | 21:30 |
| vkoskiv | 910 updates :D | 21:32 |
| vkoskiv | I'd love to try and investigate the sleepy nvme issue more if it still happens on mine | 21:32 |
| minute | josch: also i just realized one can now tweak the commandline by editing that ubootenv script in /usr/share/flash-kernel ! | 21:34 |
| vkoskiv | This long overdue 910 package update is a 1601MB download. So a very very tiny macOS update, basically? | 21:37 |
| minute | vkoskiv: haha | 21:37 |
| vkoskiv | Only this doesn't take 45 minutes to install and show me a blank screen and 3 different progress bars during it | 21:37 |
| minute | josch: ok megabizarre. after running update-initramfs -u on the system (to debug initramfs), suddenly the initramfs does include the firmware and loads it... | 21:38 |
| minute | vkoskiv: so we still have to implement those :D | 21:42 |
| minute | (otherwise it's not a professional update) | 21:42 |
| vkoskiv | Yeah, and it specifically has to be a progress bar that resets at least once, around 2/3 through the process. | 21:43 |
| Boostisbetter | with the weather finally cooling down in Germany, I'll be able to take my Reform again to work. Can't bring it in the building, but on my lunch break I get to come out and take a breather using this beautiful thing to do something productive. | 21:48 |
| vkoskiv | Cooling down to what? Because ours is heading to that really chilly damp ~10C very soon | 21:49 |
| vkoskiv | Curious why no Reform in the building, if you can say? | 21:50 |
| Boostisbetter | vkoskiv: I work in a secure facility. I got my Librem 5 approved, which is crazy, but also shows how inept the comsec folks are, but there is no way I get something that looks like a laptop in the building. | 21:51 |
| Boostisbetter | vkoskiv, the cooling down is really about how much sunlight will be hitting the vehicle and heating up the cabin. | 21:52 |
| vkoskiv | Even though the Reform is much more auditable than pretty much any mainstream laptop? :D | 21:52 |
| Boostisbetter | I'd rather not have the Reform bake in there. | 21:52 |
| vkoskiv | Boostisbetter: Ah, gotcha. Gets warm quick. | 21:52 |
| Boostisbetter | > Even though the Reform is much more auditable than pretty much any mainstream laptop? :D | 21:53 |
| Boostisbetter | that doesn't matter. wifi and bluetooth are the no nos. Classified material, etc. | 21:53 |
| vkoskiv | Yeah. I've spent a lot of time in buildings with those little lockers for all electronic devices. Kind of nice to just be without a phone sometimes. | 21:54 |
| Boostisbetter | it is, probably helping me not get cancer or something. | 21:54 |
| + murphnj (~murph@user/murphnj) | 21:56 | |
| minute | josch: ok it doesn't work because > W: unable to deduce target machine for reform initramfs-hook | 22:04 |
| vkoskiv | Interesting, mine fails to freeze 2 processes when trying to suspend, and then comes back up | 22:05 |
| vkoskiv | Looks like maybe one of them was udev-worker? | 22:06 |
| vkoskiv | minute: https://pastebin.com/4Sty2Atn | 22:09 |
| vkoskiv | Those appeared a short while after the logs that happened right as I tried to suspend twice | 22:09 |
| vkoskiv | I'm assuming it tries to format it a bit nicer for debugging, not sure | 22:10 |
| minute | josch: my theory is that mmdebstrap runs update-initramfs _before_ the --chrooted-customize-hook that sets the MACHINE runs | 22:14 |
| minute | josch: you then execute flash-kernel, but that doesn't update the initramfs | 22:15 |
| vkoskiv | I think I copied an unrelated part of the kernel log | 22:29 |
| vkoskiv | I just got it without suspending | 22:30 |
| vkoskiv | Yeah, these crypto tasks are regularly hanging | 22:30 |
| minute | vkoskiv: ah yeah this is a known problem. you need cryptomgr.notests in kernel commandline to get rid of this | 22:32 |
| + mjw (~mjw@ip-185-104-136-31.ptr.icomera.net) | 22:36 | |
| vkoskiv | Swapped molex -> laird. 2GB transfer over WiFi to my desktop took 1:49 with molex, 1:42 with laird | 22:44 |
| vkoskiv | Improvement! Will feel around a bit to see how much of an improvement I find | 22:44 |
| vkoskiv | Now it suspended, after another reboot. Let's see if I can get it back up | 22:45 |
| vkoskiv | minute: Here is the log of the suspend fail, it doesn't happen every time though. https://pastebin.com/GL9b828z | 22:50 |
| vkoskiv | I had one failed one like this, a successful suspend but unknown wake, and then another failed suspend which is this log | 22:50 |
| vkoskiv | I didn't get to the kernel log on the successful-ish one | 22:51 |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@129.110.242.224) | 22:57 | |
| minute | vkoskiv: as i said, known problem with broken cryptomgr tests | 22:58 |
| minute | vkoskiv: i suggest to edit /etc/default/flash-kernel and add cryptomgr.notests to LINUX_KERNEL_CMDLINE_DEFAULTS | 23:00 |
| minute | vkoskiv: then run update-initramfs -u | 23:00 |
| minute | josch: ok, i can confirm the whole initramfs firmware deployment rube goldberg machine works now finally. | 23:03 |
| minute | so we have a release candidate for ls1028a system image now! | 23:10 |
| + jacobk (~quassel@129.110.242.224) | 23:17 | |
| + erle (~erle@2a02:8109:da40:c4:b072:40f0:91a8:4e58) | 23:19 | |
| Boostisbetter | woot! | 23:20 |
| minute | time for feierabend | 23:21 |
| - mjw (QUIT: Ping timeout: 258 seconds) (~mjw@ip-185-104-136-31.ptr.icomera.net) | 23:43 | |
| violet | working caam seems to cut down the cpu usage of full disk encryption even more than using the armv8 aes instructions | 23:47 |
| minute | violet: this should be the case yeah | 23:54 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!