| + oliverD (~Thunderbi@user/oliverd) | 00:01 | |
| - oliverD (QUIT: Quit: oliverD) (~Thunderbi@user/oliverd) | 00:10 | |
| minute | with a little addition to reform-tools/initramfs-tools/reform.conf, mntsc + pocket panel driver seem to work fine as modules in initramfs, at least with rk3588 | 00:18 |
|---|---|---|
| josch | \o/ | 00:24 |
| + oliverD (~Thunderbi@user/oliverd) | 00:37 | |
| minute | josch: i merged reform-tools changes now but idk how to satisfy the last checks in the CI | 01:20 |
| josch | minute: don't worry about it, i'll take care of it | 01:23 |
| minute | josch: nice, thx! | 01:25 |
| josch | minute: system-controller/mnt-sc.c is now another dkms module? Then it should be installed like one? | 01:40 |
| minute | josch: no | 01:45 |
| minute | josch: the source is just in reform-tools. but it's converted into a patch and built in reform-debian-packages | 01:46 |
| minute | josch: https://source.mnt.re/reform/reform-debian-packages/-/blob/3875dc8bcbd173258fa30c8a9e215d50ef308278/linux/patches7.1/sc/1000-mnt-sc-driver.patch | 01:46 |
| minute | i now have a script that creates this patch from the source files so i don't have to do that manually anymore | 01:47 |
| josch | but now we have the source in two places, no? | 01:48 |
| josch | and the building of linux in reform-debian-packages does not depend on any specific version of reform-tools | 01:48 |
| minute | josch: i don't consider the patch a source | 01:48 |
| josch | why does the source not live in reform-debian-package or, even better: upstream? :) | 01:48 |
| minute | josch: for me it's just a kludge to get the source into the kernel | 01:48 |
| minute | josch: the driver isn't upstreamable yet imho, it's not clean enough | 01:49 |
| minute | josch: and not tested in public | 01:49 |
| minute | josch: actually tested by almost noone... | 01:49 |
| josch | minute: if system-controller/mnt-sc.c lives in reform-debian-packages, then the patch can be built automatically each time the kernel gets built instead of manually triggered | 01:49 |
| minute | josch: sure, fine with me! | 01:49 |
| josch | okay, then let me first push a commit which makes the pipeline of reform-tools green | 01:50 |
| josch | and then i'll think about how to best add this to refom-debian-packages without things becoming too messy | 01:51 |
| josch | obviously, you don't want to edit a patch as "source" | 01:51 |
| minute | yeah, as i said i have a first script now that does that | 01:51 |
| josch | this would be much easier if instead of patches, we would use git branches of the kernel sources | 01:51 |
| minute | behold :D https://source.mnt.re/reform/reform-debian-packages/-/blob/3875dc8bcbd173258fa30c8a9e215d50ef308278/linux/update-linux-patches.sh | 01:51 |
| josch | yeah, i saw that and was wondering where the file comes from :D | 01:52 |
| minute | josch: yes totally. all the patches should be generated automatically | 01:52 |
| minute | josch: we even have a kernel mirror for that utopian future, but it doesn't have those patchsets imported yet https://source.mnt.re/reform/linux-dev/ | 01:53 |
| minute | josch: ideally that would have a branch corresponding to a debian kernel version + the desired patchsets on top | 01:53 |
| josch | yes and the problem with that is, that the debian packaging does not include the upstream git tree | 01:53 |
| minute | josch: and reform-debian-packages CI would just convert those into .patches for debian tooling on the fly | 01:53 |
| josch | which means that we'd have to manually create that first | 01:54 |
| minute | josch: why would we need that? | 01:54 |
| josch | how else would you get the packaging parts? | 01:54 |
| minute | josch: they're in reform-debian-packages | 01:54 |
| josch | no, the packaging parts are in Debian unstable and a tarball with both the packaging as well as linux sources get downloaded by the script | 01:55 |
| minute | josch: i'm only considering how to replace the manually made patches7.x dirs | 01:55 |
| josch | oh i'm absolutely for replacing that cludge | 01:55 |
| josch | it would be easiest when dropping the debian packaging bit | 01:55 |
| minute | josch: and in my experience if you just use linux mainline 7.1 to rebase our patches for example, those patches will also work on the debian kernel 7.1 | 01:56 |
| josch | in my experience that is not the case | 01:56 |
| josch | because debian patches the kernel too | 01:56 |
| minute | josch: yeah i assumed you would be thinking of some bad effects of that... i wonder how often that affects patch compatibility for us? | 01:57 |
| minute | josch: i guess we would need to (auto-) import those patches as well into our kernel git branch | 01:58 |
| minute | so that we could use git rebase | 01:58 |
| minute | josch: so debian doesn't have a git somewhere with those things applied? | 01:58 |
| josch | yes, and then refresh those patches when debian refreshes them or live with minor breakage on each release | 01:58 |
| minute | yeah | 01:58 |
| josch | debian ships their patches as files only | 01:58 |
| josch | they don't store the linux upstream git as part of the packaging | 01:59 |
| josch | only the ./debian folder | 01:59 |
| minute | interesting, i wonder how the maintainer(s) manage to work with that | 01:59 |
| josch | i talked with ben hutchings abotu that last year and that led to the reworking of the abiname | 01:59 |
| josch | i need to talk with him again but didn't manage to this year because exams where at the wrong time | 01:59 |
| minute | in any case, i made a little release checklist here https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/188 | 02:01 |
| minute | you can help me with the first one for sure :3 | 02:01 |
| josch | yes, lets make a new release soon | 02:02 |
| minute | yay! | 02:02 |
| minute | now testing my newest kernel build on the rk3588 pocket... lets hope it's mostly robust now | 02:03 |
| josch | there is also more data on the sd-card read/write errors of meluzzy in the forum: I sent her an sd-card by post which worked fine for me on rk3588 with motherboard 3.0. It fails for her and she sent me the logs. I'll file an issue about that with all the details but that sounds like bad hardware on her end if the same card works for me but not for her? | 02:03 |
| josch | okay, reform-tools ci is green again | 02:04 |
| - paperManu (QUIT: Ping timeout: 245 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 02:05 | |
| minute | nice! | 02:06 |
| minute | josch: hmm yeah sounds strange about that sd card issue | 02:07 |
| josch | okay, i think i have a plan on how to auto-generate the patches based on our own branches in a fork of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git -- i'll put creating a POC for that on my todo list. Only 25 exams left to grade and then i have free time. :) | 02:08 |
| minute | josch: oh very nice, very much looking forward! and i hope that grading will go smoothly :D | 02:09 |
| minute | ok, panel comes up fine on the first cold boot | 02:10 |
| josch | \o/ | 02:10 |
| minute | battery status is also there | 02:10 |
| josch | oh arch/arm/boards/mnt-reform2-rk3588/defaultenv-mnt-reform2-rk3588/init/model | 02:11 |
| minute | yeah i merged that because it looked ok to me? | 02:11 |
| josch | nice! | 02:12 |
| josch | then i'll add that to reform-check now as well | 02:13 |
| - schalken (QUIT: Quit: Leaving) (~schalken@117-118-178-69.gci.net) | 02:14 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 02:16 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 02:17 | |
| minute | ok, after a bit of trial and error i have USB-C DP -> Apple HDMI adapter image on my rk3588 pocket reform with that kernel | 02:21 |
| josch | minute: these commits were part of what you just merged, right? https://source.mnt.re/reform/reform-tools/-/merge_requests/167 | 02:21 |
| josch | minute: congrats -- that was a long ride! | 02:22 |
| minute | josch: yes @ commit, but what i merged has a newer/improved version of that, so i'll close !167 | 02:23 |
| josch | thank you | 02:23 |
| josch | (checking whether anything else should be merged before release) | 02:23 |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 02:23 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 02:23 | |
| minute | josch: and thx about the long ride... indeed very long. and it's not over | 02:24 |
| minute | josch: next week i'll have to make the DP HPD stuff more robust. but good that i can test with a completley CI built package now | 02:25 |
| josch | phew :) | 02:25 |
| josch | did you already ship motherboard 2.0? | 02:25 |
| minute | not sure | 02:26 |
| minute | :D | 02:26 |
| josch | oh no XD | 02:26 |
| minute | we shipped 7 units of pocket with MB2 to one company, but not sure if they tried them | 02:26 |
| josch | they'll work fine if you don't need dp :) | 02:26 |
| minute | and some 6 pockets were installed for shipment yesterday/today but not sure if they already used MB2 | 02:26 |
| minute | josch: well i shipped them with a prerelease kernel and fws...... :0 | 02:27 |
| josch | minute: do you happen to be using a pocket reform with qcs6490? | 02:27 |
| - a3f (QUIT: Read error: Connection reset by peer) (~a.fatoum@flummi.grey.stw.pengutronix.de) | 02:28 | |
| minute | josch: not atm, but soon again | 02:31 |
| minute | josch: like, my personal pocket reform has rk3588 | 02:31 |
| josch | okay, then you'll see the reform-check output once that happens :) | 02:31 |
| + a3f (~a.fatoum@flummi.grey.stw.pengutronix.de) | 02:31 | |
| minute | alright! | 02:32 |
| minute | think i should go home now for the weekend, high time :D and i finally have a usable pocket reform again... | 02:32 |
| josch | have a good night and rest well! _o/ | 02:33 |
| josch | huh... i cannot find this in the log. Anybody remembers why I might have put a rmmod reform2_lpc && modprobe reform2_lpc for Pocket Reform at the end of reform-hw-setup? | 02:38 |
| josch | stephano: was this maybe related to one of the module loading order differences on arch? | 02:38 |
| + paperManu (~paperManu@204.48.0.160) | 02:46 | |
| + schalken (~schalken@117-118-178-69.gci.net) | 02:49 | |
| - paperManu (QUIT: Quit: WeeChat 4.1.1) (~paperManu@204.48.0.160) | 02:52 | |
| - XgF (QUIT: Remote host closed the connection) (~quassel@2001:19f0:5001:1174:2ca2:e8e5:a24:cc6f) | 02:56 | |
| + XgF (~quassel@2001:19f0:5001:1174:5400:2ff:fef2:1fa2) | 02:57 | |
| + paperManu (~paperManu@204.48.0.160) | 02:58 | |
| josch | reform-tools 1.88: https://source.mnt.re/reform/reform-tools/-/merge_requests/173 | 03:07 |
| josch | now building system-image with it and testing it tomorrow | 03:07 |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 03:10 | |
| + lidstah2 (~lidstah@gateway/tor-sasl/lidstah) | 03:10 | |
| stephano | josch: I've noticed on Arch that reform2_lpc is not loaded correctly on boot. So I've been running a script to basically do just that... `rmmod reform2_lpc && modprobe reform2_lpc`. I can do some testing to see if removing that line from reform-hw-setup makes any difference on Arch. | 03:44 |
| stephano | I also need to properly debug the issue. The workaround is so simple I haven't taken the time to track down the issue. | 03:45 |
| - xktr (QUIT: Ping timeout: 244 seconds) (~xktr@user/xktr) | 04:22 | |
| + xktr (~xktr@user/xktr) | 04:23 | |
| - Mot (QUIT: Server closed connection) (~mot@user/mot) | 04:33 | |
| + Mot (~mot@user/mot) | 04:33 | |
| minute | reform2_lpc's days are kind of counted though ^^ | 04:59 |
| - AnimaInvicta (PART: !!unknown attribute: msg!!) (~AnimaInvi@88-169-191-86.subs.proxad.net) | 05:07 | |
| - oliverD (QUIT: Quit: oliverD) (~Thunderbi@user/oliverd) | 06:41 | |
| + mjw (~mjw@66.183.205.74) | 07:01 | |
| - paperManu (QUIT: Ping timeout: 246 seconds) (~paperManu@204.48.0.160) | 07:36 | |
| + oliverD (~Thunderbi@user/oliverd) | 07:38 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 08:12 | |
| + RandyK (~RandyK@user/randyk) | 08:12 | |
| - marty (QUIT: Ping timeout: 252 seconds) (~marty@146.70.171.155) | 08:23 | |
| + marty (~marty@143.244.47.101) | 08:25 | |
| - marty (QUIT: Ping timeout: 276 seconds) (~marty@143.244.47.101) | 08:30 | |
| + marty (~marty@static-23-234-101-23.cust.tzulo.com) | 08:32 | |
| - austriancoder (QUIT: Server closed connection) (sid152545@hampstead.irccloud.com) | 08:33 | |
| + austriancoder (sid152545@id-152545.hampstead.irccloud.com) | 08:33 | |
| - mjw (QUIT: Ping timeout: 248 seconds) (~mjw@66.183.205.74) | 09:04 | |
| - oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd) | 10:07 | |
| + oliverD (~Thunderbi@user/oliverd) | 10:10 | |
| + lislegaard (~lislegaar@2a02:6f8:2020:196:b::173) | 11:51 | |
| lislegaard | Does the Pcket REform with rk3588 chip have a built in microphone? I am able to select it in settings, but I don't actually get any sound out of it, only some noise. | 11:52 |
| josch | lislegaard: it does not have a microphone | 11:54 |
| minute | lislegaard: you can connect a headset with microphone to the headset/headphone jack and there should be a signal then | 11:58 |
| lislegaard | aha! thanks! it is not important for me, but i was just confused if i messed up something in the build :) | 11:59 |
| lislegaard | the confusion came from the fact that if i for example in gnome audio settings select "microphone - built-in audio" for the input device i can then select between configuration: "microphone, speaker" or "headphones, microphone". | 12:01 |
| lislegaard | i also se for output there is speaker, headphones and analog output. what is the "analog output"? | 12:02 |
| + AnimaInvicta (~AnimaInvi@88-169-191-86.subs.proxad.net) | 12:05 | |
| - buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org) | 12:06 | |
| + buckket (~buckket@vps.buckket.org) | 12:07 | |
| minute | lislegaard: right, that is even more confusing: it's hdmi audio :D | 12:09 |
| AnimaInvicta | Hello everyone. The trackball on my MNT Reform classic stopped working. I replaced it with a new one from MNT, but that one doesn't work either. Has anyone here seen something similar? dmesgsays "usb 2-1.4: device descriptor read/64, error -110" and lsusb shows two hubs and a keyboard but no mouse/trackball. | 12:12 |
| lislegaard | thanks! i was suspecting as much, but was like "hey. hdmi is digital", hehe! would have tested, but since i already had you here for the other question. thanks :) | 12:12 |
| lislegaard | minute | 12:12 |
| AnimaInvicta | Later, the kernal log also says "device not accepting address 6, error -110" and "device not accepting address 7, error -110". | 12:16 |
| - lislegaard (QUIT: Quit: Client closed) (~lislegaar@2a02:6f8:2020:196:b::173) | 12:18 | |
| minute | AnimaInvicta: weird. maybe cable issue? | 12:19 |
| AnimaInvicta | minute: I'll try re-seating it. | 12:20 |
| josch | minute: suppose i'm on the old kernel but the new pocket reform sysctl 2.0 firmware. The output of /sys/bus/spi/drivers/reform2_lpc/spi1.0/firmware is just binary garbage. What would be the best way to find out that the pocket reform motherboard is on the new firmware *without* having the new kernel? | 12:25 |
| AnimaInvicta | No luck. The device number is 1-1.4 now, but the rest of the error message is the same. | 12:28 |
| minute | AnimaInvicta: what if you switch the keyboard and trackball connections on the mainboard? | 12:34 |
| minute | josch: maybe it's predictable binary garbage? | 12:34 |
| AnimaInvicta | Interesting idea. Let me check. | 12:35 |
| minute | josch: how does it look in hex? | 12:35 |
| josch | minute: too late, i just installed the new kernel version. Will investigate the binary later. | 12:36 |
| josch | my display has weird stripes now | 12:36 |
| josch | and the lower half is missing -- rebooting... | 12:37 |
| minute | josch: you also have the latest reform-tools yes? | 12:38 |
| josch | minute: yes, the version in git | 12:39 |
| josch | minute: so the initramfs should contain the modules -- but i'll check to make sure | 12:39 |
| minute | josch: the stripes+missing lower half is precisely the bug that this should have fixed :DD | 12:39 |
| minute | josch: did you also update the sysctl fw again with the latest git from my branch? | 12:39 |
| josch | i did not | 12:39 |
| minute | josch: please do! | 12:39 |
| josch | aha! lets go | 12:40 |
| minute | otherwise the gpio stuff will likely crash the sysctl because of interrupt foo | 12:40 |
| minute | josch: and you have uboot, yes? | 12:41 |
| josch | minute: no, barebox | 12:41 |
| josch | would you like me to use u-boot? | 12:42 |
| minute | ah. so you warm rebooted and got the stripes in barebox, right? | 12:42 |
| josch | no, i switched off via the circle-menu zero | 12:42 |
| josch | barebox was fine | 12:42 |
| minute | josch: ah hm. | 12:42 |
| minute | josch: something didn't work as planned then, but i guess it's the sysctl fw version | 12:43 |
| josch | yes, i'm updating that now | 12:43 |
| josch | oh i'm not | 12:43 |
| josch | the latest sysctl-mb20 branch failed | 12:43 |
| josch | minute: does it fail because cppcheck fails? | 12:46 |
| minute | josch: oh huh... i didn't actually look at the CI, i just flashed locally before going home | 12:47 |
| minute | let me boot my laptop | 12:47 |
| AnimaInvicta | minute: I've switched the keyboard and trackball. The keyboard still works and there's still no sign of the trackball :^( | 12:49 |
| AnimaInvicta | Could it be the cable? | 12:50 |
| minute | AnimaInvicta: ok so the ports on the mb are ok. then it's probably the small cable yes | 12:50 |
| josch | minute: if it's quicker for you, you can also upload the uf2 somewhere instead of fixing CI | 12:51 |
| minute | josch: yeah let me take a look... i usually enjoy fixing cppcheck stuff | 12:52 |
| AnimaInvicta | That seems likely, yes. Just in case, do you know if it might be software-related (like a missing driver or something in the DTB)? (Occam's Razor says no, but it doesn't hurt to ask) | 12:52 |
| minute | AnimaInvicta: more likely the cable, you can verify it by booting a fresh system image from sd card | 12:52 |
| AnimaInvicta | OK, now I've got a way forward. Thank you for your help, minute. | 12:53 |
| minute | heh, i didn't use the 8550 mnt reform next for a day or 2 and now turning it on again, i have to say it's quite awesome lol | 12:54 |
| minute | AnimaInvicta: no problem! if yo have a DMM you could also beep the cable, maybe crimping issue, not uncommon | 12:54 |
| + potash6 (~potash@user/foghorn) | 12:54 | |
| AnimaInvicta | I'm afraid that's beyond my current knowledge :^) | 12:55 |
| minute | AnimaInvicta: ah no problem | 12:55 |
| - jnerula (QUIT: Server closed connection) (~jnerula@li1009-93.members.linode.com) | 12:56 | |
| + jnerula (~jnerula@li1009-93.members.linode.com) | 12:56 | |
| minute | josch: omg i'm thankful for cppcheck, i can't believe those errors are in the code, i was sure i'd fixed those :D | 12:56 |
| - oliverD (QUIT: Ping timeout: 248 seconds) (~Thunderbi@user/oliverd) | 12:57 | |
| josch | once again we are grateful for our static analysis tools :) | 12:57 |
| minute | 100% lol | 13:00 |
| minute | especially grateful because they're deterministic... | 13:00 |
| josch | i just cobbled together a command line doing first reform-mcu-tool bootsel pocket-sysctl-1.0 && picotool load --bus 5 --address 2 -f /path/to/uf2 && picotool reboot --bus 5 --address 2 -- should this kind of thing not be a script in reform-tools? | 13:01 |
| josch | lol lets not start with *that* topic XD | 13:01 |
| josch | minute: sysctl firmware uses the same usb vendor and product id, so it will show up everywhere (including in reform-mcu-tool) as pocket-sysctl-1.0 or "MNT Pocket Reform System Controller 1.0" -- intentional? | 13:03 |
| josch | okay, sorry i have to go afk now because family -- ttyl! | 13:05 |
| minute | josch: semi intentional because then the tools just keep working. i'm not sure if 2.0 can be updated by reform-mcu-tools etc? | 13:07 |
| minute | josch: also i think our usb id is specifically for "1.0" | 13:07 |
| minute | but i need to double check | 13:07 |
| minute | josch: also, i have a lot of scripts like that (reform-mcu-tool this that) | 13:08 |
| minute | still easier than trying to understand fwupd unfortunately | 13:08 |
| + oliverD (~Thunderbi@user/oliverd) | 13:10 | |
| - oliverD (QUIT: Remote host closed the connection) (~Thunderbi@user/oliverd) | 13:11 | |
| - potash6 (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 13:19 | |
| minute | josch: fix pushed btw | 13:19 |
| minute | and CI even passed :D | 13:20 |
| + potash6 (~potash@user/foghorn) | 13:20 | |
| + paperManu (~paperManu@204.48.0.160) | 13:36 | |
| + paperManu_ (~paperManu@204.48.0.160) | 13:38 | |
| - paperManu (QUIT: Ping timeout: 272 seconds) (~paperManu@204.48.0.160) | 13:55 | |
| + oliverD (~Thunderbi@user/oliverd) | 14:04 | |
| - lidstah2 (QUIT: Ping timeout: 252 seconds) (~lidstah@gateway/tor-sasl/lidstah) | 14:06 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 14:07 | |
| - oliverD (QUIT: Remote host closed the connection) (~Thunderbi@user/oliverd) | 14:07 | |
| josch | minute: i'm back and flash the latest sysctl.uf2 using the command i provided above. It succeeded and the last output was along the lines of "successfully rebooted..." so i guess it was all successful. But i still have display issues (display v2). Here is a potato-cam video: https://mister-muffin.de/pocket.mp4 | 14:46 |
| minute | josch: rebooted you say | 14:46 |
| minute | josch: if you reboot, this problem is expected, because it's caused by barebox | 14:47 |
| minute | josch: barebox doesn't reset the display | 14:47 |
| minute | josch: ah but it's unfortunate that linux doesn't fix that problem | 14:48 |
| minute | josch: can i get a full dmesg? | 14:48 |
| josch | minute: in the video you see that this is a cold boot | 14:49 |
| minute | josch: ah, i didn't catch that | 14:49 |
| josch | let me grab a dmesg for you | 14:49 |
| minute | josch: do you have ssh / uart to the machine as well? then i'd have some commands for you to test | 14:50 |
| josch | oh sure! | 14:50 |
| minute | josch: systemctl stop gdm3; rmmod panel-mnt-pocket-reform; rmmod mnt-sc; modprobe mnt-sc; modprobe panel-mnt-pocket-reform | 14:50 |
| josch | i can also create a reverse ssh tunnel and give *you* access :D | 14:50 |
| josch | hrm... but i guess that's not so useful because you don't see the screen XD | 14:50 |
| minute | josch: haha that won't help _that_ much in this case because i'd like to know if the display changes then | 14:51 |
| minute | josch: yes | 14:51 |
| josch | (sorry slow sd card -- boot takes ages) | 14:54 |
| josch | oh i didn't check whether the modules are in the initramfs | 14:54 |
| josch | i have to check that too | 14:55 |
| josch | oh maybe the gray on the bottom is just this "remainder" because the display was gray there for too long? | 14:58 |
| josch | checking initramfs now | 14:58 |
| josch | this is in my initramfs: usr/lib/modules/7.1.5-mnt-reform-arm64/kernel/drivers/gpu/drm/panel/panel-mnt-pocket-reform.ko.xz | 14:59 |
| josch | as well as /usr/lib/modules/7.1.5-mnt-reform-arm64/kernel/drivers/firmware/mnt-sc.ko.xz | 14:59 |
| josch | let me share a photo | 15:00 |
| josch | minute: https://mister-muffin.de/p/-2kx.jpg | 15:01 |
| josch | minute: and here is dmesg: https://mister-muffin.de/p/WZFL.txt | 15:06 |
| minute | josch: hmmm on the photo there is a lower side visible hm | 15:09 |
| minute | josch: yeah i _think_ this is a sort of burn-in from the original problem, it should go away after a while, but i usually didn't let it in this broken mode for too long | 15:09 |
| josch | when i was on holidays earlier this year, i once had it in this state for a night -- took a long time for it to go away | 15:10 |
| minute | uff, poor display | 15:10 |
| minute | [ 5.579659] mnt_sc spi1.0: [mntsc_gpio_set] 0 <- 0 | 15:10 |
| minute | [ 5.684577] mnt_sc spi1.0: [mntsc_gpio_set] 0 <- 1 | 15:10 |
| minute | that looks good | 15:10 |
| minute | this shows that mntsc is toggling the display reset pin | 15:11 |
| minute | (virtual gpio 0) | 15:11 |
| minute | uff so much spam from analogix_dp | 15:11 |
| minute | this is the top item on my list to clean that up so it realizes early that there is nothing connected | 15:12 |
| - potash6 (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 15:13 | |
| - grimmware (QUIT: Server closed connection) (sid638840@id-638840.hampstead.irccloud.com) | 15:13 | |
| + grimmware (sid638840@id-638840.hampstead.irccloud.com) | 15:13 | |
| minute | josch: but in general the dmesg looks good, panel + mntsc integration works on your system | 15:17 |
| minute | josch: can you try brightness control (slider) in gnome? | 15:17 |
| minute | stephano: can you remind me if you had your SPI lpc driver code for barebox online somewhere? | 15:18 |
| minute | josch: i'm also wondering if the linux kernel is not shutting down the panel before rebooting / not keeping the display in reset then, maybe i need to add another line to the panel driver to ensure that hm | 15:19 |
| minute | if that is ensured, maybe that'll help barebox | 15:20 |
| + potash6 (~potash@user/foghorn) | 15:37 | |
| minute | hmm, mnt_panel_power_off definitely tries to assert the reset. | 15:37 |
| minute | so theoretically the panel shouldn't work at all in barebox after a warm reset. that is strange | 15:38 |
| minute | like, who is deasserting that reset | 15:38 |
| josch | minute: i'll try again from scratch now (need to test reform-tools before release) -- would you like me to boot with u-boot or barebox or does it not matter? | 15:49 |
| - amatecha (QUIT: Server closed connection) (sid10006@user/amatecha) | 15:55 | |
| + amatecha (sid10006@user/amatecha) | 15:55 | |
| stephano | minute: for SPI in Barebox, there are two branches... | 16:16 |
| stephano | minute: Adding board support: https://source.mnt.re/stephano/mnt-reform-barebox/-/commits/radxa-cm5-board-support | 16:16 |
| stephano | minute: Debugging / " | 16:16 |
| stephano | fixing" the issues I found: https://source.mnt.re/stephano/mnt-reform-barebox/-/commits/generic-fixes?ref_type=heads | 16:16 |
| minute | josch: try uboot | 16:23 |
| - pasta161 (QUIT: Ping timeout: 260 seconds) (~pasta161@user/xha) | 16:24 | |
| + pasta161 (~pasta161@user/xha) | 16:31 | |
| stephano | The order is basically: 1. Add the node: https://source.mnt.re/stephano/mnt-reform-barebox/-/commit/dc813027f50e1902f4a85a90db997bebdd91f454 | 16:36 |
| stephano | 2. fix a bug: https://source.mnt.re/stephano/mnt-reform-barebox/-/commit/440513a6d383666e23a13a2550bb2f05ce99f196 | 16:36 |
| stephano | 3. reset the display https://source.mnt.re/stephano/mnt-reform-barebox/-/commit/1fac7fc39cead59a272da7decd554ff7eed73589 | 16:36 |
| josch | it seems that the contents of /sys/bus/spi/drivers/reform2_lpc/spi1.0/firmware are different now compared to the older sysctl 2.0 version. With the same sysctl version, the content is identical even across multiple boots. | 16:41 |
| josch | minute: display works fine now with u-boot | 16:50 |
| josch | stephano: I'll add kernel/drivers/firmware/mnt-sc.ko to backend_modules_check() in reform-check/distro/arch.sh as well, okay? | 16:55 |
| stephano | josch: yep, makes sense, thnx! | 16:56 |
| josch | minute: i think i can confirm that the problem you see in the video only works with barebox. Once I boot with it, i see the same issue. Once I flash u-boot and boot with that, it works fine again. | 17:04 |
| josch | s/only works with/only happens with/ | 17:04 |
| minute | josch: ok many thanks for confirming | 17:05 |
| josch | minute: since this issue does not seem to be related to reform-tools I'll make a new release now okay? | 17:05 |
| minute | josch: yes, thanks! | 17:05 |
| minute | stephano: your usb bugfixes are also very interesting, would you mind opening an MR? | 17:06 |
| stephano | minute: sure, the "generic-fixes" branch or both that and the "radxa-cm5..." branch? | 17:07 |
| minute | stephano: and for the display reset, the new spi driver code looks like this https://source.mnt.re/reform/reform-tools/-/blob/main/system-controller/mnt-sc.c?ref_type=heads#L333 the response parsing is probably not important for barebox at first though | 17:07 |
| minute | stephano: generic fixes first | 17:08 |
| minute | stephano: ohh now i see you're just using a script for spi! | 17:08 |
| minute | stephano: that should be even easier to adapt | 17:08 |
| minute | i basically have to send the string "(set-gpio 0 1)" and then after a delay "(set-gpio 0 1)" over spi | 17:09 |
| minute | sorry, the second one should be "(set-gpio 0 0)" of course | 17:09 |
| minute | no magic bytes, no checksums etc | 17:09 |
| stephano | i tried to keep it simple | 17:09 |
| minute | yeah that's good! | 17:09 |
| minute | the framing is ensured by the parentheses and the strict syntax | 17:10 |
| minute | before there was 1 magic byte for framing, 0xb5, now it's 1 byte at the start '(' and one at the end ')' | 17:11 |
| bremner | looks like lisp, yay! | 17:11 |
| minute | bremner: yep! | 17:11 |
| minute | so the new framing is 0x28 and 0x29 ;) | 17:11 |
| minute | so a complete command would be in hex: 0x28 0x29 0x73 0x65 0x74 0x2d 0x67 0x70 0x69 0x6f 0x20 0x30 0x20 0x31 0x29 | 17:13 |
| minute | spi -b 0 -c 0 -m 1 -f 1000000 0x28 0x29 0x73 0x65 0x74 0x2d 0x67 0x70 0x69 0x6f 0x20 0x30 0x20 0x31 0x29 | 17:13 |
| stephano | minute: https://source.mnt.re/reform/mnt-reform-barebox/-/merge_requests/4 | 17:21 |
| minute | stephano: awesome, many thanks! | 17:29 |
| stephano | o7 | 17:30 |
| josch | stephano: reform-tools 1.88 is tagged ready to be packaged for arch :) | 17:31 |
| stephano | josch: \o/ | 17:35 |
| stephano | *sigh* The AUR is down due to maintenance. We will be back soon. | 17:41 |
| stephano | Poor AUR. So many mean people out there. | 17:41 |
| + mjw (~mjw@66.183.205.74) | 19:07 | |
| + voxpelli_ (sid31634@id-31634.tinside.irccloud.com) | 19:16 | |
| + polymetis_ (sid320610@id-320610.hampstead.irccloud.com) | 19:16 | |
| + mlarkin_ (~mlarkin@syn-076-081-194-027.biz.spectrum.com) | 19:19 | |
| + elb2 (~elb@user/elb) | 19:20 | |
| + lidstah2 (~lidstah@gateway/tor-sasl/lidstah) | 19:20 | |
| - lidstah (QUIT: Ping timeout: 252 seconds) (~lidstah@gateway/tor-sasl/lidstah) | 19:21 | |
| - manis (QUIT: *.net *.split) (01a66df340@185.72.67.185) | 19:24 | |
| - voxpelli (QUIT: *.net *.split) (sid31634@tinside.irccloud.com) | 19:24 | |
| - polymetis (QUIT: *.net *.split) (sid320610@hampstead.irccloud.com) | 19:24 | |
| - mlarkin (QUIT: *.net *.split) (~mlarkin@syn-076-081-194-027.biz.spectrum.com) | 19:24 | |
| - elb (QUIT: *.net *.split) (~elb@user/elb) | 19:24 | |
| - schneider (QUIT: *.net *.split) (~schneider@irc2.xtort.eu) | 19:24 | |
| * voxpelli_ -> voxpelli | 19:24 | |
| * polymetis_ -> polymetis | 19:24 | |
| + schneider (~schneider@irc2.xtort.eu) | 19:27 | |
| - lidstah2 (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 19:28 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 19:32 | |
| - switchy (QUIT: Ping timeout: 260 seconds) (~switchy@mechboards/switchy) | 20:21 | |
| + switchy (~switchy@mechboards/switchy) | 20:26 | |
| - antti (QUIT: Read error: Connection reset by peer) (~antti@user/antti) | 21:06 | |
| + antti (~antti@user/antti) | 21:18 | |
| - jedavies (QUIT: Ping timeout: 246 seconds) (~jedavies@plan9.info) | 21:29 | |
| + jedavies (~jedavies@plan9.info) | 21:31 | |
| - antti (QUIT: Ping timeout: 257 seconds) (~antti@user/antti) | 21:34 | |
| + antti (~antti@user/antti) | 21:34 | |
| - spew (QUIT: Quit: nyaa~) (~spew@user/spew) | 22:03 | |
| minute | lol i accidentally got dp alt-mode working on my dell 27 inch at home which never worked so far | 22:08 |
| minute | a side effect of fixing some fusb302 messages that were missing message ids and pd revision | 22:08 |
| minute | the signal is not very stable though, will probably need custom internal flex cable to enhance that. | 22:09 |
| + chrcav (~chrcav@user/chrcav) | 22:37 | |
| + paperManu (~paperManu@204.48.0.160) | 23:24 | |
| - paperManu (QUIT: Ping timeout: 251 seconds) (~paperManu@204.48.0.160) | 23:35 | |
| - antti (QUIT: *.net *.split) (~antti@user/antti) | 23:40 | |
| - schneider (QUIT: *.net *.split) (~schneider@irc2.xtort.eu) | 23:40 | |
| + antti (~antti@user/antti) | 23:42 | |
| + schneider (~schneider@irc2.xtort.eu) | 23:42 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
