| - lidstah (QUIT: Ping timeout: 252 seconds) (~lidstah@gateway/tor-sasl/lidstah) | 00:01 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 00:03 | |
| - paperManu_ (QUIT: Quit: WeeChat 4.1.1) (~paperManu@204.48.0.160) | 00:05 | |
| josch | minute: i just had display v2 in rk3588 pocket reform come on fine with u-boot and reform-tools 1.87 (the old version) -- plausible? | 00:17 |
|---|---|---|
| - S0rin (QUIT: Read error: Connection reset by peer) (~S0rin@user/s0rin) | 00:32 | |
| + S0rin (~S0rin@user/s0rin) | 00:34 | |
| - AnimaInvicta (PART: !!unknown attribute: msg!!) (~AnimaInvi@88-169-191-86.subs.proxad.net) | 01:08 | |
| + mjw (~mjw@66.183.205.74) | 01:13 | |
| josch | minute: this job now has system images with your kernel from the rk3588-pocket-dp-altmode-and-sc branch | 01:30 |
| josch | https://source.mnt.re/josch/reform-system-image/-/jobs/23559 | 01:30 |
| minute | josch: maybe initramfs-tools saw that the modules (panel and mntsc) are in use and included them? | 02:44 |
| - casparvitch (QUIT: Read error: Connection reset by peer) (~casparvit@user/casparvitch) | 04:07 | |
| + casparvitch (~casparvit@user/casparvitch) | 04:37 | |
| - jn (QUIT: Ping timeout: 245 seconds) (~quassel@user/jn/x-3390946) | 04:46 | |
| + jn (~quassel@2a0a-a54a-9b7d-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 04:46 | |
| - jn (QUIT: Changing host) (~quassel@2a0a-a54a-9b7d-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 04:46 | |
| + jn (~quassel@user/jn/x-3390946) | 04:46 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 05:11 | |
| - voltaire28 (QUIT: Ping timeout: 260 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 05:12 | |
| - spew (QUIT: Quit: nyaa~) (~spew@user/spew) | 05:29 | |
| - paperManu (QUIT: Ping timeout: 264 seconds) (~paperManu@204.48.0.160) | 06:35 | |
| - AshCurry (QUIT: Ping timeout: 276 seconds) (~AshCurry@user/AshCurry) | 06:39 | |
| - rodion_goritskov (QUIT: Ping timeout: 251 seconds) (~rodion_go@mail.goritskov.com) | 07:40 | |
| + manis (01a66df340@185.72.67.185) | 07:40 | |
| + rodion_goritskov (~rodion_go@mail.goritskov.com) | 07:43 | |
| josch | minute: yes, that's what would usually happen when the initramfs is built on a machine that booted with the new kernel, the module got loaded and the initramfs gets rebuilt. But this was the *first boot* of a system image built on MNT CI. Specifically this one (you can see in the logs that this is with reform-tools 1.87): https://source.mnt.re/josch/reform-system-image/-/jobs/23552 | 09:00 |
| josch | (no need to investigate but maybe this means that people failing to upgrade reform-tools before they upgrade linux is not actually as bad as you feared?) | 09:01 |
| + amospalla (~jordi@user/amospalla) | 09:59 | |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@66.183.205.74) | 10:25 | |
| minute | josch: that would be awesome of course... i mean, the panel driver should work without mntsc because the reset gpio dependency is optional. in case the initramfs only includes the panel driver and not mntsc. but would be most interesting to list the contents of the initramfs from that build | 11:47 |
| minute | josch: otoh, did the display come on in initramfs or later? | 11:48 |
| josch | let me flash that again, investigate the initramfs and make sure to write down when exactly the display came on | 11:49 |
| minute | josch: because without disk encryption, that wouldn't be a big deal, but with disk encryption and no panel driver, people would encounter black screen at the prompt | 11:49 |
| josch | indeed | 11:49 |
| josch | in any case, people are probably upgrading to the fixed reform-tools version now | 11:49 |
| minute | yep, that's good | 11:50 |
| - murphnj (QUIT: Quit: murphnj) (~murphnj@user/murphnj) | 12:42 | |
| + AshCurry (~AshCurry@user/AshCurry) | 12:57 | |
| minute | sigh, i just showed greta the pocket reform's state of things, and the vorführeffekt kicked in and i got a striped display on first (cold) boot, after barebox. in barebox it was still ok | 13:55 |
| minute | second cold boot was fine | 13:55 |
| minute | i hope this is just because of my instance of barebox not yet doing the reset | 13:55 |
| minute | josch: wakest_ is now preparing to test your system image builds on a311d pocket and imx8mp pocket here, and we'll test both disp v1 and disp v2 | 14:25 |
| + paperManu (~paperManu@204.48.0.160) | 14:43 | |
| josch | wakest: make sure to use job 23559 and *not* the later link to job 23552 as the latter uses reform-tools 1.87 | 14:43 |
| josch | (i'm flashing that again right now) | 14:43 |
| minute | josch: ah we used the wrong pipeline (from main repo) | 14:49 |
| minute | i was wondering why it worked on a311d and then i checked and it was kernel 7.1.4 | 14:49 |
| minute | wakest_: https://source.mnt.re/josch/reform-system-image/-/jobs/23559/artifacts/browse | 14:50 |
| minute | also, we definitely got AFBC screen flicker on a311d on 7.1.4 | 14:51 |
| minute | which is annoying because i thought i had patched that out of the kernel | 14:51 |
| josch | minute: is this limited to a311d or should also show on rk3588? | 14:51 |
| minute | josch: this shows only on rk3588 | 14:51 |
| minute | josch: ah mistyped | 14:51 |
| minute | josch: only on a311d | 14:51 |
| minute | josch: it's not a GPU issue but a display pipeline issue, specific to a311d | 14:52 |
| minute | it can be patched in mesa, but i think then it also affects rk3588 negatively (can slow it down a bit) | 14:52 |
| josch | then this is wrongly tagged rk3588? https://community.mnt.re/t/display-flickering/4464 | 14:53 |
| josch | i wanted to try to reproduce it again today and didn't see it on rk3588 (and you now confirmed that i shouldn't see it) | 14:53 |
| minute | hmm my old patch is still in there https://source.mnt.re/reform/reform-debian-packages/-/blob/main/linux/patches7.1/meson-g12b-bananapi-cm4-mnt-reform2/0030-WORKAROUND-meson-plane-disable-afbc-32x8.patch?ref_type=heads | 14:54 |
| josch | if this is a kernel regression: i'm at home with holidays and could do a kernel bisect this week and see what broke it -- unless you know what you are looking for and can fix it without a bisect of course :) | 14:55 |
| minute | josch: hmm on second viewing, the issue in that thread (in the video) looks different. on a311d that we just tested, the screen just blanks out (black) for a frame or so. in the video, the image kind of stretches for a moment | 14:56 |
| josch | okay, but i cannot reproduce that either but i only tried it on classic reform so far assuming that rk3588 would be have the same -- i can try it on rk3588 pocket reform now | 14:56 |
| minute | josch: yeah on my pocket rk3588 i cannot reproduce that either | 15:06 |
| minute | maybe it's a hardware issue in the end, on their system | 15:06 |
| josch | uff :( | 15:06 |
| minute | we should ask to confirm it's there when booting from stock system image on microsd | 15:06 |
| minute | side note: pololu has basically all kind of replacement cables for pocket reform, and ships worldwide https://www.pololu.com/category/346/jst-sh-style-cables | 15:06 |
| josch | minute: about the sd-card issue of meluzzy: her sd-card arrived today from spain and i can confirm that the sd-card which doesn't work in her rk3588 classic reform mainboard 3.0 works just fine in mine and runs multiple gnome-disk benchmarks without issues. I'll follow up in the thread. | 15:08 |
| wakest_ | note: gnome defaults to 250% scale on the pocket now | 15:11 |
| wakest_ | we also still don't have nmtui | 15:14 |
| wakest_ | note: a311d display handover to hdmi works very well in gnome, but internal display brightness stays on | 15:17 |
| - MartiniMoe (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~quassel@user/MartiniMoe) | 15:20 | |
| + MartiniMoe (~quassel@user/MartiniMoe) | 15:21 | |
| minute | josch: ok, many thanks for spending your time debugging that for us :0 hmmmmm it might be an alignment issue of the rk3588 module | 15:21 |
| josch | wakest: /usr/bin/nmtui used to be shipped by the package network-manager. I do not see it in any package anymore in unstable. | 15:22 |
| minute | josch: network-manager-tui | 15:22 |
| minute | is the package :3 at least on my pc it comes up in "apt-file search nmtui" | 15:22 |
| minute | josch: it just occured to me that meluzzy's issue could be a mechanical alignment issue on the rcore in the slot. this usually impacts SD card, but we have seen that mostly in pocket QA, but it would manifest in the same way on classic | 15:23 |
| josch | funny that packages.debian.org doesn't find that when one searches for /usr/bin/nmtui in unstable. But yes, I now see that network-manager-tui ships that. minute: i'll make network-manager-tui a Depends of reform-tools, okay? | 15:24 |
| minute | josch: some of the rcores were manufactured with a slightly "off" (like 0.1mm) outline and can be moved horizontally a little bit in the slot. if it's too far off then some pins are not connected | 15:24 |
| josch | minute: okay, then let me suggest a re-seating | 15:24 |
| josch | oh wait | 15:24 |
| josch | the chip itself? | 15:24 |
| minute | josch: no, rcore (the carrier) | 15:24 |
| josch | oh okay | 15:24 |
| minute | josch: it's a PCB right? that goes in the slot. this PCB is cut out with a CNC machine | 15:24 |
| ch | josch: https://packages.debian.org/search?suite=sid&searchon=contents&keywords=nmtui ? | 15:24 |
| minute | josch: and this cut was slightly imprecise in one batch | 15:25 |
| ch | ah but exact filename fails | 15:25 |
| minute | josch: thanks @ depends @ nmtui! | 15:25 |
| ch | great | 15:25 |
| josch | ch: https://packages.debian.org/search?searchon=contents&keywords=%2Fusr%2Fbin%2Fnmtui&mode=exactfilename&suite=unstable&arch=any | 15:25 |
| ch | yeah | 15:25 |
| ch | i guess its not a filename | 15:25 |
| josch | minute: awesome, i'll forward what you said to her | 15:25 |
| josch | ch: oh... m( | 15:25 |
| ch | very literal software there | 15:26 |
| josch | well, mystery solved :) | 15:26 |
| minute | ch: josch: funny, this works https://packages.debian.org/search?mode=exactfilename&suite=sid§ion=all&arch=any&searchon=contents&keywords=%2Fbin%2Fnmtui | 15:26 |
| minute | but /usr breaks the search?! | 15:26 |
| ch | m( | 15:26 |
| josch | hahahahaha | 15:26 |
| ch | i think i dont want to know how that works | 15:26 |
| josch | ch: does anybody still maintain that web service? | 15:26 |
| minute | reverse usrmerge | 15:27 |
| josch | indeed :D | 15:27 |
| ch | i saw the maintainer at debconf, but i dont know if they actually maintain it | 15:27 |
| - potash6 (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 15:28 | |
| josch | minute: https://mister-muffin.de/pocket_boot.mp4 This is a system image just flashed to sd-card with old reform-tools (1.87) but new kernel (it contains /usr/lib/modules/7.1.5-mnt-reform-arm64/kernel/drivers/firmware/mnt-sc.ko.xz) booting with u-boot on rk3588 pocket reform with new sysctl flashed. Nothing happens for 40 seconds, then the screen becomes gray and then you see the sd-card partition | 15:30 |
| josch | growing progress bar. The resize2fs is run in the initramfs, so this is still before the system actually boots. I investigated the initramfs and it does *not* contain mnt-sc.ko. | 15:30 |
| minute | josch: but some good news, your system image build works fine on a311d, mnt-sc works surprisingly without any issues | 15:32 |
| minute | josch: ok thanks! does the initramfs contain the panel-mnt-pocket-reform driver? | 15:33 |
| minute | josch: also funny/strange that the display is (not) rotated correctly. do you have hdmi connected as well? i'm wondering what causes that | 15:34 |
| minute | ah it could be the ghost of DP | 15:34 |
| josch | minute: the initramfs contains /usr/lib/modules/7.1.5-mnt-reform-arm64/kernel/drivers/gpu/drm/panel/panel-mnt-pocket-reform.ko.xz | 15:35 |
| josch | minute: i do not a hdmi or display port or any of this other new stuff here at home, so no :δ | 15:35 |
| josch | (pocket reform was my first ever usb-c device) | 15:36 |
| bremner | lol. petition to send josch one of the 10 HDMI panels we all have in the closet | 15:36 |
| josch | hahaha | 15:36 |
| bremner | pretty sure the shipping from Canada is more than the cost of a new panel at Saturn | 15:37 |
| josch | we have some really good second hand stores here -- i'm always looking for a small-ish display there. At some point something will show up. :) | 15:38 |
| + potash6 (~potash@user/foghorn) | 15:38 | |
| josch | wakest: about the scale: we can change the scaling in the defaults MNT ships but that will of course not only affect the pocket reform and that's why it isn't done. Hardware-specific default changes could be done in the reform-setup-wizard. | 15:45 |
| minute | josch: ok thanks for confirming that the panel driver is already included automatically, that makes things easier! | 15:51 |
| minute | josch: and also cool that you were able to confirm it also works without mntsc loaded | 15:51 |
| minute | josch: yeah, either reform-setup-wizard or the more fittingly named reform-hw-setup | 15:52 |
| minute | josch: hmm but the setup wizard deploys the user profile, so that's the most fitting place, you're right | 15:52 |
| minute | josch: it could just drop a monitors.xml | 15:52 |
| minute | josch: otoh it would be interesting to know how gnome decides this ridiculous scale setting. i'm sure it looks at the physical size of the panel which is either in EDID or in our case, hardcoded in the panel driver, and it calculates the DPI probably | 15:53 |
| minute | josch: so we could adjust the physical size number a little bit to counteract this, probably | 15:53 |
| minute | (hacks upon hacks!) | 15:53 |
| minute | so, extremely surprisingly, display v1 works with all these newfangled driver on a311d pocket | 15:54 |
| minute | now we're gonna test imx8mp | 15:54 |
| josch | \o/ | 15:56 |
| - brennen (QUIT: Quit: WeeChat 2.8) (~brennen@user/brennen) | 16:02 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 16:02 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 16:02 | |
| + brennen (~brennen@user/brennen) | 16:03 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 16:13 | |
| josch | wakest: I added "network-manager (<< 1.56.0-4) | network-manager-tui" to reform-desktop-minimal. The build is currently running on reform-debian-packages pipeline and will probably miss the next mirror refresh in 47 minutes. So it will be on the MNT mirror in about 1.8 hours. | 16:14 |
| minute | josch: thank you! | 16:15 |
| minute | btw i'm torn on including the DP alt-mode stuff on rk3588 pocket .dtb already or no, as it's still quite flaky. otherwise half-working things sometimes motivate contributors to help me fix them :D | 16:17 |
| minute | s/otherwise/on the other hand | 16:17 |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 16:18 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 16:22 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 16:23 | |
| josch | minute: you mean it's quite flaky when used with db alt-mode but it continues to be stable otherwise? If yes: I say ship it | 16:23 |
| josch | If it doesn't work, then nothing is lost because it didn't work before either. | 16:23 |
| josch | But for some setups it will improve the situation and as you say, maybe somebody feels like they want to contribute. :) | 16:23 |
| - brennen (QUIT: Quit: WeeChat 2.8) (~brennen@user/brennen) | 16:30 | |
| + brennen (~brennen@user/brennen) | 16:31 | |
| - Guest29952 (QUIT: Quit: The Lounge - https://thelounge.chat) (~ndufresne@transit.collaboradmins.com) | 16:50 | |
| + Guest29952 (~ndufresne@transit.collaboradmins.com) | 16:52 | |
| minute | josch: yeah it sounded like on your system there might be a phantom 3rd display output from it? this would be confusing, esp. on mb1.0 | 16:53 |
| josch | minute: since the output was sideways, i guess there was a 3rd phantom display, yes. Would you like me to confirm? | 16:55 |
| minute | josch: it would be great to confirm this yes, because that would be a negative effect that we'd like to avoid | 16:55 |
| minute | i just booted up my pocket reform and in gnome it doesn't show any external display by default | 16:55 |
| minute | (which is good as there is none connected. but there is analogix_dp spam in dmesg that might be confusing for some.) | 16:57 |
| minute | josch: argh i just replicated this fix by you which was never merged https://source.mnt.re/reform/reform-system-image/-/merge_requests/150/diffs | 16:58 |
| josch | minute: ups no big deal -- just close that MR :) | 16:59 |
| minute | i was so annoyed that this was still not fixed that i committed directly to main :0 https://source.mnt.re/reform/reform-system-image/-/commit/f0539dddf0e709a9b0b876225e7a478680b3ee0f | 16:59 |
| minute | josch: oki :D | 16:59 |
| josch | minute: gnome says that there is an "Unknown display" with 1024x768 60Hz and 100% scale connected | 16:59 |
| minute | josch: ah ok. that's what i want to avoid. so i'll remove the DP stuff from the .dts for merging | 17:00 |
| minute | josch: and will add it back in a new branch later | 17:00 |
| minute | it's better to first get mntsc and basic mb2.0 support shipped | 17:00 |
| josch | minute: one sec, let me confirm that i'm running the latest kernel from that branch | 17:00 |
| josch | because it's weird that you don't see what i see -- maybe i'm running something different than you | 17:01 |
| minute | josch: could be mb1.0/2.0 difference. oh also, please update to my latest mntsc fw! | 17:01 |
| minute | (if you haven't) | 17:01 |
| josch | minute: pocket reform is running the sysctl 2.0 firmware | 17:01 |
| minute | well there's different versions of that :D | 17:02 |
| josch | oh | 17:02 |
| josch | ah right we talked about that | 17:02 |
| josch | yes, i flashed the latest one | 17:02 |
| minute | josch: ok! | 17:02 |
| minute | anyway, i also need to fix display v1 on imx8mp it seems | 17:03 |
| minute | display v2 and mntsc work fine on imx8mp pocket! | 17:03 |
| josch | nice :) | 17:03 |
| minute | so everything works well except display v1 on imx8mp | 17:03 |
| josch | minute: confirmed that my pocket reform is running the latest kernel build from the rk3588-pocket-dp-altmode-and-sc branch | 17:04 |
| josch | (the version string matches exactly) | 17:04 |
| minute | josch: ok thank you! i think it's random behavior then, or mb20/mb10 difference related | 17:04 |
| minute | my money is on random/timing related though. but that part of the analogix dp patches is still a mess. | 17:05 |
| josch | okay | 17:05 |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 17:07 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 17:07 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 17:08 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 17:09 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 17:11 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 17:12 | |
| josch | minute: your changes to etc/skel/.config/waybar/config still result in a difference to the call as it happens in etc/skel/.config/sway/config. Intentional? | 17:26 |
| josch | specifically, your changes are missing --width=66% --height=66% --style= | 17:26 |
| + lidstah2 (~lidstah@gateway/tor-sasl/lidstah) | 17:28 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 17:28 | |
| minute | josch: yeah it's just -W vs --width and -S vs --style or so. if you want you can improve those as well like in your original commit | 17:28 |
| josch | minute: just wanted to ask whether this those option missing was intentional or not because if it was, I would remove these options from the sway config to make them equal | 17:29 |
| minute | josch: they're not _removed_. they're the same, just different names | 17:30 |
| minute | josch: it's aliases | 17:30 |
| josch | oh | 17:30 |
| josch | sorry i misread XD | 17:30 |
| josch | ups, sorry for the noise :D | 17:30 |
| josch | i now copypasted the sway invocation to waybar to make sure that i don't get confused in the future. Both invocations are now identical. | 17:43 |
| minute | josch: ok cool. i was a bit undecided how to handle quotes vs no quotes | 17:50 |
| josch | oh no the quotes! | 17:53 |
| josch | fixing... | 17:53 |
| josch | okay, done -- will test the image later today | 17:54 |
| - lidstah2 (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 18:00 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 18:04 | |
| + AnimaInvicta (~AnimaInvi@88-169-191-86.subs.proxad.net) | 18:13 | |
| + voltaire28 (~jlafon@28.162.2.93.rev.sfr.net) | 18:21 | |
| - talos (QUIT: Quit: The Lounge - https://thelounge.chat) (~talos@2600:6c5d:f0:1fd0:dbe5:355a:9889:e630) | 18:49 | |
| + siviq (~siviq@user/siviq) | 19:39 | |
| - siviq (QUIT: Quit: Client closed) (~siviq@user/siviq) | 19:45 | |
| - voltaire28 (QUIT: Ping timeout: 246 seconds) (~jlafon@28.162.2.93.rev.sfr.net) | 19:47 | |
| + talos (~talos@2600:6c5d:f0:1fd0:177b:a943:5777:7bd0) | 20:04 | |
| + paperManu_ (~paperManu@204.48.0.160) | 20:09 | |
| + mjw (~mjw@66.183.205.74) | 20:31 | |
| - schalken (QUIT: Quit: Leaving) (~schalken@117-118-178-69.gci.net) | 20:57 | |
| + schalken (~schalken@117-118-178-69.gci.net) | 20:58 | |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 21:03 | |
| + cobra (~cobra@user/Cobra) | 21:05 | |
| - lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah) | 21:13 | |
| + lidstah (~lidstah@gateway/tor-sasl/lidstah) | 21:13 | |
| josch | wakest: confirmed -- an upgrade of reform-desktop-minimal now pulls in network-manager-tui on systems where network-manager is too new -- thank you! | 21:37 |
| josch | minute: also confirmed (with pgrep -a) the wofi command run by sway as well as by waybar is now identical (and works) | 21:37 |
| + voltaire28 (~jlafon@28.162.2.93.rev.sfr.net) | 22:33 | |
| minute | josch: very nice, thanks! | 22:40 |
| minute | the little things! | 22:40 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!