| - MyNetAz (QUIT: Remote host closed the connection) (~MyNetAz@user/MyNetAz) | 00:58 | |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.20.104) | 01:00 | |
| + Ar|stote|is (~linx@149.210.20.104) | 01:00 | |
| - mtm (QUIT: Ping timeout: 248 seconds) (~textual@47.202.75.129) | 01:03 | |
| + mtm (~textual@47.202.75.129) | 01:05 | |
| + MyNetAz (~MyNetAz@user/MyNetAz) | 01:09 | |
| + bkeys (~Thunderbi@45.134.140.153) | 01:10 | |
| - MyNetAz (QUIT: Remote host closed the connection) (~MyNetAz@user/MyNetAz) | 01:55 | |
| + MyNetAz (~MyNetAz@user/MyNetAz) | 02:06 | |
| - bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@45.134.140.153) | 02:19 | |
| + NanoCodeBug (~NanoCodeB@c-67-170-15-47.hsd1.wa.comcast.net) | 02:44 | |
| NanoCodeBug | huh | 02:45 |
|---|---|---|
| NanoCodeBug | i have somehow burnted out the MP2650 on my pocket reform | 02:45 |
| NanoCodeBug | top left corner of it is crisped (as seen within the pcb diagram), which has the inputs from the charger on it | 02:46 |
| NanoCodeBug | i was not running any custom firmware, pocket was last plugged into a 100 watt PD charger. | 02:48 |
| NanoCodeBug | battery was at 50% when it was plugged in. | 02:48 |
| - cobra (QUIT: Ping timeout: 252 seconds) (~cobra@user/Cobra) | 02:55 | |
| - paperManu_ (QUIT: Quit: WeeChat 4.1.1) (~paperManu@107.159.243.8) | 03:02 | |
| + bkeys (~Thunderbi@45.134.140.153) | 03:03 | |
| + colinsane (~colinunin@97-113-71-244.tukw.qwest.net) | 03:05 | |
| josch | bkeys: recently one of my sd-cards broke and i sent it to a special service that can recover data from broken sd-cards. When I asked them, they told me that there are literally no consumer sd-cards that they would recommend which would break less often. They told me that in their experience, all manufacturers just glue the least expensive chips into their cards in some factory in china and hope for the | 03:05 |
| josch | best because replacing defective cards is cheaper than doing proper QA... | 03:05 |
| josch | So if your sd-card just randomly decided to not function one day (happened to me) then I guess that's more or less plausible | 03:06 |
| - sevan (QUIT: Ping timeout: 272 seconds) (~sevan@user/venture37) | 03:07 | |
| bkeys | josch: Since joining the Reform crowd my hopes has always been to have edk2 on the emmc and Fedora or centos stream on the nvme | 03:08 |
| josch | i remember :) | 03:10 |
| - op_4 (QUIT: Remote host closed the connection) (~tslil@user/op-4/x-9116473) | 04:05 | |
| + op_4 (~tslil@user/op-4/x-9116473) | 04:05 | |
| josch | minute: turns out that linux failed to build thanks to a binutils upload yesterday, so i filed https://bugs.debian.org/1090999 and let the source.mnt.re CI use the binutils version from testing until it's fixed | 04:07 |
| josch | i cherry picked the commit that uses binutils from testing into https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/74 | 04:09 |
| josch | things build now and i successfully tested the resulting kernel with a vanilla system image on a311d | 04:09 |
| josch | so i'm goint to press the merge button... | 04:09 |
| josch | so reverting upstream commit d97e71e449373efbd2403f1d7a32d416599f32ac on top of what's in Debian unstable indeed did end up fixing the display issue on a311d but other than reverting this and calling it a day i cannot do much more -- maybe you or anybody else has a better idea on how to fix this for real and communicate that to kernel.org | 04:11 |
| josch | okay, now time for bed :D | 04:12 |
| josch | _o/ | 04:12 |
| - Ar|stote|is (QUIT: Ping timeout: 244 seconds) (~linx@149.210.20.104) | 04:35 | |
| + Ar|stote|is (~linx@149.210.12.213) | 04:39 | |
| NanoCodeBug | looks like the burn mark is on the IAP pin corner of the NP2650, with that pin looking particularly toasty | 05:03 |
| NanoCodeBug | if anybody has had this happen to them before i'd be curious to know. i'll probably order a mp2650 from digikey and try to swap it out myself | 05:03 |
| NanoCodeBug | ah no, the trace underneath the chip is burned through | 05:21 |
| NanoCodeBug | specifically, the trace underneath the mp2650 that bridges pin 1 and pin 22 is burned completely through | 05:22 |
| NanoCodeBug | i will have to request a replacement charging board from mnt | 05:23 |
| NanoCodeBug | hmmm | 05:37 |
| NanoCodeBug | a quick digikey calculator for 1 amp current on a trace (assuming even distribution of current between the two terminals on the mp2650) (the sysctl controller limits to 2 amps), says that the trace width should be 0.3mm, but its 0.25mm in the schematic. | 05:39 |
| NanoCodeBug | ah no, the trace is still there, the mask is just charred over on top | 05:50 |
| - colinsane (QUIT: Ping timeout: 244 seconds) (~colinunin@97-113-71-244.tukw.qwest.net) | 06:17 | |
| + reform14646 (~lenin@2601:645:c681:db90::4ebe) | 06:52 | |
| reform14646 | q | 06:54 |
| reform14646 | gb | 06:54 |
| reform14646 | reset | 06:54 |
| reform14646 | stop | 06:55 |
| reform14646 | `q | 06:55 |
| - reform14646 (QUIT: Client Quit) (~lenin@2601:645:c681:db90::4ebe) | 06:55 | |
| + jacobk (~quassel@47-186-65-73.dlls.tx.frontiernet.net) | 07:08 | |
| - amk (QUIT: Ping timeout: 260 seconds) (~amk@user/amk) | 07:39 | |
| + cobra (~cobra@user/Cobra) | 07:40 | |
| + amk (~amk@user/amk) | 07:41 | |
| - Ar|stote|is (QUIT: Ping timeout: 265 seconds) (~linx@149.210.12.213) | 07:51 | |
| + Ar|stote|is (~linx@149.210.14.192) | 07:55 | |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 08:09 | |
| + cobra (~cobra@user/Cobra) | 08:14 | |
| - chomwitt (QUIT: Ping timeout: 272 seconds) (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1) | 08:41 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 09:00 | |
| - xha (QUIT: Quit: WeeChat 4.4.4) (~xha@user/xha) | 09:20 | |
| + xha (~xha@user/xha) | 09:23 | |
| minute | NanoCodeBug: we've seen this maybe 5 times so far (would need to count all the cases to be precise). in each case we sent a fresh charging board and it was ok | 09:34 |
| minute | NanoCodeBug: so i suspected some soldering issue by pcbway on some of the charger boards | 09:35 |
| josch | there are new system images with the fixed kernel here: https://source.mnt.re/reform/reform-system-image/-/jobs/7196/artifacts/browse | 09:36 |
| josch | if anybody wants to help, please test -- it is confirmed to work on a311d and imx8mq on classic reform | 09:36 |
| minute | NanoCodeBug: but i don't have proof / couldn't investigate it deeply yet. we also have 5-10% broken keyboards from the factory, where the rp2040 has a dead gpio and needs to be replaced (but we catch that before shipping) | 09:36 |
| minute | (pocket keyboards i mean) | 09:37 |
| minute | josch: great investigative work! | 09:37 |
| josch | minute: thank you -- this time, bisection was butter smooth. Time consuming but no issues. The problem was the binutils bug that made the whole thing FTBFS in the end... kept me awake until 4:00 in the night... | 09:39 |
| minute | NanoCodeBug: i have in my notes to send you a fresh charger board, probably we can do it tomorrow :3 | 09:39 |
| minute | josch: argh | 09:39 |
| josch | always "fun" being the first one who discovers a bug in a freshly uploaded package :D | 09:40 |
| josch | in any case, the binutils issue is not up to us but | 09:41 |
| josch | the patch i reverted probably? maybe? should be solved more elegantly? | 09:41 |
| josch | it'd probably be good to communicate this finding to the kernel ML | 09:41 |
| josch | but i have no idea what i'm doing other than having found this, so maybe somebody else [tm] can take care of this | 09:41 |
| josch | minute: you might also want to re-enable the repo updating before more people get confused over the kernel package being 404 when they attempt to upgrade | 09:44 |
| + gustav28 (~gustav@c-78-82-53-236.bbcust.telenor.se) | 10:02 | |
| - jacobk (QUIT: Ping timeout: 248 seconds) (~quassel@47-186-65-73.dlls.tx.frontiernet.net) | 10:04 | |
| + jacobk (~quassel@47-186-65-73.dlls.tx.frontiernet.net) | 10:04 | |
| minute | josch: i think we should inform narmstrong in #linux-amlogic about the dsi issue and revert | 10:28 |
| minute | my hunch is that the ti dsi to edp bridge adapter needs to get changed though | 10:29 |
| minute | s/adapter/driver sorry | 10:29 |
| josch | thank you, I've pinged narmstrong with my findings | 10:38 |
| - aloo_shu (QUIT: Ping timeout: 248 seconds) (~aloo_shu@90.166.98.198) | 11:43 | |
| + aloo_shu (~aloo_shu@90.166.98.198) | 11:51 | |
| - Ar|stote|is (QUIT: Ping timeout: 244 seconds) (~linx@149.210.14.192) | 12:19 | |
| + Ar|stote|is (~linx@149.210.13.219) | 12:23 | |
| + TwoDIsBooster (~TwoDIsBoo@user/TwoDIsBooster) | 12:45 | |
| - TwoDIsBooster (QUIT: Client Quit) (~TwoDIsBoo@user/TwoDIsBooster) | 12:45 | |
| - mtm (QUIT: Ping timeout: 265 seconds) (~textual@47.202.75.129) | 13:03 | |
| + mtm (~textual@47.202.75.129) | 13:06 | |
| + paperManu (~paperManu@107.159.243.8) | 13:28 | |
| + glu_ (~glu@user/glu) | 13:45 | |
| - glu (QUIT: Ping timeout: 276 seconds) (~glu@user/glu) | 13:47 | |
| * glu_ -> glu | 13:47 | |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 16:46 | |
| - ndufresne (QUIT: Quit: Ping timeout (120 seconds)) (~ndufresne@apple.collaboradmins.com) | 17:17 | |
| + ndufresne (~ndufresne@apple.collaboradmins.com) | 17:18 | |
| - yankcrim- (QUIT: Quit: "") (~nick@gw.tetromino.io) | 17:18 | |
| + yankcrime (~nick@gw.tetromino.io) | 17:20 | |
| - yankcrime (QUIT: Ping timeout: 265 seconds) (~nick@gw.tetromino.io) | 18:21 | |
| + mjw (~mjw@gnu.wildebeest.org) | 18:49 | |
| grimmware | okay this is weird, I'm getting this on apt upgrade: | 18:50 |
| grimmware | Error: Failed to fetch https://mntre.com/reform-debian-repo/pool/main/l/linux/linux-image-6.12.5-mnt-reform-arm64_6.12.5-1%2breform20241217T122225Z_arm64.deb 404 Not Found [IP: 91.250.115.15 443] | 18:50 |
| grimmware | but like, it's right there if I view it in the browser | 18:51 |
| grimmware | other than the URI encoding of the + I can't see anything that should be causing problems... | 18:51 |
| grimmware | and I don't feel like that should cause problems either | 18:51 |
| grimmware | I would also expect someone else to have this problem because I've had it for a few days, I just marked a few packages as held so I could get on with what I was doing | 18:52 |
| josch | grimmware: this post and my answer to it: https://community.mnt.re/t/display-driver-issue/2816/14 | 18:53 |
| josch | grimmware: essentially, minute has to re-enable the repository update now that the a311d display problem is fixed | 18:53 |
| josch | grimmware: which platform are you using? | 18:53 |
| grimmware | rk3588 original display pocket | 18:54 |
| grimmware | I'm happy to just put my apt holds back in place and get on with my life | 18:55 |
| josch | grimmware: do you want to test the fixed kernel? | 18:55 |
| grimmware | I'm going to be mean and say no right now, I'm already 5 yak shaves deep | 18:56 |
| grimmware | learn more offensive windows -> install evil-winrm -> install a version of ruby that doesn't do stupid shit with ffi -> compile a ruby with libyaml built in -> can't apt upgrade | 18:57 |
| grimmware | oh wow that is actually 5 yak shaves, I was just pulling a number out of the air | 18:58 |
| josch | grimmware: no hard feelings -- i'm just looking for willing test subjects :) | 18:59 |
| grimmware | josch: think you'll still be looking after the 25th? I think I'll be feeling a little more charitable by then | 18:59 |
| grimmware | I am astounded that the ruby ecosystem on debian is still so terrible | 19:02 |
| grimmware | this shit was a problem like, 13 years ago | 19:02 |
| josch | grimmware: maybe :) But don't worry if you forget! :) | 19:04 |
| josch | grimmware: somehow i only have very few intersections with ruby. I haven't heard much of the language in the circles I develop software in. | 19:05 |
| grimmware | josch: it used to be the language du jour back when devops was new so I did a bunch in it for a while. I'm only fucking about with it now because I'm doing some hackthebox challenges and some of the tools are written in ruby. | 19:07 |
| grimmware | most of them are just shitty python which is less hassle | 19:08 |
| josch | hah :) | 19:09 |
| ch | josch: i can test on pocket rk3588 if thats useful to you | 19:09 |
| grimmware | holy shit `bundle install` just worked | 19:10 |
| grimmware | I only had to compile that version of ruby 5 FUCKING TIMES | 19:10 |
| josch | ch: i prepared system images with the new kernel here if that makes it easier for you: https://source.mnt.re/reform/reform-system-image/-/jobs/7196/artifacts/browse | 19:10 |
| josch | ch: background is that i bisected linux between 6.11 and 6.12 to find the commit that breaks the display on a311d. Reverting that commit fixes things on a311d and i'd like to be confident that the resulting kernel does not break more things. | 19:11 |
| ch | right, i saw the backlog :) | 19:15 |
| josch | ch: well then you know that the answer to "if that's useful to you" is "i bodged the last upload by not properly testing the kernel on all platforms and i want to avoid making that mistake again" XD | 19:20 |
| josch | i also didn't notice that hdmi stopped working on imx8mq but i'm fixing that right now | 19:21 |
| josch | ugh... fallocate --posix does not work with --insert-range... aaaaaaand another yak shave... | 19:29 |
| ch | josch: boots | 19:37 |
| ch | josch: font size is tiny though, i thought that got fixed | 19:39 |
| ch | josch: more details on font size: initial font size (initramfs etc) is fine, but it gets reset during boot at some point | 19:41 |
| ch | josch: rebooting however fails. i see the watchdog messages before the display turns off, but the system never comes back (display stays off) | 19:42 |
| josch | ch: and that is new behaviour? | 19:44 |
| ch | yes | 19:45 |
| josch | i'm afraid i'll not be the best person to fix the font-size thing because i cannot personally test anything | 19:45 |
| grimmware | josch: we should change the channel topic to "MNT Reform. Don't mind the yak hair" | 19:45 |
| ch | rebooting works on 6.11.10 | 19:45 |
| josch | wow | 19:45 |
| josch | another thing to bisect yay | 19:45 |
| josch | let me see if i can reproduce it on a311d classic reform with 6.12 | 19:45 |
| ch | although on 6.11.10 now it hangs at Loading essential drivers (but at least the display is on) | 19:46 |
| josch | o0 | 19:46 |
| ch | ok, dunno what happened. works on second try on 6.11.10 | 19:48 |
| ch | let me try the system image with 6.12 again | 19:48 |
| ch | one day rebooting from greetd should also be fixed :O | 19:49 |
| josch | oh no... another bug? | 19:55 |
| ch | josch: on 6.12 rebooting seems consistently broken (display stays off) | 19:56 |
| ch | rebooting from greetd never worked, so its not a new bug :) | 19:56 |
| josch | ch: is there a bug number for the greetd thing? I suppose it's not reform-specific | 19:57 |
| ch | josch: maybe its not specific; not sure what the default config is supposed to do | 20:08 |
| minute | ch: rebooting with 6.12 on rk3588 pocket? | 21:48 |
| ch | minute: yeah, that seems to hang now | 21:54 |
| minute | ch: ok, i'll try it too | 21:56 |
| minute | i've reenabled the repo update now and will apt install the new kernel :D | 21:58 |
| josch | thank you! | 22:12 |
| minute | > Linux pocki3000 6.12.6-mnt-reform-arm64 #1 SMP Debian 6.12.6-1+reform20241222T021703Z (2024-12-22) aarch64 GNU/Linux | 22:32 |
| minute | ok, now another reboot... | 22:32 |
| minute | reboot works | 22:33 |
| minute | ch: can't reproduce so far. i do have hdmi connected in addition to the dsi display | 22:34 |
| ch | hmm | 22:34 |
| ch | i was booting from sdcard | 22:35 |
| minute | me too | 22:35 |
| minute | my /boot is on microsd, system on nvme | 22:35 |
| minute | (encrypted) | 22:35 |
| minute | (i did not try the new system image) | 22:36 |
| minute | in any case, one can apt install 6.12.6 now | 22:37 |
| ch | i'll do a full upgrade on the pocket and then we'll see | 22:40 |
| ch | also broken | 22:59 |
| ch | so it seems to boot, but the display doesnt init | 23:00 |
| josch | noooooooo | 23:00 |
| ch | (after blindtyping the cryptsetup passphrase and waiting a bit, ping works) | 23:00 |
| ch | to be clear: first boot works, on reboot -> no display | 23:00 |
| josch | that sounds familiar | 23:00 |
| ch | [ 3.139388] dw-mipi-dsi2 fde30000.dsi: [drm:dw_mipi_dsi2_bind [rockchipdrm]] *ERROR* Failed to find panel or bridge: -517 | 23:01 |
| ch | ah this is always printed :( | 23:04 |
| ch | https://zeha.at/~ch/rk3588display/dmesg.good-coldboot.txt https://zeha.at/~ch/rk3588display/dmesg.broken-reboot.txt | 23:06 |
| ch | maybe something helpful is in the dmesg | 23:06 |
| ch | this actually differs: | 23:06 |
| ch | [ 3.409127] panel-mnt-pocket-reform fde30000.dsi.0: [mnt pocket reform display] read register 12a: 0x77 | 23:06 |
| ch | (good) | 23:07 |
| ch | https://zeha.at/~ch/rk3588display/dmesg.broken-reboot.txt | 23:07 |
| ch | (broken) | 23:07 |
| ch | [ 3.449163] panel-mnt-pocket-reform fde30000.dsi.0: [mnt pocket reform display] read register 12a: 0x70 | 23:07 |
| ch | minute: ^ rings a bell? | 23:07 |
| + chartreuse (~chartreus@S0106908d78501d1d.cg.shawcable.net) | 23:31 | |
| - bpye (QUIT: Quit: Ping timeout (120 seconds)) (~bpye@user/bpye) | 23:45 | |
| + bpye (~bpye@user/bpye) | 23:45 | |
| minute | ch: ah. both are normally fine | 23:47 |
| minute | ch: i.e. on this rebooted system i also get 0x70 | 23:49 |
| - jacobk (QUIT: Ping timeout: 248 seconds) (~quassel@47-186-65-73.dlls.tx.frontiernet.net) | 23:49 | |
| minute | hmm maybe not all panels behave the same, even for the same model | 23:50 |
| minute | ch: just to be sure, brightnessctl yields non-0 number, yes? :D | 23:51 |
| minute | ch: what if you do another reboot, does it stay off? | 23:52 |
| minute | (i just tried another reboot, still works... so there might be subtle differences between the panels/hw) | 23:55 |
| ch | Device 'fde30000.dsi.0' of class 'backlight': | 23:57 |
| ch | Current brightness: 129 (51%) | 23:57 |
| minute | ok | 23:58 |
| ch | second reboot doesnt help :( | 23:59 |
| minute | a relatively quick thing one could try would be to add this hack to your rk3588 dts https://source.mnt.re/reform/reform-debian-packages/-/blob/main/linux/patches6.12/imx8mp-mnt-pocket-reform/pocket-panel/0001-v5-add-multi-display-panel-driver.patch?ref_type=heads#L653 | 23:59 |
| minute | i.e. init-in-enable; in the panel node | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!