| - chomwitt (QUIT: Ping timeout: 268 seconds) (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 00:11 | |
| - buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org) | 01:00 | |
| + buckket (~buckket@vps.buckket.org) | 01:00 | |
| - cobra (QUIT: Ping timeout: 260 seconds) (~cobra@user/Cobra) | 02:06 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 02:41 | |
| + cobra (~cobra@user/Cobra) | 02:41 | |
| - bluerise (QUIT: Ping timeout: 268 seconds) (~bluerise@pc19f8c38.dip0.t-ipconnect.de) | 02:43 | |
| + bluerise (~bluerise@pc19f8fe9.dip0.t-ipconnect.de) | 02:44 | |
| - mjw (QUIT: Ping timeout: 256 seconds) (~mjw@gnu.wildebeest.org) | 02:48 | |
| - nsc (QUIT: Ping timeout: 256 seconds) (~nicolas@29-98-142-46.pool.kielnet.net) | 03:10 | |
| + nsc (~nicolas@135-98-142-46.pool.kielnet.net) | 03:12 | |
| jfred | I've been switching back and forth between my a311d Reform and an Intel Compute Card I've had lying around and the Reform is noticeably faster. Granted it's a low spec Compute Card and those are a few years old now but I didn't realize the comparison would be so stark | 04:22 |
|---|---|---|
| jfred | Though granted today I had the card plugged into I think a 5k monitor and it didn't particularly like driving that | 04:23 |
| jfred | Even on the 1080p monitor at home though it's noticeable | 04:24 |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 04:37 | |
| + akira (~akira@37.4.230.225) | 04:37 | |
| - jn (QUIT: Ping timeout: 255 seconds) (~quassel@user/jn/x-3390946) | 05:00 | |
| + jn (~quassel@2001-4dd4-92d3-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 05:00 | |
| - jn (QUIT: Changing host) (~quassel@2001-4dd4-92d3-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 05:00 | |
| + jn (~quassel@user/jn/x-3390946) | 05:00 | |
| - skipwich (QUIT: Quit: DISCONNECT) (~skipwich@user/skipwich) | 05:04 | |
| + skipwich (~skipwich@user/skipwich) | 05:08 | |
| - akira (QUIT: Ping timeout: 246 seconds) (~akira@37.4.230.225) | 06:00 | |
| + akira (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 06:01 | |
| + robin_ (~robin@user/terpri) | 06:08 | |
| - robin (QUIT: Ping timeout: 240 seconds) (~robin@user/terpri) | 06:08 | |
| * robin_ -> robin | 06:09 | |
| - mtm (QUIT: Ping timeout: 246 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 06:24 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 06:50 | |
| + akira (~akira@37.4.230.225) | 06:50 | |
| chartreuse | Decided to look into why the headphone detection doesn't work and noticed that the wm8960 alsa driver in the kernel has a set of default register values, and in register 0x30 has the HPSEL bits there set to 00 which means headphone detect via GPIO1, but the reform has it hooked up via JD2(LINPUT3), so those bits should be 10 instead | 06:52 |
| chartreuse | Now I need to find if those values are configured in a device tree, or with module parameters, or needs to be set via ioctl or such after boot | 06:53 |
| chartreuse | Trying to figure out the "new" way the sources are organized on source.mnt.re, before the kernel patches and such were in a reform2-imx8mq folder under reform-system-image | 07:03 |
| josch | chartreuse: last time, sigrid tried to fix this issue with these changes to the device tree: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/6/diffs | 07:05 |
| chartreuse | Ah reform-debian-packages. Would have thought the kernel would have been separate to the rest but I guess that makese sense | 07:05 |
| chartreuse | Did just setting those not work? Or did he never finish commiting them or such? Looks like it's setting the right bits but dunno | 07:07 |
| josch | chartreuse: are you mixing up people's pronouns? :) | 07:08 |
| josch | chartreuse: sigrid supplied the patch, i tried it and it didn't have any noticable effect | 07:08 |
| chartreuse | Probably am XD, dunno why I defaulted to him there | 07:09 |
| chartreuse | I'll give it a try again at least, maybe make sure nothing else was set wrong from the datasheet | 07:10 |
| chartreuse | Not sure why they're also changing the gpio select lines though, I think those settings are for making GPIO1 a debounced jack select which doesn't apply to the JD2 input | 07:11 |
| josch | chartreuse: since you say that those changes should theoretically fix it, i just rebased the pipeline -- in about an hour you'll have a kernel package | 07:13 |
| chartreuse | I think the values should be 2 3 3 rather than 2 2 3 from their patches though. As looking at the circuit a high on the pin should indicate speakers while low is headphones | 07:14 |
| josch | we can give it a few rounds of trial and error :) | 07:15 |
| josch | unfortunately i switched out my imx8mq for an a311d so i'll not be any help with testing :( | 07:15 |
| chartreuse | I was just going to set up a fork of the repo to test it myself, but that works too. I'll still set up my fork for other stuff, and possibly to enable the patch I did last time to enable overriding the country code on the wifi card | 07:15 |
| chartreuse | Should work on the A311d as well if you add the patch to that dts as well. It's just overriding the defaults in the kernel module. And the WM dac is on the mainboard not the CPU | 07:16 |
| chartreuse | Though I do believe 2 3 3 is the right value since the middle bits go to Additional Control 2, and the low bit is 0 for high=hp, and 1 for high=speaker, and the top bit enables the headphone switch at all | 07:17 |
| chartreuse | Oh wait no I'm wrong | 07:18 |
| josch | chartreuse: note that if this is a new fork of the repo, minute needs to assign one of the gitlab CI runners to it or otherwise they'll not build your changes | 07:18 |
| chartreuse | The jack is the other way | 07:18 |
| chartreuse | I'm fine if the CI doesn't build it as I can build it locally, since this is just a deb package it'll build | 07:18 |
| chartreuse | Just would want to leave it on there case I make pull requests (which I guess having CI for would be a good idea but eh) | 07:19 |
| josch | the linux kernel is a very, very wild deb package though :) | 07:19 |
| chartreuse | I was building it before in 5.14 or such when I don't think it even was a package XD. Besides that I've only done it on "normal" x86 debian machines to a package | 07:19 |
| chartreuse | I'm sure the runner will beat me to compiling it from scratch XD though presumably only a make dtbs would be needed if you already had a local build of the kernel done | 07:40 |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@37.4.230.225) | 07:49 | |
| + akira (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 07:50 | |
| josch | chartreuse: you could ask minute about the x86 runner which is about twice as fast as the arm64 runner | 07:52 |
| - reform7299 (QUIT: Remote host closed the connection) (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 08:09 | |
| chartreuse | Hrm, the build errored out with tar complaining about no space left on device. I have to assume it must be trying to extract the kernel in /tmp then? It went beyond my scrollback so I'll have to find it in the sh file | 08:10 |
| josch | chartreuse: what exactly are you running? | 08:17 |
| josch | the build is done with sbuild and will do the build in whatever place you set up sbuild to use | 08:17 |
| josch | in my own ~/.sbuildrc i have $unshare_tmpdir_template = '/home/josch/tmp/tmp.sbuild.XXXXXXXXXX'; because my /tmp is too small :) | 08:17 |
| chartreuse | Ah, I was just following the minimal instructions on the README, but yeah I see now with watch df -h it's stuffing /tmp | 08:20 |
| chartreuse | Never messed with .sbuildrc so I'll have to add a line like that, or move /tmp to the drive | 08:21 |
| josch | just adding $unshare_tmpdir_template is probably most simple | 08:22 |
| Zaba | if you just want to try out a change to the device tree, why build the whole kernel? you can just build a new device tree file | 08:49 |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-64-230.tukw.qwest.net) | 08:58 | |
| + colinsane (~colinunin@97-113-64-230.tukw.qwest.net) | 09:02 | |
| [tj] | minute: do you have a config file for jtag on the imx8mp? | 10:06 |
| amospalla | Orange pi released its "Orange Pi CM5", a RK3588S compatible with cm4. | 10:18 |
| - josch (QUIT: Ping timeout: 272 seconds) (~josch@mister-muffin.de) | 10:18 | |
| amospalla | I see there are other several cm4 compatible rk3588, I guess this world is a wild field. | 10:19 |
| amospalla | And even though those as stated as compatible, are not with Reform, if were people would be using them. | 10:20 |
| + josch (~josch@mister-muffin.de) | 10:30 | |
| Zaba | I mean, most likely people have never tried with them with the reform | 10:45 |
| minute | they are also pretty new | 11:12 |
| minute | [tj]: nope... whatcha trying to do? | 11:13 |
| minute | i only have jtag things for ls1028a | 11:13 |
| gsora | didn't know rockchip made a rk3588s, it seems less powerful: https://wiki.radxa.com/Rock5/RK3588_vs_RK3588S | 11:14 |
| + mjw (~mjw@gnu.wildebeest.org) | 11:42 | |
| Zaba | not really less powerful, just smaller package and fewer external interfaces | 11:53 |
| gsora | you're right! | 11:54 |
| - Gooberpatrol66 (QUIT: Excess Flood) (~Gooberpat@user/gooberpatrol66) | 11:56 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 11:56 | |
| amospalla | I think, if you don't need the extra i/o, it doesn't matter which one you use. | 12:09 |
| Zaba | yep | 12:22 |
| [tj] | minute: I'm trying to figure out why the olimex soc won't boot and I'm hoping I can get a clue from jtag | 12:24 |
| [tj] | totally hampered by my soldering iron dying so I can't put add test probes to my sd card extension | 12:25 |
| - iank (QUIT: Quit: ZNC 1.8.2+deb2+b1 - https://znc.in) (~iank@fsf/staff/iank) | 12:27 | |
| + iank (~iank@fsf/staff/iank) | 12:27 | |
| minute | [tj]: did you adapt the olimex som to reform or how are you using it? | 12:33 |
| [tj] | it is just their som on their evb right now, I wanted a development target that wasn't just the pocket for bring up. I got it to work for 1 day and now I get nothing | 12:34 |
| [tj] | I was hoping to get the reform images running on it with some minor tweaks, but that plan fell apart | 12:34 |
| nocko | Hi there, imx8mq reform2 user here. I just upgraded to 12.6 including 6.7.12-reform2-1~bpo12+1+reform20240630T152702Z from bookworm-backports. There seems to be a significant battery life regression. I don't have solid numbers yet. I can setup a test with the old kernel tomorrow. Subjectively, I used to get around 4hr. battery life, closer to two hours now. | 12:39 |
| nocko | It pulls 225mA with the keyboard and screen backlights at zero... which seems high. But I hadn't had a reason to check under the previous kernel. | 12:40 |
| nocko | No load changes, SoC temperature is low/stable. CPU is scaling down to 1.0Ghz. Nothing obviously amiss. | 12:42 |
| josch | nocko: wow that sounds rough! | 12:45 |
| nocko | Yeah, not super sure where to start looking for answers. I am not a mobile-heavy user so there's no urgency. | 12:52 |
| josch | nocko: the weird thing is how kernel 6.7 didn't cause these problems when it was in unstable | 12:52 |
| nocko | Redoing the battery life estimate, it's closer to 3.2 hours, down from 4+ (I never measured precisely). | 12:53 |
| josch | if you can, please measure to make sure that it's the kernel | 12:54 |
| josch | on the imx8mq you are in the privileged position to have the display in u-boot | 12:54 |
| josch | so you can select which kernel you want to boot with using u-boot-menu and extlinux.conf | 12:54 |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 13:24 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 13:49 | |
| + akira (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 13:49 | |
| * Guest2827 -> mjw | 13:50 | |
| + gustav28 (~gustav@c-2f35524e.019-141-67626730.bbcust.telenor.se) | 14:05 | |
| nocko | I figured out mkimage and booted back into 6.6 | 14:05 |
| nocko | Power consumption (from OLED) is within a few mA of 6.7 ... Maybe I am imagining things? I will let it run for a while and calculate runtime. | 14:05 |
| nocko | calculating similar runtime after 5% drain. 1-2% difference, but within the error bars of my measurement capabilities... | 14:20 |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-2f35524e.019-141-67626730.bbcust.telenor.se) | 14:20 | |
| nocko | I will continue to monitor, but maybe I hallucinated this? | 14:21 |
| + chomwitt (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 14:33 | |
| - akira (QUIT: Ping timeout: 252 seconds) (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 15:15 | |
| + akira (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 15:18 | |
| + mtm_ (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 15:29 | |
| * mtm_ -> mtm | 15:30 | |
| - akira (QUIT: Ping timeout: 268 seconds) (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 15:33 | |
| - chomwitt (QUIT: Ping timeout: 268 seconds) (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 16:02 | |
| + gustav28 (~gustav@c-2f35524e.019-141-67626730.bbcust.telenor.se) | 16:14 | |
| amospalla | Oh man, I just found the kbdgfx utility to load bitmaps on my pocket, this is pure dope. | 16:46 |
| amospalla | I can use the oled as much as I wish while I keep sure I don't burn it with fixed white pixels right? | 16:52 |
| + mark_ (~mjw@gnu.wildebeest.org) | 16:57 | |
| + akira (~akira@2a01:599:a26:d475:43d2:fc02:92e:fd59) | 16:59 | |
| josch | amospalla: you know about refmon by vkoskiv too, right? :) | 17:21 |
| josch | also, the ezurio-qcacld-2.0-dkms package just migrated to testing -- for those who want to use the pocket reform with bookworm, could you see if that package makes it work? | 17:24 |
| amospalla | josch: I can not find that package neither on google or community.mnt.re, could you please give me a hint to find it? | 17:26 |
| josch | amospalla: https://packages.debian.org/trixie/ezurio-qcacld-2.0-dkms | 17:26 |
| amospalla | oh sorry, I meant refmon | 17:27 |
| josch | yes? | 17:27 |
| josch | there is a section "Download ezurio-qcacld-2.0-dkms" | 17:27 |
| josch | you can also temporarily add "trixie" to your sources.list to get a gpg verified download | 17:27 |
| amospalla | oh, I see, it is included in that package, thank you :) :) | 17:28 |
| amospalla | vkoskiv: josch just told me about your refmon tool, where is it hosted? | 17:32 |
| amospalla | josch: installed, complains on modprobe https://paste.debian.net/1322308/ | 17:57 |
| amospalla | Had to install dkms from testing, it complained about dkms from stable not being on the right version. | 17:57 |
| - sir-photch (QUIT: Remote host closed the connection) (~m-hy5poy@2a01:4f8:1c1b:4579::1) | 18:11 | |
| + sir-photch (~m-hy5poy@2a01:4f8:1c1b:4579::1) | 18:12 | |
| amospalla | maybe I'm missing the package reform-qcacld2 ? | 18:16 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 18:16 | |
| - cwebber (QUIT: Quit: crossing the threshold of Lugonu's domain) (~Christine@user/cwebber) | 18:27 | |
| + cwebber (~Christine@user/cwebber) | 18:30 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@2a01:599:a26:d475:43d2:fc02:92e:fd59) | 18:53 | |
| + akira (~akira@37.4.230.225) | 18:54 | |
| josch | amospalla: sorry, you are right, the dkms version is incompatible | 18:56 |
| josch | amospalla: i need to do a rebuild | 18:56 |
| - AbortRetryFail (QUIT: Quit: reboots) (~arf@146.ip-149-56-132.net) | 19:08 | |
| vkoskiv | amospalla: I should really publish the thing, it's been stuck in the neverending "one more fix" loop for ages | 19:26 |
| + chomwitt (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 19:51 | |
| * mjw -> Guest7935 | 20:12 | |
| - Guest7935 (QUIT: Killed (tantalum.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 20:12 | |
| * mark_ -> mjw | 20:12 | |
| + Guest7935 (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 20:12 | |
| + reform25902 (~root@216.197.221.29) | 20:14 | |
| - reform25902 (QUIT: Client Quit) (~root@216.197.221.29) | 20:15 | |
| amospalla | josch: ping me when you need a tester | 20:51 |
| amospalla | vkoskiv: ah ok, I thought it was something public, I'll be glad to try it if you publish it anytime. | 20:52 |
| josch | amospalla: let me build something for you | 21:14 |
| josch | amospalla: this is built for bookworm https://mister-muffin.de/p/HXTl | 21:15 |
| josch | it's a .deb so you can curl https://mister-muffin.de/p/HXTl > ezurio-qcacld-2.0-dkms_0.0~git20230623.2cd31b6-4_all.deb | 21:15 |
| josch | <= afk | 21:15 |
| - chomwitt (QUIT: Quit: WeeChat 3.8) (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 21:22 | |
| - sevan (QUIT: Ping timeout: 264 seconds) (~sevan@user/venture37) | 22:12 | |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-2f35524e.019-141-67626730.bbcust.telenor.se) | 22:15 | |
| josch | amospalla: just to be sure, maybe 'apt-get remove --purge dkms' and re-install the version from bookworm | 22:34 |
| + sevan (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 22:53 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!