| - chomwitt (QUIT: Ping timeout: 268 seconds) (~chomwitt@2a02:587:dc14:f500:278:be15:4a20:8304) | 00:43 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 01:09 | |
| - S0rin (QUIT: Ping timeout: 250 seconds) (~S0rin@user/s0rin) | 02:31 | |
| + S0rin (~S0rin@user/s0rin) | 02:32 | |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 03:04 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 04:00 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 04:34 | |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149-210-4-72.mobile.nym.cosmote.net) | 05:31 | |
| + Ar|stote|is (~linx@149-210-4-72.mobile.nym.cosmote.net) | 05:36 | |
| + chomwitt (~chomwitt@2a02:587:dc14:f500:9e6a:c6cb:5be3:53b3) | 07:30 | |
| - ajr (QUIT: Quit: WeeChat 3.6) (~ajr@user/ajr) | 09:03 | |
| + MajorBiscuit (~MajorBisc@c-001-013-022.client.tudelft.eduvpn.nl) | 09:58 | |
| - iank (QUIT: Quit: ZNC 1.8.2+deb2+b1 - https://znc.in) (~iank@fsf/staff/iank) | 13:25 | |
| + iank (~iank@fsf/staff/iank) | 13:27 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by wielaard!~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440))) (~mark@gnu.wildebeest.org) | 13:49 | |
| * wielaard -> mjw | 13:49 | |
| + mark_ (~mark@gnu.wildebeest.org) | 13:50 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 14:14 | |
| - buckket (QUIT: Quit: buckket) (~buckket@pdp8.buckket.org) | 14:17 | |
| + buckket (~buckket@pdp8.buckket.org) | 14:21 | |
| vkoskiv | Update: My reform battery cells appear to be fine. I charged it up fully yesterday and did multiple hours of work on battery power. | 15:51 |
|---|---|---|
| vkoskiv | Protected battery holders are definitely on my shopping list, though! | 15:51 |
| vkoskiv | Hm. My battery status went away again after an apt upgrade. | 16:39 |
| vkoskiv | I do have reform2-lpc-dkms installed, but the lpc kernel module isn't loaded for some reason | 16:40 |
| vkoskiv | Do I need to enable it somehow? | 16:40 |
| vkoskiv | I used to manually compile and install it from the forked repo | 16:40 |
| josch | vkoskiv: I don't know what the migration path is, but the kernel module is now built by reform-tools via dkms | 16:41 |
| vkoskiv | How do I trigger that to happen? | 16:41 |
| josch | sudo apt install reform-tools | 16:41 |
| vkoskiv | Already installed | 16:41 |
| vkoskiv | Reinstall maybe? | 16:42 |
| josch | vkoskiv: what version? | 16:42 |
| vkoskiv | 1.11 | 16:42 |
| josch | good | 16:42 |
| josch | vkoskiv: maybe this thread starting from here can help you: https://community.mnt.re/t/lpc-spi-firmware-and-driver-prototype/1030/18 | 16:43 |
| vkoskiv | I can see a module in the directory you mentioned there. I tried installing reform2-lpc, but that package doesn't seem to exist. | 16:46 |
| vkoskiv | Trying to modprobe that ko file spits up a fatal error, it can't find it (even though it's there, seemingly?) | 16:49 |
| josch | vkoskiv: the package is called reform2-lpc-dkms -- where do you have the other package name from? | 16:50 |
| vkoskiv | From your message in the thread: "Can you try to apt update and apt install reform2-lpc to get the latest version of the package?" | 16:50 |
| vkoskiv | I also tried reform2-lpc-dkms, but that's also already installed. | 16:51 |
| josch | vkoskiv: thanks, let me fix that error | 16:51 |
| josch | vkoskiv: but you have /lib/modules/5.19.0-reform2-arm64/updates/dkms/reform2_lpc.ko, correct? | 16:52 |
| josch | and what does "lsmod | grep lpc" show? | 16:53 |
| vkoskiv | I do, yeah. Just tried reinstalling reform2-lpc-dkms too, didn't work I think. Booting. | 16:53 |
| josch | and what happens if you try to "modprobe reform2_lpc"? | 16:53 |
| vkoskiv | Before reinstalling reform2-lpc-dkms, it showed nothing. Now it shows reform2_lpc and the battery status works again. | 16:54 |
| vkoskiv | I think this might have something to do with me manually installing it before, and/or that SD -> eMMC boot migration. | 16:54 |
| josch | vkoskiv: did you reboot after installing it? | 16:56 |
| vkoskiv | Yes | 16:56 |
| josch | hrm... | 16:56 |
| vkoskiv | minute: Can I get a thumbs up/down on https://source.mnt.re/reform/reform/-/merge_requests/31 | 17:03 |
| josch | does the reform keyboard firmware have any releases or tags? Or am I supposed to just pick the current git commit and flash that? | 17:08 |
| vkoskiv | At this time the master branch is intended to be stable, yeah. | 17:09 |
| vkoskiv | But releases would be more ideal | 17:09 |
| josch | yeah, I don't know whether what is in git HEAD is always deemed to be of stable release quality or not | 17:09 |
| vkoskiv | !31 just makes it so it ignores oled draw commands from the OS when it's displaying something. | 17:09 |
| vkoskiv | I guess it's convention. *Usually* with production software, master isn't guaranteed to be stable. Tagged commits/releases are. | 17:10 |
| vkoskiv | That's also how I do it in c-ray, though nobody uses that apart from me.\\ | 17:10 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 17:32 | |
| minute | vkoskiv: can you ask me again tomorrow afternoon please? i'm kind of in a crunch to finish the cursed fpga project | 19:04 |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark_!~mark@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:28 | |
| * mark_ -> mjw | 19:29 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:29 | |
| + oomono (uid328183@id-328183.tinside.irccloud.com) | 20:12 | |
| - cwebber (QUIT: Remote host closed the connection) (~user@user/cwebber) | 20:20 | |
| - MajorBiscuit (QUIT: Ping timeout: 246 seconds) (~MajorBisc@c-001-013-022.client.tudelft.eduvpn.nl) | 20:21 | |
| vkoskiv | minute: Will do! | 20:36 |
| vkoskiv | Good luck with cursed FPGA | 20:36 |
| josch | Where is the holy inquisition when you need them for an exorcism?! | 20:40 |
| vkoskiv | Why do /dev/hidraw* devices sometimes get "messed up"? | 20:40 |
| vkoskiv | i.e. my /dev/hidraw0 just seemingly randomly stopped being a special device file and started being a normal file | 20:40 |
| vkoskiv | And now the keyboard oled is one of the other ones instead, I have to guess. | 20:40 |
| vkoskiv | hidraw1, now. | 20:41 |
| vkoskiv | Also, is there a proper way to write to them without root? Do I have to generate a user and mess with groups and stuff? | 20:41 |
| josch | vkoskiv: you can chown them to a group (usually "input") and then make your user be part of that group | 20:43 |
| josch | that chown can be done by a udev rule | 20:43 |
| vkoskiv | Thanks, I'll look into that. | 20:44 |
| josch | vkoskiv: It's possible that /dev/hidraw* becoming normal files is caused by reform-standby | 20:45 |
| josch | Maybe it can happen that at the time that reform-standby writes xLITE0 or xLITE6 to it, the device is doesn't exist yet or not anymore | 20:45 |
| josch | vkoskiv: can you reproduce the problem? | 20:45 |
| + cwebber (~user@user/cwebber) | 20:48 | |
| vkoskiv | I did suspend and resume, so that might have been it | 20:50 |
| vkoskiv | In the middle of coding, will try to repro later. | 20:50 |
| josch | If that is it, does anybody know if it's possible to write a shell script that writes to a file but doesn't create it if it doesn't exist? I.e. how to open a file for writing without O_CREAT? Is that possible? | 20:54 |
| josch | If it isn't I guess the best we can do is to clean after ourselves if the file we just wrote to turned out to be a regular file instead of a character special. | 20:54 |
| sigrid | test -c FILE for character devices | 20:56 |
| sigrid | test -c /dev/hidraw... && echo blah >> /dev/hidraw... | 20:56 |
| sigrid | (probably) | 20:56 |
| josch | sigrid: that creates a race condition | 20:57 |
| josch | between checking and writing to it, the file might disappear and then we create a regular file through the redirection | 20:57 |
| josch | the current code does a for d in /dev/hidraw*; do... | 20:58 |
| josch | so between evaluating the glob and doing the echo, /dev/hidrawXX disappears | 20:58 |
| sigrid | what's that code do? | 20:58 |
| josch | sigrid: https://source.mnt.re/reform/reform-tools/-/blob/main/sbin/reform-standby#L40 | 20:58 |
| sigrid | you need to put backlight on after the bind | 20:59 |
| sigrid | I am pretty sure that's the reason why it sometimes creates normal files | 20:59 |
| josch | aha! | 21:00 |
| sigrid | or always, which I'd always expect | 21:00 |
| josch | that makes a lot of sense, thanks | 21:00 |
| josch | indeed the resume tasks should be in reverse order of the suspend tasks which currently they are not... | 21:01 |
| josch | sigrid: thank you, nice catch! | 21:01 |
| sigrid | np, I hope I am not wrong here ^__^" | 21:03 |
| josch | hrm... this still leaves me puzzled why the glob /dev/hidraw* matches anything | 21:03 |
| sigrid | also on a note related to lpc driver. as I understood it's not setting pm_power_off? would probably make sense to do that, so the systemd shutdown hack isn't needed | 21:04 |
| sigrid | it would also work for distros that have none of that, like Void | 21:04 |
| josch | sigrid: can you create a merge request against https://source.mnt.re/reform/reform-tools/-/blob/main/lpc/reform2_lpc.c ? | 21:05 |
| sigrid | if I get to that over the weekend, sure | 21:05 |
| josch | sigrid: what does writing xPWR0 to /dev/hidraw* do exactly? | 21:09 |
| sigrid | it asks the keyboard to write to lpc over spi to shut off power rails, which lpc does | 21:09 |
| sigrid | it doesn't give the chance for the kernel to issue a full shutdown of all devices, I believe, including nvme | 21:10 |
| sigrid | hence that post in community about unsafe shutdowns on nvme, which I could also see here with 9front (but that's because 9front doesn't tell nvme it's shutting down yet) | 21:10 |
| josch | yeah, that's why I asked | 21:11 |
| sigrid | since that lpc driver module isn't built into the kernel, I think it should be able to override pm_power_off, which by default will probably be set to something that does not really work on reform | 21:12 |
| sigrid | don't think arm trusted firmware knows how to ask lpc to turn everything off, anyway | 21:13 |
| josch | sigrid: are you planning to reply to that forum post with your insights or would you otherwise mind me copypasting your messages from here to there? | 21:13 |
| sigrid | yeah, feel free to do that | 21:14 |
| josch | okay! | 21:14 |
| sigrid | should be noted that is unverified information, I am not a linux kernel dev :) | 21:14 |
| sigrid | I think reform's keyboard firmware might have kinda the same issue as trackball one - useless hid reports are being sent non-stop | 21:23 |
| sigrid | I see 17Mb is transferred on one endpoint over 2 days, compared to 13Mb for the trackball | 21:23 |
| sigrid | can't imagine how it would ever be more data on the keyboard | 21:24 |
| vkoskiv | Oled draw commands are sent as hid reports to the keyboard | 21:24 |
| vkoskiv | I'm writing things with that. I'd assume refreshing the oled at 50Hz would add up over time :D | 21:24 |
| sigrid | I am not using that :) | 21:24 |
| sigrid | the laptop is mostly idle | 21:25 |
| - Gooberpatrol_66 (QUIT: Remote host closed the connection) (~Gooberpat@user/gooberpatrol66) | 22:09 | |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@id-328183.tinside.irccloud.com) | 22:24 | |
| sigrid | nvm, keyboard fw is fine. it's 9front's usb keyboard driver that sets a pretty small idle period (32ms?) as a workaround for raspi usb being crap | 22:49 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!