| henesy | wayfire xkb_options doesnt seem to do anything either | 00:43 |
|---|---|---|
| henesy | building evremap to try that | 00:43 |
| henesy | this altgr key is driving me up the wall | 00:43 |
| josch | henesy: as a last resort you can patch the firmware | 00:44 |
| henesy | can i patch the firmware without disassembling everything and setting up a second computer to flash? | 00:44 |
| josch | on the big reform you can -- no idea about the pocket | 00:48 |
| josch | on the big reform there is a chapter in the handbook explaining the process | 00:49 |
| josch | maybe the pocket reform handbook knows more? | 00:49 |
| ch | isnt that the system controller firmware? | 00:50 |
| josch | there are 3 firmwares: lpc (system controller), keyboard and trackball | 00:51 |
| ch | minute: maybe want to update the crowd supply text here https://mntre.com/docs-pocket-reform.html ? | 00:51 |
| ch | https://source.mnt.re/reform/pocket-reform/-/blob/main/pocket-reform-keyboard-fw/pocket-hid/flash.sh?ref_type=heads seems like it should just work? | 00:59 |
| josch | maybe have another usb keyboard ready in case something fails | 01:01 |
| josch | page 74 here: https://mntre.com/documentation/pocket-reform-handbook.pdf | 01:01 |
| henesy | well for now i got altgr to be right shift thx to evremap https://gist.github.com/henesy/8be5d901875760b961b0a985e846feb8 | 01:02 |
| henesy | Hooray | 01:02 |
| henesy | rums a systemd service and works on wayfire for me | 01:02 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 01:18 | |
| ch | josch: https://community.mnt.re/t/latest-packages-no-poweroff-solved/2404/4 <- i wonder if the system image should have an mnt-meta package that is Protected: yes | 01:21 |
| josch | ch: it does -- reform-tools is exactly that | 01:22 |
| ch | but how can andy then be missing the kernel packages? | 01:22 |
| josch | ch: because you are free to remove the meta-package | 01:23 |
| ch | thats why i said Protected: yes (: | 01:23 |
| josch | but why? | 01:23 |
| ch | 'removing these packages will break booting and recovering will be a pain' | 01:23 |
| josch | but it will not | 01:23 |
| josch | removal of the meta package does not break anything | 01:24 |
| ch | thats by definition the case for a meta package | 01:24 |
| josch | yes, so why make it protected? | 01:24 |
| ch | to keep the kernel installed | 01:24 |
| ch | (without making a lot of individual packages protected) | 01:25 |
| josch | if you try to remove the last kernel, there is a big fat warning that you have to actively acknowledge first | 01:25 |
| josch | does it need to be made even harder? | 01:25 |
| josch | same if you try to remove the kernel you are currently running | 01:25 |
| josch | we are patching the Debian linux kernel, so the same protections are in place | 01:25 |
| ch | idk. it seems you have a reform-check script that needs to be run manually, and if the meta package just depends on everything and is protected, then users cant get it wrong | 01:26 |
| josch | i think for the kernel it's important to leave a bit of flexibility | 01:26 |
| josch | sometimes you read in the forum that the kernel upgrade breaks things, so you'd like to stick with kernel X | 01:27 |
| josch | so you remove the meta package | 01:27 |
| josch | and then you manually choose what kernel to install | 01:27 |
| josch | if you are running unstable, i think that is something that should not be made harder than necessary | 01:27 |
| ch | ok | 01:27 |
| josch | because regressions happen quite a bit, depending on your platform | 01:27 |
| ch | was just thinking this is a thing to make running unstable safer | 01:27 |
| josch | make it protected:yes would force you to always run the latest kernel | 01:28 |
| josch | i'd think that especially with unstable that is not always the perfect plan :) | 01:28 |
| ch | protected: yes doesnt mean you have to upgrade its dependencies | 01:29 |
| josch | you are right | 01:29 |
| josch | it's indeed a trade-off | 01:29 |
| josch | another thing that i try keeping in the back of my head is: ideally i want to make the diff against vanilla debian such that i can propose it to get added to debian proper at some point. i mean ideally, in the end, i do not want to do what i'm currently doing anymore -- i just want to be able to install plain debian and be happy :D | 01:36 |
| ch | diff to the debian kernel? | 01:39 |
| josch | for example, yes | 01:39 |
| josch | ch: see my MRs to src:linux in salsa | 01:39 |
| josch | it seems that the eleven MRs filed since 2022 were all Reform related | 01:41 |
| josch | at minidebconf in berlin this year i sat together with ben to talk about how to make the diff even smaller | 01:42 |
| - mtm (QUIT: Ping timeout: 248 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 02:02 | |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 02:05 | |
| + cobra_ (~cobra@user/Cobra) | 02:42 | |
| - cobra (QUIT: Ping timeout: 260 seconds) (~cobra@user/Cobra) | 02:42 | |
| + skipwich (~skipwich@user/skipwich) | 02:45 | |
| - skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich) | 02:58 | |
| - aloo_shu (QUIT: Ping timeout: 252 seconds) (~aloo_shu@90.166.193.183) | 02:59 | |
| + skipwich (~skipwich@user/skipwich) | 03:00 | |
| - mjw (QUIT: Ping timeout: 276 seconds) (~mjw@gnu.wildebeest.org) | 03:02 | |
| - cobra_ (QUIT: Ping timeout: 255 seconds) (~cobra@user/Cobra) | 03:05 | |
| + cobra (~cobra@user/Cobra) | 03:06 | |
| - bluerise (QUIT: Ping timeout: 252 seconds) (~bluerise@p5b211f01.dip0.t-ipconnect.de) | 04:24 | |
| + bluerise (~bluerise@p5b21118f.dip0.t-ipconnect.de) | 04:26 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 05:00 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-136-73.tukw.qwest.net) | 05:41 | |
| + colinsane (~colinunin@97-113-136-73.tukw.qwest.net) | 05:46 | |
| - colinsane (QUIT: Remote host closed the connection) (~colinunin@97-113-136-73.tukw.qwest.net) | 05:47 | |
| + colinsane (~colinunin@97-113-136-73.tukw.qwest.net) | 05:51 | |
| - Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 06:33 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 06:34 | |
| henesy | https://hackers.town/@cypnk/112998537118304514 Trackpoint reform module? | 06:39 |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 07:02 | |
| - Gooberpatrol66 (QUIT: Ping timeout: 276 seconds) (~Gooberpat@user/gooberpatrol66) | 07:04 | |
| josch | henesy: https://community.mnt.re/t/another-alternate-keyboard-idea/1915/5 | 07:50 |
| + aloo_shu (~aloo_shu@85.51.17.181) | 09:28 | |
| josch | minute: thank you for merging https://source.mnt.re/reform/reform/-/merge_requests/70 -- could you also please create a git tag 2024-07-26 and push it? Once that is in place, I can finish up the reform-firmware package. | 09:49 |
| minute | meh. did an upgrade for testing that waybar issue, and at the next cold boot display doesn't come up anymore | 11:00 |
| minute | i guess it worked the first time because that was a soft reboot | 11:00 |
| minute | aha | 11:01 |
| minute | "no display in initramfs" issue is back from the dead, at least on a311d. was able to type my disk pw invisibly and then the display came on a bit later | 11:01 |
| josch | minute: reform-tools 1.50 is quite ready. I'm wondering whether i should wait for reform-imx8mp-uboot MR 1 before releasing or rather release now? | 11:12 |
| minute | josch: oh, i now merged mr1 | 11:17 |
| josch | minute: noooo! test it first!!! XD | 11:18 |
| josch | minute: as indicated by the last comment in that MR, andypiper tried this out and failed | 11:19 |
| josch | but andy also didn't manage to connect via uart, so was not able to investigate any further | 11:20 |
| minute | josch: ah | 11:37 |
| + andypiper (~andypiper@89.36.117.58) | 11:39 | |
| minute | today and tomorrow i need to focus on some reform next milestones though | 11:39 |
| josch | okay, then i'll release reform-tools 1.50 now and we do imx8mp another time | 11:39 |
| josch | andypiper: we were just talking about your imx8mp u-boot experiences :) | 11:40 |
| andypiper | hello. I just saw @minute merged something there. | 11:40 |
| andypiper | what can I help to test today | 11:40 |
| josch | andypiper: the problem of the day is that waybar stopped working (see forum) :( | 11:41 |
| josch | and there seems to be a regression on a311d which (again) does not manage to initialize the display on time | 11:42 |
| josch | andypiper: if you want to help test stuff, you could install reform-tools 1.50 and see if anything breaks :) | 11:42 |
| andypiper | yes, waybar crashes since yesterday's OS updates I noticed. | 11:42 |
| josch | i was going to do that later today but any help is appreciated | 11:42 |
| andypiper | I was stracing it yesterday and it seemed to be some kind of GDK issue but nothing obvious had changed there | 11:43 |
| josch | we could try reverting the changes of the last upload and see if it fixes that | 11:43 |
| andypiper | I will update reform-tools shortly | 11:43 |
| andypiper | I'm working on some Mastodon tasks today but will tinker and share anything useful on the PR | 11:44 |
| andypiper | (Pocket I mean) | 11:44 |
| + bkeys (~Thunderbi@45.134.140.153) | 11:56 | |
| - mtm (QUIT: Ping timeout: 246 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 12:11 | |
| + mtm_ (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 12:11 | |
| josch | andypiper: once https://source.mnt.re/reform/reform-debian-packages/-/pipelines/2556 is finished, you could try adding this to your /etc/apt/sources.list and then upgrade and install reform-tools=1.50 | 12:25 |
| josch | deb [trusted=yes arch=arm64] https://source.mnt.re/reform/reform-debian-packages/-/jobs/5605/artifacts/raw/repo reform main | 12:25 |
| minute | i'm now upgrading libgtk-3-0t64 on my desktop pc | 12:25 |
| minute | and indeed, even my local build of waybar crashes then. | 12:25 |
| josch | we can just undo the last libgtk upload in the mntre.com repositories | 12:26 |
| andypiper | the last time waybar had similar issues reported to their GH, it was a gdk+gtk-layer thing I think | 12:26 |
| josch | are you filing a bug with a gdb backtrace with debuginfod.d.n? | 12:26 |
| andypiper | i did a bit of digging yesterday but haven't opened my Pocket yet this morning. I will do that when I get a chance. | 12:27 |
| minute | josch: how do i make sure reportbug will not fail at the end when trying to send the email? i remember it failed there a long time ago when i tried that because it wanted to use sendmail or sth | 12:32 |
| josch | minute: would you like me to upload a version of gtk3 to the mntre.com repo which has the recent upload reverted? | 12:32 |
| minute | josch: that would be amazing yes | 12:32 |
| josch | minute: reportbug will save a draft in /tmp | 12:32 |
| minute | ok thanks | 12:32 |
| josch | i think it prints the location of your mail before exiting | 12:32 |
| minute | then i'll report the bug now | 12:32 |
| josch | great, thank you | 12:32 |
| minute | ohh there's a gtk interface now | 12:34 |
| minute | josch: it's asking me about smtp servers now in the cli interface, sigh | 12:35 |
| josch | oh :( | 12:35 |
| minute | ah, i can just press enter apparently | 12:35 |
| minute | now i'm in the gui wizard | 12:35 |
| josch | minute: you can just send a mail with your regular MUA as well if reportbug fails you | 12:35 |
| minute | > 117 bug reports found | 12:35 |
| josch | minute: this pipeline should build gtk3 with the last upload reverted. Try out what comes out of it at the end and if it fixes the problem, press the merge button: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/57 | 12:39 |
| josch | (now afk doing non-reform stuff) | 12:39 |
| minute | bug submitted | 12:44 |
| minute | josch: thank you! | 12:45 |
| grimmware | So I think I know the answer to this (no.) but is there a sensible mechanism for disseminating apt holds to people who subscribe to them? Forgetting the work it would take for a moment, something like an RSS feed where the client essentially takes the metadata from the feed and holds and unholds the requisite packages at the requisite versions. | 12:56 |
| josch | grimmware: that is what apt-listbugs can do for you using the information in the debian bts | 12:57 |
| grimmware | what's the turnaround for e.g. minute's bug that they just submitted turning up in that? | 12:59 |
| grimmware | I think I need to read the apt-listbugs manpage. | 13:00 |
| josch | grimmware: when you "apt upgrade", apt-listbugs will query the bts for RC bugs for all the package versions you are about to upgrade and then display the information to you | 13:01 |
| minute | yeah i'm not sure where my bug is / if it's there yet | 13:02 |
| minute | ah, it was acked via email 8 minutes ago | 13:02 |
| josch | minute: you made the bug Severity: important | 13:03 |
| josch | so it is not important enough to show up | 13:03 |
| * Guest4206 -> mjw | 13:04 | |
| minute | josch: it was the highest that i could select in reportbug | 13:04 |
| minute | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1079292 | 13:04 |
| minute | josch: i was not offered the choice to select "grave" | 13:04 |
| josch | interesting... | 13:04 |
| grimmware | josch: I think what I'm getting at is a mechanism by which the user doesn't have to have the insight as to what the knock-on effect of the bugs actually is but can choose to trust MNT on what to not update just yet automagically | 13:04 |
| minute | FWIW i chose "novice" level because i didn't want to wade through endless forms or so | 13:04 |
| josch | maybe that's why | 13:05 |
| minute | can-i-run-apt-upgrade-on-my-reform.com | 13:05 |
| josch | grimmware: that would require a lot of work on the side of MNT | 13:05 |
| josch | for example the bug right now was found by the community | 13:06 |
| josch | grimmware: if you do not want to be subject to this sort of bugs, you could try running Debian stable or testing for a middle ground | 13:07 |
| Zaba | I mean, generally speaking, providing an additional level of quality assurance over Debian unstable is what Debian testing is for | 13:08 |
| grimmware | I'm not implying that MNT should write such a system, I'm gauging the appetite for it as to whether it's a thing that would be worth the effort for someone in the community to implement such that all MNT *need* to do is effectively say "pin x package at y version for z reason" | 13:08 |
| josch | grimmware: such pinning would require that you keep all the old package versions somewhere though :) | 13:09 |
| josch | and then a reverse dep gets built and you cannot install the old one anymore | 13:09 |
| Zaba | debian unable is just exactly what it says on the tin | 13:09 |
| Zaba | unstable* | 13:09 |
| josch | the problem is best solved by using testing which is guarded by very many QA tools to not get into a situation where things are not installable and where they work well together | 13:09 |
| josch | unstable exists so that users find bugs that the machine cannot find | 13:10 |
| Zaba | … especially because many problems are not going to be MNT-specific, even | 13:10 |
| josch | for example in this case, there is just no test which was running wayland and clicked a button to find out that it's buggy | 13:10 |
| grimmware | To clarify the context in which I'm asking these questions, my impression of MNT's goals is to push toward a more user-friendly and less "expert users" product whilst still maintaining the velocity afforded by unstable. I've done a whole bunch of work with RSS and configuration management on debian based systems in the past so I'm not trying to solve the instability for myself so much as trying to figure out if that's a | 13:13 |
| grimmware | feasible and sensible way to spend the time I have available to help the community. Personally I'm very happy to be running unstable. | 13:13 |
| Zaba | of course, the system only works as long as enough people still use unstable in some capacity to find issues and file bugs, but I think the important bit is that people usually make a conscious choice to use unstable | 13:13 |
| Zaba | one issue I had with running testing in the past is that it’s a very different vibe when you’re in a pre-release freeze and when you’re not | 13:14 |
| josch | another important aspect is that in the past, MNT needed the very, very latest software for a new SoM like the a311d or the rk3588 to run | 13:15 |
| josch | during the imx8mq days for a long time we were patching even mesa or gstreamer | 13:15 |
| josch | often things from experimental are needed | 13:15 |
| josch | things would be very different if support for the cpu modules was more mature | 13:15 |
| Zaba | (another downside of testing is of course the lack of intentional security updates) | 13:16 |
| grimmware | Cool, not enough of a problem to warrant the effort then | 13:17 |
| grimmware | What I *really* want to work on is improving the PD so it'll start charging when turned on with more chargers but that is well outside of my area of expertise and I've just moved house so all my kit is in disarray :P | 13:18 |
| minute | grimmware: are you already on the updated sysctl fw that fixed a lot more chargers? (i think you are, right)? | 13:23 |
| grimmware | PREF1SYSR120240716normal,0,0,0 | 13:25 |
| grimmware | it didn't change anything for me, Anker, Steam Deck and Apple chargers still don't do anything unless plugged in whilst the machine is turned off | 13:29 |
| grimmware | well, the apple chargers don't seem to work at all | 13:29 |
| grimmware | Anyway, like I say I'm not set up to debug these things so imma just deal with it (and probably source a ugreen psu) until I can meaningfully contribute to the fix | 13:31 |
| grimmware | I'm mostly in here chatting right now to avoid writing slides for my 44CON talk which I really need to get done but I hate writing | 13:32 |
| [tj] | grimmware: if you read the pd spec you would have stuff for 44CON++ talk | 13:34 |
| [tj] | usb, better than bluetooth, worse than a good | 13:34 |
| grimmware | yeah, it's a space with headroom though because I would love to be able to adapt more of my shit to run off the same PSUs | 13:35 |
| josch | minute: you can install patched libgtk-3-0t64 from here: https://source.mnt.re/reform/reform-debian-packages/-/jobs/5626/artifacts/browse/changes/ | 13:38 |
| josch | minute: or you wait half an hour and then you can install it via an apt upgrade | 13:39 |
| josch | minute: if you could confirm that this fixes the waybar crash, then i can follow up on the debian bug with the details of the patch | 13:39 |
| - mtm_ (QUIT: Ping timeout: 246 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 14:02 | |
| - gsora (QUIT: Ping timeout: 252 seconds) (~gsora@user/gsora) | 14:03 | |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 14:06 | |
| minute | josch: awesome | 14:08 |
| + gsora (~gsora@user/gsora) | 14:08 | |
| josch | reform-tools 1.50 is in the repos | 14:53 |
| josch | could somebody adventurou try to upgrade and see if anything broke on their system? | 14:54 |
| grimmware | installing now | 15:21 |
| grimmware | anything you specifically want me to test after it's done? | 15:22 |
| josch | grimmware: can you run "systemctl restart reform-hw-setup" | 15:23 |
| josch | and see if that explodor if the service is still running happily? | 15:23 |
| josch | other than that, just reboot, run "sudo reform-check" and see if you spot anything out of the ordinary | 15:24 |
| grimmware | > sudo systemctl restart reform-hw-setup | 15:24 |
| grimmware | Job for reform-hw-setup.service failed because the control process exited with error code. | 15:24 |
| grimmware | See "systemctl status reform-hw-setup.service" and "journalctl -xeu reform-hw-setup.service" for details. | 15:24 |
| grimmware | gimme a sec and I'll dig in | 15:25 |
| grimmware | https://www.irccloud.com/pastebin/5gVUm1KQ/ | 15:27 |
| josch | that's just an empty page | 15:27 |
| josch | ah javascript | 15:27 |
| josch | grimmware: can you give me the output of | 15:28 |
| josch | this page does not let me copypaste... | 15:28 |
| grimmware | oof grim, I'll use a different pastebin next time | 15:28 |
| josch | i found the button for the raw text | 15:29 |
| josch | grimmware: can you show me the output of "find /opt/reform-qcacld2/" | 15:29 |
| josch | maybe to be quick, just run this: | 15:30 |
| josch | find /opt/reform-qcacld2/ | curl -F 'arg=<-' https://mister-muffin.de/paste | 15:31 |
| josch | i only have about 10 minutes left before i have to leave, so maybe we can be fast and fix this quickly :) | 15:31 |
| grimmware | https://mister-muffin.de/p/upLy.txt | 15:31 |
| josch | grimmware: but... your wifi is working, right? | 15:32 |
| grimmware | josch: yeah worst comes to the worst I can leave my pocket turned on and plugged in until tomorrow, I'm also supposed to be going to a meeting | 15:32 |
| grimmware | yeap | 15:32 |
| josch | okay, maybe lets make this non-fatal for now | 15:32 |
| josch | grimmware: what is your currently running kernel version? | 15:33 |
| grimmware | Linux fmlr 6.9.12-mnt-reform-arm64 #1 SMP Debian 6.9.12-1+reform20240801T001732Z (2024-08-01) aarch64 GNU/Linux | 15:33 |
| josch | grimmware: what happens when you do this: | 15:34 |
| josch | sudo insmod /opt/reform-qcacld2/qcacld2-6.9.12-1+reform20240801T001732Z.ko | 15:34 |
| grimmware | insmod: ERROR: could not insert module /opt/reform-qcacld2/qcacld2-6.9.12-1+reform20240801T001732Z.ko: Invalid module format | 15:34 |
| grimmware | same as the unit | 15:35 |
| josch | ugh... | 15:35 |
| josch | grimmware: what is your output of "lsmod | grep qcacld2" | 15:36 |
| grimmware | qcacld2 4542464 0 | 15:37 |
| grimmware | cfg80211 839680 1 qcacld2 | 15:37 |
| josch | grimmware: does insmod work after you rmmod-ed the loaded module? | 15:38 |
| grimmware | isn't that going to make my wifi go away? | 15:38 |
| josch | correct -- are you currently typing on your pocket? | 15:39 |
| grimmware | yep | 15:39 |
| josch | then better not touch it | 15:39 |
| josch | i'm going to upload a hotfix for now | 15:39 |
| josch | lets fix it properl later today | 15:39 |
| grimmware | yeah sgtm | 15:39 |
| grimmware | I gotta go to a meeting now anyway | 15:39 |
| grimmware | I'll see if I can get my USB tether working later too so we can test without it being an issue if wifi goes away | 15:40 |
| josch | grimmware: there are 140 more people in this channel who can also help, submit patches etc -- don't worry about it. Have a nice day and thank you! | 15:42 |
| minute | we launched pocket upgrade bundles https://shop.mntre.com/products/mnt-pocket-reform-upgrade-bundle | 15:54 |
| minute | josch: grimmware: one can try -f with insmod. it can crash the kernel in some cases though. | 15:55 |
| gsora | yay congrats! should've kept my pocket reservation, didn't know it could run the rk3588 | 16:09 |
| + mark_ (~mjw@gnu.wildebeest.org) | 16:19 | |
| andypiper | upgrade bundle ordered! I already have the wifi card and only went for the 16gb, but I threw some merch onto the order to support as well (I can wait for it all to be ready together) | 16:25 |
| andypiper | sorry, have been afk / not paying attention to package updates today... will get back to it when i can | 16:25 |
| - mark_ (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 16:29 | |
| minute | andypiper: nice, thank you!! | 17:01 |
| josch | minute: the commit that broke reform-hw-setup on the pocket (but a workaround should be in the repos now) was this one https://source.mnt.re/reform/reform-tools/-/commit/d405f82f5e09b2e744e1efd764c64739577a9a9e It was provided to me by somebody who said that they tested it on their Pocket Reform, so I'm surprised that it failed now. In any case, since the idea of the code is to have a fallback | 17:09 |
| josch | mechanism, it would be wrong to force the loading because that would invalidate the point of trying out fallback modules. | 17:10 |
| josch | minute: i just tried out a fresh system image with the patched gtk-3.0 package and waybar seems to work again. Did you make the same observation? | 17:14 |
| minute | josch: this patch is dangerous because it uses insmod -f | 17:15 |
| minute | josch: that can crash the kernel if one is unlucky | 17:15 |
| minute | josch: i'm sorry, i couldn't test gtk3 yet because i had too many other things to do, but i can do it now in parallel | 17:16 |
| josch | minute: that patch did not introduce the -f -- that was there before | 17:19 |
| minute | josch: that is true, but before it didn't try any kernel module... only the one with the matching name | 17:20 |
| minute | hmhm | 17:20 |
| josch | lets remove the -f in any case | 17:21 |
| minute | yeah, i hope it works without :D | 17:21 |
| josch | is insmod supposed to succeed if the module is already loaded? | 17:21 |
| josch | maybe what is missing is just a "is the module already loaded?" check | 17:21 |
| minute | this i don't know by heart | 17:21 |
| josch | yeah same :D | 17:21 |
| josch | lets investigate... | 17:22 |
| minute | downloading the libgtk3 stuff on my pocket. | 17:25 |
| minute | josch: indeed the issue is fixed. green light by me to release :3 | 17:26 |
| josch | great, i'll also follow up on that bug and bump severity | 17:26 |
| minute | thaanks | 17:28 |
| andypiper | nice | 17:28 |
| josch | okay, so when you try to insmod a module that is already loaded you get a non-zero exit status and an EEXISTS | 17:29 |
| josch | which makes sense | 17:29 |
| josch | but | 17:29 |
| josch | for grimmware, when trying to load /opt/reform-qcacld2/qcacld2-6.9.12-1+reform20240801T001732Z.ko, the result was "Invalid module format" | 17:29 |
| minute | invalid module format happens if the module version does not 100% fit the kernel | 17:29 |
| minute | i.e. built against not exactly the right headers | 17:29 |
| josch | i thought so too but grimmware said that the running kernel matched exactly | 17:30 |
| josch | and the code is trying to do exactly that | 17:30 |
| minute | there might be some slight difference that was overlooked | 17:30 |
| josch | it's using $(uname -r) to find the correct module | 17:30 |
| josch | yes | 17:30 |
| minute | i don't know what the right tooling to check that is, though | 17:30 |
| josch | for any pocket user who is reading this: on the pocket, this has a zero-exit status if your wifi is working, correct? grep --quiet '^qcacld2 ' /proc/modules | 17:37 |
| amospalla | josch: yes | 17:38 |
| josch | amospalla: thank you! | 17:40 |
| amospalla | (: | 17:40 |
| josch | minute: then i'll skipp running init_qca9377_wifi() in reform-hw-setup if the module is already loaded | 17:41 |
| minute | josch: sounds good! | 17:46 |
| andypiper | so I'm without wireless on pocket with 6.10.6 at the moment, and qcacld2 is not found to load. | 17:51 |
| andypiper | fortunately I have a wired option | 17:51 |
| sevan | lifehack: buy a stock pocket reform and then the upgrade kit to obtain a cheap RK3588 :) | 17:54 |
| josch | andypiper: that is probably after the latest reform-tools update? | 17:54 |
| andypiper | yes, i have 1.51 here now | 17:54 |
| josch | andypiper: okay, so you rebooted and your wifi didn't come back? | 17:54 |
| andypiper | correct | 17:55 |
| josch | andypiper: what happens when you try to insmod manually? | 17:55 |
| andypiper | `insmod: ERROR: could not load module qcacld2: No such file or directory" | 17:55 |
| josch | you need to give the full path like this: | 17:55 |
| josch | insmod "/opt/reform-qcacld2/qcacld2-$(dpkg-query --show --showformat '${Version}' "linux-image-$(uname -r)").ko" | 17:56 |
| josch | hah i found the bug! | 17:56 |
| minute | oh? :3 | 17:56 |
| josch | ah i did not | 17:56 |
| josch | 10 seconds of endorphine... | 17:56 |
| andypiper | `insmod: ERROR: could not insert module /opt/reform-qcacld2/qcacld2-6.10.6-1+reform20240821T152344Z.ko: Invalid module format` which I think grimmware reported as well | 17:58 |
| + mark_ (~mjw@gnu.wildebeest.org) | 17:58 | |
| andypiper | albeit different kernel version | 17:58 |
| - mjw (QUIT: Killed (tungsten.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 17:59 | |
| * mark_ -> mjw | 17:59 | |
| + Guest3164 (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 17:59 | |
| minute | andypiper: and with -f? | 17:59 |
| andypiper | checking | 17:59 |
| andypiper | worked and wifi came up | 18:00 |
| josch | minute: and that's why you put the -f in there back in the day :D | 18:00 |
| andypiper | heh | 18:00 |
| andypiper | i can share dmesg output | 18:00 |
| minute | josch: that might be 100% true | 18:01 |
| andypiper | https://paste.lol/andypiper/qcalcld2-dmesg | 18:01 |
| minute | still might be good to figure out why it's needed | 18:01 |
| andypiper | actually i just updated the paste to have the whole dmesg output from boot | 18:03 |
| andypiper | so you'll see me inserting the ethernet cable and then a gap and then inserting qcacld2 later | 18:03 |
| grimmware | re, sorry was just finishing up my meetings and then having a conversation with my partner about how much better our lives would be if I had an RCORE for my pocket reform and then ordering an RCORE upgrade bundle | 18:04 |
| andypiper | lmao | 18:05 |
| andypiper | and your life is now 150% better, with the additional 50% on arrival of the upgrade | 18:05 |
| josch | grimmware: we should all feel inspired by your example :D | 18:05 |
| josch | fun fact: the -f flag for insmod is not mentioned in its man page nor in its --help output?? o0 | 18:06 |
| grimmware | the great thing about documentation is | 18:07 |
| josch | i need that t-shirt | 18:07 |
| grimmware | yeah we just had a new person join our team who actually already worked at the company a year and a half ago and one of my coworkers was like "well at least they'll already have a bunch of context even if it's a 18 months out of date" and I said "they'll be up to date with the documentation at least" | 18:09 |
| josch | oh no XD | 18:09 |
| josch | see it like this: at least you *have* documentation -- don't ask me for my projects at work :D | 18:10 |
| grimmware | yeah sometimes I really have to take stock of the problems I have and realise that they're not actually terrible problems | 18:11 |
| grimmware | like sometimes I go "WHY IS EVERYTHING SO HARD?" and then I remember that I seek out hard problems on purpose. | 18:11 |
| grimmware | I like difficult problems, books and drinks | 18:11 |
| josch | grimmware or andypiper or both: could you apply this to your reform-hw-setup with reform-tools 1.51 and confirm that this lets you reboot and still have wifi and restart the service and the service does not error out? https://paste.debian.net/hidden/03ece6bc/ | 18:12 |
| grimmware | on it, just making sure I have 1.51 first | 18:14 |
| grimmware | josch: no 1.51 in the repos, just did an update | 18:14 |
| josch | oh dear... | 18:15 |
| josch | oh no, i forgot to click the run pipeline button... | 18:16 |
| josch | see you again in 1:45 hours | 18:16 |
| josch | grimmware: actually the patch should apply to 1.50 as well | 18:17 |
| grimmware | ack, I'll do that | 18:17 |
| andypiper | rebooting it now | 18:18 |
| andypiper | and we have wifi | 18:19 |
| andypiper | sp yes I confirm that patch works | 18:19 |
| josch | what does "systemctl status reform-hw-setup" say? | 18:19 |
| andypiper | one sec | 18:19 |
| josch | what happens when you "systemctl restart reform-hw-setup"? | 18:19 |
| andypiper | ok bear with | 18:20 |
| grimmware | I am so awful at using `patch`, what options should I be using? | 18:20 |
| grimmware | I have to relearn this every fucking time I backport a patch | 18:20 |
| josch | grimmware: patch -p1 < /path/to/file | 18:20 |
| grimmware | p1 uuuugh | 18:20 |
| grimmware | thank you | 18:20 |
| josch | if it cannot find the file it will ask you for the path | 18:21 |
| andypiper | what do you need from the status output, it's a longish response. | 18:21 |
| andypiper | Loaded: loaded (/usr/lib/systemd/system/reform-hw-setup.service; enabled; preset: enabled) | 18:21 |
| andypiper | Active: inactive (dead) since Thu 2024-08-22 17:18:42 BST; 1min 51s ago | 18:21 |
| josch | andypiper: just that it is successful :) | 18:21 |
| andypiper | Process: 678 ExecStart=/usr/sbin/reform-hw-setup (code=exited, status=0/SUCCESS) | 18:22 |
| grimmware | yeah same, it looks to be a oneshot that exited successfully | 18:22 |
| andypiper | ok let's restart it | 18:22 |
| josch | yes, it's a oneshot unit | 18:22 |
| grimmware | And I am, as you can see, still on wifi | 18:22 |
| josch | :D | 18:22 |
| grimmware | I'll also throw a reboot just in case | 18:22 |
| andypiper | restart did not complain | 18:23 |
| josch | awesome, thank you! | 18:24 |
| josch | the change you tested is now here: https://source.mnt.re/reform/reform-tools/-/merge_requests/83 | 18:24 |
| andypiper | lgtm | 18:25 |
| josch | oh wait, we can trick the mntre.com updated script | 18:26 |
| andypiper | i can always throw the AsiaRF card back in anyway but wanted to help get the base stuff ironed out (and I keep noodling with other cards in that slot) | 18:26 |
| josch | by re-running the build_custom job only | 18:26 |
| josch | and then the repo will skip 1.51 when it updates in half an hour | 18:26 |
| josch | and go straight to the fixed 1.52 | 18:26 |
| andypiper | I mean I do always see "ar6k_wlan mmc1:0001:1: firmware: failed to load qcacld2/qsetup30.bin (-2)" on my machine anyway in the dmesg but it has never been a problem with wifi working afterwards | 18:27 |
| grimmware | I also have working wifi on reboot | 18:29 |
| andypiper | i have to drop for now but will check back later | 18:30 |
| * andypiper -> andypiper[afk] | 18:32 | |
| - andypiper[afk] (QUIT: Quit: My device has gone to sleep. Zzzz…) (~andypiper@89.36.117.58) | 18:32 | |
| grimmware | man I am so excited about the rcore | 18:34 |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 18:42 | |
| + mjw (~mjw@89.205.132.233) | 19:06 | |
| - Gooberpatrol_66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 19:41 | |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 19:41 | |
| + mtm_ (~textual@2600:381:e503:9254:38ab:814b:8f31:1cc9) | 19:51 | |
| grimmware | minute: I installed the unofficial signal-desktop package that you recommended by the way (thanks for that) and it still has the issue under sway that it will launch backgrounded until you launch it again - do you have this issue and if so, do you have a workaround for it? | 19:51 |
| - mtm (QUIT: Ping timeout: 272 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 19:52 | |
| minute | grimmware: interesting, i don't think i had this issue... maybe it's a setting? | 19:56 |
| grimmware | are you running some kind of system tray? | 19:57 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 20:01 | |
| - Gooberpatrol_66 (QUIT: Ping timeout: 252 seconds) (~Gooberpat@user/gooberpatrol66) | 20:02 | |
| + gustav28 (~gustav@c-1f51205c.019-141-67626730.bbcust.telenor.se) | 20:02 | |
| minute | grimmware: yeah, that's part of waybar | 20:04 |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 20:06 | |
| - mtm_ (QUIT: Ping timeout: 246 seconds) (~textual@2600:381:e503:9254:38ab:814b:8f31:1cc9) | 20:07 | |
| - mjw (QUIT: Ping timeout: 248 seconds) (~mjw@89.205.132.233) | 20:10 | |
| grimmware | ah yeah ofc, I'm misremembering about swaybar instead | 20:44 |
| + mjw (~mjw@185.59.39.53) | 21:06 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 21:10 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 21:11 | |
| + andypiper (~andypiper@45.148.12.75) | 21:33 | |
| * andypiper -> andypiper[afk] | 21:41 | |
| - andypiper[afk] (QUIT: Quit: My device has gone to sleep. Zzzz…) (~andypiper@45.148.12.75) | 21:42 | |
| + andypiper (~andypiper@45.148.12.75) | 21:47 | |
| * andypiper -> andypiper[afk] | 21:48 | |
| - andypiper[afk] (QUIT: Client Quit) (~andypiper@45.148.12.75) | 21:48 | |
| + andypiper (~andypiper@45.148.12.75) | 21:50 | |
| - andypiper (QUIT: Client Quit) (~andypiper@45.148.12.75) | 21:52 | |
| + andypiper (~andypiper@45.148.12.75) | 21:53 | |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-1f51205c.019-141-67626730.bbcust.telenor.se) | 22:15 | |
| grimmware | oh TIL about apt-listbugs.service, that is very cool | 23:01 |
| grimmware | I've only ever used stable debian derivatives before | 23:01 |
| grimmware | oh I suppose I used testing for a while | 23:02 |
| grimmware | this is a very cool system though, I especially love that the pins are stored with the bug context in an easy to read manner in the authoritative plaintext file | 23:02 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!