| + staticbunny (~static@76-223-253-78.lightspeed.frokca.sbcglobal.net) | 00:03 | |
| ch | i wish it had homing markers. project for another day | 00:04 |
|---|---|---|
| - aloo_shu (QUIT: Ping timeout: 276 seconds) (~aloo_shu@90.166.193.48) | 00:11 | |
| + aloo_shu (~aloo_shu@90.166.193.48) | 00:19 | |
| - vagrantc (QUIT: Ping timeout: 260 seconds) (~vagrant@2600:3c01:e000:21:7:77:0:40) | 00:57 | |
| + jacobk (~quassel@2603:8080:b200:7b02:77cb:6304:f9db:dda1) | 01:08 | |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@2603:8080:b200:7b02:77cb:6304:f9db:dda1) | 01:14 | |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 01:19 | |
| + cobra (~cobra@user/Cobra) | 01:54 | |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 02:02 | |
| + cobra (~cobra@user/Cobra) | 02:10 | |
| - cobra (QUIT: Client Quit) (~cobra@user/Cobra) | 02:12 | |
| + cobra (~cobra@user/Cobra) | 02:24 | |
| jfred | It takes some getting used to because the keys are smaller but overall I really really like it | 02:31 |
| jfred | also I just had someone at a coffee shop walk past and then comment "that keyboard sounds very nice!" haha | 02:32 |
| jfred | I'm not quite at 100% of my normal typing speed on the pocket's keyboard but I've gotten pretty close | 02:33 |
| jfred | been a nice little blogging machine lately | 02:33 |
| - paperManu (QUIT: Ping timeout: 255 seconds) (~paperManu@107.159.243.8) | 03:00 | |
| - natalie (QUIT: Remote host closed the connection) (~natalie@user/natalie) | 03:31 | |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 03:42 | |
| + jacobk (~quassel@47-186-105-237.dlls.tx.frontiernet.net) | 03:48 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 04:28 | |
| - xha (QUIT: Quit: WeeChat 4.4.3) (~xha@user/xha) | 05:06 | |
| - colinsane (QUIT: Ping timeout: 248 seconds) (~colinunin@97-113-93-25.tukw.qwest.net) | 05:14 | |
| + xha (~xha@user/xha) | 05:27 | |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 05:28 | |
| - bluerise (QUIT: Quit: brb) (~bluerise@pc19f85d1.dip0.t-ipconnect.de) | 05:53 | |
| + chomwitt (~chomwitt@2a02:587:7a1a:7a00:1ac0:4dff:fedb:a3f1) | 08:11 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-93-25.tukw.qwest.net) | 08:58 | |
| xha | does anyone know a shop in the EU (germany?) that sells the M2 5mm screws in black for the pocket reform? | 09:05 |
| xha | ACTION just lost one and it took 1 hour to find it | 09:05 |
| xha | wanna be prepared for the next time that happens lol | 09:05 |
| - aloo_shu (QUIT: Ping timeout: 245 seconds) (~aloo_shu@90.166.193.48) | 09:37 | |
| + gustav28 (~gustav@c-78-82-55-219.bbcust.telenor.se) | 09:41 | |
| minute | hmm we should probably offer some screws | 09:43 |
| + aloo_shu (~aloo_shu@90.166.193.48) | 09:48 | |
| hramrach | I ordered some spare screws from China. There are probably local sellers that would have them but difficult to find | 09:48 |
| hramrach | Screws like these are commonly found in laptops so it's generic spare part that repair shops need to get somewhere | 09:49 |
| josch | xha: in the past i sometimes shopped at minischrauben.com but i also have two local shops that sell metric screws and electric components | 10:01 |
| - jacobk (QUIT: Ping timeout: 248 seconds) (~quassel@47-186-105-237.dlls.tx.frontiernet.net) | 10:04 | |
| + jacobk (~quassel@47-186-105-237.dlls.tx.frontiernet.net) | 10:05 | |
| [tj] | minute: will there be an assembly at 38c3? | 10:06 |
| hramrach | metric screws are common but only in bigger sizes | 10:06 |
| - grimmware (QUIT: Ping timeout: 252 seconds) (sid638840@id-638840.hampstead.irccloud.com) | 10:08 | |
| minute | [tj]: yes if it is accepted (so far that always worked out) | 10:10 |
| [tj] | cool | 10:10 |
| [tj] | I'd never considered an assembly might not be taken | 10:10 |
| + grimmware (sid638840@id-638840.hampstead.irccloud.com) | 10:12 | |
| amospalla | I touch type better on the pocket reform than on my old gpd pocket 2. | 10:12 |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 10:26 | |
| - colinsane (QUIT: Client Quit) (~colinunin@97-113-93-25.tukw.qwest.net) | 10:29 | |
| grimmware | xha: I bought a box of various spares from eBay but that was in the UK so ymmv | 10:34 |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 10:35 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 10:35 | |
| grimmware | I also found I was touch typing VERY quickly | 10:36 |
| josch | minute: do you remember whether the important part about imx8mp-mnt-pocket-reform/audio/0000-revert-crashy-audiomix-pm-support.patch was the pm stuff or whether it was the removal of bits in clk_imx8mp_audiomix_driver? It would be easier to carry that patch forward if it would not try to clean up dead code. | 10:40 |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 10:41 | |
| minute | josch: afaik pm stuff. we need to test to make sure of course | 10:46 |
| josch | minute: okay, then i'll change the commit to just pull out the pm stuff (removing the include of linux/pm_runtime.h just to make sure nothing calls it anymore) and then you can test it | 10:47 |
| minute | josch: cool, thank you! | 10:48 |
| ch | minute: did you get to try the fwupd stuff? | 11:03 |
| ch | i was also wondering where to get the m2 screws from. for some reason thats not a normal part :( | 11:03 |
| josch | if anybody in europe wants to receive a postcard from the beautiful city of Würzburg with a few M2 screws screw taped to it, i'm happy to send some out :) | 11:12 |
| - chomwitt (QUIT: Quit: WeeChat 3.8) (~chomwitt@2a02:587:7a1a:7a00:1ac0:4dff:fedb:a3f1) | 11:19 | |
| minute | ch: will test on device today! | 11:33 |
| ch | https://source.mnt.re/zeha/pocket-reform/-/jobs/6728 this job should produce a firmware with todays date as version, so you have something to upgrade to! | 11:36 |
| ch | 11:39 | |
| - erle (QUIT: Remote host closed the connection) (~erle@user/erle) | 11:51 | |
| + erle (~erle@user/erle) | 11:54 | |
| + mjw (~mjw@gnu.wildebeest.org) | 11:54 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-93-25.tukw.qwest.net) | 11:56 | |
| josch | minute: linux 6.12-rc6 to try: https://source.mnt.re/josch/reform-debian-packages/-/jobs/6715/artifacts/browse/repo/pool/main/l/linux/ | 12:13 |
| + paperManu (~paperManu@107.159.243.8) | 12:26 | |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 12:26 | |
| - colinsane (QUIT: Remote host closed the connection) (~colinunin@97-113-93-25.tukw.qwest.net) | 12:26 | |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 12:30 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-93-25.tukw.qwest.net) | 12:44 | |
| minute | trying to fix up my rk3588 pocket which is in a mega weird state | 12:51 |
| minute | it claims there are _no_ network interfaces (neither ethernet nor wifi) | 12:51 |
| minute | (still with 6.11.4) | 12:51 |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 12:52 | |
| minute | btw we will put M2x5 countersunk and cheese head screw set in the shop for pocket reform | 12:52 |
| vkoskiv | TIL they are called cheese head screws :D | 12:53 |
| josch | oh same o0 | 12:53 |
| josch | minute: once you are at it, how about also selling the screws with nylon threading? | 12:54 |
| minute | josch: yes they are with nylon dispensing | 12:55 |
| minute | so yeah they can be used for reform too | 12:55 |
| minute | aha, reform-hw-setup failed on the device | 12:56 |
| minute | (the service) | 12:56 |
| minute | gpioset: invalid option -- 'c' | 12:57 |
| minute | ok so we need to add a dependency of a newer version of gpioset on reform-tools | 12:57 |
| minute | i mean, my case is pretty special, but still | 12:57 |
| minute | (package gpiod) | 13:00 |
| minute | now getting an error from dpkg of reform-branding, that it wants to overwrite an existing file from reform-handbook 2.0 (html/_static/logo.svg) | 13:02 |
| minute | ok, doing apt install reform-tools reform-handbook pocket-reform-handbook reform-branding several times eventually untangles this | 13:04 |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-93-25.tukw.qwest.net) | 13:05 | |
| ch | :/ sounds like that needs a breaks/conflicts/replaces | 13:06 |
| minute | now upgrading the pocket to 6.11.7 | 13:06 |
| + colinsane (~colinunin@97-113-93-25.tukw.qwest.net) | 13:10 | |
| minute | now trying reform-flash-uboot because my console is not rotated | 13:11 |
| minute | it did something | 13:12 |
| minute | aha, console fixed | 13:12 |
| minute | font is still smol in the cryptsetup though. did i not merge that? | 13:12 |
| ch | can check cat /proc/cmdline if the font name is there, i guess | 13:14 |
| minute | it's not | 13:15 |
| minute | so reform-flash-uboot or the database that it uses needs to be updated to a newer build? cc josch | 13:15 |
| ch | (i havent tried a new uboot yet) | 13:15 |
| - mjw (QUIT: Ping timeout: 260 seconds) (~mjw@gnu.wildebeest.org) | 13:15 | |
| minute | the flashing definitely worked | 13:15 |
| minute | now on 6.11.7 and my ethernet is back also | 13:15 |
| josch | minute: latest tag on https://source.mnt.re/reform/reform-rk3588-uboot/-/tags is 2024-10-11b and that's the one that reform-tools knows about | 13:16 |
| minute | my asiarf card is not detected since a while it seems but maybe just mechanical issue | 13:16 |
| josch | ch: yes, breaks/conflicts/replaces is missing | 13:16 |
| josch | ch: this is currently caught in my upload of reform-branding to NEW and i'm re-using that | 13:17 |
| ch | ack | 13:17 |
| ch | wondering why i get network-manager-applet on upgrading | 13:17 |
| * Guest5844 -> mjw | 13:19 | |
| minute | josch: new tag 2024-11-13 pushed for reform-rk3588-uboot | 13:20 |
| josch | minute: thank you also for the hint about your gpioset not having the -c option. The -c option was indeed only added in gpiod 2.0 and this made me discover that reform-tools does not even depend on gpiod in the first place! ouch... | 13:24 |
| minute | josch: woops! | 13:24 |
| josch | the package is still there because we install it in reform-system-image | 13:25 |
| minute | ok, asiarf was just mechanical issue, card was not tightly plugged in | 13:26 |
| josch | minute: your build for 2024-11-13 failed: https://source.mnt.re/reform/reform-rk3588-uboot/-/jobs/6741 | 13:26 |
| - murphnj (QUIT: Remote host closed the connection) (~murph@user/murphnj) | 13:28 | |
| + murphnj (~murph@user/murphnj) | 13:28 | |
| minute | josch: meh lol | 13:33 |
| minute | josch: looks like i need to update patches | 13:39 |
| minute | ch: i've built and confirmed reformusbtool on the device and can confirm it works | 13:48 |
| minute | now trying to setup fwupd | 13:48 |
| minute | ./contrib/setup doesn't work because of error: externally-managed-environment | 13:50 |
| minute | guess i need apt install python3-virtualenv first | 13:51 |
| minute | yep, that helps | 13:52 |
| ch | right. you might need other packages installed too for building fwupd. i forgot to list that step (and dont know which libs you need) | 13:53 |
| minute | getting some weird issues | 13:59 |
| minute | ch: so far: libgnutls28-dev libgpgme-dev gnutls-bin valac | 14:00 |
| ch | could try apt-get build-dep fwupd | 14:01 |
| minute | > FAILED: libfwupd/Fwupd-2.0.gir | 14:01 |
| minute | ah > Couldn't find include 'Json-1.0.gir' | 14:01 |
| ch | i think the libs probably didnt change between 1.9.x and 2.0/git | 14:01 |
| minute | maybe that's libjson-glib-dev | 14:01 |
| grimmware | Okay, I think I've nailed down the keyboard issues I was having - something happens sometimes when unplugging/plugging in power that can cause the keyboard to no longer be able to talk to the sysctl. The sysctl is still fine, it's still up, but any attempt to talk to it from the keyboard times out. Unplugging/plugging in again, flashing or rebooting the sysctl with picotool, none of this helps | 14:01 |
| minute | grimmware: hmmmm brownout on the keyboard rp2040? it seems sensitive | 14:02 |
| grimmware | thing is the keyboard still works but it occasionally seems to miss keypresses | 14:02 |
| minute | also needs libflashrom-dev | 14:03 |
| grimmware | this seems to manifest differently if I'm actively polling the sysctl for accelerometer data interestingly enough | 14:03 |
| minute | > /usr/bin/ld: /home/minute/src/fwupd/venv/build/../../plugins/flashrom/fu-flashrom-plugin.c:369:(.text+0x1280): undefined reference to `flashrom_programmer_shutdown' | 14:04 |
| minute | maybe debian has too old libflashrom-dev? | 14:05 |
| ch | it worked for me in unstable :o | 14:06 |
| ch | libflashrom-dev:arm64 1.4.0-3 | 14:06 |
| [tj] | would some logging in the por handler show if the rpi2040 in the keyboard was browning out? | 14:07 |
| grimmware | minute: is it possible to cut the power supply for the keyboard from the sysctl? That would enable me to turn it off and on again without needing an external device because other than dropping a few keypresses I can still type over USB to the SoM, I just can't talk to the sysctl any more | 14:09 |
| grimmware | I put the keyboard into programming mode and rebooted it with picotool and no dice | 14:10 |
| minute | ch: i deleted fwupd and started the process again and now it worked | 14:12 |
| ch | great :) | 14:12 |
| minute | but! | 14:13 |
| minute | > cp: cannot stat '/home/minute/src/fwupd/venv/bin/../dist/share/polkit-1/actions/org.freedesktop.fwupd.policy': No such file or directory | 14:13 |
| grimmware | oh shit I managed to get it back again | 14:13 |
| grimmware | okay, putting it in programming mode and picotool reboot worked that time | 14:15 |
| minute | ch: where to get/build org.freedesktop.fwupd.policy ? | 14:15 |
| ch | uhuh | 14:15 |
| ch | i think for me ninja just builds it | 14:15 |
| grimmware | more data is needed (again) | 14:15 |
| ch | venv/dist/share/polkit-1/actions/org.freedesktop.fwupd.policy | 14:15 |
| minute | ch: ok maybe i'm still missing a dep that silently fails | 14:16 |
| ch | minute: did you try installing the deps with apt-get build-dep fwupd? | 14:16 |
| ch | libpolkit-gobject-1-dev maybe? | 14:16 |
| minute | ch: i'll try that now | 14:20 |
| minute | rebuilding | 14:25 |
| minute | ok that looks better | 14:27 |
| minute | ok this works > fwupdtool get-details ~/Downloads/sysctl.cab | 14:29 |
| minute | installing says > Specified firmware is already installed '1112' | 14:30 |
| minute | ch: i guess i should upload a new build to embargo now right, which one? | 14:31 |
| minute | i added the embargo conf and fwdupdmgr refresh works | 14:33 |
| minute | > Successfully downloaded new metadata: Updates have been published for 1 local device | 14:33 |
| ch | ah yeah, you can upload to embargo and use that for updating | 14:37 |
| ch | or you can use fwupdtool install sysctl-new.cab directly | 14:37 |
| ch | but obviously doing it via embargo is the more magic experience :) | 14:37 |
| minute | ch: gnome-firmware also works btw | 14:38 |
| minute | ch: yes, i want the magic | 14:39 |
| ch | ack | 14:39 |
| ch | haven't tested gnome-firmware, good to know it works! | 14:39 |
| ch | do you have screenshots? | 14:39 |
| ch | was wondering where the texts from the xml go | 14:39 |
| minute | yes i'll post some screenshots, will first publish 1113 if that exists so i can make even more screenshots | 14:42 |
| minute | ch: probably this one yes? https://source.mnt.re/zeha/pocket-reform/-/jobs/6728/artifacts/browse/pocket-reform-sysctl-fw/build/ | 14:42 |
| ch | yes | 14:43 |
| ch | (this has the usbpd fixes) | 14:43 |
| minute | we want those fixes :D | 14:43 |
| minute | ok, upload went fine, but > The remote metadata will be rebuilt in 17 minutes. This firmware has not yet been included in the XML catalog. | 14:44 |
| ch | o_o no clue what that means | 14:44 |
| ch | yeah, i need to spend more time on the usbpd branch and get it ready(-ish) | 14:44 |
| minute | i think we'll need to wait 15 minutes until their remote catalog is updated | 14:44 |
| minute | ch: what are you not happy with yet in the usbpd branch? | 14:45 |
| ch | there's like a ton of //FIXME and //TODO that i need to check, and i need to see if usb2.0 devices via an A-to-C adapter work (and if not, _if_ that is a regression) | 14:46 |
| minute | aha! | 14:54 |
| minute | i can test that | 14:54 |
| minute | > The remote metadata will be rebuilt in 6 minutes | 14:54 |
| ch | i've started reading the usb specs how to deal with a-to-c adapters, but haven't wrapped my head around it. if its not a regression, i could maybe just finish up the FIXME/TODOs and leave a-to-c for later | 14:57 |
| ch | (there's lots of other pd stuff to be done, but not as important) | 14:57 |
| ch | oh, maybe pd-source. i forgot what the status is there | 14:58 |
| ch | kinda amazing how many "corner" cases there are | 14:58 |
| minute | yeah | 14:59 |
| minute | there will be more with VDM :D (for alt mode) | 14:59 |
| ch | yes :> | 14:59 |
| minute | > The remote metadata will be rebuilt in 59 minutes. This firmware has not yet been included in the XML catalog. | 15:01 |
| minute | meh | 15:01 |
| - ndufresne (QUIT: Ping timeout: 252 seconds) (~ndufresne@apple.collaboradmins.com) | 15:03 | |
| ch | :( | 15:07 |
| minute | ok, so will try again in an hour ^^ | 15:10 |
| hramrach | when the number of corners is infinite the wheel is round! | 15:10 |
| amospalla | If you want feedback with Debian/Stable ask me, altough I guess fwupd may be old. | 15:24 |
| minute | currently struggling with the last hurdle in rk3588 uboot update | 15:28 |
| minute | > include/configs/rk3588_common.h:31:23: error: expected ‘}’ before ‘PARTS_DEFAULT’ | 15:28 |
| ch | fwupd in stable is too old | 15:31 |
| minute | ah bizarre, this error in the include is only thrown by dtc, not by gcc | 15:33 |
| ch | (re fwupd: well, the code didnt exist until a few days ago, so clearly stable cant have it) | 15:35 |
| minute | no i was mistaken, it's gcc but it's the spl build | 15:38 |
| minute | main uboot builds but spl doesn't | 15:38 |
| hramrach | and it did build before ?? | 15:41 |
| minute | well, lots of things were changed in the meantime | 15:45 |
| minute | aha, turns out PARTS_DEFAULT is not defined at all | 15:45 |
| hramrach | I see, it does not have the sources, it pulls a fresh copy each build | 15:45 |
| ch | lol, ploopy/canada post sent my order to australia m( | 15:45 |
| minute | i used some macros and MAKE V=1 to rerun the failing gcc command and PARTS_DEFAULT is undefined | 15:45 |
| minute | ch: oof... i wonder how often that happens | 15:46 |
| minute | ch: what did you order, trackball or trackpad? | 15:46 |
| ch | mouse actually! | 15:46 |
| minute | ch: ha! | 15:47 |
| minute | aha > The firmware is in the embargo-mntre remote. It was moved an hour ago | 16:01 |
| minute | fwupdmgr get-updates shows 1113 as available now | 16:02 |
| minute | also shows up in gnome-firmware now! i'll try to update from there | 16:02 |
| minute | installation successful!@ | 16:03 |
| ch | whoop | 16:03 |
| minute | i'll post some screenshots | 16:04 |
| minute | https://mastodon.social/@mntmn/113476289677361916 | 16:13 |
| - mesaoptimizer (QUIT: Quit: mesaoptimizer) (~mesaoptim@user/PapuaHardyNet) | 16:19 | |
| ch | nice | 16:24 |
| ch | did you see any release notes btw? | 16:26 |
| + mesaoptimizer (~mesaoptim@user/PapuaHardyNet) | 16:26 | |
| minute | ch: Description says > Development build for testing | 16:31 |
| minute | ch: sooo plugging in usb1/2 device with adapter in the charging port was still untested, right? can do that now | 16:33 |
| + mark_ (~mjw@gnu.wildebeest.org) | 16:39 | |
| minute | ch: usb device (superspeed though!) doesn't work on the port | 16:42 |
| - xktr (QUIT: Quit: leaving) (~xktr@user/xktr) | 16:42 | |
| ch | minute: usb device with adapter needs testing yes. weird that superspeed dev also doesnt work, but possible. need to check when i'm back home | 16:43 |
| ch | i assume thats a superspeed+pd device you're testing with :) | 16:43 |
| minute | ch: no just a samsung usb stick/drive | 16:44 |
| ch | right | 16:44 |
| ch | next week then | 16:44 |
| minute | yeah no problem | 16:44 |
| ch | did it work before? | 16:45 |
| minute | really awesome about the fwupd integration btw | 16:45 |
| minute | ch: well, i haven't tried that in a while... it worked originally with my firmware :D | 16:45 |
| ch | :) | 16:45 |
| ch | so, the usbpd branch needs more work | 16:46 |
| ch | the fwupd changes could maybe go into main sooner | 16:46 |
| ch | needs some thought on what to do about reformusbtool. (just take it? needs maybe some doc) | 16:47 |
| minute | ch: i think it's good to have it | 16:48 |
| + xktr (~xktr@user/xktr) | 17:01 | |
| + glu__ (~glu@178.115.78.40.wireless.dyn.drei.com) | 17:23 | |
| - chrcav (QUIT: Ping timeout: 246 seconds) (~chrcav@user/chrcav) | 17:24 | |
| - glu (QUIT: Ping timeout: 252 seconds) (~glu@91.141.32.69.wireless.dyn.drei.com) | 17:25 | |
| * glu__ -> glu | 17:25 | |
| + chrcav (~chrcav@user/chrcav) | 17:26 | |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@47-186-105-237.dlls.tx.frontiernet.net) | 17:45 | |
| - nsc (QUIT: Ping timeout: 260 seconds) (~nicolas@i5C74DC67.versanet.de) | 18:24 | |
| ch | minute: do you want it to live in pocket-reform somewhere or in a separate git? if separate git, can you make one? the tool works for me. would be good if someone else tries it etc | 18:32 |
| jfred | I've been having some odd phantom keypresses when I press a few keys in the middle of my big Reform's keyboard after the rk3588 upgrade. I'm guessing some contacts are getting bridged on the underside of the keyboard since the rk3588 is a bit of a tighter fit and probably flexing the motherboard some. Waiting on an order of kapton tape to try and fix that | 18:33 |
| hramrach | jfred: do you use those 'antiflex inserts'? | 18:49 |
| hramrach | picture https://shop.mntre.com/products/mnt-reform-anti-flexing-bars | 18:50 |
| staticbunny | Shouldnt my time stay correct if my batteries havent lost charge? | 18:52 |
| Twodisbetter | amospalla: do you notice somethings that your script crashes and resuming does not work? | 18:52 |
| Twodisbetter | amospalla: it doesn't happen often, but does happen. It feels very reminiscent of how the Reform itself woudl crash when trying to resume from suspend. | 18:53 |
| staticbunny | nm need to install hwclock it seems | 19:00 |
| amospalla | Twodisbetter: I don't use my pocket that much actually (lack of time), neither psuspend, so I have not seen that. | 19:01 |
| amospalla | Try locate the relevant lines of psuspend output in `journalctl -u psuspend` | 19:01 |
| ch | staticbunny: pocket? time should stay correct if the batteries did not run out | 19:09 |
| minute | ch: the tool also works for me | 19:10 |
| minute | jfred: kapton tape sounds like a good idea | 19:11 |
| ch | nice | 19:13 |
| staticbunny | ch: yeah apparently you need to install hwclock and enable it. i forgot about that | 19:13 |
| ch | uhuh | 19:13 |
| staticbunny | https://community.mnt.re/t/pocket-reform-is-not-keeping-time-anymore/2231 | 19:14 |
| ch | if thats true then something is wrong :> | 19:14 |
| staticbunny | both my imx and rk3588 didnt keep time or have hwclock installed by default | 19:15 |
| minute | debugging the new rk3588 uboot. the bootflow now starts with scanning for EFI partitions | 19:16 |
| minute | lol it wants to boot from nvme when i eject the sd card during boot | 19:20 |
| ch | "not wrong" :D | 19:21 |
| ch | staticbunny: ttbomk on a modern distro, systemd-timesyncd or ntpd should tell the kernel that the time is in sync, and the kernel should update rtc0 on shutdown | 19:21 |
| ch | so if that doesn't work, either of these things might be broken | 19:21 |
| ch | needing hwclock is a bit 80ies :> | 19:22 |
| josch | minute: maybe it would make sense to pin the commit hashes for the repos that reform-rk3588-uboot clones? You could bump those at your convenience and not be blocked by these failures when you actually want to spend time on other stuff | 19:22 |
| minute | josch: yeah it's true, we can go back of course... just was interested if we could use the newest version | 19:22 |
| minute | but maybe i should stop going down that rabbit hole now | 19:23 |
| josch | minute: sure, fair enough! You are pinning the commit hash for the other u-boots, just not for rk3588 | 19:24 |
| staticbunny | if anyone cares intel BE700 works on my pocket | 19:31 |
| minute | staticbunny: nice | 19:37 |
| staticbunny | now im trying to get a t99w175 (snapdragon X55) modem to work | 19:40 |
| staticbunny | i didnt realize the wifi that came with the RK3588 didnt have an LTE modem | 19:41 |
| staticbunny | sms_tool | 19:42 |
| staticbunny | oops | 19:42 |
| staticbunny | ch: you're prob right about something else going on with ntpd. i checked my services and ntpd was running but i will still having to do ntpdate manually after logging in. Granted that assumes someone has network connected but even without that, you would think the local RTC would be keeping time between shutdowns as long as the battery doesnt die. | 19:53 |
| staticbunny | minute: if i can't get this modem working, can you reccomend one? | 20:00 |
| minute | staticbunny: i think EM060K-GL should work | 20:04 |
| staticbunny | thank you, i'll check it out | 20:13 |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 20:25 | |
| - digitalrane (QUIT: Ping timeout: 250 seconds) (~digitalra@user/digitalrane) | 20:27 | |
| + digitalrane (~digitalra@user/digitalrane) | 20:33 | |
| + hairu (m-uotkmd@user/hairu) | 20:39 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 20:43 | |
| + hairu (m-uotkmd@user/hairu) | 20:46 | |
| jfred | hramrach: I don't, no. Might have to order some the next time I order from the shop :) | 20:54 |
| vkoskiv | Noticed this commit, which fixes a goof I wrote when reworking the trackpad fw, and now I wonder how that was spotted :D | 20:55 |
| vkoskiv | https://source.mnt.re/reform/reform/-/commit/8cbc34e01aea3ae56624277e79a9e840e319bc42 | 20:55 |
| vkoskiv | Hoping it wasn't too long of a debug session :') | 20:56 |
| vkoskiv | Here you can see the goof going in: https://source.mnt.re/reform/reform/-/commit/d7d028c37bfd4a49f104dfcd2b6e0d8c47922e7d#b675366c1b24ebe74c59585ec7a755b188fed2b8_259_255 | 20:57 |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 20:59 | |
| + hairu (m-uotkmd@user/hairu) | 21:02 | |
| - chrcav (QUIT: Ping timeout: 260 seconds) (~chrcav@user/chrcav) | 21:05 | |
| + chrcav (~chrcav@user/chrcav) | 21:06 | |
| minute | ok the uboot repo is problematic because it looks like they rewrote the history in a way that there is no good old commit to go back to | 21:06 |
| minute | but maybe there's a way by keeping 3 patches in our repo | 21:08 |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 21:24 | |
| + hairu (m-uotkmd@user/hairu) | 21:25 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 21:27 | |
| vagrantc | getting used to the new mnt/reform keyboard layout (never really was used to the old one!) ... struggling a bit but overall love the feel so far | 21:32 |
| Twodisbetter | amospalla: it happens so infrequently that I don't really mind | 21:32 |
| vagrantc | actually got an external keyboard as well as an upgraded keyboard for the mnt/reform itself... just driving the new keyboard for real now :) | 21:33 |
| vagrantc | the homing keys are subtle in a nice way :) | 21:34 |
| amospalla | Twodisbetter: ok, if it starts annoying you anytime, send me the logs so I can check it. | 21:34 |
| vagrantc | curiously, the power required by this mnt/reform keyboard seems to be 0.049 watts! I think that's less than my usb happy hacker that i've been using | 21:52 |
| minute | vagrantc: how do you measure? some usb-c power meter? | 21:54 |
| vagrantc | usb-a power meter ... i cannot vouche for superb accuracy | 21:54 |
| vagrantc | oh, actually the happy hacker doesn't even register | 21:54 |
| vagrantc | it was the optical mouse mounted off of it | 21:55 |
| vagrantc | minute: does that sound suspiciously low? | 21:55 |
| vagrantc | with the backlight turned all the way down, it does not even register any draw ... e.g. only 0.00 amps of precision at 4.98 volts | 21:57 |
| vagrantc | at what i think is max brightness, it seems to be registering 0.06 0.07 amps at 4.98 volts | 21:58 |
| minute | > [ 0.033980] smp: Bringing up secondary CPUs ... | 21:58 |
| minute | [ 5.094282] CPU1: failed to come online | 21:58 |
| minute | ok i guess i need to debug this further tomorrow... boot is totally borked now | 21:59 |
| vagrantc | i noticed even at minimum brightness there is some light under the keys still | 21:59 |
| minute | vagrantc: oh, is that a WLED keyboard? | 22:00 |
| minute | i.e. white backlight | 22:00 |
| vkoskiv | ^ If so, there is some analysis here: https://source.mnt.re/reform/reform/-/issues/20 | 22:02 |
| vkoskiv | Though my hardware analysis should be taken with a few grains of salt, I'm still more of a programmer, I just have a scope now :] | 22:03 |
| vagrantc | minute: white is the only color i see, yes | 22:03 |
| - murphnj (QUIT: Quit: Leaving) (~murph@user/murphnj) | 22:04 | |
| vagrantc | hah. i just had an idea for a feature ... would be nice if it could actually have an option to type out the firmware stuff from the circle menu ... as reading it off the oled is kind of error prone, but typing it out would be handy for accurate cut-and-paste :) | 22:05 |
| vagrantc | e.g. the system status thing, etc. | 22:06 |
| vagrantc | huh, now even with the led lights turned on, it seems to be drawing no current | 22:09 |
| minute | sounds like problem with the measurement setup | 22:10 |
| vagrantc | ah, it really should be drawing more than that? | 22:10 |
| minute | more than 0? | 22:11 |
| vagrantc | well, more than 0.00 ... that is the degree of precision ... could be 0.009 for all i know | 22:11 |
| minute | with my usb-c power meter (not sure how good that is) i get 0.152W for full brightness on the wled keyboard | 22:12 |
| vagrantc | it's jsut a cheap little thing ... figured i'd see what it had to say | 22:12 |
| minute | and 0.005W-0.0010W at lowest brightness | 22:12 |
| minute | sorry, that's 0.005W-0.010W | 22:13 |
| minute | so with leds turned off it really uses very little power | 22:14 |
| vagrantc | at minimum, i still see some light under the keys ... | 22:14 |
| vagrantc | though it is very faint... but more than when unplugged, for example | 22:15 |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-55-219.bbcust.telenor.se) | 22:15 | |
| minute | yeah, that is a known issue with the pwm control on that model | 22:18 |
| minute | on lowest brightness setting the led pwm should be fully turned off but that never landed | 22:18 |
| vagrantc | it is workable. | 22:18 |
| vagrantc | so i missed out on the v4 generation keyboards with the wild colored LEDs? not much loss, really, other than maybe red being better for night vision | 22:19 |
| minute | these functions should be refactored (DRY) https://source.mnt.re/reform/reform/-/blob/master/reform2-keyboard-fw/backlight.c?ref_type=heads#L41 | 22:19 |
| minute | and then on the lowest setting the PWM should be turned off fully | 22:19 |
| vagrantc | it is awesome to know that if i really wanted to fix it, i could at least give it a shot :) | 22:22 |
| vkoskiv | The regression was introduced by my patch that adjusted the PWM settings to be more 'correct'. I took a careful look at the datasheet some months ago to figure that out, but didn't end up with a solid conclusion beyond those comments. | 22:23 |
| vkoskiv | vagrantc: This is likely the patch that introduced that regression: https://source.mnt.re/reform/reform/-/commit/1189670232e2a1e06d1f479f52ad31fec55e6c06 | 22:24 |
| vkoskiv | I *think* that the options I chose are more correct, and it did make the backlight controller behave more consistently, but evidently something needs to be done to resolve that zero level issue. | 22:25 |
| vkoskiv | I could try setting the original (pre-patch) clock prescales for brightness level 0, since we know that that worked. | 22:26 |
| minute | aha, fixed my rk3588 uboot... the issue was that there was still an old (2017) uboot spl on emmc and that got loaded first and it loaded the other part of uboot from sd | 22:27 |
| minute | so now i ejected the sd and interrupted that old uboot on emmc and erased itself from the cmdline | 22:27 |
| josch | vkoskiv: honestly i rather have less correct code than my display on for several months now without the option to turn it completely off... | 22:40 |
| vkoskiv | Backlight, you mean? | 22:40 |
| josch | yes | 22:41 |
| vkoskiv | I'll do another deep dive this week, then | 22:41 |
| josch | vkoskiv: if the commit that broke it is easy to revert, i can also do that locally | 22:42 |
| minute | vkoskiv: why not just turn it off on brightness level 0? | 22:42 |
| minute | i.e. disable the pwm | 22:43 |
| minute | ok so the up-to-dater rk3588 uboot is actually fine. it just didn't fit together with the ancient 2017 spl that was on the emmc on my module. the good news is that we overwrite that on all the shipping rcores, just mine didn't go through that process | 22:43 |
| minute | now i want those hours back :D | 22:44 |
| hramrach | turn it off requires a gpio connected to the EN pin | 22:51 |
| hramrach | ugh, and I have too old kicad to look at schematics | 22:52 |
| josch | i always have to grab kicad from backports to look at the schematics on debian stable XD | 22:53 |
| vkoskiv | Huh, I just checked out a51dceea594 (before my backlight changes), and I'm still seeing a dim glow at 0% brightness. | 22:54 |
| vkoskiv | Definitely running the right fw, 20230703 is shown in circle+s | 22:54 |
| vkoskiv | Could someone else with kbd v2 try the same? | 22:55 |
| josch | vkoskiv: i can do that, one sec... | 22:56 |
| vkoskiv | circle+x still works on that older version, just doesn't show the status screen on the OLED there | 22:56 |
| vkoskiv | I see the problem. It existed before as well, but the brightness at 0% is much higher than before my changes. | 22:57 |
| vkoskiv | So same issue, just slightly worse now. You know, I could have written these observations down when I was last investigating this :D | 22:58 |
| vkoskiv | Let me see if there is a bit to fully disable the PWM. Though I would think I would've tried that 5 months back | 22:59 |
| josch | vkoskiv: i cannot find my usb keyboard right now (and i cannot search for it much because everybody is asleep) so i have to defer keyboard flashing to tomorrow, sorry | 23:02 |
| vkoskiv | No worries, I'm heading to bed soon here anyway | 23:07 |
| minute | vkoskiv: thanks for looking into this again! | 23:27 |
| vkoskiv | minute: The PWM is actually disabled at brightness 0. I had to check, I didn't express that very clearly in my comment on that issue. | 23:32 |
| minute | vkoskiv: ohh ok | 23:33 |
| vkoskiv | So the problem is, we're driving 0V into pin 6 of the driver chip at zero brightness, and there is still ~6.7V on the output. | 23:33 |
| vkoskiv | And then in my comment I explain that I suspect the intent may be to use that other pin to fully disable the output, but I wasn't sure about that. | 23:34 |
| vkoskiv | I'll see if I can figure out a software workaround, maybe a really really low but non-zero PWM duty cycle might do the trick, since we know it can get dimmer than it currently does with 0% | 23:36 |
| minute | vkoskiv: that sounds like the way to go! i hope it'll work. it was silly of me not to connect EN. iirc there was a gpio shortage | 23:39 |
| vkoskiv | Yeah, there is no GPIO left if I recall | 23:43 |
| vkoskiv | Weird, the older firmware is also driving 0% duty cycle at 0 brightness. So I guess the state of that chip is different with the same output, if the PWM frequency was different before transitioning to 0%? | 23:43 |
| vkoskiv | But I did confirm, before my changes, the PWM frequency is 31kHz, which is out of spec. TI specs say min 100Hz, max 20kHz | 23:45 |
| vkoskiv | Which is kind of a weirdly low limit, 20kHz is in the range of human hearing. My changes put the PWM freq into spec, which put it further down in human hearing range, which made it louder :P | 23:46 |
| - mjw (QUIT: Killed (mercury.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 23:52 | |
| * mark_ -> mjw | 23:52 | |
| + Guest3791 (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 23:52 | |
| vkoskiv | ~3.9kHz pwm frequency with my changes. the PWM +width toggles in exactly 1 microsecond increments | 23:57 |
| vagrantc | much to my surprise, while i mostly only use the right-most space bar... there are cases where i use all three | 23:58 |
| vkoskiv | But we're below T(on) I guess, spec says 'minimum on time in one dimming cycle' is nominally 5us, but ours is in the range of 0-6 | 23:58 |
| josch | vagrantc: this is why my left space-bar is "enter" and between them i have alt and backspace :) | 23:58 |
| vkoskiv | The spec is sadly quite vague about this. I'll have to do some math to figure out a frequency where 5us is as close to 0% duty cycle as possible, I guess. | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!