| gordon1 | minute: new reform2_lpc current reporting is still inverted, just checking if that's intentional | 00:26 |
|---|---|---|
| minute | gordon1: oh, is that Inverted? | 00:26 |
| gordon1 | minute: https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-class-power judging by this, if you grep by current_now | 00:26 |
| gordon1 | >Negative values are used for discharging batteries, positive values for charging batteries and for USB IBUS current. | 00:27 |
| gordon1 | s/inverted/negated/ | 00:29 |
| minute | gordon1: ok strange how this was wrong all the time, and thanks! also, the previous "negative values are not supported" or something was just a leftover from the initial code, which i trusted without double checking :D | 00:30 |
| gordon1 | if i didn't reuse my old script i would not even notice that :D | 00:31 |
| gordon1 | (i color dicharge current as red and charge as green) | 00:31 |
| minute | gordon1: pushed a fix, does it look better now? | 00:34 |
| gordon1 | still triggers the other bug when i reload the module so i need to reboot every time =/ | 00:34 |
| minute | gordon1: sorry, which other bug? they get out of step or sth? | 00:35 |
| gordon1 | minute: https://source.mnt.re/bugs/bugs/-/issues/44 | 00:36 |
| minute | gordon1: i wonder why i don't have this issue? do you have pocket or classic reform? | 00:36 |
| gordon1 | classic reform | 00:36 |
| minute | ah ok | 00:36 |
| minute | so LPC behaves differently then | 00:36 |
| minute | it's not an issue with rp2040 and rp2350 (next) | 00:36 |
| gordon1 | yep, now it works correctly | 00:38 |
| gordon1 | with negative discharge and positive charge | 00:38 |
| gordon1 | the issue #44 is really annoying btw, i can lend my full help in sorting it out with a bit of guidance in the code | 00:39 |
| minute | thanks! and sorry for that issue! i just made a new branch to test things | 00:41 |
| minute | gordon1: can you try building from the new lpc-v2-tune-delays branch? | 00:41 |
| gordon1 | is it lpc firmware branch? | 00:42 |
| minute | gordon1: https://source.mnt.re/reform/reform-tools/-/merge_requests/144/diffs#328ba5e9fbfb90bdcf78ee91fc2aeabdd10e2b64_266_266 | 00:42 |
| gordon1 | or the module? | 00:42 |
| gordon1 | ah, okay, let me try | 00:42 |
| minute | reform-tools lpc driver | 00:42 |
| minute | the main difference between api 1 and 2 is those delays, and the other difference is the spi speed | 00:42 |
| minute | ah no, the spi speed is always 400khz | 00:43 |
| minute | so the only difference are those delays. they are gigantic for the old version | 00:43 |
| gordon1 | i mean i have a feeling also that there is something in lpc firmware itself, because i can observe similar behaviour on the battery screen on the keyboard | 00:44 |
| gordon1 | with erratic data | 00:44 |
| gordon1 | or the delays are not tuned there as well | 00:44 |
| minute | yeah that's a different protocol though, over UART | 00:44 |
| minute | we'll look into that separately... although i hadn't seen that in a while, is your keyboard fw up to date? | 00:45 |
| gordon1 | sure, but it still feels like some sort of race condition | 00:45 |
| minute | or like, from what date? | 00:45 |
| gordon1 | my keyboard fw is custom build, but i did build it from master few weeks ago | 00:45 |
| minute | ok, that shouldn't be an issue then | 00:45 |
| gordon1 | maybe a month? | 00:45 |
| jfred | hmm drat, I just had my pocket reform's keyboard reset on the latest firmware | 00:45 |
| minute | jfred: yeah, there's still a bug somewhere i guess. | 00:46 |
| jfred | I'm not 100% sure of this but it seems to happen most frequently just after powering on the machine | 00:47 |
| minute | hm! | 00:47 |
| minute | jfred: btw, when did you install that FW? because it could be that it was built against the old sdk still | 00:47 |
| gordon1 | minute: yeah, looks like version from the branch does not exhibit the issue, at least not straight away, but then there would be another issue with lpc version on boot (it is always 0 and therefore v1) | 00:48 |
| gordon1 | but reloading works | 00:48 |
| minute | gordon1: oh, that's an improvement though | 00:48 |
| jfred | minute: yesterday; I'm on commit e67da46782fcb556c4f04a595d6172f7ce80bf49 | 00:48 |
| gordon1 | that is definitely an improvement | 00:48 |
| gordon1 | also all the sysfs interactions are nearly instant now | 00:48 |
| gordon1 | just would love to see that on boot | 00:49 |
| gordon1 | i guess i can hack in some module reload tho | 00:49 |
| gordon1 | for now | 00:49 |
| minute | gordon1: ah yes, that's why i tried to minimize those delays in the first place, because that was way too slow | 00:49 |
| minute | gordon1: does it load this changed driver on boot? | 00:49 |
| minute | gordon1: or the old one maybe? | 00:50 |
| gordon1 | ok, i'll try to reboot and see | 00:50 |
| gordon1 | but it is definitely should be a new one, old one is not saved anywhere | 00:50 |
| gordon1 | ACTION lives dangerously | 00:51 |
| minute | gordon1: you deleted the old one from /lib/modules/... or? | 00:51 |
| minute | i'm installing all kind of battery tools... | 00:51 |
| gordon1 | i do install it using gentoo's emerge, it replaces the file | 00:52 |
| gordon1 | minute: https://0x0.st/KmfO.txt so first two lines - module is loaded by openrc during the boot, last line - i reloaded it myself by hand | 00:53 |
| minute | gnome-power-statistics seems to work | 00:54 |
| minute | gordon1: and how do you know which location that first one is loaded from? | 00:55 |
| minute | gordon1: custom openrc service something? | 00:55 |
| gordon1 | i don't but i'm pretty sure i have only one reform2-lpc.ko on my fs | 00:56 |
| gordon1 | okay i do have two, one for older kernel, i'm 100% sure it is not one that is loaded but i can delete it just in case | 00:57 |
| gordon1 | i also am pretty sure it wont load because it is signed with different certificate | 00:57 |
| gordon1 | but yeah, i'll delete it just in case | 00:57 |
| minute | gordon1: ok, i pushed something new to that branch, which would do the version lookup twice | 00:58 |
| gordon1 | minute: okay, now it did this https://0x0.st/Kmfg.txt | 01:00 |
| minute | gordon1: great, that's what i hoped | 01:00 |
| gordon1 | i did not reload it this time by hand | 01:00 |
| - erle (QUIT: Quit: K-lined) (~erle@user/erle) | 01:01 | |
| gordon1 | is it an lpc firmware quirk? | 01:02 |
| gordon1 | minute: also just for the record, that's how i "fixed" battery screen flicker in the keyboard firmware http://0x0.st/KrOb.diff | 01:04 |
| gordon1 | i bet there is a better way, but it works for now | 01:04 |
| + elb` (~elb@68.133.31.194) | 01:06 | |
| - synnfynn (QUIT: *.net *.split) (~synnfynn@user/synnfynn) | 01:06 | |
| - chrcav (QUIT: *.net *.split) (~chrcav@user/chrcav) | 01:06 | |
| - violet (QUIT: *.net *.split) (~vi@user/meow/violet) | 01:06 | |
| - shdw (QUIT: *.net *.split) (~shdw@static.218.156.216.95.clients.your-server.de) | 01:06 | |
| - pastaonirc (QUIT: *.net *.split) (~pastaonir@user/xha) | 01:06 | |
| - elb (QUIT: *.net *.split) (~elb@68.133.31.194) | 01:06 | |
| - lexik (QUIT: *.net *.split) (~lexik@93.185.97.218) | 01:06 | |
| - kfx (QUIT: *.net *.split) (~kfx@wopr.sciops.net) | 01:06 | |
| - yqshao (QUIT: *.net *.split) (~yqshao@user/yqshao) | 01:06 | |
| + kfx (~kfx@wopr.sciops.net) | 01:06 | |
| + lexik (~lexik@93.185.97.218) | 01:06 | |
| + yqshao (~yqshao@user/yqshao) | 01:07 | |
| + synnfynn (~synnfynn@user/synnfynn) | 01:07 | |
| gordon1 | minute: oh, now i disconnected the charger and erratic values from lpc returned, it reports random battery capacity from time to time | 01:07 |
| + pastaonirc (~pastaonir@user/xha) | 01:07 | |
| + violet (~vi@user/meow/violet) | 01:07 | |
| gordon1 | and while it is actually charging the battery, not idling at 100% it also reports random values from time to time | 01:08 |
| + chrcav (~chrcav@user/chrcav) | 01:08 | |
| minute | gordon1: ok, can you crank up those delays? setting to 10,10,10 for example | 01:08 |
| minute | gordon1: just trying to find an area that works and then we can dig into the specifics later | 01:08 |
| gordon1 | okay | 01:08 |
| minute | we know that 50,50,50 works for you but that's excessive i think | 01:09 |
| minute | (these are milliseconds) | 01:09 |
| minute | we could also put a sysfs attribute in the driver to change these values on the fly... | 01:10 |
| + shdw (~shdw@static.218.156.216.95.clients.your-server.de) | 01:10 | |
| gordon1 | 10,10,10 does occasionally report erratic values but by ~order of magnitue less than 5,5,5 | 01:11 |
| gordon1 | trying 15,15,15 | 01:11 |
| gordon1 | minute: it feels like it is doing it in bursts, i put 50,50,50 and doing cat voltage_now, and it does bunch of 0s then return back to normal reporting | 01:14 |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org) | 01:14 | |
| minute | gordon1: hmmmm, sounds like i'll need to do more debugging on the lpc side. but what happens if you mess with spi speed? that 400000 value, lets say half it? | 01:16 |
| gordon1 | minute: so i did while true ; do cat /sys/class/power_supply/BAT0/voltage_now >>/tmp/data; done | 01:16 |
| gordon1 | https://0x0.st/KmfF.txt | 01:16 |
| gordon1 | isn't it specified in device tree? | 01:16 |
| minute | gordon1: ok, this is also how the stuff i did when developing this, and doing statistics | 01:16 |
| gordon1 | or is it hardcoded in the module also? | 01:16 |
| minute | gordon1: it can be set in the driver (max speed) | 01:17 |
| gordon1 | that's 4000000 not 400000 btw | 01:17 |
| minute | gordon1: but... what if you cat the version? | 01:17 |
| minute | oh right, 4 mhz it was. | 01:17 |
| minute | gordon1: i.e. are all values broken in this way (so, spi is broken), or does it report 0v sporadically | 01:17 |
| minute | (a bit hard to discern if the wrong value is 0) | 01:18 |
| gordon1 | voltage, capacity and current are definitely broken | 01:19 |
| gordon1 | yeah, looks like all of them, i tried to do charge_full_design and charge_empty (which are not supposed to change, arent they?) | 01:20 |
| minute | ah you're in class power supply | 01:21 |
| gordon1 | minute: looks like reducing the speed to 1 MHz and keeping delays as 15,15,15 make it stable both during charge and discharge | 01:21 |
| minute | can you go to /sys/class/spi_master/spi1/spi1.0/ | 01:22 |
| minute | and cat firmware in a loop ? | 01:22 |
| minute | gordon1: interesting @ 1mhz! | 01:22 |
| gordon1 | ah, got 0 zero during few minutes on voltage_now | 01:22 |
| gordon1 | so maybe a bit slower? | 01:22 |
| minute | if you cat the stuff in /sys/class/spi_master/spi1/spi1.0/ it's more direct | 01:23 |
| gordon1 | are we loosing much if we reduce the spi speed? | 01:23 |
| minute | probably not, because it's only a few bytes | 01:23 |
| minute | it was so that originally the speed was too _low_ | 01:23 |
| minute | so making it faster made the communication stable on pocket reform | 01:23 |
| minute | but also those big delays being necessary with lpc (classic) means that the lpc fw code still should be touched again to handle spi in a timer interrupt instead of in the main loop which is "cooperative" and can have varying delays | 01:25 |
| gordon1 | okay, so 500 khz and 15,15,15 is running now with no 0s atm, i'll leave it for 10 mins and after that we can call it good i guess | 01:25 |
| minute | gordon1: ok, can you also check the "firmware" file in /sys/class/spi_master/spi1/spi1.0/ please? | 01:26 |
| gordon1 | do the same while true routine? | 01:26 |
| minute | yes | 01:27 |
| minute | it is easier to detect distortions in that string than in variable data | 01:27 |
| gordon1 | >22222222 30_R1 20250701 | 01:28 |
| gordon1 | ok i assume it is not correct | 01:28 |
| gordon1 | or that | 01:28 |
| gordon1 | >MREF2LPC MMMMMMMM 20250701 | 01:28 |
| minute | the repeating characters show some timing issue yeah | 01:29 |
| gordon1 | (maybe just adding crc on both uart and spi is a way to go!) | 01:29 |
| minute | that's would be possible, but it would be papering over the issue without understanding the root cause | 01:30 |
| gordon1 | okay so reducing speed and increasing delays does improve but not fix things completely | 01:30 |
| gordon1 | but yeah, seems the firmware string thing triggers it quite a bit more often | 01:31 |
| minute | gordon1: does that also happen with the 50,50,50 delays? | 01:31 |
| gordon1 | let me try, 500 khz 50,50,50 now | 01:32 |
| minute | btw i just realized i have a classic reform with rk3588 here as well. we're talking rk3588, right? | 01:32 |
| gordon1 | yep | 01:33 |
| gordon1 | yes, rk3588 | 01:33 |
| gordon1 | yep, it does happen with 500 khz 50,50,50 | 01:33 |
| gordon1 | > 30_R1 20250701 | 01:33 |
| minute | ok, booting my reform up to check if it's still ok there | 01:34 |
| gordon1 | it does not happen when the charger is idling tho | 01:34 |
| gordon1 | so it's either because lpc is busy with something during this times, or EMI? (i hope it's not EMI) | 01:35 |
| minute | what do you mean by idling? | 01:35 |
| minute | i think "busy" is realistic | 01:35 |
| gordon1 | i mean when it is connected to charger and charged at 100% | 01:35 |
| minute | gotcha | 01:35 |
| minute | btw do you have the OLED battery screen open in parallel or no? | 01:36 |
| gordon1 | no, i do not | 01:36 |
| minute | ok good | 01:36 |
| gordon1 | and i also have the fix so i won't be able to observe, but i can confirm that oled screen flicker happened during the same conditions, i.e. charging or discharging the battery | 01:36 |
| gordon1 | but not when it was fully charged | 01:36 |
| minute | oh interesting | 01:37 |
| gordon1 | i am pretty sure i remember that correctly | 01:37 |
| minute | > [ 89.828989] reform2_lpc spi1.0: version: 0 () | 01:37 |
| minute | (on my reform) | 01:37 |
| gordon1 | nice, reproducible then | 01:37 |
| minute | yes, that's good | 01:38 |
| minute | but with this default (slow) mode, the firmware string is stable for me | 01:38 |
| gordon1 | i got only two incorrect firmware version strings so far during last 5 mins, so it is definitely almost does not happend with 50,50,50 delays | 01:38 |
| minute | (stock driver) | 01:38 |
| minute | ok, i didn't let it run for 5 minutes yet... | 01:39 |
| gordon1 | running it as while true ; do cat /sys/class/spi_master/spi1/spi1.0/firmware | grep -Ev '^MREF2LPC 30_R1 20250701$'; done | 01:39 |
| minute | nice | 01:40 |
| gordon1 | do you want me to put back original keyboard fw for debugging or are we fine with the existing clues so far? | 01:40 |
| minute | ok, reloading the driver indeed gives > [ 468.973200] reform2_lpc spi1.0: version: 20250701 (20250701) | 01:40 |
| minute | but the fw string looks stable for me hm. | 01:40 |
| minute | MREF2LPC MMMMMMMM 1 | 01:41 |
| minute | 33333333 2LPCMMMM MMMMMMMM | 01:41 |
| minute | ok, spurious issues confirmed | 01:41 |
| gordon1 | here we go | 01:41 |
| minute | gordon1: entirely reproducible! | 01:41 |
| minute | gordon1: thanks for your help tracking this down, i'm gonna write these details in my todo list and i see lpc fw fixes in my future... | 01:42 |
| gordon1 | nice | 01:42 |
| gordon1 | i'll find a suitable delay and speed for me when it is not too annoying meanwhile | 01:43 |
| gordon1 | tbh 500 khz @ 15,15,15 didn't look too bad | 01:43 |
| minute | oh, diagnostic tool https://github.com/mojyack/pocket-reform-lpc-diag | 01:45 |
| gordon1 | oof, nice | 01:45 |
| minute | anyway, gotta watch my tv show and hit the hay, more tomorrow | 01:49 |
| gordon1 | gn | 01:49 |
| - n_to (QUIT: Quit: quitidiquit) (~n_to@2a03:4000:6:3662:24b1:57ff:fec6:76c1) | 02:00 | |
| - Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 02:00 | |
| + n_to (~n_to@2a03:4000:6:3662:24b1:57ff:fec6:76c1) | 02:00 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 02:04 | |
| spew | which show? | 02:04 |
| - paperManu_ (QUIT: Ping timeout: 250 seconds) (~paperManu@64.58.44.160) | 02:39 | |
| + reform24775 (~keko@2600:1700:6630:b6e0::43) | 02:40 | |
| * reform24775 -> korgakoth | 02:40 | |
| - korgakoth (QUIT: Changing host) (~keko@2600:1700:6630:b6e0::43) | 02:41 | |
| + korgakoth (~keko@user/korgakoth) | 02:41 | |
| - korgakoth (QUIT: Client Quit) (~keko@user/korgakoth) | 02:42 | |
| minute | spew: alias | 02:46 |
| - paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@64.58.44.160) | 03:07 | |
| - timide (QUIT: Ping timeout: 244 seconds) (~timide@user/timide) | 03:09 | |
| - dok (QUIT: Ping timeout: 265 seconds) (~dok@2a01:e0a:12:d860:3696:e5bc:8411:9bc5) | 03:10 | |
| + dok (~dok@2a01:e0a:12:d860:3696:e5bc:8411:9bc5) | 03:53 | |
| + timide (~timide@user/timide) | 03:58 | |
| - vkoskiv (QUIT: Ping timeout: 245 seconds) (~vkoskiv@109-204-186-7.cust.valoonet.fi) | 04:18 | |
| + vkoskiv (~vkoskiv@109-204-186-7.cust.valoonet.fi) | 04:30 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 05:46 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 06:10 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 06:22 | |
| - chomwitt (QUIT: Ping timeout: 265 seconds) (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 06:32 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 06:34 | |
| + RandyK (~RandyK@user/randyk) | 06:34 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 06:59 | |
| - hramrach (QUIT: Ping timeout: 252 seconds) (~hramrach@nat2.prg.suse.com) | 07:36 | |
| + hramrach (~hramrach@nat2.prg.suse.com) | 07:49 | |
| - synnfynn (QUIT: Quit: gn) (~synnfynn@user/synnfynn) | 08:02 | |
| - wickedshell (QUIT: Ping timeout: 248 seconds) (~wickedshe@c-174-56-77-83.hsd1.nm.comcast.net) | 09:14 | |
| + mjw (~mjw@178.230.122.127) | 10:42 | |
| - mjw (QUIT: Remote host closed the connection) (~mjw@178.230.122.127) | 10:51 | |
| + mjw (~mjw@178.230.122.127) | 10:51 | |
| - mjw (QUIT: Read error: Connection reset by peer) (~mjw@178.230.122.127) | 10:58 | |
| + mjw (~mjw@89.205.167.88) | 10:59 | |
| + thelounge2708 (~thelounge@148.168.138.88.rev.sfr.net) | 11:05 | |
| - mjw (QUIT: Ping timeout: 260 seconds) (~mjw@89.205.167.88) | 11:11 | |
| + mjw (~mjw@89.205.167.88) | 11:13 | |
| - mjw (QUIT: Remote host closed the connection) (~mjw@89.205.167.88) | 11:20 | |
| + mjw (~mjw@89.205.167.88) | 11:21 | |
| - mjw (QUIT: Ping timeout: 256 seconds) (~mjw@89.205.167.88) | 11:54 | |
| + MartiniMoe (~moe@p200300eecf1ede469b2ac4c17de74025.dip0.t-ipconnect.de) | 11:56 | |
| - chomwitt (QUIT: Ping timeout: 255 seconds) (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 12:07 | |
| josch | minute: what are your thoughts about a new reform-tools release? Are the things you recently pushed into main ready? Do you want to finish something else (like the lpc dkms fixes) before releasing? Current list of changes since the last release a month ago: https://source.mnt.re/reform/reform-tools/-/merge_requests/145https://source.mnt.re/reform/reform-tools/-/merge_requests/145 | 12:11 |
| josch | ah my middle mouse key switch is borked again | 12:11 |
| josch | https://source.mnt.re/reform/reform-tools/-/merge_requests/145 | 12:11 |
| - buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org) | 12:25 | |
| + buckket (~buckket@vps.buckket.org) | 12:25 | |
| + robin_ (~robin@user/terpri) | 12:25 | |
| - robin (QUIT: Read error: Connection reset by peer) (~robin@user/terpri) | 12:25 | |
| + kensanata1 (~alex@user/kensanata) | 12:25 | |
| - irimi1 (QUIT: Read error: Connection reset by peer) (47a15995ae@2a03:6000:1812:100::12aa) | 12:25 | |
| - [tj] (QUIT: Read error: Connection reset by peer) (sid609767@id-609767.lymington.irccloud.com) | 12:25 | |
| - theesm1 (QUIT: Read error: Connection reset by peer) (2cbdf4b38a@2a03:6000:1812:100::11c8) | 12:25 | |
| - jfred (QUIT: Read error: Connection reset by peer) (c9b58a1025@libera/sponsor/jfred) | 12:25 | |
| - burley (QUIT: Read error: Connection reset by peer) (fc32d5b940@user/burley) | 12:25 | |
| - spew (QUIT: Read error: Connection reset by peer) (e9e8135d87@user/spew) | 12:25 | |
| - romi (QUIT: Ping timeout: 245 seconds) (bd30729973@user/romi) | 12:25 | |
| - Zaba (QUIT: Ping timeout: 245 seconds) (80b9b4b35e@2a03:6000:1812:100::116) | 12:25 | |
| - jbmorley_ (QUIT: Read error: Connection reset by peer) (sid21953@id-21953.tinside.irccloud.com) | 12:25 | |
| - polymetis (QUIT: Read error: Connection reset by peer) (sid320610@id-320610.hampstead.irccloud.com) | 12:25 | |
| - kuno (QUIT: Read error: Connection reset by peer) (sid299442@id-299442.helmsley.irccloud.com) | 12:25 | |
| - austriancoder (QUIT: Read error: Connection reset by peer) (sid152545@id-152545.hampstead.irccloud.com) | 12:25 | |
| + burley (fc32d5b940@user/burley) | 12:25 | |
| + irimi1 (47a15995ae@2a03:6000:1812:100::12aa) | 12:25 | |
| + spew_ (e9e8135d87@user/spew) | 12:25 | |
| + jfred (c9b58a1025@libera/sponsor/jfred) | 12:25 | |
| + jbmorley_ (sid21953@id-21953.tinside.irccloud.com) | 12:25 | |
| + theesm1 (2cbdf4b38a@2a03:6000:1812:100::11c8) | 12:25 | |
| + [tj] (sid609767@id-609767.lymington.irccloud.com) | 12:25 | |
| - voxpelli (QUIT: Read error: Connection reset by peer) (sid31634@id-31634.tinside.irccloud.com) | 12:25 | |
| - noam (QUIT: Ping timeout: 248 seconds) (81879d1ffa@2a03:6000:1812:100::dfc) | 12:25 | |
| - chorc (QUIT: Quit: ZNC 1.9.1 - https://znc.in) (~chorc@user/chorc) | 12:25 | |
| + austriancoder (sid152545@id-152545.hampstead.irccloud.com) | 12:25 | |
| + kuno (sid299442@id-299442.helmsley.irccloud.com) | 12:25 | |
| + polymetis (sid320610@id-320610.hampstead.irccloud.com) | 12:25 | |
| - arjan (QUIT: Ping timeout: 245 seconds) (~arjan@user/arjan) | 12:25 | |
| - amatecha (QUIT: Ping timeout: 245 seconds) (sid10006@user/amatecha) | 12:25 | |
| - ajr (QUIT: Read error: Connection reset by peer) (sid609314@user/ajr) | 12:25 | |
| + voxpelli (sid31634@id-31634.tinside.irccloud.com) | 12:25 | |
| + chorc (~chorc@user/chorc) | 12:25 | |
| * spew_ -> spew | 12:25 | |
| + ajr (sid609314@user/ajr) | 12:25 | |
| + amatecha (sid10006@user/amatecha) | 12:25 | |
| + eschaton_ (eschaton@2600:3c01::f03c:91ff:fefd:5d92) | 12:25 | |
| - dodo (QUIT: Remote host closed the connection) (~dodo@user/dodo) | 12:25 | |
| - eschaton (QUIT: Read error: Connection reset by peer) (eschaton@2600:3c01::f03c:91ff:fefd:5d92) | 12:25 | |
| - kensanata (QUIT: Read error: Connection reset by peer) (~alex@user/kensanata) | 12:25 | |
| * kensanata1 -> kensanata | 12:25 | |
| + noam (81879d1ffa@2a03:6000:1812:100::dfc) | 12:25 | |
| + Zaba (80b9b4b35e@2a03:6000:1812:100::116) | 12:25 | |
| + romi (bd30729973@user/romi) | 12:25 | |
| + dodo (~dodo@user/dodo) | 12:25 | |
| - potash (QUIT: Read error: Connection reset by peer) (~potash@user/foghorn) | 12:26 | |
| + potash (~potash@user/foghorn) | 12:27 | |
| + paperManu (~paperManu@64.58.44.160) | 12:32 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 12:42 | |
| - MartiniMoe (QUIT: Ping timeout: 250 seconds) (~moe@p200300eecf1ede469b2ac4c17de74025.dip0.t-ipconnect.de) | 12:51 | |
| + gustav2 (~gustav@c-78-82-55-162.bbcust.telenor.se) | 13:02 | |
| - dodo (QUIT: Quit: dodo) (~dodo@user/dodo) | 13:27 | |
| + dodo (~dodo@user/dodo) | 13:27 | |
| + potash9 (~potash@user/foghorn) | 13:27 | |
| - potash (QUIT: Read error: Connection reset by peer) (~potash@user/foghorn) | 13:27 | |
| * potash9 -> potash | 13:27 | |
| - chomwitt (QUIT: Ping timeout: 260 seconds) (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 13:37 | |
| elb` | I think my pocket reform (3588) doesn't always identify the attached nvme in time | 14:04 |
| elb` | sometimes when the kernel boots it gets stuck waiting for cryptsetup to find and set up the encrypted volume | 14:05 |
| elb` | I can't tell exactly why that is, but my suspicion is that it isn't identifying that the underlying encrypted volume is available | 14:06 |
| elb` | power cycling usually fixes it immediately, somtimes I have to power cycle a time or two | 14:06 |
| elb` | then it will boot normally for a half dozen or more boots in a row | 14:06 |
| elb` | (or however many; I'm not suggesting that once it boots it's more reliable, but rather that it's far less frequent than 50% of the time that this happens) | 14:07 |
| elb` | I also sometimes get the blank screen at boot that some people mention on community.mnt.re, but I haven't yet had a uart hooked up when this happens | 14:08 |
| josch | elb`: yes, i have the same issue. If i remember correctly the problem is not with nvme (do you have uart output showing that this is a problem with nvme?) but a problem with emmc which sometimes does not init properly. I think I talked about this with m.inute who confirmed that this problem exists. | 14:09 |
| elb` | I haven't captured the uart output, no, but in the former case the emmc inits _well enough_ that it has booted the kernel and loaded the initramfs | 14:09 |
| elb` | in the latter case the emmc not initing seems very plausible | 14:09 |
| josch | ah indeed, your problem is a different one than the one i have | 14:10 |
| elb` | (to be very clear I experience two things at boot which may be unrelated: one is a blank screen and no visible progress, the other is boot to initrd and an inability to find the nvme for decryption) | 14:10 |
| josch | elb`: which nvme do you have? there were problems with some in the past | 14:10 |
| elb` | I have a wd blue | 14:10 |
| josch | okay, same here | 14:10 |
| elb` | I previously had a sk hynix that froze the entire system, frequently | 14:11 |
| elb` | on moderately heavy I/O loads it would almost always freeze after just a minute or three | 14:11 |
| josch | ah i remember the sk hynix problem but i forgot it was you who had it | 14:11 |
| josch | it broke with reform-migrate | 14:11 |
| elb` | correct | 14:11 |
| elb` | but then I did some just general I/O testing and found that it was repeatable | 14:11 |
| elb` | (it might be helpful to have a structured way to store various notes about hardware/bugs/etc.; the community forum and this IRC both seem ... difficult) | 14:12 |
| josch | elb`: i think the problem with blank screen and no visible progress might be the emmc issue. The other where you see the initrd failing to find nvme is a new one that i haven't heard of yet. | 14:12 |
| * elb` -> elb | 14:12 | |
| elb | it doesn't happen always, but it happens often enough that I might be able to catch a boot log on uart console | 14:12 |
| josch | elb: yes, the "structured" way is to file this as a bug report. Forum threads are indeed suboptimal as a bugtracker :) | 14:12 |
| sigrid | I have wd blue sn570 1tb on a classic reform with rk3588 and had absolutely no issues so far. however, I run alpine and not the official reform image | 14:13 |
| + synnfynn (~synnfynn@user/synnfynn) | 14:13 | |
| elb | oh I don't mean bug reports per se, but like notes about what hardware works and what doesn't and what the various evidence is | 14:13 |
| josch | elb: bug tracker: https://source.mnt.re/bugs/bugs/-/issues/ | 14:13 |
| josch | now afk as i have to reboot to investigate https://community.mnt.re/t/suddenly-blank-screen-at-boot/3808/12 ttyl | 14:14 |
| elb | oh, I created a source.mnt.re account but it's never been approved (as far as I know) | 14:17 |
| minute | elb: i wonder if some nvmes take too long to initialize?! | 14:17 |
| minute | elb: what's your source username? i can approve it | 14:17 |
| elb | elb | 14:18 |
| minute | sigrid: i also use WD exclusively basically | 14:18 |
| minute | and that's also what we sell at the moment | 14:18 |
| elb | minute: that's possible, but if that's the case I woudl expect one of the retries to eventually find it | 14:18 |
| elb | I guess if it takes too long in the sense that the kernel gives up trying to identify it | 14:18 |
| minute | elb: ah sorry, that's right, if it needs power cycles... | 14:18 |
| minute | btw i wrote down an internal "pocket pain list" today | 14:19 |
| [tj] | minute: feet! | 14:19 |
| minute | [tj]: ahaha you mean rubber feet? i recently attached some on mine! but good, i forgot about that, will go in | 14:19 |
| [tj] | I have cork feet | 14:19 |
| elb | I put rubber feet on my pocket but they fell off immediately | 14:19 |
| [tj] | having to put a foot on the logo suchs | 14:19 |
| minute | elb: approved! | 14:20 |
| [tj] | https://usercontent.irccloud-cdn.com/file/xyhs0VEn/1758802824.JPG | 14:20 |
| elb | the adhesive they came with stuck fine to the reform but came free of the silicone :-\ | 14:20 |
| elb | minute: thank you | 14:20 |
| [tj] | I used cork furniture stops | 14:20 |
| elb | I'll try to get some more information about the boot sequence where it doesn't find the nvme and file a report | 14:21 |
| minute | ok, we have some feet that stick well... we should just always include them | 14:21 |
| gordon1 | speaking of nvmes, i found this thread https://community.mnt.re/t/adding-a-delay-to-prevent-unsafe-shutdowns-for-nvme-drive/1207/2 | 14:23 |
| - BAndiT1983_ (QUIT: Quit: No Ping reply in 180 seconds.) (~quassel@46.101.193.235) | 14:23 | |
| minute | gordon1: ah that's outdated since we use SPI + poweroff handler in the lpc driver | 14:23 |
| minute | gordon1: back then it was a bash script that was executed before the kernel shut down everything completely | 14:24 |
| gordon1 | bizzare thing, on rk3588 reform i have no issues with unsafe shutdown incrementing with poweroff/reboot, | 14:24 |
| gordon1 | yes | 14:24 |
| gordon1 | it has been fixed | 14:24 |
| gordon1 | however on my rock5b the issue is there, and i don't think it does anything differently, except triggering PMIC for a reboot instead of LPC | 14:25 |
| elb | oh man I installed firmware-linux-free and it slowed down boot _so much_ | 14:25 |
| + BAndiT1983 (~quassel@46.101.193.235) | 14:25 | |
| gordon1 | wondered if there is more to that than just doing poweroff with SPI driver | 14:25 |
| elb | I guess I can't be sure that was the difference as I also generally updated, but it seems likely based on the messages that are printing during boot | 14:25 |
| gordon1 | maybe some devicetree change with power sequencing or something? | 14:26 |
| gordon1 | as far as i know firefly and rock5b use the same PMIC... | 14:27 |
| gordon1 | maybe somebody did some investigation there before it was fixed using lpc driver? | 14:30 |
| minute | elb: interesting, i would like to know which firmware slows down the boot | 14:36 |
| + paperManu_ (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 14:39 | |
| - Twodisbetter (QUIT: Ping timeout: 252 seconds) (2cc0e4ea1c@irc.cheogram.com) | 15:09 | |
| + adam78 (~adam@2001:ee0:4b72:3d40:955:12f1:a4bd:1e7f) | 15:10 | |
| - adam78 (QUIT: Client Quit) (~adam@2001:ee0:4b72:3d40:955:12f1:a4bd:1e7f) | 15:10 | |
| + adamtc (~adamtc@2001:ee0:4b72:3d40:955:12f1:a4bd:1e7f) | 15:11 | |
| elb | minute: I rebooted and it Got Better (?) | 15:12 |
| elb | it wasn't just boot, it was ... everything | 15:12 |
| elb | but dmesg said nothing interesting that I could see | 15:12 |
| elb | what had confused me was it was still loading devices and reporting their initialization (including firmware update) long into the boot process, which is normally mostly done very early | 15:13 |
| + erle (~erle@user/erle) | 15:34 | |
| - thelounge2708 (QUIT: Quit: The Lounge - https://thelounge.chat) (~thelounge@148.168.138.88.rev.sfr.net) | 16:22 | |
| + chomwitt (~chomwitt@2a02:85f:9a74:ce00:1ac0:4dff:fedb:a3f1) | 16:39 | |
| + Guest23 (~Guest23@184.22.120.202) | 16:57 | |
| - Guest23 (QUIT: Client Quit) (~Guest23@184.22.120.202) | 16:57 | |
| - synnfynn (QUIT: Quit: leaving..) (~synnfynn@user/synnfynn) | 17:15 | |
| + wickedshell (~wickedshe@c-174-56-77-83.hsd1.nm.comcast.net) | 17:27 | |
| gordon1 | okay, am i the only one who finds the experience to zoom in/out in online maps using trackball a completely miserable experience? | 17:32 |
| gordon1 | i'm pretty sure it's because the trackball reports both vertical and horizontal scroll wheels at the same time | 17:32 |
| gordon1 | so i wonder if maybe it will be a good idea to have one of the scroll buttons to have only vertical scroll axis reported instead of both | 17:33 |
| gordon1 | gonna try it | 17:33 |
| [tj] | my ploopy trackball only does one axis for scroll | 17:34 |
| gordon1 | oh, i'm talking about mnt reform classic one precisely, forgot to mention | 17:34 |
| josch | gordon1: yes, this works. I have the bottom left button for up/down and the bottom right button for up/down/left/right | 17:37 |
| josch | because usually i want up/down only | 17:37 |
| josch | and usually i use my left thumb to press the button and my right thumb to move the wheel | 17:37 |
| josch | only sometimes i want both pan and wheel and then i have to swap my thumbs | 17:37 |
| gordon1 | would be nice to be able to configure it | 17:38 |
| josch | gordon1: though i'm afraid my patch will likely not help you because you probably have a trackball v2, no? | 17:38 |
| gordon1 | i mean without reflashing | 17:38 |
| gordon1 | need to think about it | 17:38 |
| erle | i sometimes miss a scrollwheel. even knowing that every click of the wheel is only a single mouse event. | 17:38 |
| gordon1 | i probably do | 17:38 |
| erle | it requires less concentration. | 17:38 |
| gordon1 | it's fine, i think i'll manage to patch v2 fw | 17:39 |
| gordon1 | just need to figure out how to flash it | 17:39 |
| erle | oh, regarding patches, i should make one to fix that weird keyboard thingy i encountered. | 17:39 |
| erle | i mean i have one, but i should submit it | 17:39 |
| gordon1 | also need to address the usb suspend issue | 17:41 |
| gordon1 | but that's when i receive my spare trackball... | 17:41 |
| - adamtc (QUIT: Quit: Client closed) (~adamtc@2001:ee0:4b72:3d40:955:12f1:a4bd:1e7f) | 17:49 | |
| wickedshell | [tj]: interesting, my ploopy adept does 2 axis, sometimes it's great, sometimes it's terrible :P | 17:59 |
| erle | josch since you wrote that you tried hibernating before and it did not come up again, any useful notes? namely, things you tried that didn't work out to fix that? | 17:59 |
| josch | erle: with i.mx8mq? | 18:00 |
| erle | josch yes | 18:38 |
| + mjw (~mjw@195.23.216.83) | 18:47 | |
| + AnimaInvicta (~AnimaInvi@88-120-179-216.subs.proxad.net) | 19:11 | |
| josch | erle: the last time i tried hibernate on i.mx8mq must've been more than 4 years ago -- no notes :) | 19:22 |
| + colinsane (~colinunin@97-113-128-166.tukw.qwest.net) | 19:27 | |
| minute | funky https://apps.gnome.org/Bustle/ | 19:41 |
| minute | erle: if you press both wheel buttons, it enters sticky wheel mode. maybe that can help for longer reading sessions. | 19:42 |
| erle | minute thank you, good to know! | 19:44 |
| josch | minute: do you have thoughts about a new reform-tools release (see my message to you above in the log). Do you want me to hold off the next release for more changes to the lpc dkms module? | 19:45 |
| minute | josch: i think the changes i pushed to main are good to release early, the timing stuff needs more debugging and shouldn't delay the release | 19:48 |
| josch | thank you! | 19:48 |
| erle | minute i really like the “press two non-adjacent modifiers of the same type to lock them” pattern. neo2 has this for caps lock, so you never do it by accident. | 19:53 |
| erle | (press both shift keys for CAPS) | 19:54 |
| josch | well... "never" ;) | 20:02 |
| + synnfynn (~synnfynn@user/synnfynn) | 20:16 | |
| elb | ahh, I remember the olden days when shift+shift was normally input language switch :-) | 20:29 |
| elb | (maybe it still is in some locales?) | 20:29 |
| erle | elb i think something like shift+space or ctrl+space is used nowadays | 21:04 |
| + ephase (~ephase@2a01:e0a:2a:5300:8af3:6216:8fce:7058) | 21:05 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 21:14 | |
| + MartiniMoe (~moe@p200300eecf1ede4678904cc09fd8011c.dip0.t-ipconnect.de) | 21:15 | |
| elb | that's used in some east asian input systems for changing between options _within that input system_ | 21:26 |
| elb | shift+shift was for changing languages that don't have an internal variety of methods | 21:26 |
| elb | like latin versus cyrillic | 21:27 |
| erle | elb i just checked, ibus uses super + space to switch between input methods | 21:52 |
| erle | > You have signed up successfully. However, we could not sign you in because your account is awaiting approval from your GitLab administrator. | 21:54 |
| erle | guess i should do the patch business during business hours | 21:54 |
| ch | at least once ;) | 21:55 |
| erle | it makes sense. better defeat spam at the source! | 21:55 |
| - ephase (QUIT: Quit: WeeChat 4.6.3) (~ephase@2a01:e0a:2a:5300:8af3:6216:8fce:7058) | 21:56 | |
| + helgoman_ (~helgoman@178.38.121.62) | 22:00 | |
| josch | i just hope that m.inute does not receive as many registration attempts as i see in my own forgejo instance (80-90 per day) | 22:01 |
| - gustav2 (QUIT: Quit: Quit) (~gustav@c-78-82-55-162.bbcust.telenor.se) | 22:15 | |
| erle | josch since you are using neo2, does this patch (made using git-format-patch) look like the proper bugfix for my layers issue? https://mister-muffin.de/p/s9QT.diff | 22:23 |
| erle | i think you can pipe this into git-am to apply it to your current branch | 22:25 |
| josch | erle: yes, i can try this out tomorrow | 22:25 |
| erle | thx | 22:25 |
| erle | do you have keyboard variant 3, whatever that is? | 22:25 |
| erle | (i suspect it is the one with fewer keys) | 22:25 |
| josch | erle: i have 2, 3 and 4 | 22:26 |
| josch | i'm typing this on v2 and i'm thinking about putting my v2 keyboard into the rk3588 classic reform instead of the full-rgb v4 keyboard | 22:26 |
| josch | havig two keys between the space keys is just very nice | 22:27 |
| erle | i guess you could make a similar patch for the newer ones then. i only have 1 internal and 2 external ones with 12 keys in the bottom row, so i guess they are all the same hardware (but evidently different firmware, one is still not on neo since i acquired it from someone else). | 22:27 |
| josch | erle: yes, but the firmware is entirely different | 22:27 |
| erle | oh yeah, that was why i actually bought a second external one from someone else | 22:27 |
| josch | i actually don't know why the two got merged into one | 22:28 |
| josch | i think the new key is a 2U key and there would actually be space for two keys even in v4 | 22:28 |
| josch | and even electronically, there is just some non-connected traces now | 22:28 |
| erle | the external keyboard with the 12 keys in the bottom is the best external keyboard i have ever had. but granted, usually my external keyboards were “i spent 10 euros in a dutch krimskrams store” stuff. | 22:28 |
| josch | if i ever learn how to use kicad and how to send a project to jlcpcb or similar i think i want to mod the v4 keyboard to have two keys between the spacebar keys... | 22:29 |
| erle | i think my next quest is to figure out where the sysrq key is mapped | 22:30 |
| erle | so i can do better kernel debugging | 22:30 |
| josch | erle: you mean print screen? | 22:31 |
| erle | josch Druck/SAbf i think | 22:32 |
| josch | erle: in the old firmware (the one you have) the key is KEY_PRINT_SCREEN. In the v4 firmware it's KEY_SYSRQ | 22:33 |
| josch | erle: by default there is no such key. I have it below the circle key. | 22:33 |
| erle | well i have backspace there | 22:38 |
| josch | backspace is to the left of the right spacebar as i've found i press it *a lot* :D | 22:39 |
| josch | here is my diff: https://paste.debian.net/hidden/81814aa3/ | 22:39 |
| josch | the v4 layout is not final -- i'm not happy with ctrl being in the bottom left | 22:39 |
| erle | i press ctrl with my left thumb on the reform kbd. helps that C-x C-v C-c are all a single left-hand action | 22:41 |
| josch | yes, that's what i do with the v2 keyboard but i cannot do that anymore if i move ctrl all the way to the left | 22:42 |
| josch | i have to find a new home for my alt key which used to be between the space keys... | 22:43 |
| - paperManu_ (QUIT: Ping timeout: 264 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 22:46 | |
| - bail (PART: !!unknown attribute: msg!!) (585c46d254@2a03:6000:1812:100::f91) | 22:47 | |
| erle | one thing where the horizontal and vertical scrolling are very annoying to be on the same input device is playing 0ad | 22:53 |
| erle | you gave to set the renderer to OpenGL (ARB) and disable shadows but then it works nicely | 22:53 |
| erle | the issue is that mousewheel up/down is zoom and mousewheel left/right is rotation | 22:53 |
| erle | (unless i am mistaken) | 22:53 |
| - potash (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn) | 23:02 | |
| + potash (~potash@user/foghorn) | 23:07 | |
| minute | i think you should be able to use input-remapper to work around these things | 23:22 |
| erle | josch i made an issue so i don't end up spamming this channel anymore with testing/discussion https://source.mnt.re/reform/reform/-/issues/23 | 23:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
