2024-11-28.log

+ mjw (~mjw@gnu.wildebeest.org)00:33
- mtm (QUIT: Ping timeout: 255 seconds) (~textual@47.202.75.129)01:04
- vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20)01:05
+ mtm (~textual@47.202.75.129)01:06
+ bkeys (~Thunderbi@45.134.140.153)01:06
- chomwitt (QUIT: Ping timeout: 276 seconds) (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1)01:09
- mjw (QUIT: Ping timeout: 260 seconds) (~mjw@gnu.wildebeest.org)01:17
+ mjw (~mjw@gnu.wildebeest.org)01:24
- bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@45.134.140.153)01:51
+ bkeys (~Thunderbi@45.134.140.153)01:51
- aloo_shu (QUIT: Ping timeout: 245 seconds) (~aloo_shu@90.166.98.209)01:55
- bkeys (QUIT: Ping timeout: 252 seconds) (~Thunderbi@45.134.140.153)01:56
+ aloo_shu (~aloo_shu@90.166.98.209)02:05
+ bkeys (~Thunderbi@45.134.140.153)02:20
- paperManu (QUIT: Ping timeout: 276 seconds) (~paperManu@107.159.243.8)02:54
- jn (QUIT: Ping timeout: 252 seconds) (~quassel@user/jn/x-3390946)02:56
+ jn (~quassel@user/jn/x-3390946)02:57
- bkeys (QUIT: Ping timeout: 255 seconds) (~Thunderbi@45.134.140.153)03:06
+ jn_ (~quassel@user/jn/x-3390946)03:09
- jn (QUIT: Ping timeout: 276 seconds) (~quassel@user/jn/x-3390946)03:09
- hairu (QUIT: Ping timeout: 260 seconds) (m-uotkmd@user/hairu)03:33
+ hairu (m-uotkmd@user/hairu)03:46
- jn_ (QUIT: Ping timeout: 276 seconds) (~quassel@user/jn/x-3390946)03:53
+ jn (~quassel@user/jn/x-3390946)03:53
- mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org)03:59
+ jn_ (~quassel@2a0a-a549-fc5e-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de)06:40
- jn_ (QUIT: Changing host) (~quassel@2a0a-a549-fc5e-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de)06:40
+ jn_ (~quassel@user/jn/x-3390946)06:40
- jn (QUIT: Ping timeout: 248 seconds) (~quassel@user/jn/x-3390946)06:41
+ chomwitt (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1)06:52
- jn_ (QUIT: Ping timeout: 260 seconds) (~quassel@user/jn/x-3390946)10:07
+ jn (~quassel@user/jn/x-3390946)10:07
- chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1)11:25
ryukazouminute: I facing some charging problem again, both my battery show 3.7 3.7 and battery percentage is 45%. Does this mean battery monitor is messed up again?12:13
ryukazouDo I need to drain all the battery to recalibrate the battery again?12:16
ch3.7 is not full if its a pocket12:22
ryukazouOnce the battery percentage reach 45% the charging drop to 0.05V12:24
ryukazouI’m not sure what happened12:24
ryukazouOkey new behaviour, when l plug charger into pocket, the oled will suddenly jump to 3.9 4.1 and charging is back to 1.07v, but after 30 second it will drop back to 3.7 3.7 and 0.05v.12:28
chwhich voltage does the charger handshake?12:32
chmaybe try a different charger12:32
+ paperManu (~paperManu@107.159.243.8)12:34
ryukazouch: this is before plug into charger https://i.imgur.com/EJnClH0.jpeg12:37
ryukazouThis is after https://i.imgur.com/tJaphD3.jpeg12:38
ryukazouAfter plug into the charger it become -1.09A12:41
ryukazouI have some typo in previous message, it should be A not V12:42
chstrange that the batteries report different voltages12:42
chmaybe thats why it tries slow charging12:43
ryukazouDoes that mean I have faulty battery?12:43
chdont know, sorry12:46
+ mjw (~mjw@gnu.wildebeest.org)12:47
ryukazouminute: maybe you would have so ideal.12:47
ryukazouIn the mean time I will try drain all the battery to see maybe it will fixed the problem.12:49
minuteryukazou: hmmmm it sounds like charger just stops charging but unclear why12:56
chjosch: re gcc-riscv64-unknown-elf #1086462 - did you see the O bug #1015208 ? (maybe just file the rm bug?)12:56
joschch: ooooh that's how it works! tracker.d.o says "orphaned" in "action needed" but when i looked at d/control, it still has Keith as the maintainer12:58
joschi now see that it's the "O:" bug where the authoritative information comes from12:58
chnever trust any metadata anywhere ;(12:59
chyeah12:59
chas authoritative as it can be i guess12:59
ryukazouminute: battery report 3.4 and 3.2 right not but after plug into charger battery report 3.8 and 4.112:59
ryukazouShould I create a post on forums, maybe other has same experience.13:02
minuteryukazou: so -1.09A is a healthy charging current. but after 30 seconds that stops?13:02
ryukazouminute: yes13:02
minuteryukazou: when it stops does it immediately go to 3.7V or stay at 4.x for a while?13:02
hramrachI got cables that light up when PD is negotiated, and a charger that displays the negotiated parameters, and then I can tell if something is charging or not. Without that it's very difficult. The development snapshots of the sysctl firmware dump PD messages to a serial port so you could see the status there as well if you have that.13:02
- mtm (QUIT: Ping timeout: 245 seconds) (~textual@47.202.75.129)13:02
+ gustav28 (~gustav@c-78-82-54-68.bbcust.telenor.se)13:02
minuteyeah these PD monitoring usb cables with a little display are very useful ryukazou 13:03
ryukazouminute: okey I will get one13:04
+ mtm (~textual@47.202.75.129)13:05
amospallaryukazou: my pocket has a combined voltage of 8.09v/100%, 7.10v/50%, 6.00v/0%.13:20
ryukazouamospalla: mine show 6.47V, the two battery icon show 3.4 and 3.2, battery is on 14%13:25
amospallaryukazou: for me, 6.47V is at 3% .13:26
amospallaThese values are from my pocket being new, after I calibrated it a few times.13:26
amospalla14% -> 6.80V for me.13:27
amospallahttps://paste.debian.net/1337270/13:29
- jn (QUIT: Ping timeout: 252 seconds) (~quassel@user/jn/x-3390946)13:29
+ jn (~quassel@2001-4dd3-c765-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de)13:29
- jn (QUIT: Changing host) (~quassel@2001-4dd3-c765-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de)13:29
+ jn (~quassel@user/jn/x-3390946)13:29
ryukazouamospalla: this is machine on and didn’t plug into charger https://i.imgur.com/RyBCitj.jpeg13:32
ryukazouThis is machine on and plug into charger https://i.imgur.com/kcNQ1D2.jpeg13:32
amospallaHave you tried doing one or two full charge-discharge cycles?13:33
ryukazouDone it once few days age13:34
ryukazouAgo13:35
ryukazouEverything still fine yesterday13:35
amospallaI dunno what may be happening then :/13:35
+ bkeys (~Thunderbi@45.134.140.153)13:36
minuteryukazou: can you make a close up photo of your charger board? there were a (low) number of boards where the charger IC developed an issue (but usually the device doesn't charge/turn on at all then)13:37
ryukazouI’m draining my battery right now to do charge-discharge cycle again, this is the only thing I could try before usb c pd monitor arrive tomorrow.13:38
minuteryukazou: ok!13:38
ryukazouminute: I will give you the photo one I got home, I didn’t bring my screwdriver out today.13:39
- bkeys (QUIT: Ping timeout: 276 seconds) (~Thunderbi@45.134.140.153)14:14
+ chomwitt (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1)14:24
- aard (QUIT: Ping timeout: 260 seconds) (~bwachter@edna-edison.lart.info)14:40
+ aard (~bwachter@217.11.60.132)14:49
- mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org)14:54
* Guest6251 -> mjw15:11
+ andreas-e (~Andreas@2001:861:c4:f2f0::c64)15:31
ryukazouminute: the picture of my charging board https://i.imgur.com/5ROPsk5.jpeg16:09
- kensanata (QUIT: Quit: Ping timeout (120 seconds)) (~alex@user/kensanata)16:44
- gustav28 (QUIT: Quit: Ping timeout (120 seconds)) (~gustav@c-78-82-54-68.bbcust.telenor.se)16:44
+ kensanata (~alex@user/kensanata)16:44
+ gustav28 (~gustav@c-78-82-54-68.bbcust.telenor.se)16:44
+ tadeus_brick (tadeustad@psifactor.pl)16:44
- dodo (QUIT: Quit: dodo) (~dodo@user/dodo)16:45
+ dodo (~dodo@user/dodo)16:45
- TadeusTaD (QUIT: Read error: Connection reset by peer) (tadeustad@user/TadeusTaD)16:46
+ aloo_shu_ (~aloo_shu@85.51.16.96)17:10
minuteryukazou: thanks, that looks ok!17:11
- aloo_shu (QUIT: Ping timeout: 265 seconds) (~aloo_shu@90.166.98.209)17:12
* aloo_shu_ -> aloo_shu17:12
- chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1)17:13
- glu (QUIT: Ping timeout: 260 seconds) (~glu@91.141.35.187.wireless.dyn.drei.com)17:24
ryukazouminute: I re-flash the firmware from August 26th and everything seem become fine, now my pocket reform can charge to 100%17:32
minuteoh wow ok17:34
minutecc ch 17:34
minuteryukazou: thanks for trying that!17:34
chuhuh?17:34
chryukazou: which firmware did you have before?17:35
ryukazouSame firmware from August 26th17:35
ryukazouNot sure what fixed the problem, maybe discharging-charging cycle or re-flash the firmware. I done both.17:36
chdid you try a different charger btw?17:37
+ glu (~glu@91.141.63.216.wireless.dyn.drei.com)17:37
chi guess we need some way of dumping the charging/usb/... state out of sysctl thats not adhoc printf to get to these problems17:38
ryukazouNo, I use the same power bank to charge my pocket reform while going home, it reach 100% when I got home17:38
chah power bank17:41
chwould be very interesting what voltage that gives17:41
ryukazouSo originally my battery stuck at 45%, when I do the discharging the pocket reform actually shutdown when battery reach 1%, so I think my battery gauge is showing correctly.17:42
chyeah i think you had a charging problem, not a gauge problem17:42
ryukazouI using this one https://www.amazon.com/Dell-PW7018LC-Notebook-Power-Bank/dp/B079GXGQPY?th=117:43
ryukazouCould overnight charging potentially be the problem17:44
chSupported output 20 V, 3.25 A - 15 V, 3 A - 9 V, 3 A - 5 V, 3 A17:45
chthat should work at least17:45
ryukazouI do leave my pocket reform plugged for a day yesterday.17:46
- paperManu (QUIT: Quit: WeeChat 4.1.1) (~paperManu@107.159.243.8)17:49
ryukazouch: do you need the information of my power bank from PD monitor?17:53
ryukazouI will post it after the PD monitor arrived17:54
chryukazou: yeah would be interesting, especially if you can get it into the broken state again17:55
ryukazouch: okey I will leave my pocket reform plugged overnight17:56
minutei finally got the production pocket reform display v2s to work... i was pulling my hair over why they wouldn't work in pocket but on the vendor supplied hdmi adapter18:37
minuteturns out there is some kind of eeprom in the display panels that is not programmed on the batch displays, but it was programmed on the proto18:37
minutei was lucky that many months ago i already extracted/converted a long init sequence that sets all those tables and i put that back into the driver and the display sprang to life18:37
minutealso i found a way to identify the display in the driver via a mipi dcs command18:38
minuteso we can have a unified panel driver. now i just need to make sysctl be as flexible and support brightness control creation in the reform2-lpc driver18:41
spewwell done minute!18:43
hramrach\o/18:52
choh nice18:56
chminute: so it can be the same sysctl fw?18:57
minutech: hopefully. i'm thinking of a way for the lpc driver to check which display model is active. for that i have to communicate between the two drivers somehow18:58
minuteor userland has to turn on the backlight control in the lpc driver but that's not that elegant i guess18:58
+ mark_ (~mjw@gnu.wildebeest.org)18:59
minute(for display v2, backlight is in the sysctl fw; for display v1, it is in the panel driver)18:59
minuteright now what already works is that the panel driver's backlight control is only created on demand if the correct panel is recognized18:59
chi wonder what the 'correct' way with the dtbs would be, but i have no idea19:00
chah but you'll have reform2-lpc act as a new backlight driver, if sysctl indicates it knows how to drive the backlight?19:02
minutei think the simplest semi-correct way would be to export a function from the panel driver (sth like mnt_pocket_reform_get_display_panel_version()) and use that in the lpc driver. i wonder if that could be weakly linked so it would be a nop if the panel driver is not there19:02
chthen you dont need to talk between both?19:02
minutech: currently sysctl wouldn't need to know by itself which display is there19:02
minutesysctl-fw that is19:02
chah19:02
minuteonly lpc driver needs to learn this19:02
hramrachtechnically you could turn LPC driver into MFD and bind to a backlight definition in the fw somewhere19:03
minutethe fw shouldn't be different for the 2 panels19:03
hramrachit does not have to be. The backlight function would be enabled only if the DT matches19:04
hramrachhm, but you probably do not want a separate DT either19:04
minutethere wouldn't be a different dts either19:04
minuteexactly19:04
* mjw -> Guest942919:04
- Guest9429 (QUIT: Killed (iridium.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c)19:04
* mark_ -> mjw19:04
minutei want it to be as plug and play as possible19:04
minutei think i'll try a weak function to query the display version from lpc and implement the same function strongly in panel driver19:05
chuboot could provide the DT and patch it on boot (please no)19:05
minutenono :D19:05
+ Guest9429 (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c)19:05
minutei only need those 2 drivers to exchange 1 byte, that's the job19:05
hramracharen't weak functions resolved at build time?19:06
minuteyeah19:07
minutei would want it to be weak because lpc driver also has to work on non-pocket platforms19:07
hramrachand probe order is random so you cannot rely on the function to be available at LPC probe time19:09
hramrachyou could depend on LPC from the display driver but it's out-of-tree19:09
hramrachmaybe a reason to move it into the kernel :)19:10
minutei can use a sysfs attr19:11
hramrachIt still hs the problem that if thre is no module dependency it may not be available19:15
minutenot a big deal... brightness control will only be there if lpc module is loaded, otherwise not. and the panel driver is always compiled into the kernel, so it will always be there before lpc module probes19:15
hramrachit will be there but may not be probed19:16
hramrachBTW a Chinese clone https://www.crowdsupply.com/shenzhen-tianmao-technology/genbook-rk3588 :)19:18
minuteyeah19:18
minutethey even copied some of our blog post headlines in the first update post19:18
minutebut i complained about that to CS and they got changed19:18
minuteah no the headlines are still there19:19
minutehttps://www.crowdsupply.com/shenzhen-tianmao-technology/genbook-rk3588/updates/using-the-genbook-rk358819:19
minute> Linux-y Stuff19:19
minute> GenBook RK3588 Goes to Work19:19
minutesee also https://www.crowdsupply.com/mnt/pocket-reform/updates/pocket-reform-goes-to-work :D19:20
hramrachdid not realize they went so far with the cloning19:22
minutewell, maybe that's another aspect of open source!!19:22
minuteit would be great if we also could just put money on the photocopier19:22
minutethen i could take a little vacation...19:23
hramrachWith SoMs available pretty much anyone can make their own notebook when they have access to the other parts, there is eg. https://www.clockworkpi.com/19:25
minutesure, that's fine19:26
hramrachonce somebody comes up with the idea and implements it it's clear it's doable. No need to directly plagiarize19:26
minuteyeah that just feels weird19:27
minuteok i now have a file /sys/devices/platform/soc@0/32c00000.bus/32e60000.dsi/32e60000.dsi.0/mnt_pocket_reform_panel_version20:07
minutethat contains a 2 for display version 220:07
minutenow i need to investigate how to get a more platform-independent path to this20:07
minute/sys/bus/mipi-dsi/drivers/panel-jdi-lt070me05000/32e60000.dsi.0/mnt_pocket_reform_panel_version points to the same file20:09
minuteah, but drivers are not supposed to _access_ sysfs.20:18
hramrachMaybe make an udev script for this. It gets notifications when things are loaded or unloaded so it could copy that value from one file to another or something. In the kernel getting these notifications from a driver is unlikely to be pretty20:18
minuteyeah20:18
minutewe have reform-hw-setup. it could reload reform2-lpc with a module parameter if this file is found20:19
minuteto clarify, the default brightness without any fw/driver support is 100%20:20
minuteok brightness control via lpc driver works21:12
minuteif loaded with: insmod reform2-lpc.ko enable_backlight_control=121:13
hramrachI think udev rule would be good for this, it does not depend on random service starting, applies whenever the driver is loaded, even if the user reloads it for whatever reason21:15
minutehramrach: do you have some example how i could achieve that? i'm not very experienced with udev rules yet21:16
minutebasically the logic is:21:16
minuteif content of ./devices/platform/soc@0/32c00000.bus/32e60000.dsi/32e60000.dsi.0/mnt_pocket_reform_panel_version == 221:16
minutethen modprobe reform2-lpc enable_backlight_control=121:17
minute(and rmmod reform2-lpc if already loaded)21:17
hramrachif it was not required to reload the lpc it would be easy21:17
hramrachwith driver reload there is quite a bit of potential for endless udev loop21:18
hramrachI haven't seen the code so don't know how problematic it would be to change the value dynamically21:20
hramrachor if it would need to be changed at all, or if whatever LPC does for the backlight control it could do anyway with screen v1, just to no effect21:21
- andreas-e (QUIT: Quit: Leaving) (~Andreas@2001:861:c4:f2f0::c64)21:33
digitalraneis it possible to do that check in tge module?22:05
digitalrane*the22:05
digitalranethen the option could just be used to force enable it22:06
- V (QUIT: Quit: No Ping reply in 180 seconds.) (~v@ircpuzzles/2022/april/winner/V)22:06
digitalranei guess tricky because the module is not bound to the panel, could get messy22:07
+ V (~v@ircpuzzles/2022/april/winner/V)22:08
- GNUmoon (QUIT: Ping timeout: 260 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon)22:08
- shdw (QUIT: Ping timeout: 255 seconds) (~shdw@static.218.156.216.95.clients.your-server.de)22:12
- gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-54-68.bbcust.telenor.se)22:15
minutethe sysctl's pwm brightness must not be used with display v1 because it would pwm that display's enable pin, causing it to turn off. if it was that simple i wouldn't have written here about the required mechanism.22:22
chit all sounds like being between a rock and a hard place22:23
minutenot at all from my pov, i'm happy that things are working finally22:24
minutejust wanted to bounce around some ideas here22:25
minutethe default behaviour is harmless too, full brightness and you just can't turn it down. and enabling the feature will give you the control22:25
chnice22:26
minutei think toggling it via a sysfs attribute at runtime would also be possible22:26
minutei.e. not requiring module reload22:26
+ cobra_ (~cobra@user/Cobra)22:37
- cobra (QUIT: Ping timeout: 265 seconds) (~cobra@user/Cobra)22:37
chi need to dig into the kernel driver model one day and see how its parts are all supposed to work together22:45
+ chomwitt (~chomwitt@2a02:587:7a14:db00:1ac0:4dff:fedb:a3f1)22:45
midfavilaoutta curiosity will the mnt reform mainboard ever be available for purchase in the shop?23:12
joschmidfavila: "ever"? Sure, it was in the past: https://shop.mntre.com/products/mnt-reform-motherboard-2523:14
joschso i guess future revisions will be sold that way as well -- i cannot think of a reason why they woudl not23:15
midfavilayeah i thought the old version had been but i didn't see it last i checked23:16
midfavilawasn't sure if i was misremembering or what23:16
midfavilai'm desperate to get a replacement mainboard so i can use my reform again instead of my old x86 laptop ;w;23:17
joschmidfavila: what continent are you on?23:20
digitalraneminute, i was just offering another perspective23:21
midfavilajosch, NA23:22
joschmidfavila: i'm in germany -- i could've sold you mine but shipping will be steep...23:23
midfavilayeah 23:23
midfavilai'm just waiting on a warranty is all. it's hard to be patient ;w;23:23
joschmidfavila: me too, that's why i have a spare motherboard :D23:29
- nsc (QUIT: Ping timeout: 255 seconds) (~nicolas@i5C74DF9D.versanet.de)23:31
+ nsc (~nicolas@i5C74DF51.versanet.de)23:33
midfavilafair23:33
midfavilai'm hoping to move to europe in a few years so maybe then i'll be able to go to berlin myself instead of getting everything shipped23:33
josch\o/23:36
midfavilai'll finally be able to experience the archaeotech known as "passenger trains"23:38
midfavilacan't wait23:38
josch:D23:40
joschtrains are awesome23:40
midfavilatheyre pretty cool23:40
midfavilai just wish canada had a functional train network23:40
joschme and my reform will be in one for 4 hours tomorrow :)23:41
midfavilait takes 18 hours best case to travel 400km from one major city to another and costs 100$ for the cheapest fare23:41
midfavilaACTION dies23:41
midfavilacan't even bring a bike with you or anything23:42
+ paperManu (~paperManu@107.159.243.8)23:53

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!