2025-11-28.log

+ paperManu (~paperManu@142.169.16.42)00:05
- Guest86 (QUIT: Quit: Client closed) (~Guest86@104.228.148.232)00:28
+ Guest29 (~Guest29@104.228.148.232)00:29
chohi00:40
- paperManu (QUIT: Ping timeout: 240 seconds) (~paperManu@142.169.16.42)00:50
chjosch: do you also have the 'not working' usb-c thingie with you?00:56
+ paperManu (~paperManu@107.159.15.124)01:17
- Guest29 (QUIT: Quit: Client closed) (~Guest29@104.228.148.232)01:23
+ paperManu_ (~paperManu@107.159.15.124)01:33
- paperManu (QUIT: Ping timeout: 240 seconds) (~paperManu@107.159.15.124)01:43
- paperManu_ (QUIT: Ping timeout: 240 seconds) (~paperManu@107.159.15.124)02:47
- xktr (QUIT: Ping timeout: 250 seconds) (~xktr@user/xktr)04:03
+ xktr (~xktr@user/xktr)04:04
+ elb (~elb@2600:4041:6682:ea00:c6a7:e77f:1e0b:37ab)04:06
erle> having the keys light up when you press them and trail your keypresses slightly is lovely05:00
erlepassword input just got a lot funnier :305:00
- erle (QUIT: Quit: K-lined) (~erle@user/erle)05:01
jfredyeah that's why I don't use keyboard lighting features like that haha05:01
jfredI do sometimes like some of the more flashy ones like ripples from each keypress. You could work it out if you look closely enough, but if someone's watching me type my password that closely I likely have problems anyway05:03
- lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah)06:13
+ lidstah (~lidstah@gateway/tor-sasl/lidstah)06:16
- timide (QUIT: K-Lined) (~timide@user/timide)06:24
joschch: what do you think? ;)07:10
josch(i'll have it in my backpack but as you say it only "not working")07:16
joschwait, it actually really works -- even after first attaching the hub and only then the cable to the AC07:50
joschit just takes over a minute for the power status to change to "charging" after pluggin AC into the hub...07:50
joschokay, this behaviour has nothing to do with the pocket -- same thing with the classic reform with ubb-pd to dc adapter in the middle07:57
joschgood i checked :D07:57
- _justin_kelly71 (QUIT: Quit: The Lounge - https://thelounge.chat) (~justinkel@user/justin-kelly/x-6011154)10:39
+ _justin_kelly71 (~justinkel@user/justin-kelly/x-6011154)10:42
- _justin_kelly71 (QUIT: Client Quit) (~justinkel@user/justin-kelly/x-6011154)10:45
+ _justin_kelly71 (~justinkel@user/justin-kelly/x-6011154)10:47
grimmwareminute: thinking about this rtos-for-sysctl thing, do you have any opinions on what you would *like* to use were it supported on the new sysctl? Reason I ask is because obviously it would be preferable that we don't further fragment the software stack, and it would be great to be able to get on and experiment with the target platform on the rp2040 if it's actually available10:55
grimmwareif rtos is potentially the way to go for the low power states, on the pocket we're essentially going to be blocked by making this decision for the next which kinda sucks because the sysctl power draw is likely to become the biggest power issue given it will limit how useful suspend/hibernate are10:56
grimmwarethe runtime of this thing with bigger batteries is pretty incredible, I'd be thrilled if it didn't drain so much whilst turned off10:57
joschtrue but then my11:22
joschbut then my thinkpay from work also does not surive two days in suspend so is proprietary hardware really that much better?11:23
- rodion_goritskov (QUIT: Ping timeout: 246 seconds) (~rodion_go@mail.goritskov.com)11:25
+ rodion_goritskov (~rodion_go@mail.goritskov.com)11:28
grimmwareMy GPD P2 max lasts FOREVER11:31
+ gidzit (~gidzit@gidzit.org)11:37
Zabaif your thinkpad doesn’t survive for two days then it’s probably not suspending fully correctly11:38
- _justin_kelly71 (QUIT: Quit: The Lounge - https://thelounge.chat) (~justinkel@user/justin-kelly/x-6011154)11:51
amospallach: how can I retrieve sysctl logs from my pocket?12:19
chamospalla: see the command in here: https://community.mnt.re/t/battery-meter-only-goes-to-about-40/3937/6 12:20
amospallaI just powered it on, put it to charge (sysctl shows aggretated Voltage and negative current), a few minutes later I disconnected the charger, and sysctl shows 0.000A and 0.00V.12:20
grimmwareminute: did you do much profiling for other gpu frequencies with the power saving stuff?12:25
amospallach: https://file.amospalla.es/misc/pocket_reform_sysctl_battery.txt, showing zero values, then attached the charger and showed real values, then disconnected the charger and went back to zero values.12:27
+ paperManu (~paperManu@107.159.15.124)12:28
chok, thats a real proble12:28
chm12:28
amospallaJust being curious, what does it mean?12:28
chhmm12:28
amospallaWell, don't want to take you time, please ask if I can do anything else.12:31
chi'll have to read the mp2650 datasheet how this can happen12:34
chwhat the messages mean is that, on battery, the mp2650 doesn't report any data12:34
chbut clearly some communication with it works, otherwise charging wouldnt work12:34
chamospalla: can you plug in a dumb power source that supplies 5V and no usb-pd? like an old usb "battery bank"?12:35
chand see what the logs say then12:35
amospallaI should be able to, let me find the cables.12:35
grimmwareare there any convenience ways of monitoring the gpu usage (preferably like top/htop/btop)12:36
amospallach: dumb power 1A 5V source: https://file.amospalla.es/misc/pocket_reform_sysctl_battery2.txt12:45
chso weird. i'm missing the 'CHG' flag in the log (= chg_en=1), yet the mp2650 reports chg_stat=212:50
- paperManu (QUIT: Ping timeout: 264 seconds) (~paperManu@107.159.15.124)12:52
chcan you reset the sysctl?12:53
amospallach: mechanically right?12:55
cheither that or with reform-mcu-tool12:55
amospallach: done12:56
chanything different?12:56
amospallaI'll repeat everything I did before, let me check.12:57
amospallach: it is behaving correctly now.13:01
chok. i guess there's a need to write the mp2650 config regularly or something13:02
+ gustav25 (~gustav@c-78-82-53-204.bbcust.telenor.se)13:02
amospallathank you for checking :). Ask if need anything.13:03
minutegrimmware: RTOS does _not_ make PM easier.13:25
minutegrimmware: if anything, RTOS will make things harder as there are more abstractions to deal with. it _could_ make things more stable and extensible, esp. if more features are added.13:26
minutegrimmware: zephyr doesn't have any support for PM functions on rp2350 implemented yet, but one of their devs was immediately interested in implementing it when i told my usecase13:27
minutei experimented more with deep sleep functions yesterday and it was still crashing but i think i just need to vendor the pico-extras/sleep code and remove things like stdio_flush in there13:28
minutegood that we already vendored usb stdio13:28
minutethe tricky thing is to disable all background tasks that are implemented with timer IRQs, for example tinyusb doesn't have an API for _stopping_ it13:29
minuteso we might need to implement that13:29
minuteanyway, hope that today i'll be able to get to rp2350 deep sleep + reboot on wake (this should also mostly work on rp2040)13:30
minutegrimmware: perfetto afaik @ gpu, or gallium hud13:34
grimmwareah fair13:48
- mjw (QUIT: Ping timeout: 240 seconds) (~mjw@gnu.wildebeest.org)14:01
* Guest8453 -> mjw14:13
grimmwareHmm, it appears that other people on the internet have been using /sys/class/devfreq/fb000000.gpu/load to monitor gpu load but it does not appear to be in the reform kernel14:28
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:43
+ behinus (~behinus@2a02:3030:665:5160:f905:d96d:e9d8:cdb7)14:58
- behinus (PART: !!unknown attribute: msg!!) (~behinus@2a02:3030:665:5160:f905:d96d:e9d8:cdb7)14:58
+ jordi__ (~jordi@79.117.156.55)15:05
- pomel0 (QUIT: Ping timeout: 240 seconds) (~pomel0@user/pomel0)15:06
grimmwareI really don't know enough to say for sure but the stuff around devfreq in this patch stack looks kind of promising despite being for an unrelated SoC15:07
grimmwarehttps://lkml.org/lkml/2025/9/5/76815:07
- kxtells (QUIT: Ping timeout: 246 seconds) (~jordi@user/kxtells)15:07
- swivel (QUIT: Ping timeout: 260 seconds) (~swivel@shells.gnugeneration.com)15:48
+ swivel (~swivel@shells.gnugeneration.com)15:49
+ pomel0 (~pomel0@user/pomel0)16:29
+ wielaard (~mjw@gnu.wildebeest.org)16:55
+ FirefoxDeHuk (~FirefoxDe@user/FirefoxDeHuk)17:04
grimmwareugh, I'm still getting the problem that one of my cells is higher voltage than the other so I can't charge beyond a certain % (12 in this case) because it causes the charger to drop to just supply the current draw and not charge the batteries17:04
grimmwareI don't have the stuff with me to balance the cells :(17:04
grimmwaremind you I haven't done a firmware update in a while17:13
grimmwareI should do that thing17:13
- f_ (QUIT: K-Lined) (16abab341f@postmarketOS/funderscore)17:14
+ f_ (16abab341f@postmarketOS/funderscore)17:14
- FirefoxDeHuk (QUIT: Quit: Client closed) (~FirefoxDe@user/FirefoxDeHuk)17:15
minutepocket keyboard standalone usb-c adapter is here https://mastodon.social/@mntmn/11562822027874349017:22
grimmwarethat's very cute :)17:24
grimmwareyeah firmware update hasn't changed anything.17:25
joschwhat could i try on pocket reform to get the speaker working with kernel 6.15 (trixie). i have run reform-hw-setup from reform-tools git and pavucontrol shows that output devices is set to speakers and the volume bar is moving, indicating that something should be played17:28
grimmwarejosch: alsamixer will likely expose some mute control that's hidden in pavucontrol17:32
joschbut alsamixer should be set up fine by reform-hw-setup17:32
grimmwarecheck it!17:32
grimmwarecos there's a whole bunch of shit under the other sound cards that pavucontrol doesn't even touch17:33
joschnothing is muted (nothing is set to MM)17:34
joschyes but those controls are initialized (or at least should be) by reform-hw-setup. anyways, nothing is muted17:35
minutegrimmware: this is with charger board v2 right?17:46
grimmwareminute: yeah17:47
grimmwareI had no problem charging it up to full overnight17:48
grimmwarethey are new batteries which I didn't check the voltage on beforehand but they've already been for a very long full discharge about twice17:48
grimmwarewell, not totally full but down to like 3%17:49
+ jogu (~jogu@user/jogu)17:53
+ FirefoxDeHuk (~FirefoxDe@user/FirefoxDeHuk)18:05
- gidzit (QUIT: Ping timeout: 246 seconds) (~gidzit@gidzit.org)18:11
- FirefoxDeHuk (QUIT: Quit: Client closed) (~FirefoxDe@user/FirefoxDeHuk)18:12
- mjw (QUIT: Killed (copper.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)18:35
* wielaard -> mjw18:35
+ Guest420 (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)18:35
- buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org)19:03
+ buckket (~buckket@vps.buckket.org)19:04
- jogu (QUIT: Remote host closed the connection) (~jogu@user/jogu)20:38
+ jogu (~jogu@user/jogu)20:39
+ synnfynn (~synnfynn@user/synnfynn)20:52
joguStill been poking away at trying to pin down the USB instability on my V4 keyboard. Does anybody know how the crystal used in the RP2040 based schematics was chosen? Curious if that could be a contributing factor since it deviates from the reference design21:02
joguI've been reading through a GitHub issue thread for another keyboard that sounds like a very similar issue: https://github.com/foostan/crkbd/issues/265#issuecomment-247407228621:06
- sterni (QUIT: ) (~quassel@user/sterni)21:10
+ sterni (~quassel@user/sterni)21:10
joguI'm testing out shielding with copper tape around the MCU, crystal, and USB traces near the top of the PCB to see if that helps as well 21:11
joguMy initial impression is that the shielding is helping, but it's so sporadic it's hard to be certain21:12
joguAnd even with it, I'm still seeing CRC errors/bus resets in the logs21:13
joguOver on the rpi forums I see an engineer from there mentioned that the ESR max of 50R as being important https://forums.raspberrypi.com/viewtopic.php?t=31493521:18
joguCurrent crystal on the reform keyboard is 100R, not sure how much that matters21:20
minutehave next motherboard current draw (from 1 of 2 bat packs) down to 2mA in rp2350 sleep mode now21:28
minutejogu: yes, the crystal is not ideal, do you want to try to replace it with the recommended one?21:29
joguminute: yeah, I was thinking of trying the reference crystal + capacitors to see if there's any difference but I'm not able to do that kind of soldering haha21:39
minutejogu: ah ok. we recently did such experiments here on some pocket keyboards where the oscillator wouldn't start. it's quite rare though. most of the time the rp2040 has issues. but it's not known that this happens on the "big" keyboard21:40
joguminute: Interesting, I did see that discussion on usb instability on the pocket reform keyboard disconnecting and it sounds identical to what I'm seeing21:50
joguIn the past, it used to be more of a "hard" disconnect that required I reset the MCU / re-plug the keyboard but lately it's manifesting as the last key repeating but it recovers without reseting 21:51
minutejogu: do you have rk3588, and since when, and do you have up to date uboot?22:14
- gustav25 (QUIT: Quit: Quit) (~gustav@c-78-82-53-204.bbcust.telenor.se)22:15
+ leony (~leony@2a02:8109:f083:bb00::5ef2)22:34
- paperManu (QUIT: Ping timeout: 240 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)22:44
- pomel0 (QUIT: Ping timeout: 244 seconds) (~pomel0@user/pomel0)22:58
+ pomel0 (~pomel0@user/pomel0)22:58
- pomel0 (QUIT: Ping timeout: 240 seconds) (~pomel0@user/pomel0)23:05
+ pomel0 (~pomel0@user/pomel0)23:06
- pomel0 (QUIT: Ping timeout: 240 seconds) (~pomel0@user/pomel0)23:10
+ pomel0 (~pomel0@user/pomel0)23:10
+ paperManu (~paperManu@107.159.15.124)23:24
+ paperManu_ (~paperManu@107.159.15.124)23:24
- pomel0 (QUIT: Ping timeout: 240 seconds) (~pomel0@user/pomel0)23:45
+ pomel0 (~pomel0@user/pomel0)23:50
- pomel0 (QUIT: Ping timeout: 256 seconds) (~pomel0@user/pomel0)23:54
- paperManu (QUIT: Quit: WeeChat 4.7.0) (~paperManu@107.159.15.124)23:57

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