2025-12-02.log

kfxminute: wilco, thanks00:10
- pomel0 (QUIT: Ping timeout: 246 seconds) (~pomel0@user/pomel0)00:11
+ pomel0 (~pomel0@user/pomel0)00:12
+ BAndiT1983 (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000)00:14
+ anuejn (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000)00:14
+ se6astian (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000)00:14
- chomwitt (QUIT: Ping timeout: 245 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)00:52
- cobra (QUIT: Ping timeout: 265 seconds) (~cobra@user/Cobra)01:30
+ cobra (~cobra@user/Cobra)01:32
- plomlompom (QUIT: Remote host closed the connection) (~plom@139.59.138.73)01:45
+ plomlompom (~plom@139.59.138.73)01:46
- AnimaInvicta (PART: !!unknown attribute: msg!!) (~AnimaInvi@88-120-179-216.subs.proxad.net)01:54
- mjw (QUIT: Ping timeout: 240 seconds) (~mjw@gnu.wildebeest.org)02:28
- kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:05
+ kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:08
- kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:08
+ kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:08
- paperManu_ (QUIT: Ping timeout: 265 seconds) (~paperManu@107.159.15.124)03:10
- paperManu (QUIT: Ping timeout: 240 seconds) (~paperManu@107.159.15.124)04:02
+ spew (~spew@user/spew)04:22
- spew (QUIT: Quit: nyaa~) (~spew@user/spew)05:24
+ chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)07:28
- RandyK (QUIT: Ping timeout: 272 seconds) (~RandyK@user/randyk)08:34
+ RandyK (~RandyK@user/randyk)08:35
joschcwebber: i feel you -- my pocket reform is currently better than my classic reform (which has been my daily driver for more than 3 years) in every way *except* for the charging board (but that will be fixed soon) and the size of keyboard and screen08:50
joschit is even better than rk3588 classic reform in the sense that the pocket reform motherboard and the wifi/bt extension board allow me to add an lte card as well as wifi+bt without ressorting to any hacks08:51
joschi currently have in front of me an adapter from mini-pcie to M.2 for motherboard 3.0. Since the usb lanes of the mini-pcie on motherboard 3.0 are not connected, maybe i can solder some tiny wires to D+ and D- of the M.2 adapter... we'll see :)08:53
- Gooberpatrol66 (QUIT: Ping timeout: 246 seconds) (~Gooberpat@user/gooberpatrol66)09:18
- chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)09:19
grimmwarejosch: I'm really glad I finally put the bigger batteries in, feeling like I can walk away from this thing when it's unplugged is really nice09:48
joschi just hope the batteries stay healthy XD09:50
joschi'm still suspicious :)09:50
joschbut yes, i sometimes forget the thing in my backpack and when i get it out it's still on09:51
+ xandyv (~xandyv@user/xandyv)09:57
+ xandyv2 (~xandyv@user/xandyv)10:01
- xandyv (QUIT: Ping timeout: 256 seconds) (~xandyv@user/xandyv)10:03
* xandyv2 -> xandyv10:03
chjosch: wtf. is that fixed on newer hw?10:07
chjosch: (rtc)10:07
gordon1https://fahrplan.events.ccc.de/congress/2025/fahrplan/ yey10:07
grimmwarejosch: yeah, I made a deal with myself that if I installed them I'd also get on and order a transparent acrylic bottom plate so I could keep an eye on them heh10:08
grimmwareI also just think it's a nice nod back to the classic reform10:09
joschch: i know no details10:10
joschyeah, i think that would be nice :)10:11
grimmwareheh, bluetooth range is actually pretty good10:13
minutejosch: ch: we don't use that rtc10:31
minuteall reform series have an nxp rtc chip10:32
gordon1CONFIG_RTC_DRV_RK808=y oops 10:33
gordon1time to recompile my kernel10:33
grimmwareminute: the old "thanks but I brought my own" strategy ;)10:35
grimmwareworks with airplane food and apparently RTCs10:35
minutegrimmware: has saved me a bunch of times :D10:36
josch:)10:38
joschminute: do you know whether LS1028A also checks for u-boot on a different offset? With the offset it is currently using, it is clashing with the GUID partition table10:39
minutejosch: no idea, would need to check the TRM. 10:39
joschi'm thinking whether we still need MBR partition table support in mkimage.sh to create working images for LS1028A10:40
joschcurrently the MR kicks out MBR support but maybe it has to stay if LS1028A does not check at a later offset10:41
+ Gooberpatrol66 (~Gooberpat@user/gooberpatrol66)10:47
grimmwareminute: out of interest do you have numbers for how many of each SoM are in circulation?11:16
minutejosch: meh :/11:18
minutegrimmware: not from the top of my head11:18
- Ar|stote|is (QUIT: Ping timeout: 246 seconds) (~linx@149.210.3.83)11:18
minutethe most circulated one is probably zynq-7020 :D11:20
joschminute: i'm also meh but i'd put in the work to keep support for MBR (together with the resulting complexity) in mkimage.sh unless you object and tell me that you want to throw out LS1028A images11:20
minutebut there's no reform image for it :D11:20
minutejosch: ok, i'm still on board with supporting the module11:21
joschdo you happen to have a link where i could check myself whether maybe another offset would also be valid?11:21
joschno idea where i'd look but maybe i could look it up myself11:21
+ Ar|stote|is (~linx@149.210.3.83)11:23
+ mjw (~mjw@gnu.wildebeest.org)11:38
- XYZ_ (QUIT: Read error: Connection reset by peer) (~XYZ@89-24-56-199.nat.epc.tmcz.cz)11:57
+ XYZ (~XYZ@89-24-56-199.nat.epc.tmcz.cz)12:02
- XYZ (QUIT: Remote host closed the connection) (~XYZ@89-24-56-199.nat.epc.tmcz.cz)12:12
+ paperManu (~paperManu@107.159.15.124)12:29
- xandyv (QUIT: Quit: WeeChat 4.6.3) (~xandyv@user/xandyv)12:34
+ paperManu_ (~paperManu@107.159.15.124)12:47
+ mark_ (~mjw@gnu.wildebeest.org)12:54
- mjw (QUIT: Remote host closed the connection) (~mjw@gnu.wildebeest.org)12:54
+ gustav25 (~gustav@c-78-82-54-128.bbcust.telenor.se)13:02
minutejosch: sorry, i'm still on the move, could you ping me again in the afternoon?13:32
minutei can get you the docs13:33
- pomel0 (QUIT: Ping timeout: 260 seconds) (~pomel0@user/pomel0)13:35
+ pomel0 (~pomel0@user/pomel0)13:35
joschminute: i can ping you tonight -- i have $dayjob until then :)13:38
joschThank you!13:38
- paperManu (QUIT: Ping timeout: 256 seconds) (~paperManu@107.159.15.124)14:04
- paperManu_ (QUIT: Ping timeout: 252 seconds) (~paperManu@107.159.15.124)14:09
- potash1 (QUIT: Ping timeout: 264 seconds) (~potash@user/foghorn)14:11
- jordi (QUIT: Ping timeout: 265 seconds) (~jordi@193.pool85-56-250.dynamic.orange.es)14:13
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:48
- mark_ (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org)14:50
* Guest5854 -> mjw14:59
- pomel0 (QUIT: Read error: Connection reset by peer) (~pomel0@user/pomel0)15:00
+ pomel0 (~pomel0@user/pomel0)15:00
- pomel0 (QUIT: Ping timeout: 260 seconds) (~pomel0@user/pomel0)15:10
+ pomel0 (~pomel0@user/pomel0)15:10
gordon1minute: is my understanding correct that side port covers are powder coated mild(?) steel so countersinking screw holes will potentially expose it to rust?15:20
- nsc (PART: !!unknown attribute: msg!!) (~nicolas@i5C74DF7B.versanet.de)15:24
+ chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)15:28
minutegordon1: yes that's what they are, but they could also be stainless steel and it's my understanding that powder coating can be done after countersinking. the companies where we got them made just didn't have that capability15:45
minutenowadays i think jlc can do this too, for example15:45
gordon1sure, i'm just checking if i can conuntersink it myself or do i need to make brand new covers for that15:46
minutegordon1: ahh got it16:26
+ wielaard (~mjw@gnu.wildebeest.org)16:26
minutegordon1: i'm sure you could treat the steel somehow after the fact, but i'm not a steel expert (yet)16:27
joschminute: ping for link _o/16:27
josch:)16:27
gordon1it's gonna be hard because it is likely will be damaged by the screw, powder coat is an option but then i need to strip it all and re-coat it16:28
minutejosch: do you have an nxp account?16:28
minutejosch: if so, pdf is here https://www.nxp.com/webapp/Download?colCode=LS1028ARM16:28
gordon1minute: tbh i'm just looking for an excuse to give a try of my idea to make FR4 covers with black soldermask and embed a wifi antenna into it16:30
joschminute: I wanted to avoid creating an NXP account because it means agreeing to a whole lot of things which I'm actually not following. For example I'm distributing their binaries without permission.16:30
minutejosch: ok, hang on16:31
joschaccording to the license you are only allowed to distribute their binary blobs together with a product but reform.debian.net does not sell any nxp products :)16:31
gordon1just need to find some patch antenna design i can steal16:31
minutehttps://www.scribd.com/document/806870300/LS1028ARM16:31
joschi tried to contact them several times about that and i was told that this issue was being looked into but then no replies to my follow ups16:32
joschbecause of this restriction, their binaries cannot even go into the non-free section of the Debian archive. They are more than non-free because they are unredistributable. :(16:32
joschminute: thank you!! <316:32
minutejosch: what i know is: service processor loads RCW from sd card16:32
minutejosch: oh interesting16:34
minutejosch: > If the Service processor detects that the RCW loading failed, it initiates RCW loading from a second16:34
minuteRCW source. If boot fails from the RCW loaded at offset 0 or 0x1000 (in case of SD), the boot ROM16:34
minutesearches for RCW at 8 MB (8 MB + 0x1000 for SD) offset on the device.16:34
minutejosch: looks like you can try offset 8MB + 0x100016:34
joschaha!16:34
joschminute: thank you so much yet again! :D16:34
joschls1028a might survive for just a bit longer <316:35
minutejosch: and with the RCW you can do all kinds of hacks to load stuff16:36
minuteif this pdf was not behind a login wall, this would be one of the best documented cortex-a72 processors made...16:36
joschincredible to think that somebody wrote up more than 4k pages of docs... wow16:45
minuteprobably a team of people ;)16:47
sigriddon't they _always_ write gigantic docs for those, just not making them all public?16:47
minuteyeah. what a shame. they should put them in libraries :D16:47
minutethe register table stuff is of course generate, hopefully from the HDL of the design...16:48
minutes/generate/generated16:48
- mrbcmorris (QUIT: Ping timeout: 250 seconds) (~mrbcmorri@1513413-static.lxtnkya3.metronetinc.net)17:05
+ mrbcmorris (~mrbcmorri@1513413-static.lxtnkya3.metronetinc.net)17:05
+ XYZ (~XYZ@89-24-56-199.nat.epc.tmcz.cz)17:37
- pomel0 (QUIT: Ping timeout: 260 seconds) (~pomel0@user/pomel0)17:52
+ pomel0 (~pomel0@user/pomel0)17:52
- mrbcmorris (QUIT: Ping timeout: 265 seconds) (~mrbcmorri@1513413-static.lxtnkya3.metronetinc.net)17:58
amospallaThe MNT Pocket Reform Charger Board V2 is an upgrade over the stock one right?17:59
+ mrbcmorris (~mrbcmorri@1513413-static.lxtnkya3.metronetinc.net)18:01
sigridgot the hardware crypto on rk3588 running, though not sure how useful it is. collabora tree doesn't have that patch - had to fix it for the new kernel version: https://git.sr.ht/~ft/aports/commit/ce308b3dffd8ab0cccb60b292ffdb35c0ad2e05f18:14
gordon1sigrid: can you do cryptsetup benchmark -c aes-xts just to compare?18:27
sigridgordon1: how to I run it with/without hardware offloading?18:29
gordon1hmm, it just used hw crypto when the module is loaded and algos listed in /proc/crypto for me18:30
gordon1maybe there is a flag, let me check18:30
gordon1wow it's amazing how there is pretty much 0 information about what to put in -c for cryptsetup18:41
sigridguess I can downgrade and make a run without any specific cipher again18:42
gordon1sigrid: i guess you can do bench with and without hwcrypto module loaded, or of that is not compiled as module, just bench from hwcrypt kernel suffice, i guess we all have the same rk3588, aren't we?18:42
sigridif you can do a run on your rk3588 that'd be faster18:43
gordon1you mean with swcrypto?18:44
sigridyeah18:44
gordon1sure18:44
kfxminute: ping re: port cover templates18:45
gordon1https://0x0.st/KyXl.txt kernel 6.17.1018:46
gordon1last time i tried hwcrypto on rk3588 the speed was abysmal, around 50 MB/s18:47
gordon1but i probably did something wrong18:47
gordon1so want to check18:47
sigridit's worse than yours: http://okturing.com/src/16294/body18:48
gordon1yeah oh well, not gonna turn that on then18:48
sigridbut I also have no idea why. I will downgrade and do a run18:48
gordon1i wonder why tho18:48
gordon1like there's clearly something that we're missing18:49
- Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.3.83)18:49
gordon1but at least it is almost 10x faster than my try! :D18:49
gordon1i have aes-xts-neonbs in my /proc/crypto as aes-xts algo18:50
gordon1*xts-aes-neonbs18:51
+ Ar|stote|is (~linx@149.210.3.83)18:52
sigridok it made 0 difference, so I assume something is severely wrong with my setup18:53
sigridbut this also probably mean hw crypto offloader wasn't used at all18:53
chdo you have aes in /proc/cpuinfo?18:53
sigridyes18:53
gordon1there is CONFIG_CRYPTO_BENCHMARK but i have no idea how to use it yet18:56
sigridgordon1: can you post your "grep 'driver.* aes' /proc/crypto"?19:04
gordon1driver       : aes-generic19:06
gordon1driver       : aes-arm6419:06
gordon1driver       : aes-ce19:06
gordon1i wonder if -ce is CRYPTO_DEV_ROCKCHIP (not CRYPTO_DEV_ROCKCHIP2)19:06
gordon1i should probably disable that19:06
sigridwell, I have no aes-arm64. that would explain the slowness19:07
sigridbut it doesn't explain how to fix it :(19:07
gordon1i can give you my kernel config, you can check the diff19:07
sigridyes plz19:08
gordon1but if i would guess, it's zcat /proc/config.gz | grep CONFIG_CRYPTO_AES_ARM6419:08
gordon1sigrid: http://0x0.st/Ky8p.txt19:09
chhttps://www.kernelconfig.io/config_crypto_aes_arm64_ce thats probably arm-ce19:09
sigridoh lol yeah I missed one of them19:10
sigrid# CONFIG_CRYPTO_AES_ARM64 is not set19:10
gordon1i do not think that i have some specific patches for device tree, so you can assume those are the same 19:10
sigridok, thanks. I will rebuild with that one enabled and CRYPTO_DEV_ROCKCHIP disabled19:10
kfxthe dsi in port moved on reform motherboard 3.0 and my ribbon cable is too short :(19:53
- mjw (QUIT: Killed (uranium.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)20:13
* wielaard -> mjw20:13
+ Guest4413 (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)20:13
sigridI think a longer cable is available, kfx. I had that issue with older mb and minute mentioned that20:13
sigridalthough I also managed to fit it very tightly and it works20:14
gordon1and version without dsi reqruires new rcore board, doesn't it?20:16
kfxmine is 50mm, mouser has a 100mm but I need to find a magnifying glass to count pins20:22
gordon1i meant without the cable20:22
rwa_gordon1: you use a self compiled kernel with custom options?20:30
rwa_the kernel shipped with debian stable doesn't show 'aes-arm64' in the output20:31
kfxI think I'm counting 33 pins here20:34
minutegordon1: no, you can also remove a resistor from the MB20:43
minutegordon1: to use DSI with RCORE < R220:43
minutegordon1: highly recommended vs the hdmi adapter20:45
minutekfx: wait waaat20:45
kfxminute: between the 'first release' motherboard and motherboard 3.0, different position for the dsi in port20:46
kfxI can order a longer one but I am old and my eyes are broken so it took me a while to count pins20:48
minutekfx: ohh first release. i think it already changed in the second release to accomodate LS1028A or so20:48
minutekfx: oh i'm sorry to hear... but you found the cable?20:48
minutekfx: online i mean20:48
minutekfx: about the port covers, checking now20:49
kfxminute: I think this cable will work?  https://www.mouser.com/ProductDetail/Wurth-Elektronik/68763310000220:50
- XYZ (QUIT: Read error: No route to host) (~XYZ@89-24-56-199.nat.epc.tmcz.cz)20:50
minutekfx: the specs are correct but i think it should be opposite side contacts20:53
minutekfx: can you check the cable that you have? it should have opposing side contacts?20:53
+ XYZ (~XYZ@89-24-56-199.nat.epc.tmcz.cz)20:53
kfxit does, thanks for catching that20:54
minute(a same side contact cable will flip the contacts. which i sometimes use. but not everywhere)20:54
kfxand 33 pins is correct?20:54
minutekfx: yes.20:54
minutekfx: 0.5mm pitch as well20:54
kfxhooray I got it right20:55
kfxeyesight confirmed correctable20:55
minuteyeah20:56
minutekfx: https://source.mnt.re/reform/reform/-/commit/0c9df046d077c3d0cbca6b3400fac804b6b0bb2c20:56
minutekfx: or maybe more useful https://source.mnt.re/reform/reform/-/tree/master/reform2-case-ports?ref_type=heads20:57
minuteSTEP viewer for gitlab would be cool...20:57
kfxminute: thanks!21:07
- XYZ (QUIT: Ping timeout: 255 seconds) (~XYZ@89-24-56-199.nat.epc.tmcz.cz)21:07
+ XYZ_ (~XYZ@78-80-115-228.customers.tmcz.cz)21:07
- pomel0 (QUIT: Read error: Connection reset by peer) (~pomel0@user/pomel0)21:08
+ pomel0 (~pomel0@user/pomel0)21:09
elbran my pocket reform down to 1% and it rode there for maybe half an hour without brownouts with the charger v2 board and reform-power-daemon21:11
+ vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50)21:11
grimmwareniiiiice21:11
grimmwarefirefox is the deterministic test, without reform-power-daemon it'll still shit the bed round there21:12
elbfirefox was running, but not doing much21:12
grimmwareI think it'll survive with less conservative settings on the GPU 21:12
elbI went to codeberg, advent of code, and mastodon21:12
grimmwareyeah even loading a page would cause it to spike GPU and cause a brownout for me21:13
elbinteresting, mine was never that reliable21:13
elbsometimes it would brown out for no reason that I could tell, other times it'd run for days with no issues21:13
grimmwareI also have a couple of i2c accelerometers which will be drawing a smidge more power21:14
grimmwareand I think my batteries were a bit fucked too21:15
gordon1minute: oh sorry, i completely confused the two21:15
vagrantcso, trying to update to the 6.17 kernel from reform.debian.net for trixie-backports ... but reform2-lpc-dkms fails to build with 1.76-2+reform20250622... i see sid has 1.82-2, should i locally backport that? newer versions of reform2-lpc-dkms do not seem to be available on reform.debian.net or debian's trixie-backports21:15
elbmy batteries are still young and full of energy21:15
gordon1rwa_: correct, i use custom kernel config and few patches on top21:15
vagrantcor maybe i just gget lazy and pull in the version from sid and see what breaks :)21:16
rwa_vagrantc: afaik reorm-tools 1.81 should be in trixie-backports21:16
vagrantcrwa_: doesn't appear to be21:17
vagrantc(either of them, as far as i can read)21:17
rwa_vagrantc: anyway i'm using reform-tools and reform2-lpc-dkms from sid atop of debian stable21:18
vagrantcthat appears to work for now ...21:19
vagrantcwell, i didn't bother with reform-tools update ... maybe i will regret it :)21:19
rwa_hm...maybe my memory serves me wrong21:19
vagrantcalso, curious what are the implications of running an lpc firmware with mismatched batteries (e.g. current firmware has 1.8Ah hard-coded, running 2.0Ah cells, new firmware has 2.2Ah hard-coded)21:22
vagrantcat one point i noticed my cells when fully charged were noting that they were at 2.0Ah, and just stayed at 100% until it dropped below 1.8Ah ... but lately they are only charging up to 1.8Ah ...21:23
vagrantcgiven how fiddly the process for updating the firmware is ... trying to figure weather i should patch my firmware to match, or just watch the mismatch :)21:24
vagrantcoh, it seems to be trickle-charging past the 1.8Ah mark ... all the way up to 1.81Ah, and was at 1.809Ah just a bit ago21:29
vagrantcpresuming that's not just a margin of error ...21:29
vagrantci guess the question is ... does it "learn" that the cells are degraded at some point?21:35
vagrantcor at least, not at the maximum spec that the module knows about21:35
vagrantcseems like a bit weird to hard-code it ... is there any way the reform2-lpc-dkms module could hadnle the situation where the lpc hard-coded value and cells are mismatched?21:37
cwebberhey vagrantc I'd love to talk to you more about the Guix stuff on the rk3588 and how I might be able to help out21:37
vagrantcwill just try with the default 2.2Ah21:37
cwebberI'm about to switch full time, I hope, to the Pocket Reform as my daily driver machine21:37
vagrantccwebber: yay!21:38
vagrantccwebber: i'm only testing on the early generation mnt/reform2 motherboard with an early generation rk3588 module ... so testing any other variants would be great :)21:38
vagrantccwebber: and nice talk the other day, by the way :)21:39
vagrantcneed to test that the patches still apply on newer kernel variants, and add a few more kernel configuration options (e.g. support for KVM) that are misisng in the guix kernel build21:40
vagrantccwebber: i am also tempted to get a build farm setup to at least always build the packages for me that are relevent for my mnt/reform setup ...21:41
vagrantccwebber: and for my adventures today, i'm just building the firmware from debian, but ... one of these days i might scrape together time to work on building it on guix21:42
vagrantccwebber: lots to talk about :)21:42
arjanvagrantc: that reminds me I enable some useful kernel options in my own guix system config for the pocket that I still need to upstream21:51
kfxordered one from mouser and one from digikey, we'll see who gets here first21:52
vagrantcthe dreaded battery connectors are always the part that i spend the most time on any time i have to reset my mnt/reform ...21:59
vagrantci can never simultaneously see, feel and pinch them reliably21:59
vagrantcusing two small flathead screwdrivers seems to work, but i always worry about accidentally shorting something22:01
cwebberthanks vagrantc :)22:06
cwebberah yeah22:06
cwebberit's a miracle I didn't short them before reading the thread about how easy it is to short them22:06
vagrantcoh weird, with the new LPC firmware with 2.2Ah ... it apparently assumes it is full now, rather than recalibrating? or I have to drain it to calibrate it?22:11
vagrantcit used to just display ??? for a while until the battery was topped off22:12
- gustav25 (QUIT: Quit: Quit) (~gustav@c-78-82-54-128.bbcust.telenor.se)22:15
elbcwebber: when can I get a "laptop for sickos" sticker22:26
- pomel0 (QUIT: Ping timeout: 264 seconds) (~pomel0@user/pomel0)22:35
+ pomel0 (~pomel0@user/pomel0)22:36
elbcwebber: I've been daily driving my PR for a semester of lectures and most of my course "stuff" and it's worked out very well; I'm not running Guix, however, I'm running the stock debian unstable22:40
minuteelb: oh btw! do you have the newest reform-tools version with reform-power-daemon installed already?22:41
minuteelb: asking because you had those reset issues, and that should prevent them22:41
elbminute: I do, and today I ran down to 1% and hung out there for a long time with no struggles22:41
elbminute: it's been a little while since I've been able to do that (since my batteries were brand spanking new)22:42
minuteelb: ok awesome to hear22:42
elbI'm running r-p-d and the charger v2 now22:42
minuteelb: ok great22:42
minuteelb: this can still all be fine tuned, as the daemon is really quite primitive, but it validates the whole idea of keeping the current in check22:43
elbit seemed good for my battery life, too22:43
minuteelb: yes the battery life should be improved too22:43
elbalthough that's hard to tell because the new charge controller doesn't really know the batteries yet22:43
kfxwonder if this stuff could be bolted onto tlp...22:43
minutekfx: yes totally, but i have yet to try and understand tlp, i hadn't used it ever before22:44
kfxit's a bit of a monster22:44
minutekfx: so i wanted to first make something that i know works and the next step will be integrating that with some standard thing22:44
kfxabsolutely22:44
minuteotherwise it would have been delayed too long and i didn't want people to be frustrated by the brownouts22:45
kfxmy question was more to do with whether tlp is amenable to stuff like this22:45
kfxI'm happy with shell scripts, personally :)22:45
minutei've only looked briefly at it but it looked like that, i.e. if you do apt-file show tlp (or so) there's a bunch of vendor/hardware related profile files in there22:45
kfxcool22:46
minute/usr/share/tlp/bat.d/05-thinkpad /usr/share/tlp/bat.d/15-lenovo /usr/share/tlp/bat.d/30-samsung /usr/share/tlp/bat.d/40-sony etc22:46
minutelol it's totally also just shell scripts https://github.com/linrunner/TLP/blob/main/bat.d/40-sony22:46
kfxhahah22:47
kfxwell that answers that question22:47
minutelmao https://github.com/linrunner/TLP/blob/main/func.d/10-tlp-func-cpu22:47
minute"lets integrate with the industry standard" i mean "pile of shell scripts"22:47
kfxa time-honored tradition22:48
minute"gpu frequency limits" https://github.com/linrunner/TLP/blob/main/func.d/45-tlp-func-gpu#L4322:48
minuteextremely funny to me that it's basically the same stuff, just more of it because it was gestating for a longer time22:49
- Svp (QUIT: Quit: Gateway shutdown) (~svp@2002:4f07:f0bd:0:95e7:dc62:c203:a24)23:26
+ paperManu_ (~paperManu@107.159.15.124)23:37

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