2026-04-14.log

joschit happens in build-all.sh00:00
minutejosch: ah, i thought that was used only for CI00:00
joschyou can run it locally without gitlab CI as well00:00
joschwhen you build locally you can pass any value to cmake -DKBD_HID_FW_REV=....00:01
joschmaybe the problem is coupling the api version to the keyboard firmware version string. This limits the version that user's custom builds can have. Should the information about what the keyboard supports not be transported independent on the version string? Overloading the version with this meaning works well if there is a single vendor (and then you have predictable versioning) but not if anybody can 00:05
joschflash anything they want, no?00:05
minutejosch: yes, i agree. we should decouple it. for v3 i anyway need to shorten/shuffle around the fw strings because byte 8 is now taken by the checksum00:06
joschminute: in any case, i think for now, your fix (not treating small version numbers as v3) seems to be exactly what fixes the issue because if the version string cannot be parsed, it will probably come back as zero or something.00:07
minutejosch: thanks for confirming!00:09
joschyet another area where the pocket beats the classic: flashing keyboard and sysctl is just a breeze on the pocket thanks to fwupd00:14
minutejosch: oh yeah00:20
minutebuuut... i just noticed, and still need to measure, that my classic boots and esp. gets to gdm faster than my pocket, even if they both have rk3588. hm.00:21
josch(and the pocket is connected to power?)00:23
josch(asking because i recently wanted to play mario kart and even after staring at the thing for half an hour my brain did not connect the fact that it was on battery power to the graphics suddendly being a total slide show :D)00:27
- cli (QUIT: Ping timeout: 265 seconds) (~m-vsauiy@user/cli)00:36
minutejosch: oh :D00:40
minutegdm start on my pocket takes around 10-15 seconds on my pocket and 4 seconds on my reform (but just counted in my head, no precise measurement)00:49
minuteyeah, on battery or not makes no difference00:51
+ oliverD (~Thunderbi@user/oliverd)00:54
minutethere's some weirdness about an "fstrim" process since a while00:54
minutethere's an fstrim process in D state (uninterruptible sleep?)00:55
- oliverD (QUIT: Ping timeout: 268 seconds) (~Thunderbi@user/oliverd)00:58
minuteah the disk is just very slow somehow00:59
minutesudo apt remove systemd-userdbd takes a long time in "Reading database"01:00
minuteok, encrypted nvme read speed (gnoem disks) is 2.1GB/s on my classic, and around 400MB/s on my pocket 01:05
minuteah well i guess that's just pcie gen 2 x101:09
minutebut also, classic reform is still on kernel 6.17.8.01:10
- Guest7913 (QUIT: Ping timeout: 252 seconds) (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)01:11
minutehm, either it's pocket mb2.0 power issues for the ssd, or that ssd has developed some issues after all01:23
+ Guest7913 (~mjw@82-217-174-174.cable.dynamic.v4.ziggo.nl)01:27
+ pomel0 (~pomel0@user/pomel0)02:26
- pomel0 (QUIT: Read error: Connection reset by peer) (~pomel0@user/pomel0)02:51
+ pomel0 (~pomel0@user/pomel0)02:56
- pomel0 (QUIT: Remote host closed the connection) (~pomel0@user/pomel0)03:07
+ oliverD (~Thunderbi@user/oliverd)03:17
- oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd)03:22
+ oliverD (~Thunderbi@user/oliverd)03:22
- oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd)03:23
- paperManu_ (QUIT: Ping timeout: 255 seconds) (~paperManu@204.244.186.187)03:45
+ mark_ (~mjw@gnu.wildebeest.org)03:50
+ paperManu_ (~paperManu@79.127.134.34)03:52
- mjw (QUIT: Ping timeout: 256 seconds) (~mjw@gnu.wildebeest.org)03:52
- paperManu_ (QUIT: Ping timeout: 244 seconds) (~paperManu@79.127.134.34)03:58
+ paperManu_ (~paperManu@79.127.134.34)04:01
- kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de)04:08
+ kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de)04:09
- wickedshell (QUIT: Ping timeout: 276 seconds) (~wickedshe@2601:8c0:c7c:3572:e0b7:3a9d:f96:c311)04:13
+ pomel0 (~pomel0@user/pomel0)04:18
- paperManu (QUIT: Ping timeout: 245 seconds) (~paperManu@204.244.186.187)04:38
- mark_ (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org)04:39
- paperManu_ (QUIT: Read error: Connection reset by peer) (~paperManu@79.127.134.34)04:45
- ephase (QUIT: Read error: Connection reset by peer) (~ephase@82.66.198.11)05:01
+ ephase (~ephase@82.66.198.11)05:24
+ wickedshell (~wickedshe@2601:8c0:c7c:3572:dfdc:c9cb:c532:a406)06:20
joschminute: i feel (no measurements) that gdm came up much quicker about a year ago (and there is even a forum thread where somebody laments the same thing) and it is odd that it takes so many seconds of "idling" for a simple textbox for the username to show up on the screen... "modern" guis i guess...06:46
+ oliverD (~Thunderbi@user/oliverd)06:53
+ cli (~m-vsauiy@user/cli)07:04
henesykfx minute - thx for the screen reco07:05
joschminute: linux 7.0 is out and collabora has a new patch to slow down emmc -- is that interesting for us? https://gitlab.collabora.com/hardware-enablement/rockchip-3588/linux/-/commit/6ba621e7766d5c1db3b5d7a82f874049c46f0ba607:09
- oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd)07:16
+ _justin_kelly4 (~justinkel@user/justin-kelly/x-6011154)08:11
- _justin_kelly (QUIT: Ping timeout: 246 seconds) (~justinkel@user/justin-kelly/x-6011154)08:12
* _justin_kelly4 -> _justin_kelly08:12
- aelius (QUIT: Ping timeout: 244 seconds) (~aelius@user/aelius)09:06
+ aelius (~aelius@user/aelius)09:19
joschsigrid: good news! The commits from Detlev Casanova's branch add-vdpu381-and-383-to-rkvdec-v9 got to be part of Linux 7.0 :)09:59
+ midfavila (midfavila@otaku.sdf.org)10:06
midfaviladoes anyone know if there are status updates on bnys' group buy? nothin since the start of march10:07
midfavilai would be really interested in an anodised bottom plate since my acrylic one is pretty beat up10:08
midfavilaalternatively i might just pay for one from a local place depending on the cost...10:08
midfavilai have to wonder if that would ever be an optional upgrade MNT would sell10:09
midfavilai would pay for an alu bottom plate 10:10
- wickedshell (QUIT: Ping timeout: 246 seconds) (~wickedshe@2601:8c0:c7c:3572:dfdc:c9cb:c532:a406)11:27
- buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org)12:17
+ buckket (~buckket@vps.buckket.org)12:18
+ paperManu (~paperManu@204.244.186.187)12:24
+ paperManu_ (~paperManu@204.244.186.187)12:25
+ wickedshell (~wickedshe@2601:8c0:c7c:3572:dfdc:c9cb:c532:a406)13:35
sigridjosch: amazing13:37
- paperManu_ (QUIT: Ping timeout: 246 seconds) (~paperManu@204.244.186.187)14:04
- paperManu (QUIT: Ping timeout: 272 seconds) (~paperManu@204.244.186.187)14:07
+ mark_ (~mjw@gnu.wildebeest.org)14:08
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:49
- mark_ (QUIT: Ping timeout: 272 seconds) (~mjw@gnu.wildebeest.org)14:49
joschminute: linux 7.0 is in Debian experimental. Could you please merge https://source.mnt.re/reform/qcacld2/-/merge_requests/14 so that I can finalize https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/17514:50
- jn (QUIT: Ping timeout: 244 seconds) (~quassel@user/jn/x-3390946)14:56
+ jn (~quassel@user/jn/x-3390946)14:57
* Guest7913 -> mjw15:04
- paperManu (QUIT: Ping timeout: 264 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)15:35
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)15:36
+ stephano (~stephano@71.238.14.13)17:23
- _justin_kelly (QUIT: Quit: The Lounge - https://thelounge.chat) (~justinkel@user/justin-kelly/x-6011154)18:12
+ _justin_kelly4 (~justinkel@user/justin-kelly/x-6011154)18:15
+ mark_ (~mjw@gnu.wildebeest.org)18:32
- pomel0 (QUIT: Ping timeout: 244 seconds) (~pomel0@user/pomel0)18:54
+ siviq (~siviq@user/siviq)19:16
- siviq (QUIT: Client Quit) (~siviq@user/siviq)19:18
minutejosch: thanks for the MRs! qcacld merged19:19
minutejosch: related: we just discussed here recently that it might save us a looot of trouble to build qcacld2 and especially reform2-lpc as part of the kernel package... i'm sure we discussed this back in the day but don't remember why we didn't do that19:20
minutejosch: this rings a bell, but it's for rock5b, not for rk3588 in general. not sure if that applies to firefly... did anyone have corruption issues with emmc yet?19:22
minutejosch: sorry, link for the last message: https://gitlab.collabora.com/hardware-enablement/rockchip-3588/linux/-/commit/6ba621e7766d5c1db3b5d7a82f874049c46f0ba619:23
minutemidfavila: i think the main issue about the plate is the thickness. it should be reduced/hollowed out a bit vs the acrylic plate, otherwise it's super heavy19:23
midfavilahmm. that's fair. 19:24
midfavilai don't mind the chunkiness though :D19:24
midfavilait's reassuring19:24
minutemidfavila: so someone needs to open a CAD program and do it. the ordering/manufacturing is easy19:24
minutemidfavila: if you hollowed it out a bit from the inside it would look the same on the outside19:24
- marty (QUIT: Ping timeout: 244 seconds) (~marty@static-23-234-102-239.cust.tzulo.com)19:27
+ marty (~marty@146.70.168.182)19:29
kfxsigrid and I made a direct patch for the lpc driver https://git.sr.ht/~ft/aports/tree/reform/item/community/linux-rk3588/9002-add-reform-lpc.patch19:30
kfxcontrolled with CONFIG_EC_REFORM2_LPC in the config19:30
midfavilaminute: yeah makes sense19:30
midfavilamaybe i'll get a local machine shop to see if they can mill a plate for me then 19:31
midfavilabe a good little weekend project19:31
kfxI had a friend mill me a bottom panel from G10 (pcb material) and it works pretty great19:31
midfavilanice :D19:32
sigridI wanna try wood for this19:33
midfavilaoo that'd be cool19:33
midfavilayou could use an iron to burn some cool patterns into it or something 19:34
minutehttps://www.crowdsupply.com/mnt-research/mnt-station19:36
minutekfx: awesome! that's kinda what we need, but perhaps it could be dynamically picked from the reform-tools repo cc josch 19:37
midfavilaholy moly new product 19:37
midfavilavery cool19:37
minuteyeah has been gestating for a longer time on and off19:37
wickedshellminute: was just looking at that, any chance it will launch (or later be available) sans processor module? For those who have old modules to drop in it would be nice. (But maybe too niche for crowdsupply campaign?)19:38
midfavilai've been watching, very excited for it19:38
minutewickedshell: yes without mainboard even19:38
wickedshellminute: excellent! <319:38
midfavilai've been meaning to get a replacement mobo for my laptop reform since the barrel jack is starting to go so i can put that one in the desktop enclosure and fiddle around with it19:38
minutewickedshell: there will be case, mb, module separately i think19:38
midfavilaget an LS1028a and figure out how to hook up an external kepler card or whatever the last blobless cards nouveau supports were based on... nice little setup19:39
+ pomel0 (~pomel0@user/pomel0)19:45
minutemidfavila: yup!19:50
- mark_ (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org)20:10
f_very exciting :D20:19
midfavilalast thing i'll need to totally ditch my phone and other proprietary hardware and software forever would be something like sqfmi's beepy... well, with a different SoM, but still20:27
midfavilaunfortunately they're in some sort of limbo status and the derivatives on the market have been closed, in violation of the GPL and CERN OHW20:28
midfavila:\20:28
f_just install mainline linux :p20:29
midfavilait's the principle of the matter >:c20:30
midfavilai won't give money to plagiarists and thieves20:31
midfavilaalso the kernel needs special drives for some of the input devices and those were modified so i can't really install mainline unfortunately20:31
midfaviladrivers even wah20:31
f_I mean, you have a phone already?20:36
midfavilaa second-hand blackberry that's twelve years old :p20:37
f_neat :D20:38
midfavilayeah it's an old passport that a friend of mine had laying around at the time. hardware is still fine, software is just too locked down to be useful20:39
midfavilaat least for some things20:39
midfavilafor quick notes or an on the go ssh session or quickly checking wikipedia or something like that it's fine. 20:39
minutemidfavila: i'm in a similar boat, that's why we're working on that 7 inch tablet :D20:48
midfavilayeah i saw that, it looks really cool20:50
midfavilai'm needing something a bit smaller though, and with a thumb board, hence sqfmi's beepy20:51
midfavilathey seem to be the most compact, even if there are other drawbacks20:51
f_7 inch is a bit small these days I'd say20:52
f_one of the early tablets I ever used was 7 inch :D20:52
minuteyeah it's meant to fit in a roomy pants pocket or jacket pocket :D20:52
f_and w/modem?20:53
minutemaybe more a phablet 20:53
midfavilasee i'm a total scatterbrain and i'm always losing stuff so i have this little bag i clip on my belt20:53
minutemodem pluggable20:53
f_neat!20:53
midfavilaand i just keep everything i need in that20:53
midfavilaunfortunately even the larger pocket is too small for a 7" device20:53
f_by the way I love how you're recycling on existing hardware designs rather than starting from scratch20:53
f_well. Manufacturers often don't start from scratch for new devices, but it feels like they're doing a brand new thing by copy-pasting20:54
minutef_: thx! i would like to converge things even more but that can also get complicated, and sometimes it's interesting to have different approaches compete in the field (for example, we have 3 different battery charging architectures now lol)20:54
f_it doesn't feel like it's like an MNT Reform transformed into a tablet20:54
f_minute: with that I also really love the MNT Station idea, especially as someone who sometimes recycles old desktop computers to use as servers :D20:55
minutef_: yay!20:55
f_and as of late it seems I'm also now recycling old android junk20:57
f_really, keep it up!20:57
- pomel0 (QUIT: Ping timeout: 246 seconds) (~pomel0@user/pomel0)20:58
minutethx!20:58
+ pomel0 (~pomel0@user/pomel0)20:59
joschminute: I can add sigrid's patch to the kernel build but i wonder how the migration path would look like for users who have the dkms module installed right now.21:31
joschmaybe the dkms module could be changed to gracefully refuse to load if EC_REFORM2_LPC is not "n"?21:32
midfavilahttps://github.com/antmicro/arvsom21:36
midfavilapotentially interesting21:36
midfavilastarfive jh7110-based cm-4 compatible som21:37
midfavilaofficial website says it's still wip but who knows if it's still being worked on with the repo's last commit being a while ago...21:38
minutejosch: don't add it yet. there will be a new version with different compatible soon-ish (kcxt was working on refactoring it a bit)21:40
minutemidfavila: i have a banapi risc-v cm5 som here, it gets me to uboot in reform21:40
minutebananapi even21:40
midfavilaneat :D21:40
minutebut didn't have time to explore further21:41
minutejust can say it works electrically :D21:41
midfavilai've been fiddling with a D1 in another device i have and aside from X11 being a bit slow it's a perfectly capable little guy21:41
midfavilamakes me happy21:41
orvaThose 7inch tablets have been only ones that I have liked. Well, only tablets I have used are ones I got from previous employer, one Nexus 7 Android one and then some iPad years later21:57
orvaThe 7inch form factor was actually usable with one hand and really nice size for reading books / manga21:58
+ mark_ (~mjw@gnu.wildebeest.org)22:07
- mjw (QUIT: Killed (molybdenum.libera.chat (Nickname regained by services))) (~mjw@82-217-174-174.cable.dynamic.v4.ziggo.nl)22:08
* mark_ -> mjw22:08
+ Guest3529 (~mjw@82-217-174-174.cable.dynamic.v4.ziggo.nl)22:15
minuteorva: yeah those are the kind of ideas for use cases i also have :D 22:38
- paperManu (QUIT: Read error: Connection reset by peer) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)23:00
+ paperManu (~paperManu@204.244.186.187)23:29
+ paperManu_ (~paperManu@204.244.186.187)23:29

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