2026-05-28.log

- paperManu (QUIT: Ping timeout: 249 seconds) (~paperManu@104.163.176.82)00:06
minuteradxa cm5 pocket almost fully working now, except sound, checking that00:11
minutedsi panel works, hdmi works, ssd and modem work...00:11
minute(ethernet, sd card, emmc too)00:13
minuteok, audio (i2s) lines are incompatible. but i can fix that on rcm5 of course00:16
+ paperManu (~paperManu@142.169.16.88)00:28
+ Gooberpatrol66 (~Gooberpat@user/gooberpatrol66)00:34
chany ideas on why the keyboard doesnt send keys to the host after plugging? (only sometimes obviously)00:46
chreplug fixes it00:46
chshould probably see about flashing it00:47
- thejevans (QUIT: Remote host closed the connection) (~m-7r3qil@user/jevans)00:52
- paperManu (QUIT: Ping timeout: 257 seconds) (~paperManu@142.169.16.88)00:53
+ paperManu (~paperManu@142.169.16.88)00:53
- vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50)00:53
joschminute: reform-tools 1.87 is tagged in source.mnt.re git00:57
joschI also packade it and uploaded it to Debian unstable00:57
joschtomorrow morning i can re-run the reform-debian-packages pipeline to push it to the MNT Debian repository as well00:58
joschand then we'll see which bugs i missed :)00:58
- Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.0.44)01:00
+ Ar|stote|is (~linx@149.210.0.44)01:09
- paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@142.169.16.88)01:14
+ paperManu (~paperManu@142.169.16.88)01:15
minutejosch: many thanks!01:16
minutech: the new standalone kbd that you got?01:16
chyeah01:17
chit has fw 20250701/B201:17
joschminute: you said that the kernel cmdline for quasar is wrong but you didn't say yet what the correct one looks like01:17
chso the symptom is basically: plugin; keyboard powers up; host doesnt react to keypresses01:22
chstill have to try another (!mac) host01:22
minutech: ok very weird. oled menu responds?01:42
chyes01:42
minutech: do you have some dmesg or other usb debugging equivalent on that exotic machine?01:43
minutech: maybe cable issue?01:43
chthats the right way to call it :)01:43
chlets see01:43
chIOUSBHostDevice@01100000: IOUSBHostDevice::start_block_invoke: device descriptor fragment is invalid01:47
chhmm01:47
chAppleUSBIORequest: AppleUSBIORequest::complete: device 1 (IOUSBHostDevice@01100000) endpoint 0x00: status 0xe00002ed (transaction error): 0 bytes transferred01:47
chusb-drd1-port-hs@01100000: AppleUSBHostPort::terminateDevice: destroying 0x0000/0000/0000 (IOUSBHostDevice): enumeration failure01:47
chnot super insightful01:47
chhttps://paste.debian.net/hidden/96879bdf so much log output, so little insight01:50
chhm yeah, now the host really hates the keyboard 01:54
- paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@142.169.16.88)02:09
+ paperManu (~paperManu@142.243.253.17)02:11
+ thejevans (~m-7r3qil@user/jevans)02:11
- talos (QUIT: Ping timeout: 246 seconds) (~talos@2600:6c5d:f0:1fd0:fafe:f166:270b:2348)02:12
- paperManu (QUIT: Ping timeout: 276 seconds) (~paperManu@142.243.253.17)02:31
minutech: hm maybe the usb descriptors have some bug... maybe you're the first on to try it on arm apple02:31
minutech: does it work as a scsi disk in programming mode? (there's a dip switch)02:32
chhmm no02:39
chthats even more dead02:39
minutech: do you have another computer to cross check? pocket reform? :D02:57
+ paperManu (~paperManu@104.163.176.82)02:57
+ paperManu_ (~paperManu@204.244.197.237)03:08
+ talos (~talos@2600:6c5d:f0:1fd0:ca6c:8b95:1be0:301b)03:18
- paperManu_ (QUIT: Ping timeout: 265 seconds) (~paperManu@204.244.197.237)03:26
- mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org)04:00
- stephano (QUIT: Quit: Textual IRC Client: www.textualapp.com) (~stephano@71.238.14.13)04:07
- paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@104.163.176.82)08:52
chis https://source.mnt.re down?09:42
Asmadeuslooks up to me09:44
joschch: works for me -- try from a different IP. The domain recently got swamped by "AI" bots, so tens of thousands of IPs got blocked09:44
choh no09:45
chindeed works from a remote09:45
- b0 (QUIT: Read error: Connection reset by peer) (~b0@user/b0)09:52
- johl (QUIT: Read error: Connection reset by peer) (~johl@wikidata/Jens-Ohlig)09:52
+ b0 (~b0@user/b0)09:54
+ johl (~johl@wikidata/Jens-Ohlig)09:55
- marty (QUIT: Ping timeout: 252 seconds) (~marty@23.234.102.55)10:00
joschminute: reform-tools 1.87 is in the MNT repositories10:32
joschi added a bunch of required changes to reform-system-images already to make it build with the new reform-tools: https://source.mnt.re/reform/reform-system-image/-/merge_requests/152 (mostly stuff for QUASAR support)10:33
joschthe thing still fails to build because flash-kernel does not have support for QUASAR in its all.db file10:33
joschi know that quasar will not be using boot.scr and reform-tools 1.87 should contain all pieces necessary to have systems without flash-kernel10:34
joschbut before experimenting further with that, i temporarily patched flash-kernel with the entries for the six quasar platforms10:34
joschdoing that also avoids more hacks/workarounds for missing flash-kernel stuff in reform-system-image10:34
joschthe new reform-debian-packages pipeline with patched flash-kernel with QUASAR support is running right now, so in 1.5 hours when the changes are in the MNT Debian repo me (or anybody else) can re-try the failed reform-system-image pipeline from MR!15210:37
joschMR!152 is essentially cherry-picking everything from MR!151 except the removal of flash-kernel and the uki building10:42
- RandyK (QUIT: Ping timeout: 252 seconds) (~RandyK@user/randyk)10:52
+ RandyK (~RandyK@user/randyk)11:10
- chorc (QUIT: Quit: ZNC 1.9.1 - https://znc.in) (~chorc@user/chorc)11:18
+ chorc (~chorc@user/chorc)11:19
joschugh... what's the problem now? It worked the last run of reform-debian-packages... Why does the src:linux build break now like this?11:30
joschI will be away for class for the next 2 hours, if somebody is able to see what the problem with linux-config-7.0 (= 7.0.9-1+reform20260526T130158Z) is, I'd be very happy: https://source.mnt.re/reform/reform-debian-packages/-/jobs/2127011:31
amospallaHi, any pocket user, when charging, your pocket charges your battery intermitently? It goes like this on mine, every few seconds stops and starts charging again.11:48
amospallaI replaced batteries and keeps doing it, maybe I should buy the new pocket charging board?11:48
cararemixedDid source.mnt.re just go down or did I get blocked? Just cloning a couple repos (was going to look at the handbooks and consider adding something about working with the sources because I think it could be better documented) but I get timeouts now.11:51
minutecararemixed: can you dm me the external IP you used? then i can take a look at work11:59
cararemixedThanks. I had most of what I wanted. I was cloning the next handbook to compare with the other two and think about how they'd either be updated OR consider writing a hackers handbook.12:02
cararemixedJust all the little things I've started figuring out on my own as I do things like customize my firmware or look through the mnt kernel patch stack.12:03
cararemixedAlso might be good to document alternative distro and OS setup stuff there.12:03
- RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk)12:04
+ RandyK (~RandyK@user/randyk)12:04
minuteright12:08
* shdw_ -> shdw12:12
+ paperManu (~paperManu@204.244.197.237)12:20
+ paperManu_ (~paperManu@104.163.176.82)12:20
+ mjw (~mjw@gnu.wildebeest.org)12:43
rick_amospalla where do you see it? In the logs? Or through the desktop symbol? If its not tooo often its normal.. I see it in the reform-power-logs a few switches every hour. Do you already have the v2 power board?12:46
rick_amospalla did you also try a different usb charger?12:47
amospallarick_: it does it every few seconds, I can see the pocket light go to blue and to red again. I see it on `sensors` and `acpi`: charging state, current goes up and down.12:48
amospallaI have the original charging board, and because the new one improves on this, I thought mine could be failing, or just changing it could solve the problem. I discarded the batterries, just replaced them yesterday.12:49
rick_Yeah okay, that's pretty often ^^v12:50
amospallaoh, changed the carger, feels better now :S :)12:50
chminute: didnt notice yesterday that there are two dipswitches. with the left one flipped it does connect as the rp2 'scsi'12:53
amospallathank you rick_ , I'll double check next days with another charger.12:58
chbootrom mode comes up pretty reliably also on the mac13:06
chhmm13:06
- talos (QUIT: Ping timeout: 245 seconds) (~talos@2600:6c5d:f0:1fd0:ca6c:8b95:1be0:301b)13:10
minutech: yep, the left one is the programming switch. the right one is the standalone switch13:11
minutech: but good, then the hw seems fine and it's "just" the firmware. what if you flash picosdk hid keyboard example or so?13:12
+ talos (~talos@2600:6c5d:f0:1fd0:8d6d:7326:574:cec3)13:16
- iank (QUIT: Quit: ZNC 1.8.2+deb2+deb11u1 - https://znc.in) (~iank@fsf/staff/iank)13:25
+ iank (~iank@fsf/staff/iank)13:29
- mjw (QUIT: Ping timeout: 256 seconds) (~mjw@gnu.wildebeest.org)13:38
* Guest6886 -> mjw13:40
+ wielaard (~mjw@2001:1c06:2486:4600:6fcb:4b0e:fedd:fb4d)13:51
- paperManu (QUIT: Ping timeout: 273 seconds) (~paperManu@204.244.197.237)14:25
- paperManu_ (QUIT: Ping timeout: 276 seconds) (~paperManu@104.163.176.82)14:35
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:51
- wielaard (QUIT: Ping timeout: 245 seconds) (~mjw@2001:1c06:2486:4600:6fcb:4b0e:fedd:fb4d)15:48
+ stephano (~stephano@71.238.14.13)16:15
joschminute: ERROR: Uploading artifacts as "archive" to coordinator... 413 Request Entity Too Large  id=21278 responseStatus=413 Request Entity Too Large status=413 token=glcbt-6416:20
joschhttps://source.mnt.re/reform/reform-system-image/-/jobs/2127816:20
+ wickedshell (~wickedshe@2601:8c0:c7c:3572:a4bd:38de:9adb:2280)16:31
- buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org)17:00
+ buckket (~buckket@vps.buckket.org)17:01
joschreform-system-any builds fine: https://source.mnt.re/josch/reform-system-image/-/jobs/2128117:07
[tj]minute: can you check on order R930798377 17:15
minute[tj]: checked, this is waiting for the new batch of 16gb rcore shipments that will come end of june17:52
minutejosch: ok trying to remember where to tweak this18:00
minutejosch: maximum artifacts size is currently 32768 MB18:01
minutejosch: "maximum artifacts archive size" is 5242880 bytes, not sure what an artifacts archive is though18:02
joschminute: i'll add some computation to the .gitlab-ci.yml computing the sum of image sizes then we know for sure what amount we are talking about18:04
chiirc there is an option in the gitlab omnibus config18:30
ch(because nginx in front also needs the client body size limit increased)18:31
chgitlab.rb in grmls install has: nginx['client_max_body_size'] = '8000m'18:31
minuteright, let me check that too18:33
+ marty (~marty@146.70.168.215)18:42
minuteso far i'm blocking 1003467 IPs, cracked the million19:20
joschminute: it says 36G total here: https://source.mnt.re/reform/reform-system-image/-/jobs/2128419:34
^alexwe have a few cute tricks to block scanners19:40
^alexbut it relies on the bsd blocklist service19:41
^alexwhich i _think_ got ported to linux and iptables recently?19:41
^alexanyway we have it rigged up to nginx such that if you request, like, WordPress garbage, or other stuff we've never had installed, or if you make a request to a vhost that we know doesn't serve HTTP (a.mx.se30.xyz, or our DNS hostnames), you can get into the firewall with alacrity19:42
^alexand with the kinds of bots that crawl the certificate transparency logs, making an obvious-trap vhost is our next thought19:43
^alex:319:43
minute^alex: i'm using ipset so maybe that's similar19:45
minute^alex: i'm using a similar strategy, i'm scanning the access logs for patterns that only bots use and then they get added to a sorted list and loaded into ipset19:45
^alexah ours is a daemon that proactively notifies https://github.com/zoulasc/blocklist19:46
^alexhttps://fossil.se30.xyz/ratrap/dir19:46
^alexnever tested it on linux, so idk what'd be involved in that19:47
^alexthis mechanism is a little more direct since it doesn't involve scanning access logs19:47
minutejosch: maximum artifacts size bumped to 48GB now. but i need to restart the gitlab because i also changed some other setting19:59
minuteah, i can also update gitlab to 19.0.120:00
minute(done)20:29
+ siviq (~siviq@user/siviq)20:32
- siviq (QUIT: Client Quit) (~siviq@user/siviq)20:35
ch  Serial Number: RP204020:46
chsure :>20:46
minutech: hehe... watcha doing20:49
chjust tried the keyboard flash script from the 2026-04-21 release, and thats what the keyboard reports as serial afterwards20:50
chi guess theres something missing in the usb descriptor20:51
chwill look in a while, but for now i want to see if the newer build just fixes my issue20:51
minutech: but the kbd works then on the mac or...?20:51
chyes it works now. seems like theres something weird going on. wouldnt be the first time :)20:52
- aelius (QUIT: Remote host closed the connection) (~aelius@user/aelius)20:52
ch(fwupd would be cool too)20:52
+ aelius (~aelius@user/aelius)20:53
joschminute: reform-system-any boots fine on rk3588 pocket reform, setup wizard starts, gnome starts, reform-check output is fine20:55
minutejosch: very nice! did the pipeline work now?20:57
minutei also did some gardening on the server to prevent the disk filling up with backups20:57
[tj]minute: thanks20:58
joschminute: no, i built reform-system-any only -- see my link further up in the chat log from 17:0721:02
+ wielaard (~mjw@2001:1c06:2486:4600:6fcb:4b0e:fedd:fb4d)21:03
- wielaard (QUIT: Ping timeout: 245 seconds) (~mjw@2001:1c06:2486:4600:6fcb:4b0e:fedd:fb4d)21:09
minutejosch: ahh ok21:17
+ wielaard (~mjw@gnu.wildebeest.org)21:25
* mjw -> Guest856721:25
- Guest8567 (QUIT: Killed (iridium.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:4600:5952:3a9:6e0e:555a)21:25
* wielaard -> mjw21:25
+ Guest8567 (~mjw@2001:1c06:2486:4600:5952:3a9:6e0e:555a)21:25
ch        Serial Number String:   3 "DE650815A7611037"21:42
chbetter21:42
chhttps://source.mnt.re/reform/reform/-/merge_requests/100 and https://source.mnt.re/reform/reform/-/merge_requests/101 please21:45
chi suppose the kbd should also report a higher power draw, the usb-c thing i have claims the kbd  uses 350mA with defaults21:46
chup to 700mA on max brightness21:46
joschcase insensitive filesystems in 2026??21:47
chalways been there ;)21:47
joschi'd also only trust a filesystem for which i could write the driver myself on a good weekend :D21:48
chhahah21:48
chapfs is probably not that easy21:48
joschwait, this is a macos thing??21:48
joschi thought we were talking fat3221:48
+ siviq (~siviq@user/siviq)21:51
- siviq (QUIT: Client Quit) (~siviq@user/siviq)21:56
- robin (QUIT: Quit: Leaving) (~robin@user/terpri)22:00
- Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66)22:16
joschminute: job succeeded, 36 G of artifacts uploaded: https://source.mnt.re/reform/reform-system-image/-/jobs/2128723:04
joschthis is missing qcs8550 images for pocket and next23:05
joschmaybe the opposite should happen: keep platform specific images for some platforms (like imx8mp) and just build reform-system-any for rk3588 and qcs*23:06
- paperManu (QUIT: Ping timeout: 248 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)23:13
+ paperManu (~paperManu@204.244.197.237)23:30
+ paperManu_ (~paperManu@104.163.176.82)23:31
+ Gooberpatrol66 (~Gooberpat@user/gooberpatrol66)23:31
minutejosch: yep, that makes sense23:43
+ Staff_Jess (~Staff_Jes@99.56.89.7)23:49
Staff_Jess[Network Announcement] Libera staff requests a respectful minute of silence in honor of Harambe's memory on the 10th anniversary of His death. R.I.P. Harambe (May 27, 1999 - May 28, 2016)23:50
- Staff_Jess (QUIT: Remote host closed the connection) (~Staff_Jes@99.56.89.7)23:50
chminute: could you assign a runner to zeha/reform, so the PRs get runs?23:59

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