2024-08-23.log

joschminute: system images for reform-next: https://source.mnt.re/reform/reform-system-image/-/jobs/5669/artifacts/browse00:01
- Sario (QUIT: Ping timeout: 604 seconds) (sario@libera/staff/owl/sario)00:01
minutejosch: niiice00:34
- robin (QUIT: Ping timeout: 276 seconds) (~robin@user/terpri)00:50
- andypiper (QUIT: Quit: bye) (~andypiper@45.148.12.75)00:54
+ robin (~robin@user/terpri)01:06
+ Sario (sario@libera/staff/owl/sario)01:12
- mtm (QUIT: Ping timeout: 252 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net)02:03
+ mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net)02:06
+ robin_ (~robin@user/terpri)02:26
- Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66)02:28
+ Gooberpatrol66 (~Gooberpat@user/gooberpatrol66)02:28
- robin (QUIT: Ping timeout: 260 seconds) (~robin@user/terpri)02:28
- mjw (QUIT: Ping timeout: 252 seconds) (~mjw@185.59.39.53)02:30
* robin_ -> robin02:32
- robin (QUIT: Remote host closed the connection) (~robin@user/terpri)02:58
+ robin (~robin@user/terpri)03:07
+ robin_ (~robin@user/terpri)03:09
- robin (QUIT: Ping timeout: 276 seconds) (~robin@user/terpri)03:13
* robin_ -> robin03:19
- xktr (QUIT: Remote host closed the connection) (~xktr@user/xktr)03:45
+ xktr (~xktr@user/xktr)03:46
+ Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66)04:00
- Gooberpatrol66 (QUIT: Ping timeout: 260 seconds) (~Gooberpat@user/gooberpatrol66)04:02
- cobra (QUIT: Ping timeout: 246 seconds) (~cobra@user/Cobra)04:27
+ cobra (~cobra@user/Cobra)04:35
- skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich)05:00
+ skipwich (~skipwich@user/skipwich)05:05
- skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich)05:14
+ skipwich (~skipwich@user/skipwich)05:16
- skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich)05:22
+ skipwich (~skipwich@user/skipwich)05:25
- robin (QUIT: Ping timeout: 255 seconds) (~robin@user/terpri)05:34
- Gooberpatrol_66 (QUIT: Ping timeout: 276 seconds) (~Gooberpat@user/gooberpatrol66)05:40
+ robin (~robin@user/terpri)05:44
+ Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66)06:35
- Gooberpatrol_66 (QUIT: Client Quit) (~Gooberpat@user/gooberpatrol66)06:35
+ Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66)06:35
+ Gooberpatrol66 (~Gooberpat@user/gooberpatrol66)07:27
- Gooberpatrol_66 (QUIT: Ping timeout: 260 seconds) (~Gooberpat@user/gooberpatrol66)07:28
joschminute: i just had somebody complain that https://source.mnt.re advertises being reachable via ipv6 but is then not reachable via it09:52
joschmore specifically, tracepath 2a01:4f8:c012:ada7::1 succeeds but then curl -6 https://source.mnt.re does not09:55
minutejosch: hmm ok, can look into it when i'm at the company10:00
joschthe person who reported it is on an ipv6-only network10:03
joschi'm on an ipv4-only network but i can test this by ssh-ing into debian infra if needed10:04
minutei'm too old to understand ipv6 so i will look if i can remove the dns entry10:05
joschoh no i sometimes think the same XD10:07
joschi think^Whope it might get better once my home or work network can even do ipv6 in the first place10:07
joschat uni, our office computers even have a public ipv4 address because universities just own huge address blocks...10:08
minutewow :D10:08
minuteone can't paste a ipv6 address into chromium's address bar it seems?10:09
joschclearly, google doesn't want you to use address bars anymore but just go to the place you want by googling for it ;)10:11
minutei think i fixed it now for ipv610:16
minutemntmn@reform:~$ curl -6 https://source.mnt.re10:16
minute<html><body>You are being <a href="https://source.mnt.re/users/sign_in">redirected</a>.</body></html>10:16
joschhaha!10:16
joschyes, this looks different now (from my ipv6 debian host)10:17
minutethis was an issue with the gitlab default config10:17
joschoh?10:17
minutegitlab has a default setting for nginx['listen_address'] in gitlab.rb but it is somehow wrong. setting it explicitly worked10:19
minutelisten_addresses10:19
joschgood that the fix was relatively easy10:19
minuteyeah10:20
joschtells you something about ipv6 adoption if the issue was open for that long :D10:20
minuteyeah...10:20
joschminute: since you are currently mucking with gitlab: if you create a  2024-07-26 tag for the reform/reform repo, then i can work on the reform-firmware package for Debian :)10:20
joschyou merged https://source.mnt.re/reform/reform/-/merge_requests/70 (thank you!) but apparently MRs do not add tags10:21
minuteohh10:22
minutejosch: you want 2024-07-26 and not 2024-08-23 yes?10:22
joschminute: yes, because that is what the MR you merged put into the firmware10:22
joschlike, the string you can see on the OLED10:23
minuteok, tag pushed10:23
joschthank you!10:23
chsource.mnt.re works for me over ipv610:23
minute(i didn't actually go to the company yet, still having breakfast, _but_ i have a capable rk3588 laptop!)10:23
chand has worked fine for the last weeks10:23
minutech: hmm weird10:23
joschminute: i just got confirmation from the reporter that ipv6 works now :)10:23
xhaminute: btw, you have to put the ipv6 address in brackets in URL's, for example [fe80::1] instead of fe80::110:24
joschch: according to "curl -6" on coccia.d.o it did not work before10:24
ch'oh well' :)10:24
joschminute: and i was told to pass on their thanks :)10:24
minutexha: ohhh now that works10:25
minutejosch: they're welcome :D10:25
josch:)10:25
joschminute: the CI pipeline for the tag you created failed because i failed to add an entry for the new version to reform2-keyboard-fw/CHANGELOG.md -- unless you object, i'm going to ignore this until the next tag10:28
minuteok :310:29
minutefrom #linux-rockchip: > <diederik>It actually happened ... rk3588 support in TF-A was merged \o/10:30
joschhah, i just had a conversation about rk3588 firmware over in #debian-arm :D10:31
joschnow we juuuuust have to figure out whether we are allowed to distribute the blobs this time round... ;)10:32
minuteoh :D10:32
joschyeah, i got some unhelpful comments along the lines of "just add some 32 MB SPI flash onto your board and put your blobs there!"10:32
minutelol10:34
minutejust decompile the blobs with ghidra and let chatgpt rewrite them in a different style, then recompile, ... profit10:37
chdoes tf-a include vendor blobs then?10:38
ch(or what was the redist problem about?)10:38
minutei think normally blobs are mixed in during some uboot build step10:38
minuteon most modern (fast) socs that i've seen except layerscape, there's some ddr4/5 training blob10:40
minuteunrelated: does bunnie/xobs' precursor have some sort of forum/community? i couldn't find anything. i was curious what people do with the device10:43
minuteor IRC10:43
Twodisbetterminute: I have one, and I am also kind of interested in the same question. 10:55
minuteTwodisbetter: did you do anything interesting with it?10:55
minutecan one use it as a messenger, for example, or for irc?10:55
Twodisbetterminute: apparently it has a signal client integrated10:58
minuteoh10:58
Twodisbetterthe whole thing is basically stuck in the TTL10:58
Twodisbetterbut I have just not had the time to figure it out. 10:58
TwodisbetterIt is very nicely built and solid though10:59
gsoraminute: i contribute to betrusted/precursor, lmk if you want to know anything!11:02
gsorawe usually hang out in the project's matrix space11:02
gsorai developed a PoC for an IRC client a couple years back i think? i did it mainly to get accustomed with the environment, and it also served for networking stack debugging11:08
gsorai use mine as a secure vault, kind of like a secure enclave on steroids11:09
gsoraand i'm working on the side on a bespoke TEE implementation that (if everything goes well) will run wasm applets11:09
minutecool11:11
minutethat's neat @ wasm11:12
gsorayeah, there's potential with that thing11:12
minutewhat kind of messengers can one interact with atm?11:12
gsorathe most advanced at the moment is sigchat: https://github.com/betrusted-io/sigchat11:12
minuteneat11:13
gsorai did not try it myself because i'm not interested that much in the messenger part 11:13
gsorathere has been a substantial effort by the authors though, i've watched it going from nothing to something that actually interacts with signal11:14
gsorafingers crossed, if you see someone at FOSDEM 2025 with an mnt reform and a precursor, it's me :^)11:18
minutegsora: cool :D i should go to fosdem again!11:19
gsorathis might be my first one :) never been a huge conference person, but this time i might have some friends with me11:20
+ mjw (~mjw@185.59.39.53)11:40
grimmwaremorning :)11:49
josch_o/11:58
joscheverybody happy with their reform-tools 1.52? no further regressions? :)11:58
amospallaI upgraded today, rebooted, everything ok.12:02
amospallareform-check says I can update my uboot. Is that safe?12:03
joschamospalla: what platform?12:07
amospallajosch: on a pocket.12:12
amospallaToday I discovered I can install upower and it will giver me the remaining estimated battery time.12:13
joschamospalla: uh oh, you should not get this message -- can you pastebin the complete output of "sudo reform-check" somewhere?12:14
amospallajosch: https://paste.debian.net/1327292/12:20
amospallaMaybe I cutted down too much text, tell me if so.12:20
joschoooh /boot/flash.bin12:21
joschthat is just a file12:21
joschit will not flash anything12:22
amospallajosch: it then asks to execute /usr/sbin/reform-flash-uboot, which says that will  "Download latest u-boot for the current platform and update /boot/flash.bin" "Then flash u-boot to DEVICE which can be one of the short-hands sd or emmc."12:23
amospallareform-check checks if files equal, else asks the user to run reform-flash-uboot, which runs parted (to do checks) and dd (to do raw update).12:25
joschamospalla: no, reform-flash-uboot in your case will (should) only update /boot/flash.bin12:27
amospallaoh, thank you very much12:28
joschso in your case it should tell you that on your platform it cannot flash u-boot12:28
amospallathank you again12:31
joschif it doesn't work like that, please ping me again :)12:31
amospallajosch: I just ran it without arguments, it show help and exits. Help includes "Without arguments only /boot/flash.bin will be updated unless --offline is given.", but I ran it without any argument and it did not update, but show help.12:34
amospallaI'm not forcing this tool, scares me a bit these kind of upgrades.12:35
amospallajosch: https://paste.debian.net/1327294/12:35
- mjw (QUIT: Ping timeout: 272 seconds) (~mjw@185.59.39.53)12:39
joschamospalla: thank you, it seems you found a bug!12:44
joschthe pocket is the first platform where reform-flash-uboot should do nothing, so bugs are expected i guess :)12:44
minutebtw other people are running into the gtk3/waybar crash issue https://github.com/Alexays/Waybar/issues/354212:49
amospallaoh nice.12:51
amospallathank you12:51
- mtm (QUIT: Ping timeout: 246 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net)14:02
+ mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net)14:06
- bkeys (QUIT: Quit: bkeys) (~Thunderbi@45.134.140.153)14:21
+ bkeys (~Thunderbi@45.134.140.153)14:22
+ andypiper (~andypiper@45.148.12.75)14:38
minutejosch: there is a little typo in your debian bug reply, > Contorl: forwarded -115:03
minute(not sure if important)15:03
joschit is, thank you :)15:05
joschfixed, thank you!15:06
minutenp!15:15
minuteheh, wlmaker is in debian unstable15:19
+ skipwich_ (~skipwich@user/skipwich)15:31
- skipwich (QUIT: Ping timeout: 252 seconds) (~skipwich@user/skipwich)15:32
* skipwich_ -> skipwich15:32
+ skipwich_ (~skipwich@user/skipwich)15:44
- skipwich (QUIT: Ping timeout: 276 seconds) (~skipwich@user/skipwich)15:48
* skipwich_ -> skipwich15:48
- skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich)15:50
+ mjw (~mjw@185.59.39.53)15:54
- andypiper (QUIT: Ping timeout: 260 seconds) (~andypiper@45.148.12.75)16:01
+ skipwich (~skipwich@user/skipwich)16:16
- skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich)16:19
+ skipwich (~skipwich@user/skipwich)16:20
- skipwich (QUIT: Remote host closed the connection) (~skipwich@user/skipwich)16:22
+ skipwich (~skipwich@user/skipwich)16:25
^alexis wlmaker what it sounds like? a wayland compositor that's like WindowMaker?16:25
ch> Wayland compositor inspired by Window Maker16:33
chscreenshot on https://github.com/phkaeser/wlmaker looks good16:33
jfredoh that does look fun16:37
mhoyejosch: would you happen to have a link to current-best SD boot image for the first-gen iMX8 reform?16:46
mhoyeI've done _something_ to this machine but I'm not sure what.16:46
+ andypiper (~andypiper@45.148.12.75)17:54
joschmhoye: i'm afraid the CI does not keep known-good images, so the latest image *is* the best image18:23
joschmhoye: what issue are you encountering?18:24
mhoyeStill sorting that out. 19:26
andypiperACTION TIL about wlmaker... nostalgia19:26
mhoyeI'll tell you that drilling out my heat sink and permanently installing a serial adapter is proving to be quite prescient, because this thing does _not_ want to display anything after uBoot19:27
- andypiper (QUIT: Quit: My device has gone to sleep. Zzzz…) (~andypiper@45.148.12.75)19:44
minutetrying the CI built pocket rk3588 image with > setenv bootargs keep_bootcon earlycon=uart8250,mmio32,0xfeb50000 clk_ignore_unused loglevel=719:49
minutebut output stops after > [    1.641498] Run /init as init process19:49
+ andypiper (~andypiper@45.148.12.75)19:51
minuteah, the dts still has hdmi1 in there which is, as we know, not functional anymore19:52
minutenaja, something for another day19:52
- ch (QUIT: Ping timeout: 252 seconds) (~ch@user/meow/ch)19:53
- andypiper (QUIT: Client Quit) (~andypiper@45.148.12.75)19:53
- Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66)19:56
+ Gooberpatrol66 (~Gooberpat@user/gooberpatrol66)19:56
minutenow trying with vendor uboot and extlinux.conf, it proceeds further19:57
minutebut hangs at hdmi, expected19:57
Twodisbetterminute: do you see any light on the horizon for the suspend situation on the pocket? No pressure or anything. I know it is a tough nut to crack. The Pocket is awesome just as it is already. 19:57
minute> Resizing the filesystem on /dev/mmcblk1p2 to 7296768 (4k) blocks.20:01
minuteok, now it also boots with the mainline uboot... needed a different dtb20:05
joschminute: you are using the latest system image? Because i had a weird failure where at first boot it would *not* do the resizing...20:07
minutejosch: yes20:07
- robin (QUIT: Ping timeout: 246 seconds) (~robin@user/terpri)20:10
+ robin (~robin@user/terpri)20:11
joschhuh okay, then it was just a me-problem -- that's good20:12
mhoyewhile I'm sorting this out, I would like to reiterate that the manual that shipped with the first Reform is just so good.20:12
josch+120:14
joschalso: fun things that one finds during a copyright review:20:14
joschreform2-lpc-fw/tools/lpcrc/lpcrc.exe20:14
joschPE32 executable (console) Intel 80386 (stripped to external PDB), for MS Windows, 5 sections20:14
josch:D20:14
mhoyeFor some reason you've reminded me of an old photo of a famous hip-hop artist leafing through a high-roller wad finding a one dollar bill.20:16
mhoyeWith an incomprehending look on his face. "How did this get in here... wait what the hell even _is_ this?"20:17
joschwant to see even more wild things?20:18
joschin the reform git repos i mean20:18
joschhttps://source.mnt.re/reform/reform/-/blob/master/reform2-lpc-fw/src/drivers/rf/nfc/pn532/mem_allocator/bget.c#L2920:19
josch"This is the book of the generations of Adam.   In the day that God created man, in the likeness of God made he him"20:19
joscho020:19
minute:020:20
joschcopyright reviews are usually super boring but the reform git is a gold mine :D20:21
joschminute: i can compile lpc firmware just fine without bget -- would you like a merge request removing it?20:35
+ Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66)20:41
- Gooberpatrol66 (QUIT: Ping timeout: 260 seconds) (~Gooberpat@user/gooberpatrol66)20:41
+ ch (~ch@user/meow/ch)21:13
minutejosch: yes, would like21:44
mhoye[    0.705380] simple-framebuffer fe540000.framebuffer: probe with driver simple-framebuffer failed with error -1221:47
mhoye]21:47
mhoyeThat's where I lock up.21:58
joschwant more fun? there is this and other libraries in the reform-lpc source: ./reform2-lpc-fw/cmsis/libs/libarm_cortexM0l_math.a21:59
joschjust some blobs21:59
joschthey are part of cmsis and reform2/board_reform2.h needs LPC11Uxx.h from within it, so i guess i have to figure out how to build this...22:00
minutejosch: oof i wanted to reduce that codebase for a long time but glad you're spelunking there22:03
mhoyejosch: At one point when I was at Mozilla, we found some ... unfortunate imagery in the repo, and it didn't take long to put together file that just listed all the images in the repo in one big page.22:04
mhoyeSome _magnificent_ skeletons in that closet lemme tell you.22:04
minute:D22:05
minutemhoye: weird, simple-framebuffer shouldn't be used normally, it's IIRC a mechanism to pass a preallocated display/framebuffer from uboot to linux22:06
- mrbcmorris (QUIT: Read error: Connection reset by peer) (~mrbcmorri@user/mrbcmorris)22:15
mhoyeYeah, I'm trying to figure out how to tell uboot not to do that.22:15
mhoyeIs there a fail-and-keep-going option in uboot? I don't need working video here, I just need the boot sequence to get far enough a long that I can ssh in.22:16
minutemhoye: normally it should contiinue booting, a single driver shouldn't prevent that22:20
minutemhoye: you already have loglevel set back to 7 i assume?22:20
minutemhoye: and console forced to serial?22:20
joschon imx8mq i think that was console=ttymxc0,11520022:25
joschminute: so, the lpc firmware compiles fine without linking against the static libraries arm_cortexM0l_math.a and RTX_CM0.a -- is it possible that the code will now blow up at runtime because something is missing that i'm not informed about during compile time?22:31
+ mrbcmorris (~mrbcmorri@user/mrbcmorris)22:35
joschon github they also just ship these blobs: https://github.com/ARM-software/CMSIS_4/tree/master/CMSIS/Lib/GCC22:35
minutejosch: hmm i can try it out next week22:41
mhoyeI've got console output on a different machine.22:43
joschminute: okay, i documented everything i have so far in https://source.mnt.re/reform/reform/-/merge_requests/71 so that it doesn't get lost22:44
minutenice22:53
mhoyeBut I don't think I can set uboot loglevels at runtime? At least, uboot docs say I can, I guess, but my uboot doesn't admit the existence of log or loglevel in its environment when I'm in it.22:55
minutemhoye: linux kernel loglevel i mean22:56
minutemhoye: it's part of the kernel commandline22:56
minutemhoye: in uboot that would go into the variable bootargs22:57
joschbut you are not guaranteed that boot.scr or extlinux.conf respects u-boot's ${bootargs}22:57
joschso maybe first make sure that which ever of the two you are using does respect ${bootargs} or set your loglevel there in the first place22:58
mhoyeLet's find out!23:23
mhoyeI think the first thing I need to do is understand what "error -12" is.23:29

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