| minute | openCL for valhall and thus rk3588 https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30088 | 00:10 |
|---|---|---|
| - natalie (QUIT: Ping timeout: 246 seconds) (~natalie@user/natalie) | 00:38 | |
| bkeys | josch: Alright, so something in the Fedora defaults must have been incompatible with the Reform config so now I'm just building a straight Fedora kernel with the exact config that Reform runs on debian | 01:29 |
| bkeys | we are getting closer | 01:30 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 01:38 | |
| - brennen (QUIT: Server closed connection) (~brennen@user/brennen) | 01:40 | |
| + brennen (~brennen@user/brennen) | 01:42 | |
| - bluerise (QUIT: Ping timeout: 258 seconds) (~bluerise@p5b0ac4d1.dip0.t-ipconnect.de) | 01:43 | |
| + bluerise (~bluerise@pc19f8d58.dip0.t-ipconnect.de) | 01:44 | |
| - swivel (QUIT: Server closed connection) (~swivel@shells.gnugeneration.com) | 01:50 | |
| + swivel (~swivel@shells.gnugeneration.com) | 01:50 | |
| - mtm (QUIT: Ping timeout: 244 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 02:02 | |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 02:05 | |
| - vkoskiv (QUIT: Server closed connection) (~vkoskiv@87-100-167-110.bb.dnainternet.fi) | 02:37 | |
| + vkoskiv (~vkoskiv@87-100-167-110.bb.dnainternet.fi) | 02:37 | |
| - schneider (QUIT: Server closed connection) (~schneider@irc2.xtort.eu) | 02:37 | |
| + schneider (~schneider@irc2.xtort.eu) | 02:38 | |
| - anzu (QUIT: Server closed connection) (~anzu@melkki.cs.helsinki.fi) | 02:49 | |
| + anzu (~anzu@melkki.cs.helsinki.fi) | 02:50 | |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 02:55 | |
| - wose (QUIT: Server closed connection) (wose@mail.zuendmasse.de) | 02:56 | |
| + wose (wose@mail.zuendmasse.de) | 02:57 | |
| violet | minute: alright removed D5 and it's still busted in the same way | 03:08 |
| violet | so that wasn't it unfortunately | 03:08 |
| violet | but it was a fun learning experience! | 03:10 |
| violet | i dont think ill get very far trying to debug this board further so i'll wait until 2.6 is out and mess with the pocket reform in the meantime | 03:12 |
| violet | and in the mean time maybe someone in seattle happens to have a spare motherboard | 03:12 |
| - jfred-linode (QUIT: Server closed connection) (quassel@libera/sponsor/jfred) | 03:15 | |
| + jfred-linode (quassel@libera/sponsor/jfred) | 03:16 | |
| - shdw (QUIT: Server closed connection) (~shdw@static.218.156.216.95.clients.your-server.de) | 03:27 | |
| + shdw (~shdw@static.218.156.216.95.clients.your-server.de) | 03:27 | |
| - nsc (QUIT: Ping timeout: 258 seconds) (~nicolas@i5C74DC43.versanet.de) | 03:53 | |
| + nsc (~nicolas@i5C74DCE7.versanet.de) | 03:55 | |
| - Manis (QUIT: Quit: Gateway shutdown) (01a66df340@84-73-19-122.dclient.hispeed.ch) | 04:04 | |
| + Manis (01a66df340@84-73-19-122.dclient.hispeed.ch) | 04:08 | |
| kop316 | minute: I have my hyper pocket reform! but it looks like there isn't a TB ssd HD in there? | 04:22 |
| - sknebel (QUIT: Server closed connection) (~quassel@v22016013254630973.happysrv.de) | 04:32 | |
| + sknebel (~quassel@v22016013254630973.happysrv.de) | 04:33 | |
| kop316 | whoops, nevermind I see it, the install just wasn't on the SSD | 04:37 |
| - yankcrime (QUIT: Server closed connection) (~nick@gw.tetromino.io) | 04:41 | |
| + yankcrime (~nick@gw.tetromino.io) | 04:44 | |
| - dozens (QUIT: Server closed connection) (~dozens@tilde.town) | 06:24 | |
| + dozens (~dozens@tilde.town) | 06:24 | |
| - eschaton (QUIT: Server closed connection) (eschaton@2600:3c01::f03c:91ff:fefd:5d92) | 06:31 | |
| + eschaton (~eschaton@li541-49.members.linode.com) | 06:31 | |
| - dominicm (QUIT: Server closed connection) (45ee0d70e9@2a03:6000:1812:100::3e6) | 06:34 | |
| + dominicm (45ee0d70e9@2a03:6000:1812:100::3e6) | 06:35 | |
| - jnerula (QUIT: Server closed connection) (~jnerula@li1009-93.members.linode.com) | 06:52 | |
| + jnerula (~jnerula@li1009-93.members.linode.com) | 06:53 | |
| - b0 (QUIT: Server closed connection) (~b0@user/b0) | 07:13 | |
| + b0 (~b0@user/b0) | 07:13 | |
| - cmahns (QUIT: Server closed connection) (8fe824803c@2a03:6000:1812:100::10cd) | 07:19 | |
| + cmahns (8fe824803c@2a03:6000:1812:100::10cd) | 07:20 | |
| - svp (QUIT: Server closed connection) (sid537750@id-537750.uxbridge.irccloud.com) | 07:23 | |
| + svp (sid537750@id-537750.uxbridge.irccloud.com) | 07:23 | |
| - kuno (QUIT: Server closed connection) (sid299442@id-299442.helmsley.irccloud.com) | 07:39 | |
| + kuno (sid299442@id-299442.helmsley.irccloud.com) | 07:39 | |
| - henesy (QUIT: Server closed connection) (d7619ffbc2@2a03:6000:1812:100::143) | 07:57 | |
| + henesy (d7619ffbc2@2a03:6000:1812:100::143) | 07:57 | |
| - jfred (QUIT: Server closed connection) (sid534649@libera/sponsor/jfred) | 08:05 | |
| + jfred (sid534649@libera/sponsor/jfred) | 08:05 | |
| - dook (QUIT: Server closed connection) (c9db4a6425@2a03:6000:1812:100::1222) | 08:25 | |
| + dook (c9db4a6425@2a03:6000:1812:100::1222) | 08:26 | |
| + chomwitt (~chomwitt@2a02:587:7a2b:1b00:1ac0:4dff:fedb:a3f1) | 09:00 | |
| - iank (QUIT: Server closed connection) (~iank@fsf/staff/iank) | 09:06 | |
| + iank (~iank@fsf/staff/iank) | 09:06 | |
| - gsora (QUIT: Server closed connection) (~gsora@user/gsora) | 09:07 | |
| + gsora (~gsora@user/gsora) | 09:07 | |
| - chomwitt (QUIT: Ping timeout: 244 seconds) (~chomwitt@2a02:587:7a2b:1b00:1ac0:4dff:fedb:a3f1) | 09:17 | |
| + chomwitt (~chomwitt@2a02:587:7a17:e800:1ac0:4dff:fedb:a3f1) | 09:20 | |
| - violet (QUIT: Server closed connection) (~vi@user/meow/violet) | 09:26 | |
| + violet (~vi@user/meow/violet) | 09:26 | |
| - pandora (QUIT: Server closed connection) (uid585533@id-585533.ilkley.irccloud.com) | 09:33 | |
| + pandora (uid585533@id-585533.ilkley.irccloud.com) | 09:33 | |
| - robin (QUIT: Server closed connection) (~robin@user/terpri) | 09:40 | |
| + robin (~robin@user/terpri) | 09:41 | |
| - amk (QUIT: Server closed connection) (~amk@user/amk) | 09:42 | |
| + amk (~amk@user/amk) | 09:42 | |
| - chomwitt (QUIT: Ping timeout: 248 seconds) (~chomwitt@2a02:587:7a17:e800:1ac0:4dff:fedb:a3f1) | 09:43 | |
| + chomwitt (~chomwitt@2a02:587:7a0f:c800:1ac0:4dff:fedb:a3f1) | 09:45 | |
| minute | violet: ok, thanks for trying! did you see there's a seattle meetup being organized? | 09:50 |
| josch | minute: already ahead of you :) https://community.mnt.re/t/mnt-seattle-meetup-pocket-edition-summer-2024/2269/4 | 09:53 |
| minute | josch: violet: hah, just saw that too | 09:58 |
| josch | completely unrelated, when is the next berlin reform meetup? ;) | 10:00 |
| chartreuse | Messed up my testing by accidentally disconnecting the charger and let the reform run flat and turn off. But haddn't had a lockup yet using brightnessctl but that was only a day or two | 10:11 |
| chartreuse | Do need to track down one more JGNE 1800 cell as one consistently runs flat before the rest. I should have just bought more than two replacements while the store I ordered them from had them | 10:12 |
| - chomwitt (QUIT: Ping timeout: 264 seconds) (~chomwitt@2a02:587:7a0f:c800:1ac0:4dff:fedb:a3f1) | 10:13 | |
| - austriancoder (QUIT: Server closed connection) (sid152545@id-152545.hampstead.irccloud.com) | 10:14 | |
| + austriancoder (sid152545@id-152545.hampstead.irccloud.com) | 10:14 | |
| + natalie- (~natalie@user/natalie) | 10:25 | |
| - _alice (QUIT: Server closed connection) (sid544964@id-544964.ilkley.irccloud.com) | 10:33 | |
| + _alice (sid544964@id-544964.ilkley.irccloud.com) | 10:33 | |
| + chomwitt (~chomwitt@2a02:587:7a0f:c800:1ac0:4dff:fedb:a3f1) | 11:15 | |
| - chomwitt (QUIT: Ping timeout: 265 seconds) (~chomwitt@2a02:587:7a0f:c800:1ac0:4dff:fedb:a3f1) | 11:30 | |
| - lexik (QUIT: Server closed connection) (~lexik@93.185.97.218) | 11:38 | |
| + lexik (~lexik@93.185.97.218) | 11:38 | |
| - Aard (QUIT: Server closed connection) (~bwachter@edna-edison.lart.info) | 11:40 | |
| + Aard (~bwachter@edna-edison.lart.info) | 11:40 | |
| + skipwich_ (~skipwich@user/skipwich) | 11:54 | |
| - skipwich (QUIT: Ping timeout: 252 seconds) (~skipwich@user/skipwich) | 11:54 | |
| + mjw (~mjw@gnu.wildebeest.org) | 12:18 | |
| - Zaba (QUIT: Server closed connection) (80b9b4b35e@2a03:6000:1812:100::116) | 12:31 | |
| + Zaba (80b9b4b35e@2a03:6000:1812:100::116) | 12:31 | |
| minute | josch: ha, good question, we should organize one at some point | 12:31 |
| minute | the next planned appearances are me @ kicon in bochum in september (i think 19th/20th sth like that) and chaos communication congress in december | 12:32 |
| josch | that is good to know, thanks! | 12:34 |
| josch | i just installed kicad 8 which got uploaded to bookworm-backports a few days ago | 12:35 |
| josch | finally i can have a look at the kicad files in the reform git again \o/ | 12:35 |
| minute | yaaay | 12:36 |
| minute | i will now try and see if i can access the nvme in a311d in uboot | 12:36 |
| minute | because i wasted many hours yesterday to make emmc work reliably in uboot... i just couldn't figure it out why it wouldn't work sometimes | 12:36 |
| minute | bizarre because the ROM can always load uboot from that same emmc | 12:37 |
| minute | so something is wrong with the driver in uboot | 12:37 |
| minute | > PCIE-0: Link up (Gen2-x1, Bus0) | 12:38 |
| minute | > 01.00.00 0x10ec 0x5765 Mass storage controller 0x08 | 12:38 |
| minute | ok :3 | 12:38 |
| josch | promising! | 12:39 |
| minute | ok, need to compile in nvme support | 12:39 |
| josch | minute: any chance for you and your team at MNT to catch some breath some time soon? There seems to have been a *lot* going on in the past weeks and it's crazy you are still pumping out content under so much pressure | 12:40 |
| + chomwitt (~chomwitt@2a02:587:7a0f:c800:1ac0:4dff:fedb:a3f1) | 12:44 | |
| minute | josch: well, that's an ongoing discussion... | 12:47 |
| minute | in the meantime the team has prepared 30 more hyper pockets that are ready for shipping :0 | 12:47 |
| minute | my main nagging issue is that we are very behind on reform shipments | 12:47 |
| minute | incl rk3588 | 12:47 |
| minute | it's a double whammy of me needing to finish respins of mb and especially rk3588 | 12:47 |
| minute | for future proofing | 12:48 |
| josch | uff... :/ i hope you do not take people critizing you and your team for taking so long to solve issue X to heart -- there is no way to do everything at once... | 12:49 |
| minute | yeah, but it has unavoidable financial ramifications too. people are cancelling their reform orders from time to time and we have to refund big sums | 12:50 |
| minute | i don't blame them. but it puts extra pressure on me to hurry up | 12:50 |
| - chomwitt (QUIT: Ping timeout: 265 seconds) (~chomwitt@2a02:587:7a0f:c800:1ac0:4dff:fedb:a3f1) | 12:53 | |
| minute | => nvme info | 12:54 |
| minute | Device 0: Vendor: 0x10ec Rev: VC2S038E Prod: 2310VC2S038E0021 | 12:54 |
| minute | nice | 12:54 |
| josch | reminds me a bit of the first 3 years with my daughter: "just a few more weeks and then we can finally breathe a bit and relax... riiight?" -- in any case i hope that you manage to find a plateau on which things end up being a bit smoother and more regular | 12:54 |
| minute | josch: thank you! i hope so too | 12:54 |
| josch | wow, nvme support in u-boot would be really nice! | 12:54 |
| minute | josch: nvme seems to work at least with a311d! | 12:55 |
| josch | i'll have to adjust reform-tools to put /boot on nvme then | 12:55 |
| minute | but i have to create a boot partition :D | 12:55 |
| - nocko (QUIT: Server closed connection) (~nock@user/nocko) | 12:56 | |
| + nocko (~nock@user/nocko) | 12:56 | |
| - mjw (QUIT: Ping timeout: 255 seconds) (~mjw@gnu.wildebeest.org) | 13:01 | |
| josch | in ./include/configs/meson64.h there is already BOOT_TARGET_NVME(func) in #define BOOT_TARGET_DEVICES, so distro_boot may find boot.scr on nvme out-of-the-box | 13:02 |
| * Guest8052 -> mjw | 13:03 | |
| minute | josch: yeah, i just noticed it tried to boot from nvme after failing emmc boot | 13:09 |
| josch | XD | 13:10 |
| minute | yes, it just loaded boot.scr | 13:10 |
| josch | woah nice! | 13:10 |
| minute | hmm but that seems to hang | 13:10 |
| josch | all that was needed was to enable nvme support in the u-boot config? | 13:10 |
| minute | josch: yeah. but it doesn't work all the way yet | 13:11 |
| minute | i need to remember the command for boot.scr tracing | 13:11 |
| minute | ah yeah, setenv xtrace | 13:12 |
| minute | oh it loads stuff, it's just very very very slow | 13:13 |
| minute | 33499648 bytes read in 37113 ms (880.9 KiB/s) | 13:13 |
| minute | > Starting kernel ... | 13:14 |
| minute | it works! | 13:14 |
| josch | \o/ | 13:14 |
| minute | takes over a minute to load kernel + initramfs | 13:15 |
| josch | that's concerning... | 13:15 |
| minute | ah, it's a general problem | 13:16 |
| minute | just benchmarked this disk in linux. read rate 754kB/s loool | 13:16 |
| minute | write rate 68.9MB/s :D | 13:16 |
| josch | o0 | 13:17 |
| minute | i didn't know that nvme can be so slow :D | 13:17 |
| noam_ | minute: in case you didn't notice, the line you posted above shows it running in PCI gen 2? | 13:20 |
| noam_ | and with an x1 link?? | 13:20 |
| noam_ | I could be wrong but that tells me it has like, 1/16th the intended bandwidth, at best? | 13:21 |
| noam_ | I could be totlly misreading ofc :) | 13:21 |
| minute | noam_: that's still 5GT/s | 13:21 |
| minute | noam_: these socs only have x1 lanes | 13:21 |
| minute | the first one that has more is rk3588 :3 | 13:22 |
| minute | gen2 x1 link should still give around 500MB/s, certainly not 1MB/s :D | 13:22 |
| minute | aha! | 13:23 |
| minute | it was because i had used an old rmc4 that didn't have pcie reset wired up | 13:23 |
| noam_ | rmc4? | 13:23 |
| minute | > 37161806 bytes read in 104 ms (340.8 MiB/s) | 13:23 |
| minute | noam_: sorry, rcm4 | 13:23 |
| noam_ | rcm4? :D | 13:23 |
| minute | noam_: https://shop.mntre.com/spree/products/390/large/rcm4-bpi-levels2.jpg?1690898139 | 13:23 |
| noam_ | Compute module? | 13:23 |
| minute | rcm4 is the reform cm4 adapter | 13:24 |
| noam_ | Ahh, old adapter? | 13:24 |
| noam_ | gotcha | 13:24 |
| minute | yeah development version | 13:24 |
| minute | that'll have to go to the trash ;__; | 13:24 |
| noam_ | :( | 13:27 |
| josch | on the plus side, i now added another column to this table: https://reform.debian.net/d-i/#manual-partitioning-for-boot | 13:27 |
| noam_ | minute: is it not usable without pcie? | 13:28 |
| minute | noam_: it is. but that requires microsd booting | 13:28 |
| minute | and i wanted to get rid of that | 13:28 |
| noam_ | Sure, but that doesn't need to be garbage? :P | 13:28 |
| josch | i always do microsd booting on my a311d :D | 13:28 |
| minute | noam_: ah, i don't want to keep too many half broken things around, it turns out to be costly to deal with them | 13:29 |
| josch | maybe a "Fundgrube" section in the shop where you sell things that still work well enough with a few limitations? | 13:30 |
| minute | noooo ahaha | 13:31 |
| minute | i mean, that's a nice romantic idea but the support for that would be a nightmare | 13:32 |
| minute | we might do this for mechanical parts though | 13:32 |
| minute | with some dings and dents | 13:32 |
| noam_ | minute: I mean I'd totally accept it on the provision that no support would be offered whatsoever :D | 13:33 |
| minute | josch: while i have your attention, do you know how one can configure boot device order in uboot? | 13:33 |
| minute | noam_: what could people do with that stuff? soldering/electronics bug hunting practice? | 13:33 |
| josch | yes, it's BOOT_TARGET_DEVICES | 13:33 |
| minute | josch: oh oh | 13:34 |
| josch | that's the distro_boot boot order | 13:34 |
| noam_ | minute: cheaper way to play with getting 9front running on the rpi4 compute module in the reform :P | 13:34 |
| minute | half-broken :D | 13:34 |
| noam_ | If it's just PCIe then it doesn't really matter :) | 13:34 |
| josch | minute: you have to define that before including config_distro_bootcmd.h, see ./include/configs/meson64.h for example | 13:35 |
| noam_ | but also also I could genuinely use bug hunting practice :P | 13:35 |
| minute | josch: ah, it's not a config setting? | 13:36 |
| + chomwitt (~chomwitt@2a02:587:7a22:1700:1ac0:4dff:fedb:a3f1) | 13:36 | |
| noam_ | Hmm, I might tape one of the trackball buttons on temporarily >_< | 13:37 |
| jfred | Woo I just got my hyper pocket shipping notification! :D | 13:37 |
| noam_ | Snapped one of the connectors when cleaning the ball, oops | 13:37 |
| minute | jfred: niiice | 13:37 |
| noam_ | not going to be able to fix it properly for a while :/ | 13:37 |
| minute | noam_: oh no, you mean the stems of the keycap? | 13:37 |
| josch | minute: yes, it's not a config setting but a board-specific hard-coded thing | 13:38 |
| noam_ | yea | 13:38 |
| noam_ | it's fine, who needs a left click anyways? :D | 13:38 |
| minute | noam_: oof oof | 13:39 |
| noam_ | I cn just leave the keycap off i guess | 13:40 |
| noam_ | button works fine :D | 13:40 |
| josch | minute: yes, for example for imx8mq it's defined in include/configs/nitrogen8m_som.h | 13:41 |
| josch | ah i had already answered, sorry | 13:42 |
| josch | the little one just came home, my brain switched off | 13:42 |
| josch | i should go afk for a bit _o/ | 13:42 |
| minute | josch: enjoy! | 13:50 |
| - mtm (QUIT: Ping timeout: 252 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 14:04 | |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 14:05 | |
| - noam_ (QUIT: Server closed connection) (81879d1ffa@2a03:6000:1812:100::dfc) | 14:07 | |
| + noam (81879d1ffa@2a03:6000:1812:100::dfc) | 14:07 | |
| grimmware | Seen a couple of people on masto get their hyper shipping notifications | 14:12 |
| grimmware | I do not have mine yet ;_; | 14:12 |
| grimmware | jfred: can’t wait to see your pocket with the custom backplate | 14:13 |
| rah | is there a limit on the current that a particular USB port on the MNT Reform 2 can provide or is there a total limit for all ports? | 14:26 |
| + ericsfraga (~user@2.28.216.207) | 14:45 | |
| + mark_ (~mjw@gnu.wildebeest.org) | 14:51 | |
| jfred | grimmware: :) | 14:51 |
| - Gooberpatrol66 (QUIT: Ping timeout: 248 seconds) (~Gooberpat@user/gooberpatrol66) | 14:53 | |
| minute | grimmware: we have here 30 more to ship out, just need to prepare customs docs etc :3 | 14:58 |
| - midfavila (QUIT: Server closed connection) (midfavila@sdf.org) | 15:02 | |
| + midfavila (midfavila@sdf.org) | 15:03 | |
| [tj] | do they still travel across the atlantic and back? | 15:08 |
| - mark_ (QUIT: Ping timeout: 272 seconds) (~mjw@gnu.wildebeest.org) | 15:15 | |
| minute | [tj]: sure they do! | 15:30 |
| [tj] | :D | 15:31 |
| minute | into mouser's texas warehouse and back | 15:31 |
| bkeys | Woo, I finally figured out how to set custom kernel configs in the Fedora kernel | 15:31 |
| gsora | wait, even for europe? o.O | 15:31 |
| [tj] | gsora: yeah | 15:32 |
| [tj] | I think all crowdsupply stuff does this | 15:32 |
| bkeys | It's building with the exact same config as what's in the system image | 15:32 |
| gsora | i don't think my precursor did that | 15:32 |
| gsora | it came from a CS EU warehouse iirc | 15:33 |
| minute | CS doesn't have a EU warehouse | 15:40 |
| minute | maybe it was routed on the way back via mouser munich? | 15:40 |
| minute | but we can't send stuff there | 15:41 |
| minute | josch: ok, i edited the boot sequence in meson64.h, very nice | 15:42 |
| minute | ok, 11 seconds to kernel output on screen from cold boot | 15:46 |
| - jjbliss (QUIT: Server closed connection) (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 15:58 | |
| + jjbliss (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 15:58 | |
| gsora | minute: oh i see, to be fair it might've been imported by CS/mouser first and then shipped to me... thankfully my reform2 has been bought from you folks directly, no messing up! | 16:07 |
| - plomlompom (QUIT: Server closed connection) (~plom@play.plomlompom.com) | 16:31 | |
| + plomlompom (~plom@play.plomlompom.com) | 16:31 | |
| - aperezdc (QUIT: Server closed connection) (~aperezdc@46.23.89.43) | 16:42 | |
| + aperezdc (~aperezdc@46.23.89.43) | 16:42 | |
| - sterni (QUIT: Server closed connection) (~quassel@user/sterni) | 16:47 | |
| + sterni (~quassel@user/sterni) | 16:47 | |
| rah | minute: is there a limit on the current that a particular USB port on the MNT Reform 2 can provide or is there a total limit for all ports? | 16:49 |
| + mark_ (~mjw@gnu.wildebeest.org) | 17:09 | |
| - Twodisbetter (QUIT: Server closed connection) (2cc0e4ea1c@irc.cheogram.com) | 17:38 | |
| - chartreuse (QUIT: Server closed connection) (~chartreus@S0106908d78501d1d.cg.shawcable.net) | 17:49 | |
| + chartreuse (~chartreus@S0106908d78501d1d.cg.shawcable.net) | 17:49 | |
| - mark_ (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 17:51 | |
| minute | i finally have a well working mouse again :0 | 17:51 |
| minute | rah: depends on whether you have mb2.0 or 2.5 | 17:52 |
| minute | rah: for mb2.5 it is limited by a fuse and by the max amps of the 5v regulator | 17:52 |
| minute | rah: mb2.0 had individual current limiters on the usb ports | 17:52 |
| bluerise | https://lore.kernel.org/ath12k/d44fdc0b-b4a7-4f36-9961-c5c042ed43df@quicinc.com/T/#u | 17:53 |
| bluerise | ARGH | 17:53 |
| bluerise | I'm going to order an RK3588 SoM now, can't be arsed with that Qualcomm shit. | 17:53 |
| minute | bluerise: wtf! | 17:54 |
| minute | rah: the fuse is 4A combined | 17:55 |
| minute | rah: and the 5V rail is designed for 5A total, but it is shared between SOM, backlight and USB devices, so your USB power budget is more likely to be between 2-3A | 17:56 |
| minute | (total) | 17:56 |
| bluerise | > Your order has been processed successfully | 17:56 |
| minute | rah: on motherboard 2.0 we had TPS2561 on the ports, which would trip at around 2A per port | 17:59 |
| - murphnj (QUIT: Server closed connection) (~murph@user/murphnj) | 18:07 | |
| + murphnj (~murph@user/murphnj) | 18:07 | |
| chartreuse | bare USB hub and some 4 pin JST connectors finally arrived so I can add more internal USB ports as needed. Though don't have any in mind *yet* | 18:20 |
| chartreuse | Did also have a thought of how to maybe add a third speaker as a "subwoofer" type setup. At least with jack detection removed I was able to drive both the headphones and speaker. I suspect I could put an amp on one of the headphone outputs, set that volume in alsamixer when in speaker mode, and then use the jack detect to also disable the amp when headphones are plugged in | 18:22 |
| chartreuse | Probably just a silly idea but kinda want to try it, got a little mono class d amp board | 18:25 |
| minute | chartreuse: neat! | 18:26 |
| minute | boombox reform | 18:26 |
| chartreuse | Heh, I was going to try first with a little 2W exciter speaker attached to the bottom case as the driver. Didn't sound too bad testing with an external lm386 based amp and aligator clips to it just holding it in place | 18:27 |
| rah | minute: I see, thanks | 18:47 |
| rah | I believe I have a 2.0 motherboard and only need 1A so it looks like I'll be OK | 18:47 |
| - cwebber (QUIT: Server closed connection) (~Christine@user/cwebber) | 18:49 | |
| + cwebber (~Christine@user/cwebber) | 18:51 | |
| minute | bluerise: thanks for grabbing a rk3588! | 18:54 |
| kop316 | If i may ask, is there a channel for MNT that is more like for support/new to reform requests? In lurking here, it looks like this is more development, and I would rather not spam this channel for things like that | 19:01 |
| minute | kop316: it's fine to ask about support stuff here! | 19:02 |
| kop316 | minute: ok, thanks! I'm still getting used to the pocket reform so there will likely be silly questions (like yesterday) asked | 19:03 |
| josch | kop316: we also have a lot of "i'm new to debian" quesitons here and i think that's fine :) | 19:15 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 19:20 | |
| + Twodisbetter (2cc0e4ea1c@irc.cheogram.com) | 19:46 | |
| minute | https://source.mnt.re/reform/reform-a311d-uboot/-/merge_requests/3 | 19:51 |
| minute | nvme as first boot medium is debatable, we could still put the sd card in front for emergency situations | 19:52 |
| chartreuse | sd or something external as first does make sense, like mine I updated the kernel to 6.9.9, and it wouldn't boot complaining about the dtb not being found (not certain if that would normally happen or if it was because I originally pointed /boot/dtb to my custom one from 6.9.7 before trying to point it back) | 19:53 |
| chartreuse | Need to try again, it might be because the 6.9.9 image didn't seem to include the non-hdmi dtb anymore (missing from /boot/dtbs/6.9.9*/ | 19:54 |
| chartreuse | Having nvme or emmc first would be fine though on the imx8mq theoretically since you can have a visible boot and prompt on there, but the a311d is blank until linux inits it right? | 19:55 |
| chartreuse | minute: Yep just checked and after installing the 6.9.9 package the new boot.scr is trying to load the non-hdmi dtb, (in /boot) /dtbs/6.9.9-mnt-reform-arm64/freescale/imx8mq-mnt-reform2.dtb. But only the -hdmi.dtb exists there. Also it's ignoring the /dtb or /dtb-6.9.9-mnt-reform-arm64 symlink | 20:01 |
| chartreuse | Since both of those point to the -hdmi.dtb | 20:01 |
| + gustav28 (~gustav@c-1134524e.019-141-67626730.bbcust.telenor.se) | 20:02 | |
| josch | chartreuse: no, because at the point where you get the prompt it already decided for the /boot partition and used extlinux.conf to load that prompt | 20:04 |
| josch | minute: what's the speed difference between putting nvme first or sd-card first? | 20:05 |
| minute | josch: maybe 1 second | 20:05 |
| chartreuse | Should be negligable | 20:05 |
| minute | i guess i'll comment out that patch or sth | 20:06 |
| chartreuse | Is there a way of maybe making it so holding down a key at boot would boot the sd card? | 20:07 |
| minute | that would require usb init which takes a while | 20:07 |
| chartreuse | Not sure why the 6.9.9 is missing the dtb when installed (linux-image-mnt-reform-arm64 from the apt repo). There was no changes besides the backlight fix to the reform-debian-packages repo since the 6.9.7 package which had no issue on my end (and was messing with locally to hack the dts for jack detect) | 20:11 |
| - digitalrayne (QUIT: Quit: じゃまたね) (~digitalra@vps-446f4f39.vps.ovh.ca) | 20:19 | |
| + digitalrayne (~digitalra@vps-446f4f39.vps.ovh.ca) | 20:19 | |
| chartreuse | No difference in the boot.scr besides the kernel version, seems like the only thing is that the linux-image package didn't install the non-hdmi dtb even though /boot/dtb-6.9.9-mnt-reform-arm64 is a symlink to the hdmi one | 20:22 |
| josch | minute: but maybe the other boot options after sd-card and before nvme add a lot more? | 20:22 |
| chartreuse | Hmm /usr/lib/linux-image-6.9.9-mnt-reform-arm64/freescale/imx8mq-mnt-reform2.dtb does exist though, not sure why the post install or whatever didn't copy that as well, or why boot.scr is defaulting to that one... | 20:26 |
| + mark_ (~mjw@gnu.wildebeest.org) | 20:28 | |
| * mjw -> Guest3536 | 20:28 | |
| - Guest3536 (QUIT: Killed (tantalum.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 20:28 | |
| * mark_ -> mjw | 20:28 | |
| + Guest3536 (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 20:28 | |
| josch | chartreuse: please ping me about this issue in about 2 hours -- we'll get to the bottom of this :) | 20:33 |
| josch | the thing that copies the dtb is flash-kernel | 20:33 |
| josch | and it uses the information in /etc/flash-kernel/machine to decide which dtb to copy | 20:34 |
| chartreuse | Alright, I think the main issue was something with reform-display-config not being run after that new kernel or such, I think maybe since I had it set to dual and updated the kernel which doesn't re-run it, or something else | 20:37 |
| chartreuse | I re-ran reform-display-config --emmc dual and now it did boot with the correct one. I'm guessing flash-kernel didn't get run after updating the kernel? It's not in postinst of the deb package | 20:38 |
| - robin (QUIT: Quit: Leaving) (~robin@user/terpri) | 20:43 | |
| josch | chartreuse: flash-kernel does not get run by the postinst of the deb package itself but by a hook of mkinitramfs in /etc/initramfs/post-update.d/flash-kernel | 21:25 |
| josch | chartreuse: theoretically you should only have to run reform-display-config once you want to switch between single and dual | 21:25 |
| josch | chartreuse: maybe "sudo reform-check" shows anything funny? | 21:25 |
| chartreuse | E: your currently loaded dtb is not the one referenced by flash-kernel | 22:13 |
| chartreuse | Which is odd since this is a boot just after changing to dual display config (again) | 22:13 |
| chartreuse | reform-check does 404 on trying to get the boundary-uboot flash.bin file as well (jobs/4745/artifacts/raw/flash.bin) | 22:14 |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-1134524e.019-141-67626730.bbcust.telenor.se) | 22:15 | |
| chartreuse | I wonder if it was something odd from me install the raw deb from the artifacts you had built before while testing the jack (before I just setup the kernel to make dtbs myself) | 22:15 |
| chartreuse | The system could have also been running a really old kernel before that from last year (6.1.0 as I don't see any previous kernels in the /boot, and this was from the older v3 system image from 2023 or such | 22:16 |
| chartreuse | Also noticed I was missing the headers package and that built some dkms module sbut that shouldn't have been the issue here | 22:17 |
| josch | yes, the concerning thing is the error you got -- it means that you have something configured in /etc/flash-kernel/machine which is not reflected in /proc/device-tree/model of your booted system | 22:18 |
| josch | in the past, this sometimes meant that the /boot as it got mounted by the system and on which flash-kernel operates is a different partition from the one that gets found by u-boot | 22:18 |
| josch | so maybe you have a boot.scr on sd-card but you are mounting the partition on emmc? | 22:20 |
| chartreuse | /boot is /mmcblk0p1 which is the emmc, and I have no sd card inserted. | 22:20 |
| chartreuse | And my nvme drive has no boot partition | 22:20 |
| josch | u-boot of imx8mq currently can not read nvme, so that is not a concern | 22:22 |
| chartreuse | I'm just manually running the commands from reform-check to see what is different | 22:22 |
| josch | if you do not have an sd-card inserted, and if your /boot is mmcblk0p1 then the mystery intensifies | 22:22 |
| josch | it will just be "Reform 2" versus "Reform 2 HDMI" | 22:22 |
| josch | probably reform-check should print this -- let me add that | 22:23 |
| chartreuse | /etc/flash-kernel/machine is "MNT Reform 2 HDMI" and /proc/device-tree/model is "MNT Reform 2" oddly... | 22:23 |
| chartreuse | I could have sworn I just checked before and it was the HDMI one before.. | 22:23 |
| josch | chartreuse: suppose you run "sudo flash-kernel" manually, what happens? | 22:23 |
| chartreuse | No errors and is using the -hdmi dtb | 22:24 |
| chartreuse | left me restart to see what it loads then | 22:24 |
| chartreuse | I wonder if I copied my custom dtb into /etc/flash-kernel/dtbs and somehow got the two backwards? Though the package should have overwritten that manual copy | 22:25 |
| josch | chartreuse: no, flash-kernel does not overwrite your dtbs in /etc (and it should not) | 22:25 |
| chartreuse | It doesn't but I think I might have when I copied the jack detect one in there to stop it overwriting my custom one. | 22:26 |
| chartreuse | Rebooting after that flash-kernel and same thing the proc one is the non-hdmi | 22:27 |
| chartreuse | I can remove my ones from /etc/flash-kernel/dtbs because maybe something is wrong with the -hdmi one in there though and use the kernel one | 22:27 |
| chartreuse | But those were copied straight from the kernel source tree for 6.9.7 after I built the jack detect | 22:28 |
| josch | chartreuse: what do you have in /etc/flash-kernel/dtbs right now? | 22:28 |
| chartreuse | imx8mq-mnt-reform2.dtb imx8mq-mnt-reform2-hdmi.dtb README.dtbs | 22:28 |
| chartreuse | Installing /etc/flash-kernel/dtbs/imx8mq-mnt-reform2-hdmi.dtb into /boot/dtbs/6.9.9-mnt-reform-arm64/freescale/imx8mq-mnt-reform2-hdmi.dtb | 22:29 |
| chartreuse | So it's picking the hdmi one from there | 22:29 |
| josch | good | 22:29 |
| josch | what do you get when you run readlink /boot/dtb | 22:29 |
| chartreuse | dtbs/6.9.9-mnt-reform-arm64/freescale/imx8mq-mnt-reform2-hdmi.dtb | 22:30 |
| josch | or readlink /boot/dtb-$(uname -r) | 22:30 |
| chartreuse | Same for the dtb-6.9.9... file | 22:30 |
| josch | huh | 22:30 |
| josch | then how does it load the non-hdmi one o0 | 22:30 |
| josch | could you maybe inspect your /boot/dtbs/6.9.9-mnt-reform-arm64/freescale/imx8mq-mnt-reform2-hdmi.dtb and have a look whether it is *really* the HDMI version? | 22:31 |
| josch | or whether the hdmi is maybe only in its name? | 22:31 |
| chartreuse | Those symlinks were pointing to the hdmi one when I first noticed the issue too, when the non hdmi one was missing from that boot directory | 22:31 |
| chartreuse | The files differ though | 22:31 |
| chartreuse | And when I was missing the non-hdmi one I had that boot issue from earlier where it was trying to load it | 22:32 |
| chartreuse | I can remove the non-hdmi one again and try and boot | 22:32 |
| chartreuse | $ strings imx8mq-mnt-reform2-hdmi.dtb | grep HDMI | 22:32 |
| chartreuse | ,MNT Reform 2 HDMI | 22:32 |
| josch | huh | 22:33 |
| josch | i'm starting to run out of ideas | 22:33 |
| chartreuse | When I looked at printenv in uboot it did say the fdt(something) was the non-hdmi one before. Let me check it again | 22:33 |
| josch | yeah, trying to remove the non-hdmi one from /etc/flash-kernel/dtbs sounds like a good next step but your flash-kernel output said it was copying /etc/flash-kernel/dtbs/imx8mq-mnt-reform2-hdmi.dtb so that should not change anything | 22:33 |
| chartreuse | I can remove the /etc ones so it uses the kernel ones too | 22:34 |
| chartreuse | Case the dtbs broke between 6.9.7 and 6.9.9 | 22:34 |
| chartreuse | Okay so in uboot's printenv fdtfile is freescale/imx8mq-mnt-reform2.dtb for some reason | 22:36 |
| josch | you figured it out \o/ | 22:36 |
| chartreuse | But how does that get set? Shouldn't reform-display-config change that? | 22:36 |
| josch | no, reform-display-config changes nothing in u-boot only /etc/flash-kernel/machine and then re-runs mkinitramfs | 22:37 |
| josch | on the contrary minute even changed u-boot such that you can no longer saveenv | 22:37 |
| josch | to avoid accidentally messing up u-boot environment | 22:38 |
| chartreuse | Because I've not manually set that uboot variable before, doesn't that variable come from somewhere in the system | 22:38 |
| josch | the u-boot variables do not come from the linux root file system but are stored in flash somewhere | 22:38 |
| chartreuse | And looking at the text part of boot.scr there's nothing overriding it there (and since the env is set before boot.scr is run) | 22:39 |
| - q66 (QUIT: Quit: WeeChat 4.0.2) (~q66@q66.moe) | 22:39 | |
| josch | maybe somebody else with imx8mq can run printenv to see how it should look like? | 22:39 |
| chartreuse | I'm running U-Boot 2018.07 MNT Reform 2023-01-25 (Jan 25 2023 - 10:32:55 +0000) so maybe I just need to flash a newer version | 22:40 |
| josch | chartreuse: here is one from the past: http://45.76.26.151/env | 22:40 |
| chartreuse | I think this was one of the first versions that had the screen and boot messages actually working | 22:40 |
| chartreuse | fdtpath=dtb-6.8.11-mnt-reform-arm64 | 22:41 |
| chartreuse | Seems weird for that env to have a hardcoded kernel version? | 22:41 |
| josch | prominently, that u-boot env by noam does not have ${fdtfile} set at all | 22:42 |
| chartreuse | I would have thought it'd point to dtb or such | 22:42 |
| noam | What about me? | 22:42 |
| noam | Ah, the newer vendor env? :/ | 22:42 |
| josch | noam: i'm currently happy that you shared your printenv output with everybody about a month ago :) | 22:42 |
| noam | <3 of course :) | 22:43 |
| josch | chartreuse: yes, i have no idea what could've set that | 22:43 |
| chartreuse | If I run strings on flash.bin in my /boot I do see that freescale/imx8mq-mnt-reform2.dtb | 22:43 |
| josch | yes, that is normal | 22:44 |
| chartreuse | I've never actually used the HDMI output and only recently tried switching to that kernel so perhaps this older flash.bin had it hard coded for some reason | 22:44 |
| chartreuse | Are you sure though? there's no freescale/imx8mq-mnt-reform2-hdmi.dtb string in that file | 22:44 |
| chartreuse | I suspect that's the line setting the fdtfile variable | 22:45 |
| josch | sec... | 22:45 |
| + Jonas___ (~Jonas@82-65-231-86.subs.proxad.net) | 22:45 | |
| chartreuse | If I check around that match the next line after that path is fdtfile, | 22:45 |
| chartreuse | MNT Reform 2.0 | 22:46 |
| chartreuse | imx8mq | 22:46 |
| chartreuse | freescale/imx8mq-mnt-reform2.dtb | 22:46 |
| chartreuse | fdtfile | 22:46 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 22:46 | |
| josch | chartreuse: yes, same on current flash.bin | 22:46 |
| josch | same order of strings | 22:46 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 22:47 | |
| chartreuse | Has anyone tried using the HDMI on a imx8mq with the current flash.bin? | 22:47 |
| chartreuse | Checking arandr I don't see a second output | 22:48 |
| chartreuse | Just wondering how many people actually upgraded their uboot boot sector that are using a imx8mq and are also using hdmi | 22:50 |
| chartreuse | Just downloaded the latest flash.bin and yeah you're right that that line is still there, perhaps this is a regresion that's over a year old | 22:59 |
| ch | shouldnt flash-kernel write a new fdtfile name into the list it generates anyway? | 23:00 |
| chartreuse | https://source.mnt.re/reform/reform-boundary-uboot/-/commit/19a4cf7978acb16fcf594530fd1922c3b90c9b79 | 23:01 |
| chartreuse | The whole uboot around that seems to be ignoring the possibility of a -hdmi.dtb existing | 23:01 |
| josch | chartreuse: but that's not a versioned one | 23:02 |
| josch | chartreuse: you have fdtpath=dtb-6.8.11-mnt-reform-arm64 | 23:02 |
| chartreuse | It sounds like for -hdmi to work it needs to be named as the non-hdmi version in the /dtbs/<kernel version>/freescale/ | 23:02 |
| josch | and it's not fdtpath but fdtfile in that commit | 23:02 |
| chartreuse | No I don't that's noams | 23:02 |
| chartreuse | I don't have a fdtpath | 23:03 |
| josch | oh a few messages ago you said fdtpath=dtb-6.8.11-mnt-reform-arm64 | 23:03 |
| chartreuse | Yeah I was quoting that from noam's as something odd | 23:03 |
| chartreuse | Since it had a hardcoded kernel version | 23:03 |
| noam | um | 23:03 |
| chartreuse | Sorry for the ping | 23:04 |
| noam | there's a nonzero chance that env wasn't entirely clean | 23:04 |
| noam | no, no | 23:04 |
| noam | don't be | 23:04 |
| noam | I never get attention!!!! | 23:04 |
| noam | /s | 23:04 |
| chartreuse | Which version of uboot are you running btw? Is it a pre Jan 2023 one? | 23:04 |
| noam | The um was me realizing I don't remember if I'd set anything manually for that :/ | 23:04 |
| chartreuse | The commit I linked is the blame for the fdtfile line that's still in the current one, which was only added in Jan 2023 | 23:05 |
| chartreuse | It feels like fdtfile and/or the boot.scr should be looking for /dtb or /dtb-${fk_kvers} rather than the full absolute path | 23:06 |
| chartreuse | (Notices the commit was done by josch XD) | 23:07 |
| josch | yes, sorry i was confused about fdtfile versus fdtpath | 23:08 |
| chartreuse | boot.scr is setting fdt path to be dtbs/${fk_kvers}/${fdtfile} IF fdfile is set, otherwise setting fdtpath to be dtb-${fkkvers}, and presumably that later would go to the correct symlink | 23:10 |
| josch | chartreuse: the magic comes later in boot.scr -- read the comment starting with "We need to set ${fdtfile} in uboot" | 23:10 |
| chartreuse | Yes, but the boot.scr fallback seems sensible? `setenv fdtpath dtb-${fk_kvers}` | 23:11 |
| josch | yes, that's what should select the symlink set up by flash-kernel | 23:11 |
| chartreuse | That's only used if fdtfile is NOT set though | 23:12 |
| chartreuse | 72 if test -n "${fdtfile}"; then | 23:12 |
| chartreuse | 73 setenv fdtpath dtbs/${fk_kvers}/${fdtfile} | 23:12 |
| chartreuse | 74 else | 23:12 |
| chartreuse | 75 setenv fdtpath dtb-${fk_kvers} | 23:12 |
| chartreuse | 76 fi | 23:12 |
| chartreuse | Or wait.. | 23:12 |
| josch | yes, it's below what you quoted | 23:12 |
| noam | chartreuse: thatwas a vendor u-boot from a module purchased a few months go | 23:13 |
| noam | ago* | 23:13 |
| noam | I don't recall the timestamp | 23:13 |
| noam | probably post-2023 | 23:13 |
| chartreuse | Also the commit comment is wrong for the first point, flash-kernel is not making that symlink | 23:13 |
| - rah (PART: !!unknown attribute: msg!!) (rah@verain.settrans.net) | 23:13 | |
| chartreuse | lrwxrwxrwx 1 root root 37 Jul 18 14:29 imx8mq-mnt-reform2-hdmi.dtb -> freescale/imx8mq-mnt-reform2-hdmi.dtb | 23:13 |
| chartreuse | There's this symlink but not a imx8mq-mnt-reform2.dtb one | 23:14 |
| chartreuse | In dtbs/6.9.9-mnt-reform-arm64/ | 23:14 |
| chartreuse | I've got a copy backed up of the boot.scr, I'm going to try commenting out the if and doing just the else case which I suspect will work | 23:15 |
| josch | chartreuse: the interesting symlink is the one set up by boot.scr as dtb-${fk_kvers} which is /boot/dtb-$(uname -r) which, in your case, points to dtbs/6.9.9-mnt-reform-arm64/freescale/imx8mq-mnt-reform2-hdmi.dtb which is how it should work | 23:16 |
| chartreuse | Yes, but that's in the else case of that if statement isn't it? For when ftbfile isn't set? Or am I reading the script wrong there | 23:18 |
| chartreuse | (Also forgot that boot.scr isn't just a text file and has a crc I need to recompute | 23:19 |
| chartreuse | I feel it's doing that top path fdtpath dtbs/${fk_kvers}/${fdtfile} which is a non-existant file assuming fdtfile is imx8mq-mnt-reform2.dtb | 23:20 |
| josch | chartreuse: you mean the else case in line 75? that one is irrelevant because fdtpath gets set in line 97 | 23:20 |
| chartreuse | I don't see fdtpath being set on line 97 (or near since I'm just opening boot.scr directly). 97 is part of the long load && load, etc line | 23:21 |
| josch | oooh? | 23:21 |
| chartreuse | The last time fdtpath is set is that line at 76 in boot.scr | 23:21 |
| josch | chartreuse: https://paste.debian.net/1323674/ | 23:22 |
| josch | you have that part? | 23:22 |
| chartreuse | https://pastebin.com/ZtKwucAJ | 23:22 |
| chartreuse | That's not part of the boot.scr I'm getting from the reform package | 23:23 |
| josch | chartreuse: your boot.scr is missing a very important bit | 23:23 |
| josch | chartreuse: you have /usr/share/flash-kernel/preboot.d/00reform2_preboot right? | 23:23 |
| josch | do you also have /etc/flash-kernel/preboot.d/00reform2_preboot ? | 23:23 |
| chartreuse | That boot scr was the one generated when I installed 6.9.9, I have a backup of my 6.9.7 one which only differs by the kernel name. | 23:24 |
| chartreuse | That file does exist with those contents in the usr share path yes, let me check the other | 23:24 |
| chartreuse | In the /etc one it's blank with just a comment to add any u-boot commands to be executed before boot | 23:24 |
| chartreuse | Likely from the v3 system image I started with | 23:25 |
| josch | chartreuse: did you not get a error about that from reform-check? | 23:25 |
| josch | chartreuse: what is your reform-tools version? | 23:25 |
| josch | chartreuse: the reform-check utility checks for precisely those things as you are not the only one who upgraded from before sysimage-v3 :) | 23:25 |
| chartreuse | https://pastebin.com/sEHTsvXX | 23:26 |
| josch | ooooh reform-check quits early because the download failed | 23:26 |
| josch | so you never get the error message | 23:26 |
| chartreuse | Reform check is failing because it can't get flash.bin (since the file isn't called flash.bin anymore IIRC in the latest builds but imx8mq-mnt-reform2-flash.bin | 23:26 |
| josch | yes | 23:27 |
| chartreuse | Okay I changed the url to match the correct filename and got a longer one | 23:29 |
| josch | chartreuse: this should be the right patch fixing the issue: https://paste.debian.net/1323675/ | 23:30 |
| chartreuse | https://pastebin.com/cgiw22xE | 23:30 |
| josch | chartreuse: these are the magic lines for you: | 23:30 |
| josch | E: /etc/flash-kernel/preboot.d/00reform2_preboot contains default content overriding /usr/share/flash-kernel/preboot.d/00reform2_preboot | 23:30 |
| josch | E: consider deleting /etc/flash-kernel/preboot.d/00reform2_preboot in favour of /usr/share/flash-kernel/preboot.d/00reform2_preboot | 23:30 |
| chartreuse | Yeah that makes sense | 23:30 |
| chartreuse | That flash.bin error was hiding those | 23:30 |
| josch | indeed | 23:31 |
| josch | nice, we found a bug and fixed it :) | 23:31 |
| chartreuse | Even if it wasn't the one I thought there was XD | 23:32 |
| chartreuse | Let's see if flash-kernel fixes it now without my override | 23:32 |
| + q66 (~q66@q66.moe) | 23:33 | |
| chartreuse | Okay no more error about the loaded dtb being wrong | 23:34 |
| chartreuse | I still only see one output in arandr but I also don't have a monitor plugged in so maybe that's something else | 23:34 |
| josch | i also made it now so that being offline does not make the script quit early | 23:35 |
| josch | could've saved a lot of time if this was already fixed :D | 23:35 |
| chartreuse | Heh might have caught it there | 23:35 |
| chartreuse | Does seem like maybe the script is being a bit strict with "E: the reform repository is not known to apt" since I have the line but without the arch=arm64, which isn't ideal but works the same | 23:37 |
| chartreuse | And the last check for the reform2_lpc.ko module isn't counting a compressed .xz of it in the same location | 23:38 |
| josch | the compressed .xz is a very new thing | 23:39 |
| josch | that one is indeed missing | 23:39 |
| josch | okay, fixed the xz thing | 23:40 |
| josch | now for the other | 23:40 |
| chartreuse | Also found another bug with the -hdmi. reform-flash-uboot is using the dtb name instead of the dtb name without -hdmi | 23:40 |
| josch | slower please | 23:40 |
| josch | otherwise it gets confusing which bug we are at | 23:40 |
| josch | about "not known to apt" | 23:40 |
| chartreuse | Ah yeah | 23:40 |
| chartreuse | I have deb [trusted=yes] http://mntre.com/reform-debian-repo reform main | 23:40 |
| josch | what output do you get if you run this: | 23:40 |
| josch | apt-get indextargets 'Created-By: Packages' 'Repo-URI: https://mntre.com/reform-debian-repo/' --format '$(RELEASE)' | 23:40 |
| chartreuse | Nothing | 23:41 |
| josch | huh | 23:41 |
| josch | that's odd | 23:41 |
| chartreuse | I am multiarch if that might break something | 23:41 |
| chartreuse | I am multiarch if that might break somethingAh | 23:41 |
| chartreuse | Ah it's http vs https | 23:41 |
| chartreuse | I've got plaintext http apt repos being used which the command and script wants https | 23:42 |
| josch | for mntre.com you should use https because http is just a http 302 to the https version | 23:43 |
| josch | so i think that error message is correct | 23:43 |
| josch | does the error go away once you fix that? | 23:43 |
| chartreuse | I guess, though maybe a bit confusing if someone doesn't spot http/https. And yep no error with https set, so not related to the architecture just the protocol | 23:46 |
| chartreuse | Probably not worth patching unless there's a way for the Repo-URI to take a wildcard protocol | 23:46 |
| josch | i don't think i want to patch it -- i think http is wrong because mntre.com is https-only | 23:47 |
| josch | i agree that it's hard to spot the http versus https difference | 23:47 |
| josch | but on the other hand, the script prints the line that it expects to see | 23:47 |
| josch | so if in doubt, one can just copypaste that | 23:48 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 23:48 | |
| chartreuse | Fair, could be another warning, but that works. I don't think I hand typed http in there so likely was how the v3 image was shipped | 23:48 |
| josch | yes, i would not be surprised | 23:48 |
| chartreuse | Shouldn't be an issue though | 23:48 |
| chartreuse | I suspect the patch https://paste.debian.net/1323675/ might have an error I just noticed with reform-flash-uboot as well. $(basename "$DTBPATH" .dtb) gives imx8mq-mnt-reform2-hdmi instead of imx8mq-mnt-reform2 | 23:53 |
| chartreuse | The uboot flash doesn't differentiate by hdmi | 23:53 |
| chartreuse | Line 81 of reform-flash-uboot does the same substitution | 23:55 |
| - Jonas___ (QUIT: Ping timeout: 252 seconds) (~Jonas@82-65-231-86.subs.proxad.net) | 23:57 | |
| josch | chartreuse: thank you there is indeed a bug but i do not think it's in reform-check or reform-flash-uboot | 23:57 |
| josch | the fact that *-hdmi exists in the first place is an ugly hack | 23:57 |
| josch | to fix this in reform-tools, i'd need to add yet another field into the machines/*.conf files containing the u-boot filename | 23:58 |
| josch | these filenames would be identical to the dtb names except for the reform 2 hdmi variant | 23:58 |
| chartreuse | Well yeah, though it exists becuase -hdmi adds a non-free blob | 23:58 |
| + jacobk (~quassel@2601:380:837e:24e0:b413:1241:5e3:eebd) | 23:58 | |
| josch | chartreuse: the u-boot binary is the same independent on what dtb linux boots | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
