| - chomwitt (QUIT: Ping timeout: 260 seconds) (~chomwitt@2a02:85f:9a08:a800:1ac0:4dff:fedb:a3f1) | 00:08 | |
| vagrantc | https://source.mnt.re/vagrantc/reform-debian-packages/-/commits/wip-linux-6.12.3x ... about to try a build... | 00:15 |
|---|---|---|
| vagrantc | the patches apply, what could possibly go wrong? :) | 00:15 |
| josch | imx8mq goes wrong all the time | 00:15 |
| josch | the hdmi cadence patch stack is super brittle | 00:15 |
| vagrantc | no worries, i won't test that! | 00:17 |
| josch | vagrantc: you won't? don't you have an imx8mq? | 00:17 |
| vagrantc | yes, but not enough to want to swap out the rk3588 :/ | 00:18 |
| vagrantc | one rabbit hole at a time, and maybe just one rabbit hole... | 00:18 |
| josch | :) | 00:18 |
| vagrantc | there is someone in the guix community with an imx8mq mnt/reform ... but felt too intimidated by the instructions to test it | 00:20 |
| vagrantc | and honestly, the compile times would be a bit hefty | 00:20 |
| vagrantc | the rk3588 makes it fine to compile natively | 00:20 |
| josch | heh i still have to set up my rk3588 classic reform for these tass | 00:22 |
| josch | *tasks | 00:22 |
| josch | still compiling everything on a311d | 00:22 |
| josch | which admittedly is a bit better than on imx8mq :) | 00:23 |
| minute | josch: great @ karlsruhe, looking fwd! | 00:30 |
| minute | josch: uefi shell of edk2 was able to read our ext4 partitions fine. | 00:31 |
| josch | :) | 00:32 |
| josch | minute: so you say that there is a way to configure edk2 to boot the existing stuff automatically without us changing something to efi? | 00:33 |
| - mjw (QUIT: Ping timeout: 268 seconds) (~mjw@gnu.wildebeest.org) | 00:33 | |
| josch | i'm in the process of rebasing our patches on linux 6.15 and am stuck here: https://salsa.debian.org/kernel-team/linux/-/merge_requests/1481#note_620648 | 00:51 |
| josch | just in case anybody here has an idea what i'm doing wrong | 00:51 |
| sad_plan | josch: did you figure out the battery stuff on your pocket? the extra batteries that is. you mentioned it in your post, and you had some issues the other day iirc | 01:06 |
| josch | sad_plan: it is possible that the issues i had stem from the sysctl upgrade which i performed at around the same time | 01:10 |
| josch | sad_plan: it might be that my unit powering off is not a hardware issue but the watch dog killing it | 01:10 |
| josch | sad_plan: i just downgraded to the april firmware and will run another test tomorrow (still not 100% charged) | 01:11 |
| sad_plan | josch: nice. looking forward to hearing more about it, when you get any further C: | 01:20 |
| josch | sad_plan: you probably read the rk3588 pocket battery runtime benchmark post in the community forum? | 01:21 |
| sad_plan | josch: yep. read the one from aether or w/e his name was again. iirc, you chimed in with some details aswell, which was very informative | 01:46 |
| vagrantc | hrmpf. drivers/gpu/drm failed to build ... so my patch refresh is sad. :( | 01:47 |
| vagrantc | entirely unsurprising... | 01:47 |
| + bkeys (~Thunderbi@173.186.16.211) | 02:17 | |
| bkeys | josch: https://copr.fedorainfracloud.org/coprs/bkeys/kernel-reform/build/9161907/ | 02:17 |
| bkeys | My initial build seems to have passed, that's with the .dtb files and reform2 kernel patches added in | 02:17 |
| bkeys | But the kernel config is the default Fedora one | 02:17 |
| vagrantc | my next bright idea is to revert the conflicting commits ... why did i not think of that sooner? :) | 02:27 |
| vagrantc | from upstream ... | 02:27 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 02:49 | |
| - xktr (QUIT: Ping timeout: 252 seconds) (~xktr@user/xktr) | 03:25 | |
| + xktr (~xktr@user/xktr) | 03:37 | |
| - nsc (QUIT: Ping timeout: 252 seconds) (~nicolas@19-96-142-46.pool.kielnet.net) | 04:01 | |
| + nsc (~nicolas@i5C74DD4E.versanet.de) | 04:03 | |
| - op_4 (QUIT: Remote host closed the connection) (~tslil@user/op-4/x-9116473) | 04:05 | |
| + op_4 (~tslil@user/op-4/x-9116473) | 04:05 | |
| - murph__ (QUIT: Quit: Leaving) (~murph@pool-108-35-93-154.nwrknj.fios.verizon.net) | 04:33 | |
| - paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@72.10.128.164) | 04:45 | |
| josch | sad_plan: Her name was Esther or Æsthr | 06:12 |
| josch | bkeys: uh fedora has a public build service? We've been dreaming about this for ages. XD | 06:12 |
| bkeys | josch: Yeah you never heard of copr? | 06:13 |
| josch | nope, i'm very uneducated when it comes to rpm land | 06:14 |
| bkeys | Yeah it exists but for RPM based distros | 06:15 |
| bkeys | So it'll build stuff, even for non-Fedora distros | 06:15 |
| bkeys | I think SUSE also has something similar | 06:17 |
| josch | yes, they run OBS (Open Build Service) which can also build Debian *.deb packages | 06:19 |
| bkeys | Yeah it's 12:30 here so I'll have to test out the kernel I built | 06:27 |
| bkeys | I have to figure out how to get my config to apply properly; the kernel built just has the default Fedora config which the Fedora ARM maintainer says he has multiple rk3588 boards running mainline Fedora kernels no problem | 06:27 |
| bkeys | But he's using uboot for everything. My thing is edk2 can boot the system image with full hardware support, so in theory there is no reason the Fedora kernel can't do the same | 06:28 |
| josch | bkeys: then did you check whether maybe all the options we have are already enabled? | 06:30 |
| josch | bkeys: you should only need the options at the bottom of the link to the config i pasted | 06:31 |
| kfx | copr is pretty great, it's basically fire and forget | 07:03 |
| - Ar|stote|is (QUIT: Quit: No Ping reply in 180 seconds.) (~linx@149.210.5.33) | 07:27 | |
| + Ar|stote|is (~linx@149.210.5.33) | 07:27 | |
| + Core1951 (~aakast@87-49-43-63-mobile.dk.customer.tdc.net) | 09:59 | |
| - aakast (QUIT: Ping timeout: 244 seconds) (~aakast@net-93-65-57-194.cust.vodafonedsl.it) | 10:01 | |
| + chomwitt (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 10:16 | |
| - helgoman_ (QUIT: Quit: Konversation terminated!) (~helgoman@178.38.121.62) | 10:23 | |
| + helgoman_ (~helgoman@178.38.121.62) | 10:23 | |
| - helgoman_ (QUIT: Quit: Konversation terminated!) (~helgoman@178.38.121.62) | 10:39 | |
| + mjw (~mjw@gnu.wildebeest.org) | 10:49 | |
| josch | minute: i just flashed the previous sysctl version onto my pocket and i got 6 hours of uninterrupted runtime from it (before it ultimately died because the battery was flat) | 12:03 |
| josch | i'll let it run for a long time again after the batteries are full again but it did not survive more than half an hour before, so this suggests that my shut-off problems are indeed not hardware related and all my unplugging/replugging might've been useless XD | 12:04 |
| - robin_ (QUIT: Read error: Connection reset by peer) (~robin@user/terpri) | 12:59 | |
| + robin (~robin@user/terpri) | 13:00 | |
| + murphnj (~murph@user/murphnj) | 13:52 | |
| sad_plan | josch: sounds about right | 14:26 |
| minute | josch: wow, thanks for testing that | 14:50 |
| minute | josch: i wonder what can cause watchdog feeding to be delayed for 1000ms | 14:51 |
| minute | (if that's the problem) | 14:51 |
| minute | josch: ok, uefi spec indeed says esp needs to be fat32, but not /boot | 14:55 |
| minute | josch: i'm not sure if edk2 is strict in regards to this spec, so we could explore that. or use an extra fat32 partition that's just /boot/efi | 14:57 |
| minute | otoh | 15:00 |
| minute | > The UEFI specification mandates support for the FAT12, FAT16, and FAT32 file systems (see UEFI specification version 2.10, section 13.3.1.1), but any conformant vendor can optionally add support for additional file systems; for example, the firmware in Apple Macs supports the HFS+ file system. | 15:00 |
| minute | otooh https://wiki.archlinux.org/title/EFI_system_partition#Typical_mount_points | 15:01 |
| + paperManu (~paperManu@72.10.128.164) | 15:03 | |
| - chomwitt (QUIT: Ping timeout: 276 seconds) (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 15:11 | |
| josch | minute: i'm currently doing option 1 (mount esp to /boot). With EFI the requirement for a separate /boot goes away because even for an encrypted setup it is sufficient for the ESP to be encrypted. | 15:30 |
| josch | Right now, I'm also storing boot.scr and extlinux.conf on the ESP so that the images are still compatible with non-efi setups. | 15:30 |
| josch | for maximum backwards compatibility one *could* do 3 partitions: the first is /boot because u-boot only searches the first partition, the second is the fat32 partition and gets mounted to /boot/efi and the third is the rootfs (optionally encrypted) | 15:32 |
| josch | that is also close to the setup created by debian-installer if one selects "separate /boot partition" and has efi booting | 15:32 |
| + chomwitt (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 16:40 | |
| minute | josch: interesting | 17:36 |
| gsora | I’m planning to experiment with different distros, and I would like to do so on with an sd card. will reform boot from it even if I have a system installed on the emmc? | 17:43 |
| josch | gsora: yes, u-boot on emmc will check the first partition on usb media, then sd-card, then emmc and pick the first with a boot.scr or extlinux.conf on it to boot | 17:51 |
| josch | gsora: so even if you have stuff on emmc, u-boot will boot from sd-card if it finds boot.scr or extlinux.conf on the first partition | 17:51 |
| - chomwitt (QUIT: Ping timeout: 265 seconds) (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 17:59 | |
| gsora | good to know! do you happen to know if I had e.g. u-boot on the sd card as well, would the cpu boot from that instead? | 17:59 |
| josch | gsora: which SoM? | 18:00 |
| gsora | I might want to mess with edk2 :^) hoping all the experience I gained working with the rk3566 will help not fscking everything up haha | 18:00 |
| gsora | josch: rk3588 | 18:01 |
| josch | gsora: rk3588 will try to find something bootable on emmc and if it finds something there, use that. It will check the SD card for u-boot if it does not find anything bootable on emmc. | 18:01 |
| josch | note that this is risky as you might soft-brick your device -- maybe ask bkeys for advice | 18:02 |
| gsora | yea thankfully I have played with rk SoCs already, I know my way around maskrom mode and the likes, it’s unfunny how easy it is to just soft brick them | 18:05 |
| gsora | my rk3566 had a physical switch to choose the boot media, but i guess that specific board was designed with it | 18:07 |
| josch | imx8mq also has such a switch -- it's really nice to have that | 18:08 |
| gsora | working with imx is always so nice | 18:08 |
| gsora | maybe imx9 will be available one day :’) | 18:09 |
| Zaba | imx9 is available, it’s just not particularly interesting to adapt for general purpose use | 19:03 |
| + chomwitt (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 19:12 | |
| - Core1951 (QUIT: Ping timeout: 252 seconds) (~aakast@87-49-43-63-mobile.dk.customer.tdc.net) | 19:22 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 19:33 | |
| + RandyK (~RandyK@user/randyk) | 19:33 | |
| + AnimaInvicta (~AnimaInvi@88-120-179-216.subs.proxad.net) | 20:16 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!