| bkeys | Well I got a Fedora 42 minimal install on my usb disk | 00:31 |
|---|---|---|
| bkeys | It can boot via ACPI so I can use that to install my custom kernel and see if that solves my woes | 00:31 |
| - AnimaInvicta (PART: !!unknown attribute: msg!!) (~AnimaInvi@88-120-179-216.subs.proxad.net) | 00:45 | |
| - chomwitt (QUIT: Ping timeout: 260 seconds) (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 02:00 | |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 02:10 | |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@72.10.128.164) | 03:17 | |
| - nsc (QUIT: Ping timeout: 260 seconds) (~nicolas@i5C74DD4E.versanet.de) | 04:01 | |
| + nsc (~nicolas@i5C74DE14.versanet.de) | 04:02 | |
| + chomwitt (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 06:45 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 08:17 | |
| + mjw (~mjw@gnu.wildebeest.org) | 10:48 | |
| + andreas-e (~Andreas@2001:861:c4:f2f0::c64) | 10:56 | |
| - andreas-e (QUIT: Client Quit) (~Andreas@2001:861:c4:f2f0::c64) | 10:59 | |
| - L29Ah (QUIT: Ping timeout: 248 seconds) (~L29Ah@wikipedia/L29Ah) | 11:49 | |
| - gsora (QUIT: Remote host closed the connection) (~gsora@user/gsora) | 12:09 | |
| + gsora (~gsora@user/gsora) | 12:09 | |
| - gsora (QUIT: Remote host closed the connection) (~gsora@user/gsora) | 12:09 | |
| + gsora (~gsora@user/gsora) | 12:09 | |
| - kensanata (QUIT: Quit: OK) (~alex@user/kensanata) | 12:11 | |
| + kensanata (~alex@user/kensanata) | 12:12 | |
| - elb (QUIT: Remote host closed the connection) (~elb@2600:4041:6671:1300:f2d5:d71d:a66b:d886) | 12:17 | |
| + elb (~elb@2600:4041:6671:1300:2598:4bda:6003:6d0d) | 12:17 | |
| - Ar|stote|is (QUIT: Ping timeout: 260 seconds) (~linx@149.210.5.33) | 12:29 | |
| + Ar|stote|is (~linx@149.210.5.33) | 12:32 | |
| + paperManu (~paperManu@72.10.128.164) | 12:54 | |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org) | 13:16 | |
| - Sario (QUIT: Quit: WeeChat 4.3.5) (sario@libera/staff/owl/sario) | 13:27 | |
| bremner | what is the current state of getting suspend working on rk3588 pocket reforms? I remember a month ago or so there was some kernel patches in progress / needing testing? Unfortunately I didn't make a note of the details. | 13:40 |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 13:54 | |
| - L29Ah (QUIT: Ping timeout: 260 seconds) (~L29Ah@wikipedia/L29Ah) | 14:03 | |
| josch | bremner: fortunately you do not need to take notes because this channel is logged: https://mntre.com/reform-irc-logs/2025-04-30.log.html#t11:15:34 | 14:04 |
| josch | you may want to ping gordon1 with a question about the current status i guess | 14:04 |
| gordon1 | i'm glad you asked, i plan to give it another shot this week, there is interesting pci PM patches from collabora | 14:05 |
| josch | awesome! i'm up for testing any patches from you as well :) | 14:08 |
| josch | minute: speaking of patches, I rebased our stuff onto 6.15 here: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/118 | 14:08 |
| bremner | hurray. Now is a good time for me to test, before I am relying on the pocket | 14:08 |
| josch | (had no time to test it yet, maybe somebody else does) | 14:09 |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 14:09 | |
| gordon1 | bremner: do you have nvme and relying on it to run linux? | 14:09 |
| bremner | gordon1: I have nvme but it is idle so far | 14:09 |
| gordon1 | if not you can test if graphics resumes correctly, i am so far focuesd on testing PCI resume since i have no graphics on my rock-5b+ | 14:10 |
| gordon1 | i had much more luck with hibernate than suspend so you can start there | 14:11 |
| gordon1 | and i don't think you need any patches for pcie then | 14:11 |
| gordon1 | bremner: do you need help with how to test it manually with all the dmesg output etc? | 14:13 |
| bremner | I can probably work that out, but if there is a document I should read, by all means. What should I test, the 6.14.6-mnt-reform-arm64 that shipped recently, or something else? | 14:14 |
| gordon1 | i'm testing recent git kernel and cherry-pick patches, if you can pick the most recent one and only patches that are vital for pocket to work - do that, if that would turn out to be too difficult - then try 6.14.6 | 14:15 |
| gordon1 | just for the record, i am still waiting for my reform classic too so i'm not experimenting on the actual device | 14:16 |
| gordon1 | bremner: https://docs.kernel.org/power/basic-pm-debugging.html this includes some interesting info, but if you would experience some issues - ping me, i can make a little tutorial out of what i do to test it | 14:17 |
| bremner | OK | 14:17 |
| gordon1 | tbh i gave up last time because i failed to fixup some patches for 6.15, but that was a month ago or so | 14:18 |
| gordon1 | i mean pcie PM patch and so like, not reform patches | 14:18 |
| gordon1 | i think reform patches do not exists for 6.15, correct me if i'm wrong | 14:18 |
| josch | gordon1: they do exist since six hours ago :) | 14:19 |
| gordon1 | oh, nice! | 14:19 |
| josch | see my message above | 14:19 |
| gordon1 | sorry, can | 14:19 |
| gordon1 | can't read lol | 14:19 |
| josch | unfortunately i'm currently doing a battery runtime test on my pocket so i cannot test it right now but only in about five hours | 14:20 |
| josch | (current uptime 06:54 and 31% battery left) | 14:20 |
| gsora | impressive runtime! | 14:21 |
| minute | josch: ok wow @ uptime | 14:22 |
| josch | yesterday i had it running for six hours while playing a 1080p movie at full display brightness :) | 14:22 |
| minute | that's amazing | 14:22 |
| josch | since i reverted to the old sysctl, the pocket does not switch off anymore | 14:22 |
| minute | ok | 14:22 |
| josch | but to be sure it was the sysctl and not my reseating everything i'll upgrade to the new version after my battery runtime tests are done | 14:22 |
| bremner | oof. one task leads to another. I guess it is time to get my debian package building setup working on the pocket | 14:23 |
| josch | bremner: currently everything is configured to work out of the box on a vanilla install by just installing sbuild, mmdebstrap and uidmap. But it's also supposed to "just work" with custom setups if it's not too crazy. So ping me if your custom setup happens not to work with reform-debian-packages. | 14:24 |
| bremner | josch: sure. | 14:25 |
| - paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@72.10.128.164) | 14:45 | |
| josch | is somebody able to spot what makes this CI job fail? https://source.mnt.re/josch/qcacld2/-/jobs/11153 | 14:45 |
| - amk (QUIT: Remote host closed the connection) (~amk@user/amk) | 15:14 | |
| * Guest1551 -> mjw | 15:19 | |
| - L29Ah (PART: !!unknown attribute: msg!!) (~L29Ah@wikipedia/L29Ah) | 15:38 | |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 15:46 | |
| + amk (~amk@user/amk) | 15:59 | |
| + andreas-e (~Andreas@2001:861:c4:f2f0::c64) | 16:02 | |
| josch | minute: can you merge https://source.mnt.re/reform/qcacld2/-/merge_requests/11 please? The pipeline fails but it failed before and it workes fine when run via reform-debian-packages. | 16:03 |
| - chomwitt (QUIT: Ping timeout: 276 seconds) (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 16:42 | |
| minute | josch: ooh, thanks | 17:15 |
| minute | josch: also, i think linux 6.15 brings hdmi audio for rk3588?! | 17:15 |
| - andreas-e (QUIT: Quit: Leaving) (~Andreas@2001:861:c4:f2f0::c64) | 17:22 | |
| + AnimaInvicta (~AnimaInvi@88-120-179-216.subs.proxad.net) | 18:09 | |
| minute | ok, my pocket finally has wifi, bt, lte and gps again :D | 18:19 |
| BoostisBetter | minute: any idea when the replacement kb for my pocket will be getting shipped? | 18:32 |
| minute | BoostisBetter: i can look into it today, do you have the support ticket number again please? | 18:37 |
| BoostisBetter | minute: no, I just have an email chain with plom. They asked for my mailing address and cell number for the carrier. | 18:50 |
| BoostisBetter | minute: there is a big long number written in the response email (service desk) I could give you. | 18:50 |
| minute | BoostisBetter: ok, i'll look into my tickets then | 18:51 |
| minute | BoostisBetter: ah yes, i have all the info i need | 18:52 |
| minute | BoostisBetter: i assume red switches? | 18:52 |
| BoostisBetter | minute: yes please. Also really super looking forward to the RK3588 upgrade 😉 | 18:54 |
| bremner | oops, plugged into the wrong usb c cable, ran the pocket reform battery down to 0 | 18:58 |
| bremner | I was like, wow the chassis is suspiciously cool... | 18:59 |
| - aloo_shu (QUIT: Quit: solar maintainance) (~aloo_shu@90.166.98.218) | 19:05 | |
| josch | minute: I need to know which antennas you are using for lte and gnss :) | 19:20 |
| josch | minute: also, my pocket just had a final uptime of 10:06 hours with display on lowest brightness and terminal idling with htop | 19:20 |
| bremner | spiffy | 19:22 |
| minute | josch: that's awesome runtime! | 19:23 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 19:28 | |
| + chomwitt (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 19:28 | |
| vagrantc | josch: so, good news, and bad news, regarding the 6.12.x patchsets ... | 19:28 |
| vagrantc | reverting https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c?h=linux-6.12.y&id=d2b58a10228a906d46155eb7c15d79f39be25b37 gets it to build again | 19:28 |
| vagrantc | but it does not boot | 19:29 |
| vagrantc | with 6.12.33 ... | 19:29 |
| BoostisBetter | josch: That is a serious good runtime! | 19:29 |
| gordon1 | double it and we don't need suspend after all! | 19:30 |
| vagrantc | reverting the above *and* using the patchsets from reform-debian-packages commit af0a461d38e13481323f061d9ff6827d1d13873b ... both builds and boots, though! :) | 19:30 |
| BoostisBetter | gordon1: my hope is that we can get hibernation working. | 19:30 |
| gordon1 | yeah, that would be nice | 19:30 |
| gordon1 | we're pretty close, i can restore the image with root in tmpfs and it works fine, just some devices are missing | 19:31 |
| gordon1 | have good feeling about 6.15 | 19:32 |
| BoostisBetter | gordon1: that is really great to hear! I mean if hibernation can work on the Reform platform in general it will close the only real missing feature IMHO. It is really impressive how far the Reform line has come! | 19:33 |
| josch | vagrantc: i am not familiar with that patch. My problems so far with 6.12 were the change of s/&hdptxphy_hdmi1/&hdptxphy1/ and now with 6.12 we also have s/&hdptxphy_hdmi0/&hdptxphy0/ | 19:39 |
| vagrantc | josch: the reverted patch is just some changes to rockchip-drm that conflict with the mnt/reform patches regarding rockchip-drm | 19:40 |
| vagrantc | josch: better would be to actually rebase the changes, but at least reverting the upstream changes allows the mnt/reform patches to do thier thing | 19:40 |
| vagrantc | josch: since it was a single patch to revert, figured it was worth poking at | 19:41 |
| vagrantc | i should try booting the borken 6.12.33 kernel with increased verbosity on the serial console to hopefully see why it fails | 19:42 |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 19:42 | |
| + mark_ (~mjw@gnu.wildebeest.org) | 19:45 | |
| josch | ACTION is still at 6.12.30 like a noob :D | 19:49 |
| vagrantc | i have not tried 6.12.30 with the newer patches, but i had it working with the older patches without problems | 19:49 |
| vagrantc | well, with the only consoel being serial console, it loads the initramfs successfully ... but then appears to hang ... | 19:50 |
| josch | minute: https://source.mnt.re/reform/reform-system-image/-/jobs/11086 "No space left on device" | 19:51 |
| minute | josch: hm, this is building on amd64 runner, intentional? | 19:53 |
| * mjw -> Guest1629 | 19:54 | |
| - Guest1629 (QUIT: Killed (copper.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 19:54 | |
| * mark_ -> mjw | 19:54 | |
| josch | oh good catch! why is that happening?? | 19:54 |
| josch | that's also why it took 2 hours :D | 19:54 |
| + Guest1629 (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 19:54 | |
| josch | hitting the retry button also creates a job on the mntre-builder | 19:55 |
| josch | i'll add the "arm64" tag to the job to force it on the ampere runner | 19:55 |
| minute | josch: i guess this is since i assigned the amd64 runner to the project recently for building the amd64 image POC? | 19:56 |
| josch | ah indeed | 19:56 |
| josch | yes, that explains it | 19:56 |
| josch | then lets add the tag to fix it | 19:56 |
| minute | josch: do i have to do sth? | 20:05 |
| minute | (right now testing mojyjack's mesa patch with firefox) | 20:05 |
| minute | huh, firefox is really faster now | 20:07 |
| josch | minute: nope, this is something i'll add to .gitlab-ci.yml | 20:07 |
| minute | josch: ah, alright | 20:07 |
| minute | ff feels at least as fast as chromium now | 20:08 |
| minute | josch: heh, i also had the reset now | 20:10 |
| bkeys | josch: So I got the patches from https://source.mnt.re/reform/reform-debian-packages/-/tree/main/linux/patches6.14/rk3588-mnt-reform2?ref_type=heads and the patches aren't applying my RPM build | 20:11 |
| bkeys | I created a patch that patches in the .dts files, and those have no issue. In my script I also removed all the patches that have MERGED in the name so none of those are being applied | 20:12 |
| bkeys | Any clue? | 20:12 |
| zeha | logs | 20:13 |
| bkeys | https://download.copr.fedorainfracloud.org/results/bkeys/kernel-reform/fedora-42-aarch64/09165983-kernel/builder-live.log.gz | 20:13 |
| josch | minute: oh phew you also seeing a reset is a big releave for me :D | 20:20 |
| minute | josch: yeah first i thought it was because my batteries were too weak, but i'm not sure that's the case | 20:21 |
| bkeys | josch: Do you know where the collabara patches are exactly? If they do the same thing as the patches in the Reform repo I can see if they conflict or nto | 20:21 |
| bkeys | *not | 20:21 |
| josch | bkeys: https://gitlab.collabora.com/hardware-enablement/rockchip-3588/linux | 20:22 |
| josch | bkeys: they have branches for 6.12, 6.13 and 6.14. For 6.15 I used what they have in rockchip-devel | 20:22 |
| bkeys | Is it gonna be patch files or should I just package this kernel instead of upstream? | 20:22 |
| josch | bkeys: i guess that depends on how fedora kernel packaging works -- don't forget that you still need the MNT dtb and a handful of patches on top of the collabora stack for the second hdmi and stuff | 20:23 |
| bkeys | My script puts the DTB stuff in there no problem, it's the patches in my link that are conflicting | 20:24 |
| bkeys | Well Collabara's stuff looks like a fork of the Linux kernel so I can likely use it instead of upstream for now | 20:25 |
| bkeys | If I just use the collabara kernel what other patches do I need to fully enable hardware on the Reform? | 20:26 |
| josch | bkeys: yes, they regularly rebase on top of linus' tree | 20:26 |
| josch | bkeys: it's in the rk3588 directory in the reform-debian-packages git | 20:26 |
| bkeys | https://source.mnt.re/reform/reform-debian-packages/-/tree/main/linux/patches6.14/rk3588-mnt-reform2?ref_type=heads | 20:27 |
| bkeys | These? | 20:27 |
| josch | yes | 20:27 |
| bkeys | Well these are what's conflicting with my upstream kernel | 20:27 |
| bkeys | I'll see if they clash on the collabara kernel | 20:27 |
| josch | bkeys: above you said that you removed the patched that said MERGED -- don't do that, they are important | 20:28 |
| bkeys | Okay, I'll try this out | 20:29 |
| josch | bkeys: in that directory, all patches numbered below 100 are from the collabora linux6.14 branch -- the ones 1000 and beyond are from MNT | 20:29 |
| bkeys | Okay so I will ignore the below 100 patches | 20:29 |
| josch | if you are already on collabora's linux6.14 branch, then yes | 20:30 |
| bkeys | Yeah I'll switch it over to that | 20:31 |
| + bkeys1 (~Thunderbi@173.186.16.211) | 20:47 | |
| - bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@173.186.16.211) | 20:48 | |
| * bkeys1 -> bkeys | 20:48 | |
| + bkeys1 (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 20:57 | |
| - bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@173.186.16.211) | 20:58 | |
| * bkeys1 -> bkeys | 20:58 | |
| + bkeys1 (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 21:03 | |
| - bkeys (QUIT: Ping timeout: 276 seconds) (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 21:05 | |
| * bkeys1 -> bkeys | 21:05 | |
| - bkeys (QUIT: Ping timeout: 260 seconds) (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 21:24 | |
| - digitalrane (QUIT: Ping timeout: 250 seconds) (~digitalra@user/digitalrane) | 21:32 | |
| + digitalrane (~digitalra@user/digitalrane) | 21:34 | |
| + bkeys (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 21:53 | |
| minute | hmm, pico-sdk repeating timers are peculiar | 21:57 |
| minute | aha https://github.com/raspberrypi/pico-sdk/issues/751#issuecomment-1061325845 (cc zeha if curious about alarm pools / irq priorities) | 21:58 |
| + chomwitt (~chomwitt@2a02:85f:9a1d:4100:1ac0:4dff:fedb:a3f1) | 22:00 | |
| minute | on rp2350 this becomes > irq_set_priority(timer_hardware_alarm_get_irq_num(timer_hw, 2), 0xc0); // larger number is lower priority | 22:12 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 22:21 | |
| minute | it's not actually a good idea to do i2c stuff in an interrupt when the interrupted code can also do i2c stuff | 22:25 |
| + Sario (sario@libera/staff/owl/sario) | 22:48 | |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 22:50 | |
| josch | minute: now that we have logelevel back to the default, pocket reform boot shows tons of debug messages and backtraces: https://community.mnt.re/t/the-latest-update-full-upgrade-and-boot-issue/3489/10 | 22:59 |
| josch | minute: i'm sure much of this output is just unintended debugging printfs which are not supposed to be there but went unnotuced because we ran with loglevel=3 for so long? | 22:59 |
| + bkeys (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 23:11 | |
| - bkeys (QUIT: Ping timeout: 272 seconds) (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 23:25 | |
| + bkeys (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 23:26 | |
| + bkeys1 (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 23:30 | |
| - bkeys (QUIT: Ping timeout: 252 seconds) (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 23:32 | |
| * bkeys1 -> bkeys | 23:32 | |
| + paperManu (~paperManu@72.10.128.164) | 23:38 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 23:40 | |
| + hairu (m-uotkmd@user/hairu) | 23:41 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 23:51 | |
| + aloo_shu (~aloo_shu@90.166.98.218) | 23:51 | |
| + hairu (m-uotkmd@user/hairu) | 23:52 | |
| murphnj | josch, I set the loglevel back to 3, and it seems to be back to "normal" (not reporting all of those errors) | 23:55 |
| josch | murphnj: nice! | 23:55 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!