| ch | BoostisBetter: it should be in the normal branch | 00:56 |
|---|---|---|
| ch | but it doesnt fully (mostly) support usb devices on the charging-enabled port | 00:56 |
| ch | which is what i wanted to fix at some point | 00:57 |
| ch | but life is getting in the way | 00:57 |
| - Gooberpatrol_66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 01:00 | |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 01:01 | |
| - mjw (QUIT: Ping timeout: 272 seconds) (~mjw@gnu.wildebeest.org) | 02:31 | |
| - paperManu (QUIT: Ping timeout: 276 seconds) (~paperManu@107.159.71.33) | 02:43 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 03:38 | |
| - Gooberpatrol_66 (QUIT: Ping timeout: 276 seconds) (~Gooberpat@user/gooberpatrol66) | 03:39 | |
| - nsc (QUIT: Ping timeout: 248 seconds) (~nicolas@242-99-142-46.pool.kielnet.net) | 03:48 | |
| + nsc (~nicolas@50-98-142-46.pool.kielnet.net) | 03:49 | |
| - ptrc (QUIT: Remote host closed the connection) (~ptrc@ptrc.gay) | 04:30 | |
| + ptrc (~ptrc@ptrc.gay) | 04:30 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 05:03 | |
| - ex-parrot (QUIT: Quit: _b) (~fincham@user/ex-parrot) | 06:39 | |
| + ex-parrot (~fincham@user/ex-parrot) | 06:39 | |
| + gustav28 (~gustav@c-78-82-55-198.bbcust.telenor.se) | 10:02 | |
| BoostisBetter | ch: I understand! THanks for your efforts! | 11:02 |
| + pandora (uid585533@id-585533.ilkley.irccloud.com) | 11:23 | |
| + mjw (~mjw@gnu.wildebeest.org) | 12:00 | |
| + paperManu (~paperManu@107.159.71.33) | 12:00 | |
| - Ar|stote|is (QUIT: Ping timeout: 276 seconds) (~linx@149.210.12.55) | 12:34 | |
| + Ar|stote|is (~linx@149.210.13.15) | 12:38 | |
| - L29Ah (QUIT: Ping timeout: 248 seconds) (~L29Ah@wikipedia/L29Ah) | 13:20 | |
| minute | interesting (rk3588 emmc related) https://github.com/armbian/build/pull/7921 | 13:32 |
| - pandora (QUIT: Quit: Connection closed for inactivity) (uid585533@id-585533.ilkley.irccloud.com) | 13:33 | |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 13:45 | |
| - Ar|stote|is (QUIT: Ping timeout: 260 seconds) (~linx@149.210.13.15) | 14:13 | |
| + Ar|stote|is (~linx@149.210.24.235) | 14:17 | |
| BoostisBetter | I wasn't aware that the RK had issues using the emmc. Glad a solution seems to have been found. | 14:24 |
| minute | BoostisBetter: it doesn't have, per se | 14:27 |
| minute | i'm currently thinking about rk3588 suspend/resume support... i would really like to see it happen | 14:58 |
| minute | i found an interesting thread about clockwork pi's journey to fix this on rk3399 https://forum.clockworkpi.com/t/a06-suspend-to-ram-is-working/9263 | 14:58 |
| gordon1 | minute: isn't it be easier to make hibernate work? at least you don't have to work on ATF | 14:59 |
| minute | gordon1: both are interesting. hibernate is more interesting for the laptops definitely | 14:59 |
| gordon1 | ah, you're thinking about touch device as well, yes | 15:00 |
| minute | gordon1: yeah, and it might also benefit the pocket to have regular quick sleeps | 15:00 |
| minute | (for battery runtime) | 15:00 |
| gordon1 | did someone test how much it consumes in s2idle? | 15:00 |
| minute | no, it's not clear if it successfully enters sleep | 15:00 |
| gordon1 | s2idle is not really sleep and it usually always works | 15:01 |
| gordon1 | it just turns everthing it can off and stops the scheduler | 15:01 |
| gordon1 | but does not shut down cpus | 15:01 |
| BoostisBetter | minute: that is GREAT news! Suspend and using less than <100mA would be great. But like gordon1 I would probably prefer hibernation as well. Still if the stuff already figured out here makes suspend easier I would be happy with that as well! | 15:02 |
| BoostisBetter | As a matter of fact, the only real reason I haven't sprung for the RK on the Pocket is the suspend story and the lack of bluetooth. The Next is a bigger platform and just easier to mitigate. | 15:03 |
| minute | gordon1: i see. btw where you the one who already experimented with hibernate here? | 15:03 |
| BoostisBetter | So I think it is wonderful that this remains a focus for ya. | 15:03 |
| minute | s/where/were | 15:03 |
| minute | BoostisBetter: for bt there is a plan as well now. | 15:03 |
| gordon1 | minute: yes, i am, i can give it another try if you want, i just need to schedule some time to put the little board i have offline | 15:03 |
| gordon1 | i did a thing but lost all the nvmes on the way | 15:04 |
| minute | gordon1: ohh. i wonder if i could support your hacking in some way, i.e. provide you with more hardware | 15:04 |
| minute | about suspend, it might make sense to reach out to other rk3588 stakeholders like collabora and radxa | 15:05 |
| BoostisBetter | minute: That is awesome to hear! The RK on the Pocket is a dream for me. I am happy to hear that things are getting more and more mature there! | 15:05 |
| gordon1 | i since invented an ugly hack for nvme rescan for a different reason, so it is possible that it would fix nvmes after a suspend https://raw.githubusercontent.com/gordon-quad/mkinitramfs-ll/refs/heads/master/hooks/2-rescan-nvmes | 15:05 |
| minute | gordon1: oh this is very reminiscent of stuff we did on the imx8mq for sleep back in the days. we had to remove and rescan the pci wifi | 15:05 |
| gordon1 | minute: let me give it another try and see how comfortable i feel in hacking the rk3588 in mainline, i just don't want to commit to something that i couldn't pull off | 15:06 |
| minute | gordon1: the root filesystem on nvme would not like this though right... i guess we could look at patching the kernel to do this instead after the rootfs is frozen | 15:07 |
| minute | gordon1: sure thing | 15:07 |
| minute | also i'm embarassingly not sure if we're on the open-source tf-a already | 15:07 |
| minute | ( https://www.collabora.com/news-and-blog/blog/2024/02/21/almost-a-fully-open-source-boot-chain-for-rockchips-rk3588/ ) | 15:07 |
| gordon1 | root fs is not a big deal tbh, i have a suspend hack for my x230 that does chroot in ramfs before the sleep to remove the encryption keys from the memory, so kernel wakes up without rootfs available, but can be brought back | 15:08 |
| gordon1 | your script for u-boot builds it :) | 15:09 |
| gordon1 | i mean tf-a | 15:09 |
| gordon1 | https://source.mnt.re/reform/reform-rk3588-uboot/-/blob/main/build.sh?ref_type=heads#L7 | 15:09 |
| gordon1 | but there is no suspend support there | 15:10 |
| minute | gordon1: ah good, thanks :D | 15:14 |
| minute | gordon1: that sounds extremely nifty @ ramfs chroot | 15:14 |
| minute | gordon1: looks like there is suspend support though? https://gitlab.collabora.com/hardware-enablement/rockchip-3588/trusted-firmware-a/-/blob/rockchip/plat/rockchip/rk3588/drivers/pmu/pmu.c#L756 | 15:19 |
| gordon1 | i remember checking the patches from your previous link for rk3399 and one thing that stood up was ram suspend that isn't there for 3588 | 15:20 |
| gordon1 | let me find it again | 15:21 |
| minute | https://gitlab.collabora.com/hardware-enablement/rockchip-3588/trusted-firmware-a/-/blob/rockchip/plat/rockchip/rk3588/drivers/pmu/pmu.c#L1254 | 15:21 |
| minute | gordon1: ah i see, maybe they had some explicit PSCI calls for it? | 15:21 |
| gordon1 | that's what i'm talking about https://gitlab.collabora.com/hardware-enablement/rockchip-3588/trusted-firmware-a/-/blob/rockchip/plat/rockchip/rk3399/drivers/dram/suspend.c?ref_type=heads | 15:23 |
| gordon1 | tbh rk3588 has no dram in its platform at all so maybe it handled differently | 15:24 |
| minute | aha, this is the dram part ok | 15:27 |
| minute | as far as i understand so far the call for entering sleep is rockchip_soc_sys_pwr_dm_suspend | 15:27 |
| gordon1 | that's probably for cortex-m0 so we're not really concerned about that https://gitlab.collabora.com/hardware-enablement/rockchip-3588/trusted-firmware-a/-/tree/rockchip/plat/rockchip/rk3399/drivers/m0/src?ref_type=heads | 15:28 |
| gordon1 | just checking the mentions of suspend | 15:29 |
| minute | ddr sleep for rk3588 seems to be done here https://gitlab.collabora.com/hardware-enablement/rockchip-3588/trusted-firmware-a/-/blob/rockchip/plat/rockchip/rk3588/drivers/pmu/pmu.c#L896 | 15:29 |
| gordon1 | i wonder if you have to have something on cortex-m0 to make it sleep properly | 15:29 |
| gordon1 | ah so there is a chance | 15:30 |
| gordon1 | interesting that author of the original rk3399 post mentions that hibernate does not work | 15:45 |
| gordon1 | minute: do you have any rk3588 suspend attempt logs at hand? | 15:49 |
| minute | btw if you look at the original commit to tf-a by rockchip, suspend/resume is advertised https://gitlab.collabora.com/hardware-enablement/rockchip-3588/trusted-firmware-a/-/commit/e3ec6ff4b24c7daa4dfa82709c23a22829947160 | 15:51 |
| minute | gordon1: i don't have logs atm but could make some a bit later today | 15:54 |
| minute | in any case it's great we have open source tf-a, so one can also put debug prints etc in there | 15:55 |
| - Ar|stote|is (QUIT: Ping timeout: 260 seconds) (~linx@149.210.24.235) | 15:56 | |
| minute | it might also make sense to do tests in initramfs first | 15:57 |
| minute | (with minimum extra hw/drivers) | 15:57 |
| gordon1 | yep, i'm gonna try it later today too | 16:01 |
| grimmware | minute: you mentioned recently that you had a patched build of Blender working on the rk3588 - is that trivially reproduceable or is it $WORK? | 16:13 |
| - GNUmoon (QUIT: Ping timeout: 264 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 16:14 | |
| minute | grimmware: hmm i think it's not so much work, but needs a while to build | 16:14 |
| minute | like 15 minutes? not exactly sure | 16:14 |
| grimmware | that's no problem! I'd love to take a pop at it because I've got some travel coming up and it'd be a great chance for me to get some modeling done for some of my projects | 16:16 |
| grimmware | any notes or build scripts or anything would be greatly appreciated | 16:16 |
| josch | minute: we could throw your blender patch into reform-debian-packages, no? | 16:17 |
| minute | josch: we could! grimmware: it's not super usable though and can crash the gpu... so nothing you'd like to use productively atm | 16:20 |
| minute | grimmware: i guess the steps are 1. check out blender git main. 2. apt build-dep blender 3. mkdir build; cd build; cmake-gui ..; toggle lots of stuff off that you don't need and that just waste compile time 4. generate makefiles (i recommend ninja) 5. apply patch 6. ninja -C build | 16:21 |
| minute | (7. ninja -C build install is also required to copy some stuff around in build dir) | 16:22 |
| minute | grimmware: patch http://dump.mntmn.com/20250315-mnt-blender-panvk.patch.txt | 16:22 |
| minute | i also use mesa built from git for this. not sure if there are panvk things in there that are required that debian mesa doesn't have yet | 16:23 |
| grimmware | perfect, that's enough for me to get on with :) Probably give that a pop tomorrow | 16:23 |
| grimmware | I'll try without mesa git first and see how I get on just as a data point | 16:24 |
| minute | cool | 16:24 |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 16:27 | |
| - Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 16:51 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 16:52 | |
| - L29Ah (PART: !!unknown attribute: msg!!) (~L29Ah@wikipedia/L29Ah) | 17:06 | |
| BoostisBetter | minute: is the wifi/bt card you are using in the Reform Next what you were referring to by a solution for BT on the Pocket? | 17:18 |
| BoostisBetter | minute: is the wifi/bt card you are using in the Reform Next what you were referring to as a solution for BT on the Pocket? | 17:19 |
| minute | BoostisBetter: kind of, but i will change both | 17:19 |
| minute | to another m.2 key e slot | 17:19 |
| BoostisBetter | very cool! | 17:19 |
| BoostisBetter | minute: speaking of rabbit holes! I just stumbled on the mnt (you) mastadon feed and have been looking at all the awesome stuff you have been posting. I love it! The reform desktop is super dope! The pocket based tablet is awesome! So cool if that rabbit hole led to a bigger touch screen in the pocket. Anyway thanks for being so open with all the stuff you are working on! | 17:25 |
| BoostisBetter | minute: and I am now educated on exactly what you mean with the m.2 solution you are working on! | 17:26 |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 17:27 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 17:27 | |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 17:33 | |
| - Gooberpatrol66 (QUIT: Ping timeout: 248 seconds) (~Gooberpat@user/gooberpatrol66) | 17:33 | |
| - GNUmoon (QUIT: Ping timeout: 264 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 17:41 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 17:41 | |
| - cow321 (QUIT: Ping timeout: 252 seconds) (~deflated8@user/meow/deflated8837) | 18:09 | |
| - xktr (QUIT: Ping timeout: 252 seconds) (~xktr@user/xktr) | 18:44 | |
| + xktr (~xktr@user/xktr) | 18:45 | |
| minute | grimmware: modeling and rendering (on CPU) mostly works but opening material preview crashes the GPU, and also going to the shading tab | 18:48 |
| - Gooberpatrol_66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 19:23 | |
| + Gooberpatrol_66 (~Gooberpat@user/gooberpatrol66) | 19:23 | |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 19:27 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 19:40 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 19:41 | |
| - L29Ah (PART: !!unknown attribute: msg!!) (~L29Ah@wikipedia/L29Ah) | 19:57 | |
| + cow321 (~deflated8@user/meow/deflated8837) | 20:09 | |
| - Gooberpatrol_66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 20:09 | |
| + colinsane (~colinunin@97-113-143-86.tukw.qwest.net) | 20:37 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 20:40 | |
| - colinsane (QUIT: Ping timeout: 260 seconds) (~colinunin@97-113-143-86.tukw.qwest.net) | 20:51 | |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 20:58 | |
| + Ar|stote|is (~linx@149.210.25.83) | 21:26 | |
| BoostisBetter | video playback (youtube) in brave and clapper do not work anymore. They seem to be just staging forever but never starting. Has anyone else had this issue? | 21:49 |
| minute | i wouldn't use that browser, so no idea. do you see the same problem with mpv? | 21:54 |
| BoostisBetter | let me try that real quick | 21:54 |
| BoostisBetter | that fails saying it can't open the source file | 21:57 |
| BoostisBetter | it feels like it is system wide.clapper doesn't work either. I am wondering if I accidentally uninstalled a component necessary for playback. | 21:59 |
| BoostisBetter | as for the browser I would say it is chromium based. | 21:59 |
| minute | sounds weird | 22:00 |
| minute | can ffprobe identify the file correctly? does ffplay play it? | 22:00 |
| BoostisBetter | I just installed the chroium flatpak and youtube is showing the same isue. | 22:03 |
| BoostisBetter | I think I accidentally uninstalled a dependency or something. | 22:04 |
| BoostisBetter | I installed plasma and then uninstalled it | 22:05 |
| - Ar|stote|is (QUIT: Ping timeout: 260 seconds) (~linx@149.210.25.83) | 22:05 | |
| minute | well you'll have to be systematic about it to analyze what's wrong... | 22:06 |
| minute | btw most of the time video playback doesn't work in browsers is because of an audio issue | 22:07 |
| minute | i hadn't had that since the days of pulseaudio though. i.e. since pipewire audio has been working mostly just fine | 22:08 |
| josch | BoostisBetter: yes, i've had issues with mpv+yt-dlp in the past few days | 22:09 |
| BoostisBetter | ok, same with clapper? | 22:09 |
| josch | but i assumed the restrictive uni net would be the problem | 22:09 |
| BoostisBetter | I went through all the dependencies and all seem to be present | 22:10 |
| + Ar|stote|is (~linx@149.210.32.29) | 22:10 | |
| josch | BoostisBetter: i cannot test anymore because it's working right now | 22:10 |
| BoostisBetter | it is not the end of the world to me, but I went to peer tube, I went to the MNT page and none of the videos work there either. | 22:11 |
| BoostisBetter | so it seems global to the system | 22:11 |
| minute | BoostisBetter: have you tried rebooting? | 22:12 |
| BoostisBetter | haha, no. Let me try that real quick | 22:12 |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-55-198.bbcust.telenor.se) | 22:15 | |
| BoostisBetter | minute: your audio suggestion was exactly it. I was trying to get bluetooth headphones to work and accidentally switched the audio sink. | 22:19 |
| BoostisBetter | thank you very much! | 22:19 |
| minute | ah. | 22:20 |
| minute | glad it worked out | 22:20 |
| - Ar|stote|is (QUIT: Ping timeout: 252 seconds) (~linx@149.210.32.29) | 22:21 | |
| BoostisBetter | indeed | 22:24 |
| BoostisBetter | I don't watch a lot of youtube, but when I do it is drool over the Reform Next. hahaha | 22:24 |
| + Ar|stote|is (~linx@149.210.0.85) | 22:25 | |
| minute | hehe | 22:39 |
| - L29Ah (QUIT: Ping timeout: 276 seconds) (~L29Ah@wikipedia/L29Ah) | 22:47 | |
| + L29Ah (~L29Ah@wikipedia/L29Ah) | 23:25 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!