| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 00:00 | |
| dustfinger | minute: when I try to clone the reform-system-image I am getting a permission deneid. shouldn't I be able to just clone that repo? i.e. git@source.mnt.re:reform/reform-system-image.git. | 00:00 |
|---|---|---|
| dustfinger | I think the problem is probably on my end with how I have all my ssh keys configured maybe. I will figure it out. I can see the repo is public, so I should be able to clone it. I have to head out for a bit now. Will look at this again later... | 00:07 |
| minute | dustfinger: clone with https url | 00:20 |
| minute | dustfinger: the git/ssh url is only for registered accounts | 00:20 |
| minute | dustfinger: just select the other url from the dropdown | 00:20 |
| jfred | You know it is also interesting that the diode that failed was in the buck/boost charger circuit. The stock charger outputs 24V while the cable I was charging from outputs 20V - which is still within-spec per the handbook but potentially noteworthy given the failure… | 00:39 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 00:40 | |
| jfred | Datasheet says it's used "during periods of reverse current inhibit operation" which sounds like something that might happen more when charging from lower voltages. Just a wild guess though, I haven't actually looked too closely at this circuit | 00:44 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 01:06 | |
| - mtm (QUIT: Ping timeout: 252 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:02 | |
| - qwer (QUIT: Ping timeout: 260 seconds) (~qwer@78-80-112-49.customers.tmcz.cz) | 03:37 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 03:52 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 04:10 | |
| - dustfinger (QUIT: Ping timeout: 252 seconds) (~user@d75-159-228-218.abhsia.telus.net) | 05:03 | |
| + reform11357 (~masop@67-43-243-35.fidnet.com) | 05:29 | |
| * reform11357 -> morrig | 05:30 | |
| + klardotsh (~klardotsh@98.97.35.142) | 06:14 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 06:18 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 06:46 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 06:48 | |
| morrig | I've come to the conclusion that I'm just a bit too dumb for the imager. I've been trying and reading for fixes but I just havent found one yet. IDK what I'm doing wrong here. | 07:13 |
| josch | morrig: what's the error message? | 07:14 |
| morrig | i have the newest version of the imager and i try to run it but it just tells me that a chroot fails and exits without ever making an image. I've been trying to figure this out for the last week but I'm just at a loss. Sorry if the fix is obvious. | 07:14 |
| morrig | Get:14 https://mntre.com/reform-debian-repo reform/main arm64 reform-handbook all 1.1 [264 MB] | 07:14 |
| morrig | Fetched 374 MB in 9min 25s (661 kB/s) | 07:14 |
| morrig | E: Failed to fetch https://mntre.com/reform-debian-repo/pool/main/l/linux/linux-image-5.19.0-reform2-arm64_5.19.6-1%2breform1_arm64.deb 404 Not Found [IP: 91.250.115.15 443] | 07:14 |
| morrig | E: Failed to fetch https://mntre.com/reform-debian-repo/pool/main/l/linux/linux-image-arm64_5.19.6-1%2breform1_arm64.deb 404 Not Found [IP: 91.250.115.15 443] | 07:14 |
| morrig | E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing? | 07:14 |
| morrig | E: run_chroot failed: E: apt-get -o Dir::Bin::dpkg=env -o DPkg::Options::=--unset=TMPDIR -o DPkg::Options::=dpkg -o DPkg::Chroot-Directory=/home/masop/reform-system-image/reform2-imx8mq/mmdebstrap.agsuj6kGEa --yes install -oAPT::Status-Fd=<$fd> -oDpkg::Use-Pty=false e2fsprogs u-boot-tools sudo kbd bash-completion net-tools parted iputils-ping console-data htop console-setup xwayland rsync systemd-sysv cryptsetup xterm gpgv bsdmainutils mesa-utils | 07:14 |
| morrig | I: removing tempdir /home/masop/reform-system-image/reform2-imx8mq/mmdebstrap.agsuj6kGEa... | 07:14 |
| morrig | E: mmdebstrap failed to run | 07:14 |
| josch | morrig: if you paste more than three lines into IRC, please use a pastebin | 07:15 |
| morrig | will do my bad | 07:15 |
| morrig | im still learning how to do all this. im used to discord. | 07:15 |
| josch | morrig: linux-image fails to download because we currently are investigating a problem with the latest image our CI built | 07:15 |
| josch | morrig: this is a known problem on which we have been working since sunday -- it will hopefully be fixed soon | 07:15 |
| morrig | ahhh ok | 07:16 |
| josch | morrig: so this is not something that you did wrong but something that we have to fix first before we can build new images again | 07:16 |
| morrig | thank you, sorry about the bugging | 07:16 |
| josch | no problem, that's why this channel exists :) | 07:16 |
| morrig | im very happy you guys are working on this. I would like to help if i can ever get profficient enough. I love everything this project stands for | 07:17 |
| josch | minute: do you think it would be a good idea to create a second staging repo (similar to the one you already created) to which CI can output, such that in situations like this one we keep the old working versions available and only move stuff from staging to the real one manually every once in a while? | 07:27 |
| morrig | exit | 07:30 |
| - morrig (QUIT: Quit: Leaving) (~masop@67-43-243-35.fidnet.com) | 07:30 | |
| + indefini[m] (~indefinim@2001:470:69fc:105::1e2a) | 08:13 | |
| - wielaard (QUIT: Ping timeout: 268 seconds) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 09:02 | |
| + spiral (~root@spiral.sh) | 09:32 | |
| spiral | hey, I accidentally flashed my standalone keyboard with non-standalone firmware, and now I can't get back to flashing mode to flash new firmware. how can I fix this? | 09:33 |
| spiral | it looks like by trying to perform operations on the system controller and then pressing reset while the display shows "waking up LPC", the keyboard goes into DFU mode, but dfu-programmer shows "no device present" | 09:40 |
| spiral | ah, dfu-programmer on arch/ubuntu are broken. I got an alpine live usb and was able to flash it from there | 09:57 |
| josch | spiral: what is the dfu-programmer bug? | 10:01 |
| spiral | it doesn't recognise the keyboard when it's in flashing mode | 10:03 |
| josch | spiral: and the alpine version has it fixed? | 10:04 |
| spiral | yes | 10:04 |
| josch | Ah it seems that ubuntu only has dfu-programmer 0.6.1 while alpine ships 0.7.2 -- lets fix this | 10:06 |
| josch | minute: did you build your own version of dfu-programmer to be able to flash the keyboard? You are also a Debian user, right? | 10:10 |
| minute | josch: i used dfu-programmer from debian | 10:16 |
| josch | hrm... then it's strange that it didn't work for spiral | 10:17 |
| minute | spiral: did you use sudo? | 10:17 |
| minute | a common mistake is to not use sudi | 10:17 |
| minute | sudo | 10:17 |
| minute | the error message is misleading then | 10:18 |
| spiral | yes I am running dfu-programmer as root | 10:18 |
| minute | spiral: how does the device present itself in lsusb? | 10:19 |
| minute | maybe you erased the usb bootloader? | 10:19 |
| spiral | Bus 001 Device 019: ID 03eb:2ff4 Atmel Corp. atmega32u4 DFU bootloader | 10:20 |
| minute | that's fine | 10:20 |
| minute | what's the error from dfu-programmer? | 10:20 |
| spiral | just `dfu-programmer: no device present.` | 10:20 |
| minute | that's usually what i get when not being root / not having permission, maybe there's another reason it can't find the usb device | 10:21 |
| minute | do you have another computer you could try it with? | 10:21 |
| spiral | i tried on a few different computers, was able to flash only under alpine | 10:22 |
| minute | aha | 10:22 |
| minute | so it works with alpine? | 10:22 |
| spiral | yep | 10:22 |
| minute | ok so it's a software issue | 10:22 |
| spiral | fwiw joshc: arch also has 0.7.2 and it doesn't work | 10:22 |
| josch | then it's probably not the dfu-programmer version... nevertheless, the version in Debian should be made up-to-date (I'm on it) | 10:23 |
| spiral | ah | 10:23 |
| spiral | fwiw, running dfu-programmer in a privileged alpine docker container works correctly, on the same host where it throws the error | 10:24 |
| minute | interesting | 10:25 |
| minute | maybe something about libusb | 10:25 |
| minute | or could there be other protections at play? something like selinux? wildly guessing | 10:26 |
| spiral | hm there's a comment in dfu-programmer source mentioning it throwing that error because something about atmel changing interface class (arguments.c:513) | 10:38 |
| spiral | but that doesn't seem related | 10:38 |
| spiral | oh, well, compiling dfu-programmer myself does work under arch | 10:43 |
| spiral | and now flashing with arch's own dfu-programmer also works | 10:45 |
| spiral | I'm so confused | 10:45 |
| spiral | user error somehow, maybe? | 10:45 |
| + dustfinger (~user@d75-159-228-218.abhsia.telus.net) | 11:03 | |
| josch | spiral: what happens if you copy the /usr/bin/dfu-programmer to your ubuntu or arch and run it there? | 11:14 |
| josch | spiral: if that works, then it's not dfu-programmer itself but something about your system (like libusb, selinux etc) | 11:14 |
| + chomwitt (~chomwitt@2a02:587:dc16:fe00:36b8:e455:625c:e00b) | 12:40 | |
| minute | josch: uart debugging the 5.19 build now | 13:13 |
| josch | nice :) | 13:14 |
| minute | my guess is the panel is not probed | 13:14 |
| minute | we repeatedly get > [ 7.734638] [drm:ti_sn_bridge_probe [ti_sn65dsi86]] *ERROR* failed to create panel bridge | 13:14 |
| minute | this points at the ti sn probe being deferred | 13:14 |
| minute | waiting for the panel, which doesn't appear | 13:15 |
| minute | root@reform:~# lsmod | grep panel | 13:16 |
| minute | panel_edp 32768 0 | 13:16 |
| minute | drm_display_helper 118784 2 panel_edp,ti_sn65dsi86 | 13:16 |
| minute | drm_dp_aux_bus 20480 2 panel_edp,ti_sn65dsi86 | 13:16 |
| minute | drm 442368 11 drm_cma_helper,gpu_sched,drm_kms_helper,imx_dcss,panel_edp,etnaviv,nwl_dsi,drm_display_helper,ti_sn65dsi86,mxsfb | 13:16 |
| minute | perhaps another mixup of which panel driver is responsible now | 13:17 |
| minute | there was one transition from panel-simple to panel-edp before iirc | 13:17 |
| minute | ah: | 13:18 |
| minute | [ 6.328282] pwm-backlight backlight: unable to request PWM | 13:18 |
| minute | [ 6.333797] pwm-backlight: probe of backlight failed with error -22 | 13:18 |
| minute | [ 330.129168] of_pwm_get(): can't parse "pwms" property | 13:19 |
| minute | looks like a change in dts? | 13:19 |
| minute | (this happens if i rmmod and modprobe pwm_bl) | 13:19 |
| minute | might be this: | 13:22 |
| minute | [ 330.124313] OF: /backlight: #pwm-cells = 3 found 2 | 13:22 |
| minute | i don't get it though, as we have [ 330.124313] OF: /backlight: #pwm-cells = 3 found 2 | 13:23 |
| minute | sorry, copypaste fail. we have pwms = <&pwm2 0 10000>; | 13:24 |
| minute | i dumped the loaded device tree, and pwm2 is not in there | 13:26 |
| minute | well, it is actually in there. phandle 0x5e | 13:29 |
| minute | josch: this is likely the culprit https://github.com/torvalds/linux/commit/15ca3f00a1e3a56811371a63bf0c067603226eae | 13:32 |
| minute | they even fixed our dts upstream! | 13:32 |
| josch | minute: should i revert that patch in our 5.19 kernel? | 13:33 |
| minute | no | 13:35 |
| minute | i'm just fixing our dts | 13:35 |
| minute | josch: https://source.mnt.re/reform/reform-debian-packages/-/commit/ce0ca34ff1af3e3410374fd914df2fd20edb318e | 13:36 |
| minute | (chassis-type is another thing they added, not critical but i'm just including it while i'm at it) | 13:36 |
| josch | oh cool | 13:36 |
| minute | fixed wrong whitespace in next commit | 13:36 |
| josch | lets see what happens :) | 13:36 |
| minute | yeah! | 13:37 |
| - mtm (QUIT: Ping timeout: 268 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 14:03 | |
| minute | we have a new build on staging, will test | 14:26 |
| minute | hmm, still no display | 14:33 |
| minute | that error is gone though | 14:34 |
| minute | there's no /dev/fb0 | 14:34 |
| minute | [ 6.476065] imx-dcss 32e00000.display-controller: failed to get pll_phy_ref clock | 14:36 |
| minute | probably IMX8MQ_CLK_PHY_27MHZ is not there/patch is not 100% correc t | 14:37 |
| josch | minute: the problem is probably mnt3004-MNT-Reform-imx8mq-add-PHY_27M-clock.patch | 14:38 |
| josch | that was the patch that I wasn't sure about when rebasing the patches | 14:39 |
| minute | will check | 14:42 |
| minute | josch: in my monolithic 6.0 test kernel, i by accident did not include any of those clk27 patches... and it still worked. so i'm gonna check what they did in 6.0 | 14:49 |
| minute | or maybe it just worked because i was reusing the old dtb | 14:50 |
| + wielaard (~mjw_@83-87-18-245.cable.dynamic.v4.ziggo.nl) | 14:52 | |
| minute | ahhh the difference is that i did not apply the hdmi-dcss patch in my test | 15:03 |
| - mechs (QUIT: Ping timeout: 252 seconds) (~mechs@190.218.136.138) | 15:14 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by wielaard!~mjw_@83-87-18-245.cable.dynamic.v4.ziggo.nl))) (~mark@gnu.wildebeest.org) | 15:14 | |
| * wielaard -> mjw | 15:15 | |
| + mark__ (~mark@gnu.wildebeest.org) | 15:15 | |
| + mechs (~mechs@190.218.136.138) | 15:26 | |
| - Boostisbetter (QUIT: Ping timeout: 260 seconds) (4a410829d7@irc.cheogram.com) | 15:38 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 15:39 | |
| minute | the error doesn't make a lot of sense to me. the clock itself is there (it's just a fixed 27mhz clock) root@reform:/sys/kernel/debug/clk# cat /sys/kernel/debug/clk/hdmi_phy_27m/clk_rate | 15:43 |
| minute | 27000000 | 15:43 |
| minute | it must be failing to get this clock for another reason | 15:43 |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw_@83-87-18-245.cable.dynamic.v4.ziggo.nl) | 15:45 | |
| minute | interestingly, clock-hdmi-phy-27m at the end of dt dump from the running kernel does not have a phandle | 15:46 |
| minute | aha! > [ 2.028063] i.MX clk 303: register failed with -17 | 15:50 |
| minute | that's the one | 15:50 |
| + mjw (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 15:51 | |
| minute | that's an error thrown by drivers/clk/imx/clk.c | 15:52 |
| josch | at least all the errors seem to be properly reported :) | 15:56 |
| minute | > #defineEEXIST17/* File exists */ | 15:57 |
| minute | might be hitting "check to see if a clock with this name is already registered" in clk.c | 15:59 |
| minute | ohhhh maybe it got added to dtsi? | 16:01 |
| minute | yes https://github.com/torvalds/linux/blob/7d0d3fa7339ed5a06d6608b7cde9f079eba62bb1/arch/arm64/boot/dts/freescale/imx8mq.dtsi#L66 | 16:01 |
| minute | so maybe we can just reference this by phandle and drop the extra hw clock 303 | 16:02 |
| minute | this might do the trick https://source.mnt.re/reform/reform-debian-packages/-/commit/30e935c36ca6f8980350f25d8c7715769382a066 | 16:06 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 16:09 | |
| - hl (QUIT: Quit: ZNC - https://znc.in) (~hl@user/hl) | 16:22 | |
| + hl (~hl@user/hl) | 16:30 | |
| josch | minute: when you add or remove patches you also have to adjust this list: https://source.mnt.re/reform/reform-debian-packages/-/blob/main/linux/build.sh#L89 | 16:44 |
| josch | (I just did that and pushed) | 16:44 |
| josch | minute: I could also make this script just apply all the patches from the linux/patches directory but for that to work we'd have to rename them in alphanumeric order so that the script knows in which order to apply them | 16:44 |
| minute | josch: ooohh | 16:45 |
| josch | would you like me to do the renaming? maybe by prefixing them with 0001, 0002, ...? | 16:45 |
| minute | thanks! | 16:45 |
| minute | that would be great, yes... | 16:45 |
| minute | lets see if this one works | 16:45 |
| josch | yes, i'll only do the renaming once we have a working kernel again | 16:46 |
| + monkeybusiness (~monkbusy@user/monkeybusiness) | 17:08 | |
| bkeys | I got about another week and my Reform will be repaired and hopefully shipped back | 17:11 |
| josch | \o/ | 17:13 |
| josch | minute: kernel built successfully | 17:15 |
| josch | three times is a charm! :D | 17:15 |
| minute | oh :0 gonna try it then! | 17:18 |
| minute | bkeys: are they exchanging the edp chip? | 17:19 |
| minute | alright. new errors | 17:24 |
| minute | [ 7.419065] phy phy-30a00300.dphy.2: Failed to configure PHY with invalid PHY mode: 0 | 17:24 |
| minute | [ 7.419076] nwl-dsi 30a00000.mipi-dsi: [drm:nwl_dsi_bridge_mode_set [nwl_dsi]] *ERROR* Failed to configure DSI phy: -22 | 17:24 |
| minute | we have a panel and framebuffer though | 17:24 |
| minute | but no display output of course. | 17:24 |
| bkeys | minute: Yes I sent them everything you said when you troubleshooted it | 17:27 |
| bkeys | Once I get it back working I will write a post on the community endorsing them | 17:27 |
| minute | josch: possibly we miss https://github.com/torvalds/linux/commit/69ed3dd6be9cf5d75252940b9a927dff4bab7860 | 17:29 |
| minute | confirming | 17:29 |
| minute | why is 5.19 so buggy | 17:29 |
| minute | yeah, we're hitting this, it got introduced in 5.19 https://github.com/torvalds/linux/commit/3fbae284887de24c83e88b1ebbb0a5fe8dbe9ac1#diff-6d31c533aa3108fa98483a2f53b05458cb8483171b77e0d6caa04f74e950a995R454 | 17:35 |
| minute | but setting the mode in nwl-dsi was introduced only in 6.0 | 17:35 |
| minute | absurd | 17:35 |
| minute | here we go again https://source.mnt.re/reform/reform-debian-packages/-/commit/43887f9f5d205af34ac86c772fc17b850669aaed | 17:44 |
| bkeys | I think what I will do is if I get a booting Fedora system and the Fedora kernel doesn't support the Reform well then I will make a .rpm for the v3 kernel y'all use until stuff gets mainlined | 17:47 |
| minute | has someone tried this already on reform? https://docs.waydro.id/usage/install-on-desktops#install-waydroid | 17:53 |
| - monkeybusiness (QUIT: Remote host closed the connection) (~monkbusy@user/monkeybusiness) | 18:01 | |
| + monkeybusiness (~monkbusy@103.87.94.35) | 18:02 | |
| eery | bkeys: the kernel Fedora ships doesn't work | 18:17 |
| eery | minute: I tried, but the v3 sysimage kernel was missing some stuff it needed, so it never "booted" properly | 18:18 |
| - monkeybusiness (QUIT: Changing host) (~monkbusy@103.87.94.35) | 18:19 | |
| + monkeybusiness (~monkbusy@user/monkeybusiness) | 18:19 | |
| eery | this was using the Fedora COPR-packaged version, so maybe it assumed something specific to the Fedora kernel that Waydroid itself doesn't strictly need | 18:19 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 18:21 | |
| eery | guess it would probably work on debian since it'd build the extra modules with dkms | 18:22 |
| minute | eery: hmm! there is a waydroid debian repo | 18:22 |
| minute | yeah | 18:22 |
| minute | josch: finally, display is on | 18:24 |
| bkeys | eery: Does it at least boot to a tty session? | 18:26 |
| bkeys | If so I can SSH into it and install the v3 kernel via a custom kernel rpm | 18:27 |
| bkeys | I'd put it up on copr for everyone to benefit from | 18:27 |
| minute | did some quick tests with 5.19: sway, clapper with h264 decode, minetest with sound, chromium (with ozone), wifi, sleep/wake, all works | 18:28 |
| sknebel | \o/ | 18:29 |
| minute | (booted from encrypted nvme) | 18:29 |
| minute | will try hdmi next. there could still be dragons | 18:29 |
| minute | 4K hdmi display has wrong colors (green instead of black), otherwise computer works fine | 18:38 |
| minute | josch: i'm gonna release the latest build to the default repo | 18:41 |
| + reform2836 (~sevan@host-92-23-203-220.as13285.net) | 18:41 | |
| - reform2836 (QUIT: Client Quit) (~sevan@host-92-23-203-220.as13285.net) | 18:43 | |
| flowy | ACTION cheers | 18:50 |
| josch | minute: YES!! \o/ | 18:57 |
| minute | :D | 18:57 |
| josch | ACTION throws confetti at minute | 18:57 |
| minute | thanks! | 18:57 |
| josch | lets hope 6.0 will be a nicer experience XD | 18:57 |
| minute | i think it will | 18:57 |
| josch | minute: could you add some patch headers (just From: and a short Subject:) to the patches in linux/patches that do not have that yet? Then later tonight I'll restructure the patching stuff such that you just need to drop it into linux/patches for the script to pick it up. | 18:59 |
| josch | minute: otherwise I guess I put you for From: and the patch name as Subject: ? | 18:59 |
| josch | s/patch name/patch filename/ | 18:59 |
| minute | the following ones are by me: | 19:00 |
| minute | 0001-nwl-dsi-fixup-mode-only-for-LCDIF-input-not-DCSS.patch | 19:00 |
| minute | 0005-pci-imx6-add-support-for-internal-refclk-imx8mq.patch (based on nxp kernel, though, but doesn't matter -- i can remove the bulk of it soon, and rewrite the patch i think) | 19:01 |
| minute | mnt4001-lcdif-fix-pcie-interference.patch | 19:01 |
| minute | mnt4002-imx-gpcv2-wake-smccc.patch <- this adds some commented code that can be removed | 19:02 |
| minute | and that's it, everything else has From: and Subject: | 19:02 |
| + natalie- (~natalie@user/natalie) | 19:09 | |
| - natalie (QUIT: Ping timeout: 240 seconds) (~natalie@user/natalie) | 19:10 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark__!~mark@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:11 | |
| * mark__ -> mjw | 19:11 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:12 | |
| josch | minute: thank you! | 19:14 |
| vkoskiv | minute: Great work with the kernel patching! Updating very soon | 19:25 |
| vkoskiv | I seem to recall I need to manually mount my SD card to /boot for the upgrade to work | 19:25 |
| vkoskiv | I'm still booting off an SD card, but using internal nvme for data | 19:25 |
| vkoskiv | Yees, I found the instructions from the irc logs. From june. | 19:27 |
| vkoskiv | I'll sort this out at some point, but this works for now. | 19:28 |
| flowy | sounds like you might just need to fix yer fstab | 19:31 |
| minute | yep, in the fstab it shall go | 19:31 |
| flowy | i've been quiet in the channel for a while but that's basically because all has been well with my reform for the past months. use it on a daily basis and it's my primary computer when i'm travelling | 19:33 |
| minute | flowy: great to hear | 19:33 |
| minute | the new battery board (final version to test will arrive here this week) will add a bit of additional robustness | 19:34 |
| flowy | i'm currently on a trip in finland and have been making days worth of audio recordings in the forest | 19:34 |
| flowy | i use the reform to sort through the files | 19:34 |
| minute | oh cool | 19:34 |
| minute | flowy: which audio software do you use for that? | 19:35 |
| flowy | also, the built-in SD reader has been handy for this | 19:35 |
| flowy | i bring the reform to the forest and transfer files off in the field, then put the SD back into the recorder | 19:36 |
| vkoskiv | The SD card that ships with Reform is 32GB, right? | 19:36 |
| vkoskiv | And the internal emmc is 16 | 19:36 |
| jfred | Woo new battery board! :D | 19:36 |
| vkoskiv | I think the last time I encountered this I fixed my fstab, so /boot is indeed coming from the SD | 19:36 |
| vkoskiv | It's coming from mmcblk1p1. Trying to make sure that is indeed the SD card | 19:36 |
| flowy | i've just been using audacity to preview files, but then i use a combination of other programs like sox ffmpeg csound and pure data for filtering | 19:36 |
| vkoskiv | flowy: Whereabouts in Finland? | 19:37 |
| vkoskiv | I'm based in Helsinki. Reform meetup? :D | 19:37 |
| flowy | ah cool! close to fiskars | 19:37 |
| flowy | unfortunately i'm not sure i'll have time anymore but that could have been fun | 19:38 |
| vkoskiv | That's a fun place, I haven't visited (I think) | 19:38 |
| Boostisbetter | I've always wanted to go to Finland. | 19:38 |
| Boostisbetter | One of these days I will | 19:38 |
| flowy | vkoskiv: you're finnish? | 19:38 |
| vkoskiv | Yes. | 19:38 |
| flowy | yeah i am liking my visit here a lot. only briefly went through helsinki | 19:39 |
| flowy | can't say i'm extremely well travelled when it comes to remote parts of europe but actually the forests here are the first i've seen that remind me of where i'm from, pacific north west / coast of BC | 19:40 |
| flowy | love the rockiness | 19:40 |
| flowy | vkoskiv: actually i might have a bit of time on thursday | 19:44 |
| flowy | am going back to helsinki for the ferry and have a few hours before | 19:45 |
| vkoskiv | Woo! Let's see if we can arrange something | 19:52 |
| vkoskiv | Just coffee/hot beverage of choice and nerding out with reforms | 19:53 |
| vkoskiv | Sounds fun | 19:53 |
| vkoskiv | I work from 8-16 usually, but I have flexible hours. | 19:53 |
| flowy | nice! | 20:00 |
| Boostisbetter | I'm kind of jelly. | 20:03 |
| eery | bkeys: No internal display without the patches, not sure if the system comes up otherwise. Beyond the kernel I think there'd be a lot of extra work in adding support for the reform's fork of u-boot, versus just managing it manually | 20:06 |
| + kfx (~kfx@wopr.sciops.net) | 20:30 | |
| josch | eery: I recently tried 5.18 from Debian without our patches and you get an error from the kernel when attempting to boot | 20:36 |
| - buckket (QUIT: Quit: buckket) (~buckket@pdp8.buckket.org) | 20:38 | |
| flowy | Boostisbetter: i'm sure a germ meet-up can happen. i'd be into that too, in berlin anyways. | 20:40 |
| + buckket (~buckket@pdp8.buckket.org) | 20:41 | |
| Boostisbetter | flowy: Berlin would be cool. They do Superbooth there and MNT is from there as well. | 20:52 |
| vkoskiv | flowy: You're based in Germany? | 21:01 |
| bkeys | eery: Yeah that is what I am thinking, I will build the v3 kernel and package it into a .rpm | 21:02 |
| bkeys | The uboot should be separate from the Fedora cause I am using the .iso and not the .img | 21:02 |
| vkoskiv | So my fstab was sorted, I just upgraded and I'm running 5.19 now. | 21:02 |
| vkoskiv | Is there a changelog of what improvements happened to etnaviv? | 21:02 |
| vkoskiv | I guess I can go look at the kernel git logs | 21:02 |
| eery | bkeys: How do you intend to use the ISO image with Reform? | 21:05 |
| - klardotsh (QUIT: Ping timeout: 268 seconds) (~klardotsh@98.97.35.142) | 21:13 | |
| - monkeybusiness (QUIT: Ping timeout: 252 seconds) (~monkbusy@user/monkeybusiness) | 21:19 | |
| bkeys | My plan is to have the /boot partition and the custom uboot on the emmc and load the rest of the OS on the nvme | 21:26 |
| bkeys | So dd the iso onto a USB drive or some other bootable medium to do the initial installation, and set the installer to do said partition scheme | 21:26 |
| bkeys | It should "just work" since josch's uboot has the distroboot functionality, so I have an EFI implementation | 21:27 |
| eery | uh, I'm 99% sure the unmodified installer isn't going to work, so at a minimum you'll need to generate a new ISO with the custom reform kernel | 21:30 |
| bkeys | Or I could just run the installer on a different ARM machine so the drive has everything but the /boot partition and copy the /boot partition over | 21:31 |
| bkeys | I believe I can configure the installer to install custom packages, which could include my reform kernel | 21:32 |
| eery | any reason you want to run anaconda instead of just copying the files from raw image over? | 21:32 |
| bkeys | Having an encrypted install would be nice | 21:32 |
| bkeys | I don't really care how it gets done, I just want to use Fedora on my Reform when I get it back | 21:33 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 21:33 | |
| eery | well, that would definitely make installation more complicated | 21:38 |
| bkeys | Yeah, although push comes to shove it'd be fine without it since most of my actual data is on a network share on my power9 server | 21:39 |
| bkeys | I think that it can definitely be done | 21:40 |
| eery | I figure if you had a working fedora-like kernel and initramfs you could get it working by mimicking the default anaconda setup of a LUKS encrypted LVM volume and pass rd.luks.uuid in the kernel args to get it to boot | 21:40 |
| bkeys | That or I could just use anaconda and then copy the installation over the way I do a .img file | 21:41 |
| eery | and hope the various UUIDs don't change | 21:41 |
| bkeys | The only thing is I would copy the /boot partition onto the eMMC which can be done by booting from the SD card | 21:42 |
| eery | not sure if any of those are dependent on hardware identifiers | 21:42 |
| bkeys | I don't believe so | 21:42 |
| bkeys | I just copy /boot over and modify the fstab | 21:42 |
| eery | you'd probably want to tune the LUKS configuration to try and squeeze out some performance | 21:44 |
| eery | or does luks/dm-crypt make use of whatever AES acceleration is on the A53? | 21:47 |
| eery | I just have this feeling FDE would be dreadful on the Reform | 21:49 |
| eery | encrypting /home or something is the compromise I think I'd consider once I have an nvme drive | 21:50 |
| + monkeybusiness (~monkbusy@103.87.94.35) | 21:50 | |
| - monkeybusiness (QUIT: Changing host) (~monkbusy@103.87.94.35) | 21:57 | |
| + monkeybusiness (~monkbusy@user/monkeybusiness) | 21:57 | |
| josch | eery: https://community.mnt.re/t/does-drive-encryption-cause-the-reform-to-slowdown/871 | 21:57 |
| eery | josch: oh nice - makes me wonder how much worse it'd perform with the CPUs pegged at 100%, like when ... browsing half the JS-laden internet in chromium :| | 22:04 |
| eery | I just feel like it'd take a noticeable hit in "time for things to happen after pressing a button" | 22:05 |
| josch | eery: are you attempting to read from and write to your disk at the same time as you access JS-laden internet? | 22:05 |
| eery | I mean yes, so far as both firefox and chromium lean heavily into disk caches | 22:06 |
| eery | I don't have a good idea for the average disk load a typical page reads but I know both browsers keep fairly enormous caches on drive and seem to make heavy use of them to avoid fetching media and the like | 22:07 |
| eery | so I'd guess the combo of limited memory for a block cache + slow CPU + lack of AES accelerator (?) + full CPU utilization would be unpleasant | 22:08 |
| josch | eery: and browsers will call fsync all the time or where would the slowdown come from? | 22:12 |
| josch | eery: even if browsers write to disk a lot, that writing will not be synchronous but will be written to disk whenever there is time to do so. | 22:12 |
| + klardotsh (~klardotsh@98.97.35.142) | 22:16 | |
| minute | eery: i use chromium on encrypted nvme all the time. | 22:30 |
| vkoskiv | Speaking of chromium, I use Firefox for privacy reasons, but I never figured out how to configure that as a default | 22:34 |
| vkoskiv | I spent some time on the problem a few months ago, but I just never got Firefox to open when I, for instance, open a link from xfce4-terminal | 22:35 |
| minute | vkoskiv: that can be tricky, there are several sources of truth depending on the framework the application is using. xfce4-terminal is gtk3 | 22:39 |
| minute | one thing to try: > xdg-settings set default-web-browser | 22:41 |
| josch | vkoskiv: what does your /usr/bin/x-www-browser point to? | 22:41 |
| minute | or maybe first `get` that | 22:41 |
| josch | if x-www-browser and gnome-www-browser point to the browser you want it's probably what minute said | 22:44 |
| vkoskiv | josch: It points to /etc/alternatives/x-www-browser, which points to /usr/bin/firefox | 23:01 |
| vkoskiv | minute: xdg-settings did the trick! | 23:02 |
| eery | josch: minute: cool, glad to hear it's not actually bad | 23:02 |
| vkoskiv | I'm around 87.2% certain I tried that when I last messed around with this, but it worked now. | 23:02 |
| minute | Sep 06 23:08:48 reform waydroid[2596]: [23:08:48] modprobe: FATAL: Module ashmem_linux not found in directory /lib/modules/5.19.0-reform2-arm64 | 23:11 |
| minute | S | 23:11 |
| minute | we have binder_linux but we don't have ashmem_linux | 23:11 |
| minute | apparently upstream removed ashmem? | 23:12 |
| eery | oh, fun | 23:13 |
| eery | the COPR module I use to install waydroid has the notice "NOTE: Kernels 5.18.18 to 5.19.5 are broken" | 23:13 |
| eery | s/module/repo | 23:13 |
| eery | although same module also notes it doesn't require ashmem | 23:14 |
| minute | oh https://github.com/waydroid/waydroid/blob/02d2b113c6031cc41f610ce338dfd6a7a4f2a316/debian/changelog#L3 | 23:14 |
| minute | yeah | 23:14 |
| eery | something else breaks tho | 23:15 |
| minute | i get a blue window so far | 23:16 |
| minute | changing ro.hardware.gralloc to default in /var/lib/waydroid/waydroid_base.prop gives me some graphics | 23:25 |
| minute | but i guess it's not accelerated | 23:25 |
| minute | i get a lock screen but it's crashing all the time "Trebuchet keeps stopping" | 23:27 |
| - mtm (QUIT: Ping timeout: 260 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 23:27 | |
| vkoskiv | My love for my Reform has significantly increased recently. I now use it more than my main desktop. | 23:34 |
| vkoskiv | I forgot how significant sleep/suspend was to my workflow. | 23:34 |
| vkoskiv | Still haven't had it fail once since it started working. | 23:34 |
| minute | cool | 23:36 |
| bkeys | See I want to have a power9 workstation and Reform as my laptop | 23:36 |
| - monkeybusiness (QUIT: Ping timeout: 244 seconds) (~monkbusy@user/monkeybusiness) | 23:57 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!