| - S0rin (QUIT: Ping timeout: 268 seconds) (~S0rin@user/s0rin) | 01:08 | |
| + S0rin (~S0rin@user/s0rin) | 01:13 | |
| - S0rin (QUIT: Ping timeout: 252 seconds) (~S0rin@user/s0rin) | 01:26 | |
| + S0rin (~S0rin@user/s0rin) | 01:27 | |
| - wielaard (QUIT: *.net *.split) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 02:00 | |
| - mjw (QUIT: *.net *.split) (~mark@gnu.wildebeest.org) | 02:00 | |
| - Asmadeus (QUIT: *.net *.split) (~asmadeus@240b:13:8c80:d300:e:98c:8000:d300) | 02:00 | |
| - Boostisbetter (QUIT: *.net *.split) (4a410829d7@irc.cheogram.com) | 02:00 | |
| - jjbliss (QUIT: *.net *.split) (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 02:00 | |
| - lexik (QUIT: *.net *.split) (~lexik@171.25.222.217) | 02:00 | |
| - minute (QUIT: *.net *.split) (~mntirc@softboy.mntmn.com) | 02:00 | |
| - iank (QUIT: *.net *.split) (~iank@fsf/staff/iank) | 02:00 | |
| - ggoes (QUIT: *.net *.split) (~gregf@fsf/staff/ggoes) | 02:00 | |
| + Asmadeus (~asmadeus@240b:13:8c80:d300:e:98c:8000:d300) | 02:00 | |
| + lexik (~lexik@171.25.222.217) | 02:00 | |
| + mjw (~mark@gnu.wildebeest.org) | 02:00 | |
| + jjbliss (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 02:00 | |
| + minute (~mntirc@softboy.mntmn.com) | 02:01 | |
| + iank (~iank@fsf/staff/iank) | 02:02 | |
| - mtm (QUIT: Ping timeout: 244 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:04 | |
| + ggoes (~gregf@fsf/staff/ggoes) | 02:04 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:47 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 03:47 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 04:02 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 04:02 | |
| + robin (~robin@user/terpri) | 05:09 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 08:21 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 08:22 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 08:30 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 08:30 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 09:05 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 10:07 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 10:07 | |
| - buckket (QUIT: Ping timeout: 248 seconds) (~buckket@pdp8.buckket.org) | 11:16 | |
| + buckket (~buckket@pdp8.buckket.org) | 11:21 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 12:41 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by wielaard!~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440))) (~mark@gnu.wildebeest.org) | 12:42 | |
| * wielaard -> mjw | 12:42 | |
| + mark__ (~mark@gnu.wildebeest.org) | 12:42 | |
| eery | okay, so screensharing/webRTC from Xwayland is actually usable | 13:14 |
|---|---|---|
| eery | the output stream has a very low framerate but unlike xdg-desktop-portal-wlr it doesn't bottleneck the compositor | 13:15 |
| minute | eery: interesting. what are you using, chromium? | 13:40 |
| - buckket (QUIT: Quit: buckket) (~buckket@pdp8.buckket.org) | 13:40 | |
| + buckket (~buckket@pdp8.buckket.org) | 13:44 | |
| - mtm (QUIT: Ping timeout: 268 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 14:03 | |
| eery | yep, tested with discord running inside chromium (inside a rooted Xwayland), will probably try teams later today | 14:24 |
| minute | eery: ah, sounds good. rooted xwayland ftw | 14:35 |
| eery | I've been using wayfire and it's awesome, I've got a 3x3 square of workspaces, throw a couple Xwayland instances up, some FreeRDP/VNC sessions, and just zoom in/out between them | 14:38 |
| eery | nice and snappy even on the gc7000 | 14:38 |
| minute | oh, i really need to spend some time with wayfire | 14:38 |
| minute | do you use a dock or launcher or something? | 14:39 |
| eery | also, waypipe ssh -- <host> Xwayland -shm is a fun trick | 14:39 |
| eery | nope, well, except for rofi | 14:39 |
| eery | I've bound super + ctrl + space to launch foot terminals and made a plugin to place new windows directly under the cursor | 14:40 |
| eery | personally I dial down the effects/animations and think it's just a nice stacking compositor | 14:42 |
| minute | nice | 15:03 |
| minute | can it also do sth like half screen tiling with some shortcuts? | 15:04 |
| eery | yep, it does both horizontal/vertical splits and 1/4 tiling, with both keyboard and drag-to-edge shortcuts | 15:20 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 15:21 | |
| - bkeys (QUIT: Client Quit) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 15:22 | |
| eery | which is really the most tiling I did in i3 for years, not many situations where further splits help (IMO) | 15:27 |
| eery | currently I feel it's best to keep terminals around the same width and just shuffle them around, temp. maximizing when I need the space | 15:29 |
| minute | same | 16:02 |
| minute | i think wayfire could be a really good intermediate level desktop for us | 16:03 |
| minute | like, for people that would like the comfort of gnome, but gnome is really too heavy | 16:03 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 16:08 | |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 16:09 | |
| eery | With a good dock or GUI launcher it probably could be | 16:10 |
| eery | it does ship with one, but it's not particularly amazing | 16:10 |
| minute | ok | 16:13 |
| Boostisbetter | I really love sway. I find it to be the perfect middle ground | 16:36 |
| rah | how usable is vanilla Debian on an mnt-reform? | 18:06 |
| minute | rah: what exactly do you mean by that? | 18:08 |
| minute | Boostisbetter: yeah but i know people who can't remember the key combos | 18:08 |
| rah | minute: if one installs Debian downloaded from debian.org, is any functionality missing compared to MNT's "system images"? | 18:09 |
| - jjbliss (QUIT: Remote host closed the connection) (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 18:15 | |
| + jjbliss (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 18:16 | |
| rah | minute: does that make sense? | 18:23 |
| Boostisbetter | minute: it helps that you can define your own key combos if you want. | 18:28 |
| + MajorBiscuit (~MajorBisc@62-71-185.netrun.cytanet.com.cy) | 18:31 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 18:37 | |
| Boostisbetter | rah: there is some customs stuff needed for the Reform, I imagine that is the stuff you need. The plan is to make the Reform image match mainline though, and I know that there is work on that front which has been done. System image 3, for example, has made a lot of progress on that front. | 18:41 |
| rah | Boostisbetter: what custom stuff? | 18:42 |
| Boostisbetter | minute: I know this has been discussed before, but just one more time, hibernation does not work on the Reform, correct? | 18:42 |
| Boostisbetter | minute: and there is no hope of it ever working, right? | 18:42 |
| rah | or perhaps a better question would be: what won't work on vanilla Debian? | 18:42 |
| Boostisbetter | rah: no idea, but things like the suspend script come to mind. | 18:43 |
| Kooda | Video output doesn’t work on mainline yet, iirc | 18:44 |
| rah | O_o | 18:47 |
| rah | I see | 18:47 |
| minute | i think it does | 18:48 |
| minute | but not hdmi | 18:48 |
| minute | afaik LCDIF based internal display is in mainline | 18:48 |
| minute | Boostisbetter: i don't know the current status of hibernation | 18:49 |
| minute | Boostisbetter: there's no real technical reason why it should _not_ work | 18:49 |
| Boostisbetter | minute: that is good to hear. I love hibernation. It is my preferred workflow tool. Glad to see that there might be light at the end of the tunnel for the Reform / Pocket Reform. | 18:50 |
| minute | rah: i think probably josch and vagrantc can answer the debian question better | 18:50 |
| kfx | mainline linux supports the reform now? | 18:54 |
| minute | kfx: since a longer time, but no hdmi support | 18:55 |
| minute | kfx: https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/freescale/imx8mq-mnt-reform2.dts | 18:56 |
| kfx | very nice, glad to see it | 18:56 |
| minute | and also no support for the mpcie slot because there's a tiny thing missing to enable internal refclk | 18:56 |
| jjbliss | So, I think I've broken my Reform's SOM. I posted on the forums a few days ago, but no advice there yet, so I thought I'd see if anyone here has any troubleshooting suggestions. | 19:16 |
| jjbliss | A few weeks back, when switching to the V3 image I switched to SD to try the new image and then back to eMMC to run things from there. When switching back, the screwdriver I was using to flip the boot selection switch on the SOM slipped and knocked into the board. No visible damage, but since then, the reform will not boot, nothing showing on the screen. | 19:16 |
| jjbliss | My assumption is that I damaged the SOM. Unless there is something I could have messed up on the eMMC that prevents it from booting even from SD? I guess I don't have a perfect understanding of the SOM's boot process. | 19:17 |
| - MajorBiscuit (QUIT: Quit: WeeChat 3.5) (~MajorBisc@62-71-185.netrun.cytanet.com.cy) | 19:18 | |
| minute | jjbliss: emmc is not needed to boot from sd | 19:18 |
| jjbliss | That's what I thought | 19:18 |
| minute | jjbliss: so probably som damage yeah :/ anything visible upon close inspection? | 19:18 |
| minute | jjbliss: the som has a green led, does it still turn on? | 19:19 |
| jjbliss | LED still turns on | 19:19 |
| jjbliss | I've tried hooking up a serial connection, but that doesn't give me anything either. | 19:19 |
| minute | jjbliss: if you can live with 2gb ram for a while, you could get one from mouser https://eu.mouser.com/ProductDetail/Boundary-Devices/Nit8MQ_SOM_2r16e?qs=B6kkDfuK7%2FANVHg5JpGi9A%3D%3D | 19:20 |
| minute | jjbliss: 4gb has 24 weeks lead time at the moment | 19:20 |
| jjbliss | That might be worth it. | 19:20 |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark__!~mark@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:21 | |
| * mark__ -> mjw | 19:21 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:21 | |
| minute | we're working on a bunch of alternate cpu module options, but the nitrogen8m som is the tried and tested one | 19:21 |
| jjbliss | Yeah, I'm definitely interested in trying some of the new options, but it would be nice to get the Reform working sometime soon. | 19:22 |
| jjbliss | Just to check, anything special I'll need to do to get the 2gb SOM working? Or should it just plug in and work? | 19:27 |
| minute | jjbliss: i think you need u-boot from jacquelines branch, one moment | 19:40 |
| minute | jjbliss: https://source.mnt.re/reform/reform-boundary-uboot/-/merge_requests/7 | 19:40 |
| jjbliss | Thanks! | 19:40 |
| + ajr (~ajr@user/ajr) | 19:43 | |
| josch | rah: i'm maintaining the shell scripts that build the reform system image and i'm also maintainting all the patches of Debian packages that we ship in our own repo -- what exactly do you want to know? | 20:57 |
| rah | josch: what doesn't work on vanilla Debian | 21:31 |
| josch | rah: the internal display doesn't come up without our kernel patches that have not been mainlined yet | 21:31 |
| rah | josch: what about things that aren't kernel bound? | 21:33 |
| josch | rah: like what? | 21:33 |
| rah | josch: you tell me | 21:34 |
| josch | you are asking which things would not work on vanilla debian, right? well, the only thing we patch is the kernel, so there should be no differences for things that aren't kernel bound | 21:34 |
| rah | 17:43 < Boostisbetter> rah: no idea, but things like the suspend script come to mind. | 21:35 |
| rah | so suspend works? | 21:35 |
| rah | I mean, pm-utils? | 21:35 |
| josch | rah: not for me -- but that has nothing to do with vanilla or not | 21:35 |
| minute | rah: we have a systemd hook that turns of wifi before suspend because of a kernel/hw hook | 21:35 |
| minute | s/hook/bug | 21:36 |
| minute | ah and also hantro_vpu | 21:36 |
| josch | rah: what minute said is correct but that's just a config option -- maybe i misunderstand what you mean by "vanilla"? | 21:36 |
| rah | josch: I want to use Debian from debian.org, not from MNT | 21:38 |
| rah | I'd like to know what won't work if I were to do that | 21:38 |
| josch | rah: if you like i can prepare you a system image that only contains packages from debian.org and no package from the mnt repos | 21:38 |
| josch | or i can explain to you how you can do it yourself | 21:39 |
| rah | O_o | 21:39 |
| josch | hm? | 21:39 |
| rah | 20:38 < rah> josch: I want to use Debian from debian.org, not from MNT | 21:39 |
| josch | yes? | 21:39 |
| rah | not an image from MNT containing packages from debian.org | 21:40 |
| josch | do you want to use debian-installer to install it? | 21:40 |
| rah | yes | 21:40 |
| josch | this can theoretically work, but needs further work on making flash-kernel work correctly | 21:41 |
| josch | i haven't managed to get d-i working yet | 21:41 |
| josch | but i did manage to start d-i from a usb stick | 21:41 |
| josch | d-i will also install a system -- the result is just not bootable | 21:41 |
| josch | the reason is, that d-i needs to set up a boot.scr from u-boot and currently it doesn't do that on the reform | 21:42 |
| rah | I see | 21:43 |
| vagrantc | but that's just because there's no flash-kernel entry for mnt/reform2 ? | 21:44 |
| vagrantc | e.g. it will otherwise work? | 21:44 |
| vagrantc | i also noticed there's an xwayland package last i looked with patches ... | 21:44 |
| josch | vagrantc: there is a flash-kernel entry for the reform but i have no idea how the platform autodetection works in debian-installer | 21:44 |
| vagrantc | i think i "downgraded" that to the version from debian and wayland stuff still seemed to work ... maybe at reduced performance? | 21:45 |
| vagrantc | josch: it just runs flash-kernel | 21:45 |
| vagrantc | rah: you could add the flash-kernel entry manually while debian-installer is running | 21:45 |
| vagrantc | rah: or add a debconf preseed hook to create it | 21:45 |
| vagrantc | but you'd still get no LCD | 21:46 |
| vagrantc | so serial console only | 21:46 |
| josch | what you'd need to add manually would be to write "MNT Reform 2" to /etc/flash-kernel/machine | 21:46 |
| vagrantc | josch: oh, there is an entry? but it just doesn't detect correctly? | 21:46 |
| josch | vagrantc: yes, i don't know how flash-kernel would do the auto-detection | 21:47 |
| vagrantc | josch: it looks at /proc/device-tree/model (or the equivalent /sys entry maybe) | 21:47 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 21:48 | |
| vagrantc | looks like /proc/device-tree is a symlink these days /proc/device-tree -> /sys/firmware/devicetree/base | 21:48 |
| vagrantc | josch: and that is populated from the model property in the .dtb | 21:49 |
| josch | how does it know the right dtb? because u-boot passes the correct path? | 21:49 |
| minute | btw vagrantc josch why would there be no LCD? | 21:50 |
| minute | because it is supported in mainline | 21:50 |
| josch | minute: maybe but mine doesn't come up with vanilla linux | 21:50 |
| minute | aha, so that's a bug then | 21:50 |
| minute | that we should probably address upstream :| | 21:50 |
| josch | which is apparently fixed by one of the patches we carry | 21:50 |
| vagrantc | minute: mainline as of what? | 21:52 |
| vagrantc | 5.19? 5.18? | 21:52 |
| vagrantc | either that, or there are further configuration options needed in the kernel | 21:52 |
| minute | hmm, when was that merged... a longer time ago | 21:54 |
| josch | i'm going to try out vanilla 5.18 again tomorrow | 21:54 |
| minute | https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/arm64/boot/dts/freescale/imx8mq-mnt-reform2.dts?id=eb893e34309ce2153003f3f2a13cdbfdea939e00 | 21:56 |
| minute | authorLucas Stach <dev@lynxeye.de>2021-12-18 19:20:20 +0100 | 21:56 |
| minute | committerShawn Guo <shawnguo@kernel.org>2022-02-11 11:16:17 +0800 | 21:56 |
| minute | don't know how i can see if this went into 5.17 or 5.18 | 21:57 |
| rah | I'm not so worried about the kernel or installation, I'm more concerned with the operation of the system once installed with the latest mainline kernel | 21:57 |
| rah | so suspend, audio, etc. | 21:57 |
| minute | rah: well, one is the prerequisite for the other | 21:57 |
| - Gooberpatrol66 (QUIT: Quit: Leaving) (~Gooberpat@user/gooberpatrol66) | 21:58 | |
| josch | rah: why do you ask what vanilla can do? just curiosity? | 22:02 |
| minute | https://crowdsupply.com/mnt/pocket-reform | 22:09 |
| vkoskiv_ | minute: !!! | 22:15 |
| josch | Boostisbetter: ^ | 22:15 |
| Boostisbetter | minute: take me money already! | 22:17 |
| Boostisbetter | *my | 22:17 |
| smj | minute! congrats | 22:20 |
| smj | I've still got a MNT reform pre-order but have since purchased 3 other units second hand :) | 22:21 |
| smj | I've been sharing them with other devs | 22:21 |
| smj | August 27th I will be doing an inperson demo in Kirkland WA - if anyone is interested in details, let me know. | 22:22 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 22:22 | |
| rah | minute: which one and other are you referring to? | 22:23 |
| rah | josch: because that's what I want to use | 22:23 |
| rah | josch: is perhaps the question you want to ask: why don't you want to use my work? :-) | 22:23 |
| rah | josch: assuming so, the answer is: because it's not sustainable | 22:24 |
| rah | at some point, you and MNT will stop working on the "system images" | 22:25 |
| vagrantc | minute: git describe --contains eb893e34309ce2153003f3f2a13cdbfdea939e00 | 22:26 |
| josch | rah: that is correct. But even if I upload the script to the debian.org server today, it would not increase sustainability. But it would meet your requirement of "only debian.org" -- so i don't think that's a good requirement. Sustainability is a good one, I agree. Do you want to join us working on this to spread the work across more shoulders and make it more sustainable in the process? :) | 22:26 |
| rah | josch: I'm not prepared to work on commercial hardware for free, no | 22:27 |
| josch | rah: is there non-commercial hardware? | 22:27 |
| rah | plenty | 22:28 |
| josch | like what? | 22:28 |
| rah | everything on eBay whose manufacturers have gone out of business | 22:28 |
| rah | perhaps I should be more specific: I'm not prepared to work for free on a company's products while that company is still in business | 22:29 |
| josch | okay | 22:30 |
| vagrantc | josch: maybe the last time you tested debian-installer flash-kernel wasn't yet updated in debian? | 22:42 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 22:43 | |
| josch | vagrantc: yes, that's possible -- so i want to try it out again | 22:43 |
| vagrantc | looks like it migrated end of april | 22:44 |
| smj | The link to reform-system-img.gz seems to be incorrect on reform-system.img.gz | 22:55 |
| smj | https://source.mnt.re/reform/reform-system-image/-/releases | 22:55 |
| smj | Sorry, there. | 22:55 |
| minute | smj: we're shipping out all mnt reform shop orders over the next week | 23:01 |
| minute | (existing ones) | 23:01 |
| vkoskiv_ | Do the firmware bits have dedicated maintainers or does minute do them as time permits? | 23:03 |
| smj | minute, would it be possible to adjust #R694348337 (#1274) to whenever pre-orders for the Pocket would be? | 23:03 |
| vkoskiv_ | (Shamelessly plugging my lil fix to the kbd firmware :D) | 23:03 |
| minute | smj: all existing reform orders are already built and packed, sorry | 23:04 |
| smj | If not, that would be the 4th reform I'd have. ;) I'm sure I can find someone who can use it, but also there may be other people waiting who it should go to | 23:04 |
| smj | ahh OK, I'll have 4 then :) | 23:04 |
| minute | that's great though :D | 23:04 |
| minute | we're gonna make some more | 23:04 |
| sknebel | woah, great news! | 23:05 |
| minute | vkoskiv_: unfortunately no maintainers yet for fw... josch helps me a great deal with system image and reform-tools etc, though. | 23:05 |
| sknebel | (both pocket and existing orders shipping :)) | 23:05 |
| smj | when the pocket is available, we'll certainly work on getting Plan9 running on it as well | 23:05 |
| vkoskiv_ | My fix should be trivial to merge, when you have the time. | 23:06 |
| vkoskiv_ | I've been running it on my unit since I wrote it, and it's been working as expected. | 23:06 |
| vkoskiv_ | Though the wake function still doesn't work, but that also happened on non-modified firmware | 23:07 |
| minute | i really need to get the spi battery status thing merged | 23:07 |
| vkoskiv_ | Yeah that's a nice one, for sure. | 23:07 |
| minute | smj: awesome, counting on it ^^ | 23:07 |
| vkoskiv_ | I did have one oopsie at Assembly where I totally forgot to monitor battery status :D | 23:07 |
| minute | btw i have new battery board pcbs at the office | 23:07 |
| minute | but waiting for the chips to try em | 23:08 |
| minute | with protector for each cell | 23:08 |
| vkoskiv_ | I saw you tweet about a new idea for the design, but I don't understand electronics enough to appreciate the difference | 23:08 |
| vkoskiv_ | (w.r.t. battery protector board) | 23:08 |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 23:08 | |
| vkoskiv_ | But I'm getting them whenever available. | 23:09 |
| josch | yeah, same here :) | 23:10 |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 23:11 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 23:12 | |
| vkoskiv_ | For me #1 annoyance, pretty much the only one, is no working sleep mode. | 23:13 |
| vkoskiv_ | Years of Mac laptops allowed me to get used to never shutting the thing off :D | 23:13 |
| minute | hmm, it should mostly work though | 23:14 |
| vkoskiv_ | Beyond that the thing works great, I can get real work done on there. I only run out of RAM if I'm doing some heavy testing with my renderer, but I can offload that to my desktop over the network | 23:14 |
| minute | vkoskiv_: are you on sysimage v3? | 23:14 |
| vkoskiv_ | minute: For me the issue is still that bizarre thing where the LPC/keyboard wake function is borked | 23:14 |
| vkoskiv_ | I'm on v3, yeah. Have been the whole time I had it. | 23:14 |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 23:14 | |
| minute | ah so the keyboard stays asleep? | 23:14 |
| vkoskiv_ | As far as I know, the actual system part of it works, but the actual LPC wake thing doesn't. I've re-flashed stock firmware several times. | 23:15 |
| vkoskiv_ | I described it on the forum, let me grab a link... | 23:15 |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 23:15 | |
| minute | ok | 23:15 |
| vkoskiv_ | This message describes the exact symptoms I have: https://community.mnt.re/t/standby-suspend-to-ram-mnt-reform/538/83 | 23:15 |
| vkoskiv_ | That sequence of events happens every time. Reflashed both LPC and keyboard at least twice each. | 23:16 |
| vkoskiv_ | I think the 0% never goes up before the step 2 of the symptom happens. I can check to make sure. | 23:16 |
| vkoskiv_ | Ah, I see I wrote it down below :D | 23:16 |
| minute | vkoskiv_: ahh i see | 23:18 |
| minute | vkoskiv_: so it is probably sending the wake command, but imx is not waking | 23:18 |
| vkoskiv_ | I did go over the relevant LPC/kbd code when I encountered it, but nothing looked weird. | 23:18 |
| vkoskiv_ | That might be what's happening, yeah. | 23:18 |
| minute | we can check that | 23:20 |
| vkoskiv_ | I have the thing with me here, I can do debugging. | 23:20 |
| minute | i just checked, so: | 23:21 |
| minute | sudo screen /dev/ttymxc2 57600 | 23:21 |
| minute | and then do the wake function on the keyboard | 23:21 |
| minute | the word "wake" and a nonprintable character should appear | 23:21 |
| minute | in screen | 23:21 |
| minute | the waking up LPC 0% behavior is normal btw, i'm getting that too | 23:22 |
| minute | if the word "wake" does not appear, the uart connection between imx/lpc is busted somehow or the uart interface in the imx is busted | 23:22 |
| minute | there might be alternative wake sources though | 23:23 |
| vkoskiv_ | sec, replugging batteries | 23:24 |
| minute | to clarify, i don't want you to put the laptop into sleep mode, we just want to see if the wakeup word appears on the internal serial port | 23:25 |
| vkoskiv_ | In the screen window, I got wake<?> | 23:26 |
| vkoskiv_ | Where <?> is the diamond with a question mark | 23:26 |
| vkoskiv_ | Some unprintable byte | 23:26 |
| minute | great | 23:26 |
| minute | so the serial connection and the command etc are fine | 23:27 |
| minute | maybe there's something wrong with the sleep hook then | 23:27 |
| minute | do you have /lib/systemd/system/reform-sleep.service ? | 23:27 |
| vkoskiv_ | I do. | 23:28 |
| vkoskiv_ | Enabled as well | 23:28 |
| minute | do you have /usr/sbin/reform-standby ? | 23:28 |
| minute | and does it include hantro_off ? | 23:28 |
| vkoskiv_ | Inactive (dead) | 23:28 |
| vkoskiv_ | Let me see | 23:28 |
| vkoskiv_ | It has hantro_off in there | 23:29 |
| minute | ok | 23:29 |
| minute | lets try doing this manually | 23:29 |
| minute | rmmod hantro_vpu | 23:29 |
| vkoskiv_ | ok | 23:29 |
| minute | echo enabled > /sys/devices/platform/soc@0/30800000.bus/30880000.serial/tty/ttymxc2/power/wakeup | 23:30 |
| minute | both as root | 23:30 |
| vkoskiv_ | permission denied for the second one | 23:30 |
| vkoskiv_ | With sudo | 23:30 |
| minute | yeah, you need to be root, sudo doesn't work with redirection | 23:30 |
| vkoskiv_ | Worked when switching user, yeah | 23:30 |
| minute | because the > works in the context of your shell | 23:30 |
| vkoskiv_ | Done | 23:30 |
| minute | ok | 23:30 |
| minute | echo mem > /sys/power/state | 23:31 |
| vkoskiv_ | Off it went! :D | 23:32 |
| minute | ok, then wake via circle, space | 23:32 |
| vkoskiv_ | kbd stayed lit, I assume that's a separate (optional) step | 23:32 |
| minute | yep | 23:32 |
| vkoskiv_ | And it's back! | 23:32 |
| minute | ha! | 23:32 |
| vkoskiv_ | I got the same output as before on the oled, but it actually woke this time | 23:32 |
| vkoskiv_ | Right back to where I was | 23:32 |
| minute | ok, so the hardware is probably fine | 23:32 |
| minute | maybe systemd doesn't call the hook for some reason, or something breaks in the script? | 23:33 |
| vkoskiv_ | I assume those scripts were updated when I've been running apt upgrade | 23:35 |
| minute | yes, they also work for me | 23:37 |
| minute | how do you normally suspend? | 23:38 |
| minute | i.e. which command do you use? | 23:38 |
| vkoskiv_ | And the proper way is 'systemctl suspend', right? | 23:38 |
| minute | yeah | 23:38 |
| vkoskiv_ | That now does nothing | 23:38 |
| vkoskiv_ | Oh, scratch that | 23:38 |
| vkoskiv_ | Just took a sec. | 23:38 |
| minute | yeah i think it has a sleep 1 in there | 23:39 |
| vkoskiv_ | And now the wake also works :D | 23:39 |
| vkoskiv_ | Let me reboot and see if that's still the case | 23:39 |
| vkoskiv_ | Okay, should have tried this today but now the suspend/resume just... work? | 23:41 |
| vkoskiv_ | Maybe one of those steps above fixed it, maybe some update earlier did | 23:41 |
| vkoskiv_ | Didn't work at all before, but now it seems to :D | 23:41 |
| minute | maybe it's the hantro fix | 23:41 |
| minute | maybe you didn't have that when you last tried | 23:41 |
| vkoskiv_ | I didn't do it manually after rebooting | 23:41 |
| minute | yeah i mean it hasn't always been in the script | 23:41 |
| vkoskiv_ | Ah. | 23:41 |
| minute | we only found it a while ago | 23:41 |
| vkoskiv_ | What does removing it actually do? Somehow limit the GPU? | 23:42 |
| minute | no no | 23:42 |
| minute | it's just the hw video decoder | 23:42 |
| vkoskiv_ | Oh right, mixed it up with vivante | 23:42 |
| minute | it's suspend/wake code in the kernel is buggy and breaks wake | 23:42 |
| minute | its | 23:42 |
| vkoskiv_ | Okay, thanks anyway! I'll mess around with this more. | 23:43 |
| minute | so we remove it before sleep and plug it back in after waking | 23:43 |
| minute | similar with the wifi card, it freaks out after waking up, so we disable it and reenable it | 23:43 |
| vkoskiv_ | Yeah I noticed it dropped the wifi for a bit. | 23:43 |
| vkoskiv_ | What if I'm hw decoding a video in mpv and then suspend+resume? | 23:44 |
| vkoskiv_ | I guess I'll try :D | 23:44 |
| minute | i think the video player might stop or crash | 23:44 |
| minute | vkoskiv_: can i ask you to write a note in the thread that it works for you now? might inspire others to retry as well | 23:47 |
| vkoskiv_ | Will do! | 23:48 |
| minute | thanks! | 23:49 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!