| + jacobk (~quassel@2603-8080-b200-7b02-77cb-6304-f9db-dda1.res6.spectrum.com) | 00:02 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@2603-8080-b200-7b02-77cb-6304-f9db-dda1.res6.spectrum.com) | 00:09 | |
| - mjw (QUIT: Ping timeout: 240 seconds) (~mjw@gnu.wildebeest.org) | 00:43 | |
| + jacobk (~quassel@2603-8080-b200-7b02-77cb-6304-f9db-dda1.res6.spectrum.com) | 00:44 | |
| - Boostisbetter (QUIT: Ping timeout: 245 seconds) (4a410829d7@irc.cheogram.com) | 00:54 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@2603-8080-b200-7b02-77cb-6304-f9db-dda1.res6.spectrum.com) | 00:59 | |
| + jacobk (~quassel@2603-8080-b200-7b02-77cb-6304-f9db-dda1.res6.spectrum.com) | 01:46 | |
| - deflated8837 (QUIT: Ping timeout: 260 seconds) (~deflated8@50.53.206.61) | 01:57 | |
| + sl (~sl@contrib.inri.net) | 01:57 | |
| - jacobk (QUIT: Ping timeout: 246 seconds) (~quassel@2603-8080-b200-7b02-77cb-6304-f9db-dda1.res6.spectrum.com) | 02:02 | |
| - mtm (QUIT: Ping timeout: 245 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:03 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 02:07 | |
| + deflated8837 (~deflated8@50.53.206.61) | 02:11 | |
| + jacobk (~quassel@47-186-122-163.dlls.tx.frontiernet.net) | 02:31 | |
| - nsc (QUIT: Ping timeout: 245 seconds) (~nicolas@133-48-142-46.pool.kielnet.net) | 03:39 | |
| + nsc (~nicolas@39-98-142-46.pool.kielnet.net) | 03:40 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:09 | |
| violet | hmm, still not seeing HDMI output even after the kernel update to 6.3 | 05:19 |
|---|---|---|
| - sl (QUIT: Quit: Leaving...) (~sl@contrib.inri.net) | 05:31 | |
| violet | like the display does not even show up in /sys/class/drm the way the eDP does | 05:34 |
| violet | i found this https://source.mnt.re/reform/reform-debian-packages/-/blob/main/linux/imx8mq-mnt-reform2-hdmi.dts but this does not seem to be whats in my dtb in /boot that i installed last night from the repos | 05:59 |
| violet | because i converted my dtb back to dts and the only hdmi related thing in it was the clock | 05:59 |
| violet | oooh can the imx8mq not drive both the internal and external display simultaneously? do i need to swap device trees based on what display i want | 06:15 |
| violet | i wonder if i could pick that at bootup somehow like selecting a boot entry or holding a key | 06:22 |
| violet | oh i checked the logs, i see the "reform-display-config" script now, will try that | 06:24 |
| violet | ok well that did not break antyhing at least and its set to the hdmi dtb now but still no hdmi display. maybe im missing a module in the initramfs-tools/modules | 06:31 |
| josch | violet: you indeed have to swap the dtb u-boot loads which is done by editing /etc/flash-kernel/machine and regenerating the initramfs | 06:34 |
| josch | violet: the reform-display-config script automates that | 06:34 |
| josch | violet: can you tell me the output of this command: | 06:34 |
| josch | cat /proc/device-tree/model | 06:34 |
| violet | MNT Reform 2 HDMI | 06:34 |
| josch | okay, that's good | 06:35 |
| violet | it was MNT Reform 2 before but i just swapped the device tree to the HDMI one and rebooted | 06:35 |
| josch | then you successfully loaded the correct dtb | 06:35 |
| josch | with that dtb you can get both the internal display and hdmi at the same time | 06:35 |
| violet | no difference in flash-kernel or initramfs-tools configs from the v3 image | 06:35 |
| violet | so im not sure why i still cant see the external display | 06:36 |
| josch | "no difference in flash-kernel or initramfs-tools configs from the v3 image" -- what do you mean by that? | 06:36 |
| violet | i ran `diff` over /etc/initramfs, /etc/initramfs-tools, /etc/flash-kernel, /etc/default/flash-kernel | 06:37 |
| violet | the only difference is on my machine the flash-kernel machine is switched to MNT Reform 2 HDMI vs the v3 image rootfs has MNT Reform 2 the internal display only one | 06:38 |
| josch | that is expected | 06:39 |
| violet | right | 06:39 |
| josch | you are probably on linux 6.3 right? what could've happened is that i accidentally broke hdmi when rebasing the cadence patches | 06:41 |
| violet | yeah im on 6.3 | 06:41 |
| josch | last time i used hdmi with my reform was around the time of kernel 5.18 | 06:41 |
| violet | i can try and hack on it when i get my gentoo install up and running | 06:46 |
| violet | im not very comfortable on how to properly build and use custom kernels on debian | 06:46 |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 06:47 | |
| - klardotsh (QUIT: Ping timeout: 246 seconds) (~klardotsh@98.97.36.213) | 08:26 | |
| josch | minute: this is how a split pipeline would look like: https://source.mnt.re/reform/reform-debian-packages/-/pipelines/1043 | 09:11 |
| josch | minute: would that help? | 09:14 |
| + mjw (~mjw@gnu.wildebeest.org) | 10:50 | |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 12:02 | |
| * mark_ -> mjw | 12:03 | |
| minute | violet: how many cards in /dev/dri? | 12:36 |
| minute | btw josch i am running the new kernel 6.3.0-2-reform2-arm64, Debian 6.3.11-1+reform20230706T205126Z1 (2023-07-06) on my imx8mq machine and so far everything is fine | 12:43 |
| minute | nvme, wifi, display, sound, ethernet, mmc, usb... | 12:44 |
| minute | i will test hdmi now | 12:44 |
| minute | hmmm, `reform-display-config --emmc dual` did not have the desired effect, /proc/device-tree/model is still just "MNT Reform 2" | 12:59 |
| minute | upgrading reform-tools just in case | 13:00 |
| minute | yeah, reform-display-config is broken for me | 13:12 |
| minute | it does not switch the dtb | 13:13 |
| josch | minute: but it did change /etc/flash-kernel/machine? | 13:14 |
| minute | so /etc/flash-kernel/machine has the right entry | 13:14 |
| minute | yep | 13:14 |
| josch | does reform-check show anything funny? | 13:14 |
| minute | so flash-kernel doesn't do the job... | 13:14 |
| minute | yeah, it says "your currently loaded dtb is not the one referenced by flash-kernel" | 13:14 |
| minute | flash-kernel also replaces the boot.scr, right? | 13:15 |
| minute | funky "the reform repository is not known to apt" | 13:15 |
| minute | that's not true though, it is in /etc/apt/sources.list.d/mnt-research-debian.list | 13:16 |
| minute | josch: how does flash-kernel actually manifest the chosen dtb on /boot? | 13:18 |
| minute | there are 2 symlinks that look correct | 13:18 |
| minute | dtb -> dtbs/6.3.0-2-reform2-arm64/freescale/imx8mq-mnt-reform2-hdmi.dtb | 13:18 |
| minute | dtb-6.3.0-2-reform2-arm64 -> dtbs/6.3.0-2-reform2-arm64/freescale/imx8mq-mnt-reform2-hdmi.dtb | 13:18 |
| minute | and there is an older symlink dtb-6.1.0-reform2-arm64 -> dtbs/6.1.0-reform2-arm64/freescale/imx8mq-mnt-reform2.dtb | 13:19 |
| minute | aha. | 13:19 |
| minute | once again the problem is that $fdtfile is set in my uboot | 13:19 |
| minute | which overrides the choice made by the symlinks i guess | 13:19 |
| minute | ok, /proc/device-tree/model is correct now | 13:21 |
| minute | but i think now the hdmi module is not loaded | 13:21 |
| minute | this might be the same issue that violet is encountering | 13:21 |
| minute | something is very, very wrong | 13:23 |
| minute | cdns_mhdp_imx.ko is missing | 13:24 |
| minute | > modprobe: FATAL: Module cdns_mhdp_imx not found in directory /lib/modules/6.3.0-2-reform2-arm64 | 13:25 |
| minute | drivers/gpu/drm/imx/cdns/ is missing the source files huh | 13:29 |
| minute | it only has Makefile and Kconfig | 13:30 |
| minute | josch: something got mangled here https://source.mnt.re/reform/reform-debian-packages/-/commit/9bf2da76cc48936eee4602d2be981b02022d3f31#b51533daeec03dd6cabd81f6a9508db1a851775f_45_46 | 13:32 |
| minute | josch: it looks like you moved the cdns-imx stuff to their own files in a new subdirectory "imx/cdns", but you didn't move the source files there. what was the motivation behind this? | 13:36 |
| - mtm (QUIT: Ping timeout: 264 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:04 | |
| minute | note for later, should we ever want to search for a311d reform audio issues https://libera.irclog.whitequark.org/linux-amlogic/2023-05-26 | 14:14 |
| Boostisbetter | minute: is there anymore testing that is needed on the Pocket Reforms in their current state? | 14:30 |
| josch | minute: funny and just recently i pointed out how unlikely it is that nothing broke yet with my rebases of the cadence stuff :D | 14:36 |
| josch | minute: my reasoning was, that in 6.3 the other sources in the directory got moved into their own subdirectories | 14:36 |
| josch | minute: so the patches to the Makefiles didn't apply anymore as now the original makefile got split into multiple sub-makefiles each in its own subdirectory | 14:37 |
| josch | minute: i must've forgotten to move all the files -- i wonder how things still compiled nevertheless | 14:37 |
| minute | Boostisbetter: not sure what you mean! | 14:50 |
| minute | josch: yes, that's really strange | 14:51 |
| minute | josch: do you want to attempt a fix? | 14:51 |
| minute | josch: my guess is it compiled because the subdirectory was ignored | 14:52 |
| minute | josch: because you would also need to include this subdirectory in the makefile of the parent dir | 14:52 |
| josch | minute: i'll not get to doing that until some time during the weekend. If you manage to do it before then please go ahead. :) | 14:54 |
| minute | josch: thank you! idk if i will manage but lets see | 14:55 |
| josch | the limiting factor is the 6 hour build time of the kernel on my system :) | 14:57 |
| josch | indeed cdns_mhdp_imx.ko is missing | 14:59 |
| josch | i'll add this as a check to reform-check... | 14:59 |
| josch | the other open question is what is up with your $fdtfile variable... | 15:02 |
| josch | minute: so in reform-boundary-uboot we set fdtfile to freescale/imx8mq-mnt-reform2.dtb | 15:03 |
| josch | this is obviously wrong for the hdmi dtb | 15:03 |
| josch | but setting this is needed to boot for example debian-installer which expects some reasonable default | 15:04 |
| josch | so what we are shipping with reform-tools is a flash-kernel preboot script which checks: | 15:04 |
| josch | if test -e ${devtype} ${devnum}:${partition} ${prefix}dtb-${fk_kvers}; then setenv fdtpath dtb-${fk_kvers}; fi | 15:04 |
| josch | minute: do you have that in your /boot/boot.scr? | 15:04 |
| minute | josch: yes, but i think it is ignored if $fdtfile is baked into u-boot | 15:06 |
| josch | fdtfile is baked into u-boot yes | 15:07 |
| minute | josch: why does debian-installer care about a variable set in u-boot? | 15:07 |
| josch | at the top of boot.src $ftdfile (if it is set) is used to set $fdtpath which will be used at the bottom | 15:07 |
| minute | josch: ok, so that means that reform-display-config cannot change the dtb | 15:08 |
| josch | minute: because it uses extlinux.conf | 15:08 |
| josch | what reform-display-config does is change /etc/flash-kernel/machine and then re-runs flash-kernel which in turn changes the symlink | 15:08 |
| josch | and the preboot script then resets $fdtpath to dtb-${fk_kvers} | 15:08 |
| josch | but only if the symlink exists -- which is the case for us but not on the d-i media | 15:09 |
| minute | josch: i think the last part does not work for me | 15:11 |
| minute | $fdtfile is always used, except if it is empty | 15:11 |
| minute | if it is empty, the correct file is loaded via the symlink | 15:11 |
| josch | but why would it be empty? it is hardcoded | 15:11 |
| minute | josch: i have to set it manually to be empty | 15:12 |
| minute | josch: i.e. i stop the autoboot and do "setenv fdtfile", then it does the right thing | 15:12 |
| josch | yes, that makes sense | 15:12 |
| josch | the question is why it doesn't work for you with ftdfile set to something but it did work for me | 15:12 |
| josch | maybe something changed | 15:12 |
| josch | i didn't try out hdmi for several months | 15:12 |
| minute | hmm :3 | 15:12 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 16:10 | |
| Boostisbetter | minute: Don't take this the wrong way. I like IRC, and I enjoy this channel. I am curious though why you went with IRC over something like XMPP or Matrix? | 16:44 |
| Boostisbetter | minute: also the start page picture for mnt.re is fantastic. Well shot and a great showcase of greating looking computers and accessories. | 16:45 |
| minute | Boostisbetter: basically noone uses XMPP, and matrix did not seem very reliable for the basic job it has to do | 16:56 |
| minute | Boostisbetter: thank you!! | 16:56 |
| minute | i am lurking in some matrix channels so i might reevaluate that at some point | 16:57 |
| minute | libera has told matrix they can't bridge anymore, right? because they did a bad job with their bridge afaik | 16:57 |
| minute | i am thinking more in the direction of, maybe it would be possible for us to offer a web based client for this channel that also has searchable access to the history | 16:58 |
| Boostisbetter | > libera has told matrix they can't bridge anymore, right? because they did a bad job with their bridge afaik | 17:07 |
| Boostisbetter | I was bridging from matrix for a while but the sync was out of wack several times. I am bridging from XMPP now and it works VERY well. | 17:07 |
| Boostisbetter | minute: If you think nobody uses XMPP but people use IRC, I think you'll be pleasently surprised. | 17:08 |
| Boostisbetter | I mean I don't see a problem with using IRC as it is the lightest of the lightweight chat platforms. I'm just curioused why you chose the most obscure from my perspective. | 17:09 |
| minute | Boostisbetter: ok, maybe it is just very subjective, but i don't personally have contacts who use xmpp | 17:12 |
| sigrid | irc? obscure? heh | 17:12 |
| minute | Boostisbetter: IRC is very commonly used for development chatrooms for FOSS software and kernel stuff | 17:13 |
| minute | so that's my main usecase | 17:13 |
| minute | i'm currently writing my keynote talk for FSIC2023 that i have to give monday morning at sorbonne in paris | 17:14 |
| minute | i wish i had already written it :D | 17:14 |
| + mark_ (~mjw@gnu.wildebeest.org) | 17:15 | |
| josch | Boostisbetter: everything I do for Debian either uses IRC or email. I don't know of people in the project who use matrix or xmpp to communicate project-specific matters. | 17:16 |
| Boostisbetter | hmmm, nheko has a Matrix channel | 17:17 |
| josch | I use xmpp to talk to friends. I'm running my own server (ejabberd) and haven't yet found a matrix server that is lightweight nor a matrix client that is lightweight. Clients seem to be very web-focused instead of text-focused. | 17:17 |
| Boostisbetter | I know of that one. | 17:17 |
| Boostisbetter | josch: Nheko reborn is very lightweight given the huge amount of functionality it exposes. | 17:17 |
| Boostisbetter | josch, | 17:17 |
| josch | I'm using nheko to chat with clapper developers. It is a lot slower than my IRC or XMPP clients. | 17:18 |
| Boostisbetter | josch: I hear you regarding the bloat. | 17:18 |
| Boostisbetter | josch, do you use Dino for XMPP, it is very good. | 17:18 |
| josch | Boostisbetter: you were wondering why i never had any problems with the slow processor and 4 GB of ram, so... ;) | 17:18 |
| josch | I tried dino and ran into a lot of bugs. | 17:18 |
| josch | their last release is from march... | 17:19 |
| Boostisbetter | josch, one other thing I would say, is that like you I spent a lot of time not liking Matrix. I thought it was unnecessary and that they should have just improved XMPP. But now that i run my own instance, I can see that the memory and CPU requirement are VERY near to my old Rocket Chat server. | 17:19 |
| Boostisbetter | That is to say that it has been greatly improved, and there are a ton of really nice improvements coming. | 17:19 |
| josch | i might re-evaluate things in the future | 17:20 |
| josch | but right now everybody i chat with either uses irc or xmpp so i'm fine :) | 17:20 |
| Boostisbetter | josch, I am using Dino 4.2 and it works great on Linux. On Linux and Windows, I only use Dino. Everything else has been buggy and missing features. | 17:20 |
| Boostisbetter | josch, so what do you use for XMPP? Gajim? | 17:20 |
| josch | profanity | 17:23 |
| josch | my partner uses Conversations on android | 17:23 |
| josch | and a friend uses gajim | 17:23 |
| Boostisbetter | Sorry if I have brought this up before. I really like hearing about peoples computing preferences. And then finding out why. | 17:23 |
| josch | (and is unhappy with the new interface) | 17:23 |
| Boostisbetter | Conversations on Android is the best XMPP client currently available. | 17:23 |
| sigrid | there is no matrix client on plan 9 | 17:24 |
| Boostisbetter | web client (element) a no go? | 17:24 |
| sigrid | there are, however, clients for irc and xmpp | 17:24 |
| sigrid | no | 17:24 |
| Boostisbetter | are you sure there isn't a terminal client that can't be used? | 17:24 |
| sigrid | no | 17:24 |
| Boostisbetter | There are really a lot of matrix clients. | 17:24 |
| Boostisbetter | but i hear you, that is a bad on the matrix community. | 17:25 |
| Boostisbetter | minute: forgive me asking about this, but the MNT community is one of the brightest I know, and I am really curious about this. | 17:25 |
| Boostisbetter | I am wondering what the MNT crowd thinks about the Librety Phone from Purism (basically using the 4gb / 128gb eMMC variant that the Reform has been using, well sans the 128gb eMMC) | 17:26 |
| Boostisbetter | Personally I see it as compelling. I like that it is manufactured, for the most part, in the US (as I'm American, and I really feel that manufacturing needs to be brought back from China) despite the really high price. I admit that this price is out of reach for many people that would otherwise want it. | 17:27 |
| Boostisbetter | The other thing is that it is at its best when using Convergence, but then Phosh kind of lets you down, because it just is no where as good as Sway. | 17:27 |
| Boostisbetter | But if the phone is best as a laptop, then a Reform just makes way more sense, and that has been my experience as well. | 17:28 |
| Boostisbetter | I also REALLY love that the Reform is using plain vanilla Debian. | 17:28 |
| Boostisbetter | I really just can't wait for the Pocket. I have so many plans for that thing already. | 17:29 |
| Boostisbetter | And the Reform has a seriously timeless design. It is robustly build and meant to last. Where as the kill switches on the Librem 5 just feel like they will break very easily. And then there is the whole issue with Purism's transparency and shady business practices. | 17:31 |
| Boostisbetter | minute: and finally one serious question: When the Reform is plugged in and charging is finsihed. The LPC reports that the incoming or outgoing voltage is 0V. Does this mean that the wall adapter is powering everything and not the batteries? | 17:33 |
| q66 | the purism phone is absolutely worthless | 17:33 |
| Boostisbetter | q66: would you care to explain? I have one and i wouldn't go that far. | 17:33 |
| q66 | costs too much for what you get and is probably made by people who think bitcoin is the future | 17:34 |
| q66 | if i were to spend a bunch of money on a relatively underpowered but open device i'd much rather throw it at mnt because at least that's cool and not sketchy | 17:35 |
| Boostisbetter | q66: I can agree with that for the most part. I think crypto is a gold rush and a terrible waste of power and silicon. | 17:35 |
| Boostisbetter | q66: I will say that I kind of seperate purism developers from purism itself. From that perspective, they have done A HUGE amount of work for the linux phone as whole. That part should be acknowledged at least. | 17:36 |
| q66 | i won't separate people from the product | 17:36 |
| Boostisbetter | but like you, their shady business practices fly directly in the face of what they say they represent. | 17:37 |
| Boostisbetter | MNT has NEVER done anything like that, and has always been up front and genuine. I notice, and VERY much respect and appreciate that. | 17:37 |
| Boostisbetter | In fact if Lukas would create a Linux phone himself, I would be one of the first to buy it. | 17:38 |
| q66 | in any case even if purism was perfectly clean a $2200+VAT raspberry pi 3 is just too much no matter how you look at it | 17:39 |
| q66 | and "made in the USA" is hardly a selling point for anything | 17:39 |
| Boostisbetter | I think the cost is inflated some, but I think the cost of manufacturing it in the US instead of China is significantly more expensive. If that doesn't appeal to someone now in the US, then it doesn't make sense, and that is a problem. | 17:39 |
| Boostisbetter | $1200 will get you the Chinese made one. | 17:40 |
| q66 | which is already too much for what you get | 17:40 |
| Boostisbetter | Cost is relative, but software is the MOST expensive part of any product. | 17:40 |
| Boostisbetter | by far the most difficult as well. | 17:40 |
| Boostisbetter | so I can understand the prices to a degree, but that doesn't mean that average customer will, and that is another problem. | 17:41 |
| Boostisbetter | Still let's not belabor this. I agree with Purism not being trustworthy. The usefulness of their product is something else. I can respect your opinion on it. | 17:41 |
| Boostisbetter | I love my Reform and can't wait for the Pocket. | 17:41 |
| - mark_ (QUIT: Ping timeout: 260 seconds) (~mjw@gnu.wildebeest.org) | 17:48 | |
| Boostisbetter | josch and minute: is it really by you all as well today? | 17:57 |
| Boostisbetter | Kind of nice that the Reform is passively cooled, and doesn't really vent heat into the room like an active cooling solution would. | 17:57 |
| josch | Boostisbetter: i admit i fail to parse what you said in your second to last message :D | 18:17 |
| minute | Boostisbetter: fwiw, the imx8mq and imx8mplus modules we use are also manufactured in the USA (by boundary devices) | 18:18 |
| Boostisbetter | josch: yeah typo. I was asking if it was hot by you all today as well. | 18:18 |
| Boostisbetter | minute: well that is rad. I also really like made in Germany things as well. | 18:18 |
| minute | Boostisbetter: the laptops are assembled by us here in berlin. the metal is cut in china. the pcbs are etched and soldered in china. | 18:18 |
| minute | some of the pcbs and rework we do here of course | 18:19 |
| Boostisbetter | yep, you all have a model manufacturing and assembly model going on. I really like it. | 18:19 |
| minute | we also have assembled all of the motherboards etc by hand several times | 18:19 |
| minute | but we don't have pick and place machines, so for the runs, we outsource that | 18:19 |
| minute | we can't afford a pick and place machine and the setup time/work it requires. otherwise that would be very interesting. most of our PCBs can be also made in other places, like at aisler. we also had some PCBs made in bulgaria during the early covid outbreak in china | 18:20 |
| minute | but we don't only go to pcbway because it is more affordable (they are also by far not the cheapest pcb house in china), but also because they are very experienced and flexible | 18:22 |
| Boostisbetter | minute: I think you have it all pretty figured out nicely. | 18:31 |
| sigrid | noam: coming back to temperature question - I look at it right now and it's 49°C cpu / 50°C nvme | 18:35 |
| Boostisbetter | minute: are you attending All Systems Go 2023 in Berlin this year? | 18:35 |
| Boostisbetter | or josch | 18:40 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 18:41 | |
| josch | Boostisbetter: travelling is really hard for me because of my baby at home so for the next few years, I'm keeping any time I spend outside my home town to a minimum :) | 18:44 |
| Boostisbetter | josch, totally understand. Right on! | 18:44 |
| Boostisbetter | Is there a way to know on the Reform when you should reboot after updating, if you did it via the console? | 18:55 |
| josch | Boostisbetter: when you update the kernel you might want to (but don't have to) reboot | 18:56 |
| Boostisbetter | yeah, I know. I was just curious if there was a way to know if it would be a good thing. | 18:56 |
| josch | there are some mechanisms for desktop environments where you get a popup | 18:57 |
| Boostisbetter | yeah, not really on Sway though. I love Sway | 18:57 |
| noam | sigrid: yeah, it's normally not that hot for me either... | 19:01 |
| noam | I have a few hypotheses, but I'm not sure what the real answer is | 19:01 |
| noam | 1. Maybe it doesn't like if I leave it plugged in when it's fully charged, but... that doesn't sound right, to me | 19:02 |
| noam | 2. Maybe the desk I'm putting it on cannot dissipate heat quickly enough, so the case hits a thermal limit, and then the heat sink and then the CPU stop having anywhere to put the heat | 19:02 |
| josch | i have my reform 22/7 plugged in on top of a couch cushion | 19:03 |
| josch | 65°C | 19:03 |
| sigrid | mine's always plugged here | 19:06 |
| noam | Could be a different revision, though ;) | 19:11 |
| noam | Mine's through crowdsupply, and I think theirs are all the older stock still? | 19:12 |
| noam | The mobo has microUSB, not USB-C, for instance | 19:12 |
| noam | ...it's probably not that, though. I think it might be a combination of the desk inhibiting the heat dissipation, and the outdated trackball firmware spamming the CPU with wakeups and preventing it from ever hitting deepest idle states, and thus generating more heat? | 19:13 |
| noam | I'll have to science it | 19:13 |
| - mtm (QUIT: Ping timeout: 252 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 19:19 | |
| sigrid | mine has no usb-c | 19:21 |
| sigrid | i got it in the end of 2021 iirc | 19:21 |
| noam | Gotcha, thanks for the data :) | 19:23 |
| noam | And I didn't even have to inject malicious JS to your computer to gather it! /s | 19:23 |
| Boostisbetter | > The mobo has microUSB, not USB-C, for instance | 19:39 |
| Boostisbetter | I don't believe that USB-C is coming to the Reform platform, rather to the Pocket. | 19:39 |
| Boostisbetter | btw, does anyone have a good GUI based IRC client for the Reform? | 19:40 |
| noam | Huh, I thought the newer boards had USB-C already? Guess I was wrong /shrug | 19:44 |
| Boostisbetter | the other great thing about the Pocket is that it charges using USB C PD | 19:45 |
| noam | So, you mean, I wouldn't be able to plug it into the sun? :P | 19:49 |
| noam | which, oh! Does 100% absolutely work :D | 19:49 |
| + amospalla (~amospalla@212.231.228.113) | 19:50 | |
| Boostisbetter | solar panel through a usb c would work. | 19:51 |
| noam | I don't have USB-C, I have + and - wires | 19:51 |
| amospalla | Hi guys, is the official Pocket Reform image still supposed to run under Debian unstable? | 19:52 |
| noam | and, unlike DC, I can't just make a USB-C cable | 19:52 |
| Boostisbetter | well then on the normal Reform you can easily wire that to a barrel | 19:52 |
| amospalla | I'm planning on maybe moving my main O.S. on my current desktop, depending on what the reform will be running upon. | 19:52 |
| noam | Boostisbetter: that's what I'm saying | 19:52 |
| noam | One sec, pics are easier | 19:52 |
| noam | Boostisbetter: https://pixelhero.dev/tmp/cable.jpg | 19:53 |
| noam | I made that cable last night using a DC jack, some stranded 14AWG wire, a soldering iron, and electrical tape :P | 19:53 |
| noam | Then I just use a double-sided barrel jack to plug it in to the Reform :) | 19:54 |
| noam | USB-C is just a tiny bit harder to rig up like this :/ | 19:58 |
| noam | I do need to figure it out, though, for everything else - e.g. work laptop | 19:58 |
| minute | Boostisbetter: nope. the next thing for me after fsic2023 is the cccamp | 20:00 |
| josch | amospalla: the official reform image is using Debian unstable but i'm currently setting up a machine creating Debian stable (bookworm) images as well | 20:01 |
| noam | minute: the reform charged successfully both directly off of the solar panels (through some truly horrifying DIY-ed adaptor cables :P) and through the charge controller+battery. The Reform has been the single easiest thing to charge off of solar so far :D | 20:02 |
| noam | Once again, really really liking everything about the design, nice work :D | 20:02 |
| vagrantc | Boostisbetter: there is needrestart and needrestart-session for determining which processes you might want to restart (including rebooting) | 20:02 |
| vagrantc | Boostisbetter: it hooks into apt to nudge you at the end of an apt run | 20:02 |
| josch | noam: i don't have a solar panel (yet) but i have batteries that give me 12V or 24V DC -- works extremely well there as well. No need for the complication of USB-C but just one + and one - cable. | 20:04 |
| noam | josch: yeah, the concern with solar was the unstable voltage | 20:04 |
| noam | With the stabilizing battery, that's probably not even a concern | 20:04 |
| josch | it's so crazy cool that it works despite the unstable voltage :D | 20:04 |
| noam | yeah | 20:04 |
| noam | Boostisbetter: actually, come to think of it, I have a USB-C -> DC cable already, which was meant for going from a USB-C device to charge my thinkpad. I wonder if that will work in reverse? [e.g. run power through the DC end, use that to give 12V to USB-C?] probably not, since USB-C has that weird chip for setting voltage :/ | 20:05 |
| josch | I have an old 19 V thinkpad adapter (the rectangular plug) which works perfectly with the reform | 20:06 |
| vagrantc | i recently tried a usb-c to barrel plug adapter with the mnt/reform ... and just caused the red charging lights on the mnt/reform to blink on and off again | 20:07 |
| josch | would the reform not have the 5521 barrel jack i couldn't have repurposed old AC-adapters for new purposes | 20:07 |
| Boostisbetter | yeah, I got a usb c to barrel for the Reform, it works well also. | 20:07 |
| - amospalla (QUIT: Quit: Client closed) (~amospalla@212.231.228.113) | 20:09 | |
| vagrantc | the one i had was specced to output 20v ... maybe i just got unlucky :/ | 20:09 |
| noam | vagrantc: source port needs to be willing to output 20V | 20:10 |
| noam | and some power supplies are less than happy to do so, even when they're rated to! | 20:10 |
| noam | e.g. the portable pinepower can give 20V, and a USB-C <-> USB-C cable will give 20V to my work laptop, a Framework, just fine | 20:10 |
| vagrantc | the source port definitely supports 20v out ... | 20:11 |
| noam | but a 20V USB-C -> barrel jack cable plugged into it will st u tt t tuut utter | 20:11 |
| noam | the thinkpad will constantly beep connected/disconnected/connected/disconnected every ~6 seconds | 20:11 |
| noam | The same cable, in the desktop pine power, works fine | 20:11 |
| noam | and IIRC it works fine with the framework adapter | 20:11 |
| noam | vagrantc: it's not just about supporting 20V, it needs to also support the initial power surge IIUC | 20:12 |
| vagrantc | used on both a pinepower and a coolgear usb-c output rated for 2.4amps at 20v ... works fine with my framework | 20:12 |
| noam | Yeah, portable pinepower cannot handle 20V barrel jack | 20:12 |
| noam | I can say that empirically. | 20:12 |
| vagrantc | oh wait, now ... the coolgear is rated for way more than that ... | 20:12 |
| noam | I dunno if it's because the laptop tries pulling more than 65W initially, and the pinepower responds by hard resetting itself, thus entering an infinite loop | 20:12 |
| noam | Dual USB-C works fine, because IIUC the pinepower can *tell the laptop* how much current it's willing to give | 20:13 |
| noam | I dunno how all that negotiation garbage works, but it usually *does* | 20:13 |
| noam | Usually | 20:13 |
| noam | I've had issues with it on the framework before :/ | 20:13 |
| noam | but with a barrel jack, all bets are off | 20:13 |
| vagrantc | rated for 4.1amp at 20v ... still no dice | 20:14 |
| noam | Rather, with USB-C to barrel jack adaptors, all bets are off | 20:14 |
| josch | when usb-c works its nice | 20:15 |
| noam | I don't know if it's an issue with the PinePower, the cable, or the interaction between one or both of them and the thinkpad, or what | 20:15 |
| noam | josch: exactly | 20:15 |
| vagrantc | would be nice to find one that worked with the mnt/reform ... but probably a lot of hit or miss trials before finding one | 20:15 |
| noam | USB-C is absolutely *lovely* when it works | 20:15 |
| josch | but my main issue is that when it doesn't i have no way to diagnose anything | 20:15 |
| noam | It's just that its failure modes are horrendous | 20:15 |
| sknebel | people on the forum have listed usb-c->barrel adapters that work with the reform | 20:15 |
| josch | yup, i'm using one. it's nice when i travel with my partner and all we take to charge our things is a single usb-c wall plug | 20:16 |
| noam | I might just buy a bunch of USB-C jacks with the pins exposed so I can make my own USB-C -> barrel jacks that I can confirm 100% work with the reform | 20:16 |
| vagrantc | sknebel: ah, good idea | 20:16 |
| noam | and with everything else, in both directions | 20:16 |
| noam | This is the path that leads to me making my own PSU XD | 20:17 |
| vagrantc | i honestly forget to even look at forums ... kind of outside my usual workflow | 20:17 |
| noam | Haha, my portable 250Wh battery *refuses to charge* off of the solar battery lol | 20:22 |
| noam | It's pulling maybe 5W. Maybe. | 20:22 |
| noam | If I unplug it and plug in the reform, I can pull 40W easily; the cables are rated for at least 15A (at 12V!), and the battery _can_ charge off of the solar panels directly. I'm going to have to make a _second_ barrel jack, plug that one in parallel with the controller into the panels, and only use _that_ to charge the portable battery lol | 20:23 |
| + mark_ (~mjw@gnu.wildebeest.org) | 20:44 | |
| vagrantc | whoah. i never realized that the barrel plug on the mnt/reform seems to work with 5.5x2.1mm barrel plugs ... | 20:48 |
| vagrantc | the power supply that came with it looked more like 5.5x2.5mm | 20:48 |
| noam | ...wait, is it? | 20:50 |
| vagrantc | huh. or not. i've just been using a 5.5x2.5 adapter all this time for no reason ... worked, though | 20:51 |
| noam | phew | 20:52 |
| noam | had me worried! | 20:52 |
| minute | noam: that's great to hear @ solar | 20:59 |
| sknebel | is the charging current software-controlled? could it in theory offer a setting (e.g. in the OLED menu) to limit power usage if one wanted to charge from a weaker power source? | 21:02 |
| minute | nope, unfortunately not | 21:04 |
| minute | well | 21:04 |
| minute | one could do some sort of PWM | 21:04 |
| minute | so in the lpc firmware you could rapidly switch the charging on and off | 21:04 |
| minute | i wouldn't do this unless i had a spare motherboard to experiment with though :D | 21:04 |
| sknebel | yeaah, that sounds dicey | 21:10 |
| sigrid | I think that's usually the job of the solar/battery controller | 21:11 |
| sigrid | but I guess if the power source is very stupid, it would make sense | 21:12 |
| minute | so yeah, your power source should have its own limiter | 21:15 |
| noam | Hmmm, I'm kinda tempted to look at modding the LPC firmware to allow more control of charging from the OS | 21:15 |
| noam | So that I can e.g. write a script for 9 that kills charging at ~60% | 21:15 |
| minute | that should be quite doable | 21:18 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 21:25 | |
| sknebel | vagrantc: https://community.mnt.re/t/will-we-be-able-to-usb-power-mnt-reform/241/ was the thread I was thinking of | 21:28 |
| noam | Terrible idea for reform3: have two copies of the system controller, both equally capable of applying power to the voltage lines | 21:29 |
| noam | So that one of them can be flashed from the SoC while the other keeps it powered :P | 21:29 |
| noam | [I did say it was terrible] | 21:29 |
| minute | lol | 21:29 |
| minute | i think it would be possible with some trickery to make the LPC flash itself | 21:30 |
| vagrantc | sknebel: yeah, i've been looking over that | 21:30 |
| + klardotsh (~klardotsh@98.97.114.50) | 21:35 | |
| josch | vagrantc, noam: it's a 5.5 x 2.1 mm barrel jack: https://source.mnt.re/reform/reform-handbook/-/merge_requests/7/diffs | 21:40 |
| josch | 5.5 2.5 plugs fit as well but more loosely | 21:40 |
| vagrantc | yeah, don't know how i got the idea in my head | 21:41 |
| vagrantc | some 5.5x2.5 plugs are springy to intentionally handle this sort of thing | 21:41 |
| vagrantc | although the plug i have been using is not springy but ... works anyways, to my surprise | 21:42 |
| vagrantc | though happy to discover that 5.5x2.1 work ... as that is what i mostly have | 21:42 |
| noam | oh, we could probably do some sorta A/B flashing scheme, couldn't we? Have it split flash in half, copy an image from the host to the secondary, and then swap which one is secondary and reboot itself? Not sure we can reboot it without it taking down the main board, though :P | 21:43 |
| vagrantc | high availability DIY laptop | 21:43 |
| vagrantc | hot swappable power supplies with hot swappable batteries and hot swappable cpu and ram ... | 21:44 |
| sigrid | noam: also https://mntre.com/reform-irc-logs/2022-09-10.log.html#t09:53:39 | 21:45 |
| noam | nice, thanks | 21:46 |
| violet | minute: awake now if there's any debugging you'd like me to do on my system | 22:04 |
| violet | per your question about /dev/dri i have card0, card1, render128 | 22:06 |
| violet | and confirming i do not have cdns_mhdp_imx in my modules dir | 22:09 |
| violet | there is cdns_mhdp_drmcore.ko and cdns-mhdp8546.ko | 22:09 |
| josch | violet: thank you for confirming! | 22:09 |
| josch | violet: i'll try to find some time to fix this issue over the weekend | 22:10 |
| violet | cool! | 22:10 |
| Boostisbetter | so I have asked this question for, but I already forgot. Because I use suspend a lot I have a lot of cases where the system crashes. I have to turn it off and then turn it back on, booting from a cold start. | 22:31 |
| Boostisbetter | So far I have noticed any serious issues with this, but I am curious if this is hurting my nvme drive or messing up my system. | 22:31 |
| noam | heh | 22:37 |
| noam | Boostisbetter: my nvme reports Unsafe Shutdowns: 55 | 22:37 |
| noam | it's fiiiiine | 22:38 |
| Boostisbetter | noam: how do you check for that? | 22:38 |
| Boostisbetter | also good to here. | 22:38 |
| Boostisbetter | also good to hear | 22:38 |
| noam | It's PROBABLY fine | 22:39 |
| noam | Boostisbetter: I just cat /dev/sdN0/smart | 22:39 |
| noam | dunno how to get it on non-9 tbh | 22:39 |
| sigrid | smartctl | 22:39 |
| Boostisbetter | noam: yeah forgot you were on plan 9 | 22:40 |
| sigrid | chances are, when the laptop goes into suspend nvme caches are flushed, so I don't expect you to lose any data if wakeup did not work | 22:40 |
| sigrid | and I doubt it will cause any effect on nvme | 22:41 |
| Boostisbetter | sigrid: that is great to hear. If this was not the case, I was thinking about a possible way to prevent damage when the resume doesn't work. | 22:42 |
| Boostisbetter | Glad that it isn't really something I need to be worried about. | 22:42 |
| - klardotsh (QUIT: Ping timeout: 245 seconds) (~klardotsh@98.97.114.50) | 23:05 | |
| + klardotsh (~klardotsh@98.97.36.213) | 23:16 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 23:57 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
