| vkoskiv | I still can't get over how cool this is: https://github.com/jart/cosmopolitan | 00:55 |
|---|---|---|
| vkoskiv | x86_64 only, sadly, but it's such a wildly impressive project. | 00:56 |
| - mjw (QUIT: Quit: Leaving) (~mark@gnu.wildebeest.org) | 01:06 | |
| - Christoph_ (QUIT: Remote host closed the connection) (~Christoph@p54bf6030.dip0.t-ipconnect.de) | 01:10 | |
| - mtm (QUIT: Ping timeout: 246 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:03 | |
| nocko | josch: I am following your instructions for encrypted nvme boot from emmc. SoC switch is confirmed off... but screen blank unless the SD card is inserted. I am guessing this means bad/no uboot on emmc? I ran reform-flash-rescue... which should have done the needful. | 02:48 |
| nocko | It boots normally w/ SD card inserted and kernel+initramfs come from emmc. | 02:49 |
| nocko | Well... I don't know about that last one. /boot is mounted from EMMC. I don't have serial to check output from uboot, yet. | 02:50 |
| nocko | Pulling a fresh flash.bin from the gitlab artifacts and `dd ... of=/dev/mmcblk0 bs=1k skip=33` didn | 03:02 |
| nocko | 't help either. | 03:02 |
| nocko | It really seems to be acting like SoC switch in ON position... but it's not. | 03:02 |
| nocko | Welp, I'll break out the serial cable tomorrow and try again. Let me know if this is a thing you can explain. | 03:03 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 03:56 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 04:09 | |
| josch | nocko: this sounds similar to the effect bkeys is seeing, no? | 07:16 |
| josch | if you used reform-flash-rescue, then kernel+initrd on emmc should at least turn your display on (even if your rootfs should be configured wrongly) | 07:16 |
| josch | nocko: yes, serial output is the next thing we need because u-boot cannot turn on the screen by itself | 07:17 |
| josch | but sl is working on making u-boot turn on the internal display :) | 07:19 |
| + littlebo1eep (~alMalsamo@gateway/tor-sasl/almalsamo) | 07:47 | |
| - littlebo1eep (QUIT: Ping timeout: 240 seconds) (~alMalsamo@gateway/tor-sasl/almalsamo) | 07:54 | |
| + littlebo1eep (~alMalsamo@gateway/tor-sasl/almalsamo) | 08:45 | |
| - littlebo1eep (QUIT: Ping timeout: 240 seconds) (~alMalsamo@gateway/tor-sasl/almalsamo) | 08:53 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:03 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:04 | |
| - klardotsh (QUIT: Remote host closed the connection) (~klardotsh@98.97.119.207) | 09:44 | |
| + MajorBiscuit (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 09:58 | |
| + mjw (~mark@gnu.wildebeest.org) | 10:21 | |
| - excited-mango[m] (QUIT: Quit: You have been kicked for being idle) (~excited-m@2001:470:69fc:105::1:faee) | 11:00 | |
| - MajorBiscuit (QUIT: Quit: WeeChat 3.5) (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 11:09 | |
| nocko | It does seem similar to bkeys in some ways. But their switch is on, mine is definitely off. | 12:36 |
| nocko | Will work on serial console after I get the kids off to school. | 12:36 |
| + mark (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 13:03 | |
| + Christoph_ (~Christoph@p54bf6332.dip0.t-ipconnect.de) | 13:03 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark!~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440))) (~mark@gnu.wildebeest.org) | 13:06 | |
| * mark -> mjw | 13:06 | |
| + mark__ (~mark@gnu.wildebeest.org) | 13:06 | |
| + MajorBiscuit (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 13:13 | |
| - mtm (QUIT: Ping timeout: 260 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 14:02 | |
| - vkoskiv (QUIT: Ping timeout: 276 seconds) (~vkoskiv@89-166-62-97.bb.dnainternet.fi) | 14:03 | |
| - MajorBiscuit (QUIT: Ping timeout: 248 seconds) (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 14:04 | |
| + vkoskiv (~vkoskiv@89-166-62-97.bb.dnainternet.fi) | 14:04 | |
| + MajorBiscuit (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 14:07 | |
| bkeys | Yeah hopefully there is a resolution to my issue | 15:01 |
| - erle (QUIT: Quit: Just say no, then Putin can not legally invade your nation without your consent.) (~erle@ip5f5af7b3.dynamic.kabel-deutschland.de) | 15:52 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 16:09 | |
| vkoskiv | Has anyone done any kind of rubber foot/gasket mod on the display edge? The anodizing is starting to wear off on the front edge of the palmrest in some spots. | 16:20 |
| vkoskiv | I'm thinking either little dots of thin rubber sheet or longer strips along the top edge. Just enough to prevent aluminium on aluminium contact | 16:21 |
| vkoskiv | It looks to me like there has been some particulate between there that has then rubbed a spot both on the palmrest and the edge of the lid, but I keep this pretty clean. I don't keep sand in my apartment :D | 16:21 |
| - thum (PART: Leaving) (~thum@user/thum) | 16:23 | |
| josch | no sand in your apartment? lucky you! :D | 16:31 |
| vkoskiv | Well I mean, street dust could be likened to sand. And a lot of the filling in tarmac here is granite. | 16:33 |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 16:38 | |
| nocko | josch: uboot is loading from emmc: http://okturing.com/src/13806/body | 16:47 |
| - Christoph_ (QUIT: Remote host closed the connection) (~Christoph@p54bf6332.dip0.t-ipconnect.de) | 16:48 | |
| josch | nocko: this looks okay -- what happens if you let it boot? | 16:50 |
| nocko | The first time I let it autoboot, it did not load the kernel. I turned it off and on again and it loaded the kernel from emmc, but display is not initialized: http://okturing.com/src/13807/body | 16:51 |
| nocko | It's probably waiting for my luks passphrase. | 16:51 |
| nocko | ... to continue boot... not to initialize the display. | 16:51 |
| josch | i wonder why your output stops at 3.826444 | 16:53 |
| josch | nocko: which image did you use? Can it be that you picked sysimage-v3 before this commit: https://source.mnt.re/reform/reform-system-image/-/commit/d4e2f61a04da271562006da4a0f3474463d59160 | 16:54 |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 16:55 | |
| nocko | Currently emmc is running whatever reform-flash-rescue would have pulled at 2200 UTC yesterday. | 16:56 |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 16:56 | |
| josch | aha, let me check... | 16:56 |
| josch | nocko: can you test a theory for me? | 16:58 |
| nocko | Sure | 16:58 |
| nocko | I built reform-system-image from main yesterday as well... so it shouldn't be hard to test a newer version... | 16:59 |
| josch | load the latest sysimage-v3 onto an sd-card, boot from it and then modify /usr/sbin/reform-flash-rescue with this patch: | 16:59 |
| josch | http://paste.debian.net/1244396/ | 16:59 |
| josch | Currently, reform-flash-rescue downloads the latest successful build artifact from the sysimage-v3 branch but that branch was merged into main and so reform-tools has to be changed to download from main instead to get the latest image. | 17:00 |
| nocko | Working on it now. | 17:00 |
| josch | nocko: independent on whether this fixes things for you (it might due to the commit I referenced above) this has to be fixed... i'm working on a merge request | 17:00 |
| josch | https://source.mnt.re/reform/reform-tools/-/merge_requests/16 | 17:02 |
| josch | unfortunately, the ci system is broken right now so even if that would get merged soon, we'll not see an updated reform-tools package on the mirror | 17:03 |
| josch | nocko: basically this should do it: sed -i /sysimage-v3/main/ /usr/sbin/reform-flash-rescue | 17:04 |
| nocko | Flashing emmc now. | 17:04 |
| nocko | It looks to me like d4e2f61a04da made it into sysimage-v3 branch well before the merge. I'm not certain it'll fix my issue. Good to get it sorted for the future, though. | 17:08 |
| nocko | Still dd'ing | 17:09 |
| - MajorBiscuit (QUIT: Quit: WeeChat 3.5) (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 17:10 | |
| josch | yes I also think that d4e2f61a04da made it into sysimage-v3 but i'm uncertain whether it was processed by the CI into the latest working image or not and i have no other explanation for why your serial output suddenly stops | 17:11 |
| + MajorBiscuit (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 17:13 | |
| nocko | Ugh. Imaging seems busted. I didn't umount /boot, so there were some ext4 errors... will umount and try again. | 17:17 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 17:28 | |
| vkoskiv | Is there a way to get more verbose logging out of the wifi card? I'd like to see more about signal strength and stuff. | 17:30 |
| vkoskiv | I found random wifi antennas in a box while cleaning up and stuck them in my reform :D | 17:30 |
| vkoskiv | So far, they seem worse than the molex one. | 17:30 |
| nocko | josch: It is working now. | 17:38 |
| josch | nocko: nice!!! :D | 17:38 |
| nocko | Running reform-flash-rescue when you have emmc mounted as boot is bad... maybe there should be a check... but it's probably rare. | 17:39 |
| josch | minute: thank you for merging the reform-tools MR 16 so quickly -- now we just need a working CI so that we can rebuild the packages and the sysimage | 17:39 |
| josch | minute: it seems this fix is really necessary to prevent the problem that nocko (and maybe bkeys) was seeing | 17:40 |
| nocko | After the corruption, to avoid removing heat sink for switch, I manually executed flash-rescue steps using artifacts I built from main yersterday. | 17:40 |
| minute | what is broken with the CI again? not enough space? | 17:40 |
| minute | (sorry, i was a bit in pcb design tunnel) | 17:40 |
| josch | minute: I guess it's the space, here is the log: https://source.mnt.re/reform/reform-system-image/-/jobs/808 | 17:41 |
| nocko | josch: I also had complete serial output on the console, so it seems like the v3 branch artifacts are older than mid-April. | 17:41 |
| josch | nocko: there should absolutely be a check -- I'll take care of that | 17:41 |
| - bkeys (QUIT: Ping timeout: 256 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 17:44 | |
| josch | this also reminds me that the reform-migrate and reform-setup-encrypted-nvme scripts need an --emmc argument or otherwise they will not work or require additional action if the user wants to boot from emmc | 17:50 |
| nocko | I think I'd prefer a standard where we agree that uboot lives on emmc. Treat emmc more like BIOS w/ built-in "recovery". Then boot options could be stored in a single place (in uboot env partition) and passed to whatever booted via boot args. | 17:58 |
| nocko | Everything everywhere is a bit too many variables to track. Also, If I never have to remove the heatsink again, it'd be too soon. | 17:59 |
| nocko | Developers could continue to do their own thing, but it would leave a single pretty straight-forward path for less technical users. | 18:00 |
| nocko | Pretty much every script in the system could be simplified by agreeing on a "happy path". | 18:00 |
| josch | nocko: i would agree with you if we had display output and keyboard input in u-boot | 18:09 |
| josch | then yes, it would be like a "bios" and on bootup you could for example select where you want to boot from (emmc, sd-card, usb...) | 18:09 |
| josch | but unless we get the internal display working in u-boot, some setups might just produce a black screen and that's it and that is when we need u-boot on the sd-card | 18:10 |
| josch | thus, i'd argue that until we have display output we are stuck in a situation where you need to be a technical user and be able to access your system via serial if something goes wrong | 18:10 |
| nocko | Yeah. uboot display will be great. | 18:11 |
| nocko | I get that the current system it | 18:12 |
| nocko | is less brickable... but for less technical users the various blank screens which are equally likely on sdcard are effectively bricked. | 18:12 |
| josch | yes | 18:13 |
| nocko | I'm not very familiar with our user base. A fair number of technical users... but I wonder how many have just put their reform on the shelf... | 18:13 |
| vkoskiv | Woo! I found my UART adapter box thing. | 18:13 |
| nocko | In any case, things are coming along and uboot display will go a long way (and seems likely soon). | 18:14 |
| minute | josch: i'm trying something new: /dev/zram0 47G 24K 45G 1% /ramdisk | 18:15 |
| josch | nocko: bug sl and/or cinap_lenrek about it :) | 18:17 |
| josch | minute: oh interesting, lets see how it works :) | 18:17 |
| minute | josch: i made /make-zram.sh in case the server reboots, idk yet how to make this permanent | 18:18 |
| minute | i will once again run the ci pipeline... | 18:21 |
| josch | minute: thanks -- i don't know a standardized mechanism for doing what you do in /make-zram.sh either -- the module and options could be put into /etc/modprobe.d but to run the mkfs i guess one would need to write a new systemd unit or something... | 18:25 |
| sl | josch: i appreciate the vote of confidence, but i have neither time nor skill to get far with this atm :) | 18:28 |
| minute | josch: yeah, i thought about systemd too | 18:30 |
| minute | bluerise: would you be able to integrate cinap_lenrek's LCDIF+DSI code into our u-boot? | 18:30 |
| Boostisbetter | question: is zram really necessary with the Reform? I mean would it net us that much of a performance gain? | 18:34 |
| minute | Boostisbetter: this is about the serevr. | 18:37 |
| minute | server | 18:37 |
| minute | nothing to do with reform per se | 18:37 |
| minute | the image is built in a ramdisk | 18:37 |
| Boostisbetter | Thanks please ignore my comment. Hahaha | 18:43 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 18:48 | |
| - mjw (QUIT: Quit: Leaving) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:10 | |
| - bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 19:19 | |
| - MajorBiscuit (QUIT: Ping timeout: 246 seconds) (~MajorBisc@wlan-145-94-167-213.wlan.tudelft.nl) | 19:28 | |
| minute | josch: lol, the build almost worked... but was killed during artifact upload by 1h timeout | 19:33 |
| minute | increased timeout to 2h | 19:42 |
| - sl (QUIT: Quit: leaving) (~sl@104-59-85-219.lightspeed.iplsin.sbcglobal.net) | 19:55 | |
| * mark__ -> mjw | 20:16 | |
| + klardotsh (~klardotsh@98.97.119.207) | 20:39 | |
| cinap_lenrek | minute: uboot already has the infra in place i see | 20:43 |
| cinap_lenrek | minute: in drivers/vicdeo/mxsfs.c, thats the LCDIF driver | 20:43 |
| cinap_lenrek | tho no idea how to plug it together | 20:46 |
| - klardotsh (QUIT: Ping timeout: 246 seconds) (~klardotsh@98.97.119.207) | 20:52 | |
| josch | minute: I guess this is because we now install kde which will take a few extra minutes. | 20:54 |
| - mtm (QUIT: Ping timeout: 240 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 20:58 | |
| minute | > Pipeline has been fixed and #623 has passed! | 21:02 |
| josch | awesome! :) | 21:09 |
| josch | I now also re-triggered the reform-debian-packages pipeline. Once that is done and the mirror updated, we can run reform-system-image again to let it include the reform-tools package with the most recent fixes. | 21:11 |
| + klardotsh (~klardotsh@98.97.116.100) | 21:18 | |
| - dodo (QUIT: Remote host closed the connection) (~dodo@user/dodo) | 22:04 | |
| + dodo (~dodo@user/dodo) | 22:05 | |
| minute | josch: cool | 22:21 |
| josch | will have to wait a bit -- cross compilation in Debian unstable is currently broken because of a recent src:linux upload :) | 22:22 |
| minute | lol | 22:25 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 22:38 | |
| _E | minute: is it normal for a package to go through 11 import scans and 7 warehouse scans in one warehouse? | 22:41 |
| minute | _E: sounds like UPS | 22:43 |
| _E | by the way, I forgot to put F1 back in place before I shipped it | 22:43 |
| _E | hopefully that's not one of the impossible-to-find parts | 22:44 |
| minute | should be fine | 22:46 |
| _E | wonder how much UPS would charge to ship that fuse 😂️ | 22:47 |
| klardotsh | minute: don't want to pester too badly but did want to loop back another time on the "no eDP, HDMI works" issue. unless you think anything more trivial is a culprit, my hunch looking at the BOM and rough schematics from the handbook is likely a fried SN65DSI86IPAPQ1. if you agree with that assessment, would you recommend sending it to you (I'll pay shippingx2 and parts/fair labor) or should I find someone local here in USA-PNW who can do SMD work? | 22:48 |
| minute | klardotsh: if you are willing to pay for shipping + repair, we can do it | 22:48 |
| minute | probably the best option atm | 22:49 |
| klardotsh | minute: I think with all signs pointing to "I probably screwed it up in the first place" that's only fair, so deal :) | 22:49 |
| klardotsh | way better than waiting for Mobo rev N+1 availability (parts shortages and all) | 22:50 |
| minute | ok cool... please UPS or fedex it to the address at the bottom of mntre.com. if you need us to create a label, contact support@mntre.com. actually loop in support@mntre.com in any case, so you have a ticket. | 22:50 |
| klardotsh | minute: y'all are the best. thank you! | 22:51 |
| klardotsh | minute: to make sure I get everything in one sweep, do you need anything beyond the mobo? | 22:53 |
| - bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 22:56 | |
| vkoskiv | klardotsh: Do you have an idea of how the oopsie happened? I'm trying to not end up frying anything as best as I can :D | 22:58 |
| vkoskiv | If I'm doing something small inside, I disconnect the packs. If I'm doing more significant tweaks, I also take out the cells. | 22:58 |
| klardotsh | vkoskiv: honestly no. I don't remember enough details; I was doing a should-have-been-trivial endeavor of installing a Laird wifi antenna onto my Intel AX200 card (I forgot to order an antenna so Day 1 of assembly, my Reform was ethernet only) | 22:59 |
| klardotsh | vkoskiv: my hunch is that likely something got bumped metal-on-metal with bats attached not being careful enough, but again, no evidence to confirm that other than "my display stopped working, and the display driver (DSI chip) is right under the wifi antenna acryllic" | 23:00 |
| klardotsh | vkoskiv: so I think I'll just echo kfx 's sentiment of "as soon as you unscrew the backing plate, disconnect the bats before touching anything else in any way" (their case was a zapped RTC battery while also installing wifi antennae) | 23:02 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 23:05 | |
| * yankcrim- -> yankcrime | 23:06 | |
| sigrid | my case was dropping a screwdriver in and shortening, burning a track :D | 23:09 |
| sigrid | luckily soldering was enough in that case | 23:10 |
| - iank_ (QUIT: Quit: ZNC 1.8.2+deb2+b1 - https://znc.in) (~iank@fsf/staff/iank) | 23:12 | |
| + iank (~iank@fsf/staff/iank) | 23:12 | |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 23:15 | |
| - S0rin (QUIT: Quit: WeeChat 3.0) (~S0rin@user/s0rin) | 23:23 | |
| klardotsh | O_O yikes! | 23:23 |
| - bkeys (QUIT: Ping timeout: 246 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 23:29 | |
| josch | i fried one of my cells because when I noticed that one of the screws fixing one of the battery boards to the frame was loose, of course I was too lazy to first remove all the cell before attempting to tighten the screw with a metal screwdriver... | 23:30 |
| josch | vkoskiv: you recently expressed interest in blender -- we have another build failure of blender 2.79 (grep for "error:") https://source.mnt.re/reform/reform-debian-packages/-/jobs/814/raw | 23:54 |
| josch | vkoskiv: will you take care of this? | 23:57 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!