| ex-parrot | sigrid: are you on the newer LPC firmware yet? | 00:17 |
|---|---|---|
| sigrid | yeah I am | 00:17 |
| sigrid | it's all back to normal. I just freaked out :D | 00:17 |
| ex-parrot | yay | 00:19 |
| ex-parrot | I have felt much more comfortable with the new LPC firmware | 00:19 |
| ex-parrot | I keep paranoid checking every few days but my batteries are barely drainign now | 00:19 |
| sigrid | speaking of drain. how much does it drain during normal use? for me it shows 0.172A with lowest brightness and music playing | 00:20 |
| ex-parrot | I don't think I have a good figure off hand. I should check | 00:20 |
| ex-parrot | O | 00:20 |
| ex-parrot | I've mainly just been booting in to LFS and running vim | 00:20 |
| ex-parrot | so I'm probably at a fairly minimal draw | 00:20 |
| sigrid | init=/bin/vim | 00:30 |
| ex-parrot | I almost could | 00:31 |
| ex-parrot | the lfs init scripts are pretty basic, and I found surprisingly buggy :/ | 00:31 |
| sigrid | I found void linux to be the most easy to install and use so far, with minimal work to get a live image built for reform specificially via the Official Way | 00:33 |
| ex-parrot | I have never tried void | 00:33 |
| sigrid | *easy way | 00:33 |
| ex-parrot | tbh at some point I'll probably go back to Debian as it's what I use everywhere else (or OpenBSD when we get there) | 00:34 |
| ex-parrot | but for now LFS has been quite soothing | 00:34 |
| - mtm_ (QUIT: Ping timeout: 250 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 01:03 | |
| - nppvxexhf (QUIT: Ping timeout: 256 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 01:04 | |
| - Christoph_ (QUIT: Remote host closed the connection) (~Christoph@p54bf63e6.dip0.t-ipconnect.de) | 01:53 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 02:25 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 03:09 | |
| + mtm- (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 03:10 | |
| - GNUmoon (QUIT: Ping timeout: 276 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 05:37 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 06:36 | |
| - verx (QUIT: Ping timeout: 256 seconds) (~verx@matrix.16bit.dev) | 08:28 | |
| + verx (~verx@matrix.16bit.dev) | 08:29 | |
| - doppler (QUIT: Ping timeout: 240 seconds) (~doppler@user/doppler) | 09:28 | |
| + MajorBiscuit (~MajorBisc@c-001-001-054.client.tudelft.eduvpn.nl) | 09:31 | |
| - stjohn (PART: !!unknown attribute: msg!!) (~stjohn@cpe-68-174-13-238.nyc.res.rr.com) | 09:42 | |
| josch | sigrid: in your config-void.patch to u-boot you load boot.scr but your mkimage places it in /boot/boot.scr -- are you using a dedicated partition for /boot or are you working with symlinks? | 09:46 |
| + doppler (~doppler@user/doppler) | 10:22 | |
| + stjohn (~stjohn@cpe-68-174-13-238.nyc.res.rr.com) | 10:23 | |
| sigrid | josch: dedicated partition | 10:27 |
| josch | Yeah, I think there are some good reasons for a dedicated partition. I filed this MR https://source.mnt.re/reform/reform-boundary-uboot/-/merge_requests/2/diffs | 10:34 |
| josch | sigrid: I think everything you need for void linux is in there? | 10:34 |
| josch | I'm gonna add that u-boot searches for boot.scr in / as well as an /boot so that it understand both scenarios | 10:35 |
| sigrid | boot.scr, kernel itself, initramfs and dtb files | 10:37 |
| sigrid | for me one reason is that I can use f2fs for rootfs and fat32 for /boot | 10:38 |
| josch | sigrid: Ah I didn't mean everything you need is in /boot but everything you need is in that MR so if mntmn accepts that, then you don't have to patch it anymore. | 10:38 |
| sigrid | ah ok | 10:38 |
| sigrid | it should probably work for booting emmc -> nvme as well. i need to get a working nvme and update it for some kind of "try nvme first, fall back to sd card" | 10:40 |
| josch | Yes, that would be another good improvement. | 10:41 |
| josch | Then one u-boot binary would be sufficient for both scenarios and thanks to boot.scr we don't have to rebuild u-boot just to change kernel parameters. | 10:42 |
| + sts-q (~sts-q@212.53.219.175) | 11:38 | |
| josch | sigrid: this works for Debian -- does it also work for your setup? https://source.mnt.re/reform/reform-boundary-uboot/-/merge_requests/2/diffs | 11:41 |
| josch | it tries sdcard first, then eMMC and looks into / and /boot for boot.scr on both | 11:41 |
| - GNUmoon (QUIT: Ping timeout: 276 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 11:52 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 11:53 | |
| + mjw (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 12:09 | |
| + Christoph_ (~Christoph@p4fe73ee8.dip0.t-ipconnect.de) | 12:28 | |
| sigrid | josch: should work, yeah | 12:35 |
| - mtm- (QUIT: Ping timeout: 260 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 13:02 | |
| mntmn | josch: have now assigned digital ocean runner to your uboot repo https://source.mnt.re/josch/reform-boundary-uboot/-/jobs/621 | 13:03 |
| - mtm (QUIT: Ping timeout: 256 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 13:03 | |
| josch | Thanks! I'm going to push some fixes until the job succeeds. | 13:05 |
| mntmn | cool, thx | 13:07 |
| josch | mntmn: could you also enable the pipeline for reform/reform-boundary-uboot? Then I will adapt reform-system-image to get u-boot as the latest job artifact. | 13:19 |
| mntmn | josch: i believe i did that, too! | 13:21 |
| josch | mntmn: okay, then I think you can merge !2 now :) | 13:21 |
| mntmn | https://source.mnt.re/reform/reform-boundary-uboot/-/jobs/623 | 13:22 |
| mntmn | done! | 13:22 |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 13:35 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 13:35 | |
| josch | mntmn: thanks! My MR https://source.mnt.re/reform/reform-system-image/-/merge_requests/37 is now using that artifact and it produces a bootable image. I also replaced genext2fs with mke2fs so that the pipeline finishes faster again. | 14:10 |
| josch | Next I'll make sure that switching to the hdmi dtb and booting with root on nvme works. | 14:11 |
| sigrid | any reason to not have hdmi on by default? | 14:12 |
| josch | sigrid: I added the reason as a comment to my flash-kernel patch: https://source.mnt.re/reform/reform-debian-packages/-/blob/main/patches/flash-kernel#L23 | 14:13 |
| sigrid | hmmm.. guess I've never noticed the difference | 14:14 |
| sigrid | you've probably already noticed I made boot.scr choose a dtb based on /boot/disable_hdmi file existing (or not) | 14:16 |
| josch | Yes, I've considered that option as well. But that would mean that I would have to add support for multiple dtbs for the same board to flash-kernel. It's to tell flash-kernel that there are two different reform boards: the normal and the hdmi one and switching between them is then done by changing the machine identifier in /etc/flash-kernel/machine and re-running flash-kernel which will then put the | 14:18 |
| josch | right dtb accordingly | 14:18 |
| josch | *It's easier to... | 14:18 |
| josch | mntmn: before you merge !37, could you add https://mntre.com/reform-debian-repo ? I would like to make sure that the changes I pushed there in the past days didn't break the image. | 14:34 |
| josch | s/add/refresh/ | 14:34 |
| - erlehmann (QUIT: Ping timeout: 250 seconds) (~erle@ip5f5bd57a.dynamic.kabel-deutschland.de) | 14:47 | |
| mntmn | josch: ok, one sec | 15:04 |
| mntmn | josch: done (i hope) | 15:05 |
| josch | thanks! | 15:06 |
| josch | maybe it starts making sense to put a daily cronjob so that i don't have to bug you all the time :) | 15:06 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 15:08 | |
| + mtm- (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 15:09 | |
| mntmn | josch: yep, will do | 15:26 |
| + erlehmann (~erle@ip5f5bd57a.dynamic.kabel-deutschland.de) | 16:31 | |
| Boostisbetter | has anyone been able to widevine of the Reform? (Netflix, Amazon, Disney+, etc.) | 17:20 |
| Boostisbetter | has anyone been able to get widevine on the Reform? (Netflix, Amazon, Disney+, etc.) | 17:21 |
| - sts-q (QUIT: Ping timeout: 256 seconds) (~sts-q@212.53.219.175) | 17:38 | |
| + sts-q (~sts-q@212.53.219.175) | 17:40 | |
| sigrid | Boostisbetter: no. I don't think it's available for aarch64 linux at all | 17:40 |
| Boostisbetter | Bummer. I guess I can use it via a good vnc connection. Hahahaha | 18:16 |
| sknebel | it is. for raspberry pi people did tricks to get the chromeOS versions to run | 18:23 |
| sknebel | (and now the pi has official support afaik) | 18:23 |
| sknebel | so its not necessarily impossible | 18:23 |
| sigrid | I think in that case you kinda have to redefine the word "available" | 18:24 |
| kfx | the pi just ships th armhf build. | 18:25 |
| sknebel | chromeos is an aarch64 linux, so I'd say that counts as "available" | 18:27 |
| kfx | they explicitly call out in their 64-bit release announcement that widevine is not available and instruct users to 'apt install chromium-browser:armhf libwidevinecdm0' if you want drm | 18:27 |
| sknebel | that official support is going to be scarce is kind of given :D (actually was surprised the rpi thing happened) | 18:29 |
| kfx | 64-bit chromeos also ships an armhf widevine library. | 18:30 |
| kfx | google has never released an aarch64 build of widevine. | 18:30 |
| mntmn | i just use a PS4 as my DRM viewing device... | 18:34 |
| technomancy | yeah I keep that shit off devices I actually trust. quarantine the DRM. | 18:38 |
| kfx | yeah, I have a crappy little dell plugged into a projector, and the family netflixes on that | 18:40 |
| kfx | seemed like a better plan than handing smartphones to my kids | 18:41 |
| sigrid | looking forward to not being able to watch netflix & co, tbh | 18:41 |
| technomancy | not being able to run electron programs is <3 | 18:42 |
| kfx | we've been slowly moving to stuff provided by the local public library, but that's also webshit with widevine, sadly | 18:42 |
| kfx | at least the quality of the content is, on average, higher | 18:42 |
| technomancy | Kanopy? | 18:42 |
| kfx | technomancy: that and hoopla | 18:43 |
| Boostisbetter | mntmn: would this work in the reform? https://www.youtube.com/watch?v=bCgkzLayauA | 18:55 |
| mntmn | Boostisbetter: do you have a product/datasheet link that is not a video? kinda busy atm | 18:59 |
| Boostisbetter | mntmn: https://www.dfi.com/product/index/1455 | 19:01 |
| Boostisbetter | Perhaps | 19:01 |
| vkoskiv | mntmn: 22 watts max power draw | 19:01 |
| vkoskiv | They are running it with a passive heat sink, fwiw | 19:02 |
| + stag (~stag@li1611-13.members.linode.com) | 19:03 | |
| vkoskiv | Though it runs at 81C at heavy load. dual core, 2.6GHz base clock, 3.5 boost. 4GB of DDR4 | 19:04 |
| vkoskiv | And some radeon GPU. | 19:04 |
| stag | sorry to drop in and ask such a tired question, but if i were to order a reform in the USA, what would be the estimated shipping date? late april, or is that for the earlier crowdfunding stage? | 19:04 |
| vkoskiv | Any x86 chip in the reform is going to run hot, that's just how those chips are. | 19:04 |
| vkoskiv | The memory is fast, but only single channel. | 19:06 |
| mntmn | stag: late april sounds realistic, we are currently waiting for the first post-crowdfunding batch to arrive at mouser to see how fast it will be forwarded | 19:07 |
| - mjw (QUIT: Quit: Leaving) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 19:08 | |
| stag | sounds good | 19:11 |
| - MajorBiscuit (QUIT: Ping timeout: 256 seconds) (~MajorBisc@c-001-001-054.client.tudelft.eduvpn.nl) | 19:24 | |
| Boostisbetter | vkoskiv: mntmn: I agree. It looks like it needs a pretty beefy passive heat sink, BUT I'm sure throttling it a bit would prevent it from getting so hot. It can use up to 8gb of RAM as well. For the power envelope it is not bad. It will require some work to be down to get it to mate up with the mainbord via the card slot. | 19:52 |
| Boostisbetter | mntmn: was just watching some fosdem presentations and watched Purism's Librem 5 progress report, and they mentioned there that their suspend should be out this year. but they mention that the cpu core cannot be woken from the deepest sleep state, and they are unsure if a solution for that exists. They have a software work around but it isn't a real solution. | 20:31 |
| - sts-q (QUIT: Read error: Connection reset by peer) (~sts-q@212.53.219.175) | 20:50 | |
| josch | mntmn: the new bootloader built by the ci as well as the current state of sysimage-v3 is able to also work from nvme. I will adapt the reform-migrate accordingly. Though I think for this to work properly we need to have a dedicated /boot partition on the SD card. | 21:01 |
| josch | The problem is that currently, the bootloader will load the kernel and dtb from /boot on the sdcard and then mount nvme as /. This means that any kernel upgrade will not affect the content on the SD-Card. But we can also not mount the SD-Card as /boot because the kernel and dtb are in the /boot subdirectory on the SD-card. A solution will be to the sysimage-v3 script create a image with two partitions. | 21:03 |
| josch | The first for /boot and the second for /. What do you think? Do you see another solution? | 21:03 |
| mntmn | josch: ah, i see, it would then end up in /boot/boot | 21:04 |
| josch | Yes. | 21:04 |
| mntmn | josch: ok then, lets try the extra partition. i guess reform-init, reform-display-config, reform-boot-config (perhaps not needed anymore?) and reform-migrate will need to be touched. | 21:08 |
| + sts-q (~sts-q@212.53.219.175) | 21:10 | |
| josch | Okay. At least I'm happy nvme boot works. :) Tomorrow I'm trying the dtb switching for hdmi output. I'm meeting a friend who has a hdmi monitor. | 21:11 |
| josch | Yes, I'll also adapt the scripts so that they work with the sysimage-v3. But I'll probably only get to do that on the weekend. | 21:12 |
| mntmn | josch: totally fine, thank you for all your hard work! | 21:13 |
| josch | Thanks for making such awesome hardware! :) | 21:14 |
| mntmn | btw ruff apparently got SPI communication working between i.MX and LPC, which is wonderful (on community.mnt.re) | 21:14 |
| mntmn | this paves the way for low-overhead battery display in the OS i think | 21:14 |
| - sts-q (QUIT: Remote host closed the connection) (~sts-q@212.53.219.175) | 21:17 | |
| sigrid | maybe a better shut down mechanism as well? | 21:25 |
| + sts-q (~sts-q@212.53.219.175) | 21:42 | |
| mntmn | sigrid: theoretically, but the current one is not so bad, because the keyboard needs to know about the shutdown too | 21:54 |
| Boostisbetter | btw, is the correct way to use LPC commands by push circle and THEN pushing the other key? I have been holding down the circle key and then pushing the other key. I am wondering if that might be contributing to the crashing of it in anyway. | 21:59 |
| mntmn | Boostisbetter: yep, one after the other, not a combo | 22:00 |
| Boostisbetter | thanks! | 22:01 |
| Boostisbetter | do you think doing it the way I was could lead to it being spammed and crashing? | 22:09 |
| mntmn | i don't know! | 22:17 |
| - GNUmoon (QUIT: Ping timeout: 276 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 22:23 | |
| Boostisbetter | Yeah, I'll start using the LPC with the correct key combos, and report back if I discover anything. | 22:59 |
| Boostisbetter | The only other question I had was when the Reform crashes on resume from standby. When I do a circle + 0 to shut it down, and then a circle + 1 to turn it on, is that messing up my installation? Are there files that are getting corrupted because of that? | 23:00 |
| Boostisbetter | I don't see any evidence of that, and everything fires back up as if nothing happened. | 23:01 |
| sigrid | before entering into standby, allegdly, all caches are flushed and filesystems are synced | 23:02 |
| sigrid | you're going to loose the state of the programs that were running, though | 23:03 |
| sigrid | that might include files that weren't saved yet, for example | 23:03 |
| Boostisbetter | yeah I'm pretty good about saving everything before suspending. I've been well conditioned. | 23:04 |
| Boostisbetter | ok was just making sure I'm not corrupting system files or anything like that. | 23:04 |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 23:27 | |
| - Christoph_ (QUIT: Remote host closed the connection) (~Christoph@p4fe73ee8.dip0.t-ipconnect.de) | 23:33 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!