| vkoskiv | Cool, yeah. That works. I'll take care of the rest. | 00:00 |
|---|---|---|
| sigrid | minute: do you think lpc11u24 could talk to the imx over spi clocked at 6mhz? | 00:26 |
| sigrid | I should probably just try it :) | 00:26 |
| sigrid | or, rather, I can see the _max_ clock is set to 400kHz at the moment | 00:27 |
| sigrid | but it can be set to 6MHz, which should be enough to maybe try increasing the master clock to something higher than 400kHz which is very slow | 00:28 |
| minute | sigrid: try! not sure how the lpc is clocked internally | 00:28 |
| sigrid | I also figured out how to trigger an interrupt instead of polling spi data register in lpc fw | 00:29 |
| minute | nice! | 00:29 |
| minute | oh https://www.cnx-software.com/2022/09/08/radxa-cm5-rockchip-rk3588s-module-raspberry-pi-cm4/ | 00:32 |
| minute | that's awesome... means i don't have to make a completely new mobo for this chip | 00:34 |
| josch | The Boundary devices Nitrogen8M SOM User manual claims 4Kp60 h265 encode/decode -- how do they come up with these numbers?? | 00:43 |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 01:00 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 01:00 | |
| minute | josch: must be a mistake, there is no encoder afaik | 01:02 |
| minute | there is a h265 decoder though | 01:04 |
| josch | but no support for in from hantro, right? | 01:04 |
| minute | it's hantro g2, there's an article about it from last year https://www.collabora.com/news-and-blog/blog/2021/02/17/adding-hevc-support-for-imx-8m/ | 01:07 |
| minute | looks like there is a staging driver | 01:08 |
| minute | hantro_g2_hevc_dec.c | 01:08 |
| minute | vp9 decode support was also merged to gstreamer https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1012 | 01:13 |
| minute | ndufresne: is there hevc decode for hantro g2 in gstreamer nowadays? | 01:14 |
| - S0rin (QUIT: Ping timeout: 268 seconds) (~S0rin@user/s0rin) | 01:28 | |
| + S0rin (~S0rin@user/s0rin) | 01:36 | |
| - chomwitt (QUIT: Ping timeout: 268 seconds) (~chomwitt@2a02:587:dc16:fe00:6307:f534:81e2:947) | 01:46 | |
| - S0rin (QUIT: Ping timeout: 268 seconds) (~S0rin@user/s0rin) | 01:48 | |
| + S0rin (~S0rin@user/s0rin) | 01:48 | |
| vkoskiv | Has hibernation (suspend to disk) been investigated yet for the Reform? | 01:59 |
| - mtm (QUIT: Ping timeout: 252 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:03 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:40) | 03:04 | |
| - buckket (QUIT: Quit: buckket) (~buckket@pdp8.buckket.org) | 03:12 | |
| + buckket (~buckket@pdp8.buckket.org) | 03:16 | |
| - monkeybusiness (QUIT: Quit: Leaving) (~monkbusy@user/monkeybusiness) | 03:43 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 04:07 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 04:09 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 07:04 | |
| josch | vkoskiv: I tried it and it failed to come back 1-2 months ago. But I'm the wrong person to ask because in contrast to others in this chan my suspend is super reliable too. | 08:26 |
| josch | vkoskiv: if you want to try out hibernation you have to edit your /usr/lib/systemd/sleep.conf.d/reform-sleep.conf | 08:26 |
| josch | In that file change AllowHibernation=no to AllowHibernation=yes | 08:27 |
| + chomwitt (~chomwitt@2a02:587:dc16:fe00:b62b:4d00:b365:1fd8) | 08:32 | |
| Boostisbetter | I don't believe hibernation works. Suspend is pretty reliable though. | 09:15 |
| sigrid | since lpc11uxx supports reprogramming over uart, i wonder if the fw could jump back into bootloader (without a reset) on a spi command so the host os could update the fw - no extra computer involved | 09:53 |
| sigrid | or maybe it's possible to reflash with fw itself? | 09:59 |
| sigrid | apparently fw should be able to trigger isp over uart on its own if pio0_3 pulled low, which as i understand is always pulled high on reform for usb msc isp | 11:19 |
| Boostisbetter | would be pretty cool if possible, seems like it could be. | 11:21 |
| minute | sigrid: i'm pretty sure it should be possible to write to flash from the fw, or jump to the isp in software | 11:24 |
| - qwer (QUIT: Ping timeout: 244 seconds) (~qwer@89-24-41-218.nat.epc.tmcz.cz) | 11:32 | |
| + qwer (~qwer@89-24-41-218.nat.epc.tmcz.cz) | 11:41 | |
| + reform23389 (~sevan@2001:470:1f1d:1d6:59ef:bf82:c112:b94b) | 12:09 | |
| sigrid | minute: if i understand schematics correctly, LPC_UVBUS is pulled high, so isp reinvoke would always enable usb msc, not uart isp :/ | 12:15 |
| sigrid | and lpc probably executes the code directly from flash, so iap overwriting itself is probably not perfect | 12:17 |
| sigrid | don't think it can fit two copies of the fw | 12:17 |
| minute | sigrid: couldn't the fw writing function be copied to ram and executed from there? | 12:23 |
| sigrid | i guess so | 12:32 |
| sigrid | then it could just as well use spi | 12:34 |
| - jjbliss (QUIT: Ping timeout: 248 seconds) (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 12:54 | |
| + jjbliss (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 12:55 | |
| sigrid | might make sense to have a secondary bootloader maybe | 12:55 |
| - reform23389 (QUIT: Quit: Leaving) (~sevan@2001:470:1f1d:1d6:59ef:bf82:c112:b94b) | 13:04 | |
| + sevan (~venture37@host-92-23-203-220.as13285.net) | 13:49 | |
| - jjbliss (QUIT: Remote host closed the connection) (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 13:51 | |
| + jjbliss (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 13:56 | |
| sevan | Hi, I signed up for a gitlab account (username sevan) on source.mnt.re a while back but the account was never approved. I was wondering if someone here could help with that. :) | 13:58 |
| - mtm (QUIT: Ping timeout: 244 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 14:03 | |
| minute | sevan: i can! | 14:04 |
| sevan | excellent :) | 14:05 |
| + monkeybusiness (~monkbusy@103.87.94.35) | 14:14 | |
| - monkeybusiness (QUIT: Changing host) (~monkbusy@103.87.94.35) | 14:14 | |
| + monkeybusiness (~monkbusy@user/monkeybusiness) | 14:14 | |
| josch | grubman: remember the reform-setup-encrypted-nvme bug you reported August 20? I can confirm its existance! :D | 15:43 |
| minute | what's the bug? | 15:51 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 16:09 | |
| - chomwitt (QUIT: Ping timeout: 260 seconds) (~chomwitt@2a02:587:dc16:fe00:b62b:4d00:b365:1fd8) | 16:26 | |
| josch | if at the end, you choose to use the sd-card for booting, the script will quit (wrongly with exit 0) and tell you that /boot is in use (which is correct) so it will not do anything (which is wrong) | 16:26 |
| josch | grubman: that's what you saw, right? | 16:27 |
| - monkeybusiness (QUIT: Remote host closed the connection) (~monkbusy@user/monkeybusiness) | 16:30 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 17:23 | |
| Boostisbetter | who would have though but Vivaldi works on the Librem 5 (specifically that it is adaptive)? | 17:26 |
| Boostisbetter | Yeah nevermind, works great in the Reform, crashes the Librem 5 | 17:42 |
| grubman | josch: yeah! | 19:10 |
| grubman | i think what i did as a workaround was running the migrate script again and it booted from nvme afterwards | 19:11 |
| + chomwitt (~chomwitt@2a02:587:dc16:fe00:5010:70b4:e094:50b5) | 19:14 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by wielaard!~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440))) (~mark@gnu.wildebeest.org) | 19:37 | |
| * wielaard -> mjw | 19:37 | |
| + mark__ (~mark@gnu.wildebeest.org) | 19:38 | |
| smj | I'm trying to identify a failing battery cell. In the oled read out, how do the cells correspond with the batteries? | 19:44 |
| smj | what is the orientation | 19:44 |
| ndufresne | minute: it's merged now, https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1079 | 19:45 |
| Boostisbetter | smj: it should be detailed in the user manual | 19:45 |
| Boostisbetter | although I think there was a possible error in it, maybe? | 19:45 |
| ndufresne | minute: It was merged as soon as the hevc uAPI was acceptable as stable | 19:45 |
| Boostisbetter | probably already corrected at this point. | 19:46 |
| minute | ndufresne: oh great! do you know which version of gstreamer that went into? | 19:47 |
| ndufresne | It will be in 1.22, to be released by end of October | 19:48 |
| minute | ah, awesome | 19:48 |
| minute | we have 1.20 at the moment | 19:48 |
| smj | I think I've got it figured out by swapping things around, I'll make notes for myself. | 19:48 |
| ndufresne | I think theb uAPI stability will be in Linux 6.0 | 19:49 |
| minute | josch: cc, it looks like with gst 1.22 and kernel 6.0 we'll get hevc decode | 19:49 |
| ndufresne | We might need to tweak the version check, though it's just a warning, we check the c structure size | 19:49 |
| minute | ndufresne: do you know if imx8mplus hantro is compatible? | 19:49 |
| ndufresne | Anything with G2 chip, but sometimes I get lost with the imx8 series | 19:50 |
| minute | haha ok thanks | 19:50 |
| ndufresne | I'll say "I think so", I think it's a G1/2 combo but with a vc8000 encoder | 19:51 |
| ndufresne | But if its a vc8000 decoder, the good news is that we have a port somewhere that works, but it needs cleanup | 19:52 |
| ndufresne | Btw, I've also added GL tile support, but didn't got a chance to test that (mostly performance whose) on the etnaviv driver, I'm super curious of what we can get, it saves a lot of bandwidth | 19:54 |
| minute | nice | 19:54 |
| minute | sevan: approved | 19:54 |
| sevan | minute: thank you. I'm in :) | 19:59 |
| sevan | After I signed in, it asked my role and then showed me https://source.mnt.re/-/manifest.json | 20:00 |
| minute | sevan: lol, weird | 20:01 |
| sevan | ikr :) | 20:01 |
| - ehmry (PART: https://quassel-irc.org - Chat comfortably. Anywhere.) (~quassel@WLWB-VID145-122-227.lcom.net) | 20:02 | |
| smj | I have the older atmel trackball and newer rpi trackball and I've noticed that he rpi is much slower. Is this possible to change via firmware? | 20:07 |
| josch | grubman: thanks a lot for your bug report and sorry for having taken so long before looking into it. Now it's on my todo list to get fixed. :) | 20:13 |
| josch | ndufresne, minute: thanks a lot for the updates! If the release happens by end of october then there are good chances for hevc decoding support making it into the next Debian stable release. :) | 20:15 |
| minute | smj: i have removed the fixed acceleration in the firmware of the rp one, i recommend to just boost the speed in software. which desktop are you using, sway? | 20:17 |
| minute | smj: you can set accel_profile and pointer_accel in sway config https://github.com/swaywm/sway/wiki#libinput-config-options | 20:18 |
| ndufresne | josch: that would be amazing, getting this working on stock Linux is really the goal | 20:18 |
| smj | OK we'll have to implement that in plan9 then. | 20:21 |
| minute | smj: ah, excuse the faux pas | 20:23 |
| smj | no problem :) | 20:28 |
| smj | all our other units have the atmel device. | 20:29 |
| smj | so this is the first one with the rpi. We can adapt to this. | 20:29 |
| minute | cool | 20:30 |
| smj | right now from origin to the max diagonal is 16 swipes of the ball vs. 5 (for me) :)) so it creeps, but you're right it should be programmable. | 20:32 |
| smj | rather than firmware hardcoded, so its the right path. | 20:32 |
| minute | smj: hmm, that sounds excessive | 20:32 |
| smj | I'm adapting until we can program it, the new switches are very clicky too. | 20:33 |
| minute | i can reach the diagonally opposite corner in 3-5 swipes | 20:33 |
| smj | which is very nice. | 20:33 |
| minute | oh yeah, these are choc whites | 20:33 |
| josch | Which poor soul has been crimping all the JST-PH cables at MNT Research? I just have to crimp three new four-wire cables and I'm already going insane. XD | 20:42 |
| minute | josch: well, we quickly decided to source them from a factory after doing a lot of crimping in house ^^ | 20:44 |
| josch | I can now fully understand that decision. XD | 20:44 |
| smj | minute, do you have stats on cpu temps and what would be considered reasonable? for instance, I have one unit that gets up to 68C while another runs at 54C. | 20:52 |
| minute | smj: that's still ok (the 68), but it can help to reapply thermal paste under the heatsink and make sure the heatsink screws are fastened | 20:54 |
| vkoskiv | You have multiple Reforms? :D | 20:54 |
| smj | I just tried that, so I'm monitoring now. | 20:55 |
| smj | ummm I have 4 :) | 20:55 |
| smj | most are loaners for plan9 dev | 20:55 |
| smj | free loaners :) | 20:56 |
| grubman | josch: no worries :-) | 21:15 |
| * grubman -> grubmn | 21:56 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark__!~mark@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 21:57 | |
| * mark__ -> mjw | 21:57 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 21:57 | |
| sevan | reform hass been running off charger and compiling for the past 11.5 hours without issue. I guess I shouldn't do long sustained loads when running off battery? (there was more than 20% battery left when it switched off during compilation last night) | 22:06 |
| minute | sevan: why not? | 22:08 |
| minute | but yeah, if you don't have to charge-cycle the batteries, they'll probably last longer | 22:08 |
| sevan | I was thinking because of the battery drain when system is at full load for a long period of time. | 22:10 |
| sevan | I'm likely wrong, just trying to rationalise what happend | 22:11 |
| Boostisbetter | sevan: it is possible that your battery meter wasn't properly calibrated when your Reform powered off. But in general 11 hours of compiling and the CPU at full tilt is not possible on battery. You will see about 4 hours of battery life in that scenario. | 22:24 |
| - buckket (QUIT: Quit: buckket) (~buckket@pdp8.buckket.org) | 22:31 | |
| - sevan (QUIT: Ping timeout: 268 seconds) (~venture37@host-92-23-203-220.as13285.net) | 22:31 | |
| + buckket (~buckket@pdp8.buckket.org) | 22:34 | |
| josch | I think what sevan meant to say was that even though they had the charge plugged in, the battery was at 20% after those 11 hours as if power from AC wasn't enough and somehow batteries still got drained? | 22:43 |
| + sevan (~venture37@host-92-23-203-220.as13285.net) | 23:04 | |
| - chomwitt (QUIT: Ping timeout: 268 seconds) (~chomwitt@2a02:587:dc16:fe00:5010:70b4:e094:50b5) | 23:05 | |
| sevan | Boostisbetter: sorry, I mean I was previously compiling when running solely off battery when the laptop switched off, the past 11 hours have been running whilst connected to charger. | 23:06 |
| Boostisbetter | sevan: ok,yeah I also was saying it is possible the your battery percent was off because it was not yet calibrated. | 23:07 |
| sevan | The comment was in relation to this https://community.mnt.re/t/laptop-unresponsive/749/16 | 23:07 |
| sevan | Boostisbetter: makes sense, I hadn't thought about that. | 23:08 |
| Boostisbetter | I know when I got the new 2000mAh batteries for mine. I had to charge them, then let them drain down until laptop shut off. Then I had accurate battery percentages. | 23:09 |
| sevan | I'll give it try. I like that one of my batteries reports 3.3v and one reports 3.5v and things sort of level out :D | 23:11 |
| sevan | (all the others report 3.4v) | 23:12 |
| Boostisbetter | Yeah, mine does that as well, but my Reform has been super solid. | 23:12 |
| minute | sevan: i've been testing new protected battery boards for a while now and we have the first production pcbs at the office, so basically a matter of days or 1-2 weeks until we release them... they will give you a bit more peace of mind | 23:13 |
| sevan | minute: very cool :) | 23:14 |
| sevan | Boostisbetter: I really like mine. | 23:15 |
| josch | minute: would it be much hassle to offer replacement 3d printed keycaps for the trackball in the shop? Our ultimaker doesn't print fine enough to create good replacement keycaps. | 23:18 |
| minute | josch: it would not, i will write it in my notes | 23:18 |
| josch | awsome, then those would go into my shopping basked together with the new battery board :) | 23:19 |
| Boostisbetter | minute: yet something else to add to the MNT shopping cart. Hahaha!! | 23:20 |
| minute | nice :3 | 23:20 |
| Boostisbetter | pocket reform, camera, and battery boards. | 23:22 |
| minute | haha | 23:24 |
| minute | yeah there's a bunch of stuff coming! | 23:24 |
| josch | Did I dream of a planned ITX carrier board for the SoM? I cannot find it being mentioned in the irc logs... | 23:39 |
| minute | i might have mentioned it, also on twitter | 23:46 |
| minute | i'm not sure about the featureset though | 23:46 |
| - Nulo (QUIT: Remote host closed the connection) (~Nulo@user/nulo) | 23:52 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!