| + klardotsh (~klardotsh@98.97.112.82) | 00:44 | |
| - robin_ (QUIT: Read error: Connection reset by peer) (~robin@user/terpri) | 00:51 | |
| + robin (~robin@user/terpri) | 00:53 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 01:40 | |
| - mtm (QUIT: Ping timeout: 252 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:04 | |
| - mjw (QUIT: Ping timeout: 240 seconds) (~mjw@gnu.wildebeest.org) | 02:05 | |
| - S0rin (QUIT: Ping timeout: 245 seconds) (~S0rin@user/s0rin) | 02:26 | |
| + S0rin (~S0rin@user/s0rin) | 02:28 | |
| - gnou_liber (QUIT: Read error: Connection reset by peer) (~gnou_libe@223.pool85-50-3.static.orange.es) | 03:17 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 03:28 | |
| - nsc (QUIT: Ping timeout: 250 seconds) (~nicolas@251-49-142-46.pool.kielnet.net) | 03:29 | |
| + nsc (~nicolas@96-49-142-46.pool.kielnet.net) | 03:31 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:09 | |
| josch | c-keen[m]: what is your booting setup? /boot on emmc and / on nvme? | 04:15 |
|---|---|---|
| josch | c-keen[m]: can you boot a rescue system from an sd-card for some analysis? | 04:16 |
| josch | c-keen[m]: once you did boot from an sd-card, i'd be interested in your contents of /boot as well as /etc/fstab | 04:16 |
| josch | c-keen[m]: also, which u-boot version are you using? | 04:17 |
| josch | c-keen[m]: and since you said that this was after an upgrade, i'd also be very interested to see your /var/log/apt/term.log and /var/log/apt/history.log -- but those two might contain sensitive data so either only select the parts that are interesting for this case before pastebin them or send them to me privately | 04:23 |
| + sl (~sl@contrib.inri.net) | 04:24 | |
| - sl (QUIT: Quit: Leaving...) (~sl@contrib.inri.net) | 04:37 | |
| - jacobk (QUIT: Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) (~quassel@47-186-122-163.dlls.tx.frontiernet.net) | 04:48 | |
| - S0rin (QUIT: Ping timeout: 250 seconds) (~S0rin@user/s0rin) | 05:23 | |
| + S0rin (~S0rin@user/s0rin) | 05:52 | |
| - gnou_liber (QUIT: Ping timeout: 245 seconds) (~gnou_libe@223.pool85-50-3.static.orange.es) | 06:06 | |
| - GNUmoon (QUIT: *.net *.split) (~GNUmoon@gateway/tor-sasl/gnumoon) | 06:08 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 06:32 | |
| - gnou_liber (QUIT: Ping timeout: 240 seconds) (~gnou_libe@223.pool85-50-3.static.orange.es) | 06:38 | |
| + gnou_liber (~gnou_libe@90.174.4.76) | 06:39 | |
| - gnou_liber (QUIT: Read error: Connection reset by peer) (~gnou_libe@90.174.4.76) | 06:41 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 06:42 | |
| - gnou_liber (QUIT: Ping timeout: 245 seconds) (~gnou_libe@223.pool85-50-3.static.orange.es) | 06:55 | |
| + gnou_liber (~gnou_libe@90.174.4.76) | 07:01 | |
| - sbp (QUIT: Server closed connection) (~sbp@apache/doge/sbp) | 07:04 | |
| + sbp (~sbp@2001:19f0:6800:1102:5400:ff:fe11:39a1) | 07:04 | |
| - gnou_liber (QUIT: Read error: Connection reset by peer) (~gnou_libe@90.174.4.76) | 07:04 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 07:04 | |
| - gnou_liber (QUIT: Ping timeout: 250 seconds) (~gnou_libe@223.pool85-50-3.static.orange.es) | 07:09 | |
| josch | minute: you extended the table at https://mntre.com/modularity.html#table with the rcm4 rpi but in the text above it still says "Subsequently, we have made four additional processor modules" (but it's five additional ones) and the list below that only has four entries and misses the rpi | 07:28 |
| + jacobk (~quassel@47-186-122-163.dlls.tx.frontiernet.net) | 07:31 | |
| + Assassink786[m] (~assassin7@2001:470:69fc:105::2:ac85) | 09:12 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:20 | |
| - klardotsh (QUIT: Ping timeout: 240 seconds) (~klardotsh@98.97.112.82) | 10:03 | |
| minute | josch: thanks for spotting that! | 10:13 |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 11:01 | |
| violet | minute: ordered batteries from that site you linked and they came in, and they're all at 3.29v by my multimeter! popped them in and they appear to be working | 11:02 |
| violet | now i just need to update the battery controller firmware or w/e | 11:02 |
| - gnou_liber (QUIT: Read error: Connection reset by peer) (~gnou_libe@223.pool85-50-3.static.orange.es) | 11:04 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 11:05 | |
| - gnou_liber (QUIT: Read error: Connection reset by peer) (~gnou_libe@223.pool85-50-3.static.orange.es) | 11:11 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 11:12 | |
| + mjw (~mjw@gnu.wildebeest.org) | 11:13 | |
| minute | violet: great | 11:14 |
| violet | with the banana pi cm4 module, is the HDMI something you can swap between the HDMI port and the internal display with like a software command, or is it a device tree thing you'd need to reboot for, or is it like actually swapping the physical connection | 11:17 |
| - S0rin (QUIT: Ping timeout: 250 seconds) (~S0rin@user/s0rin) | 11:33 | |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 11:38 | |
| + S0rin (~S0rin@user/s0rin) | 11:41 | |
| * mark_ -> mjw | 11:41 | |
| minute | violet: it's in software. just it hasn't been figured out yet how to drive both at the same time. afaik it is possible though, but i didn't want to overpromise | 11:47 |
| violet | ah rad. i personally never find myself using both anyhow so that'd work for me | 11:55 |
| violet | the perf cores are rather appealing ngl | 11:56 |
| minute | violet: it is a very different experience compared to imx8mq | 11:56 |
| - plomlompom (QUIT: Server closed connection) (~plom@play.plomlompom.com) | 12:45 | |
| + plomlompom (~plom@play.plomlompom.com) | 12:45 | |
| - mtm (QUIT: Ping timeout: 250 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:03 | |
| + mark_ (~mjw@gnu.wildebeest.org) | 14:13 | |
| - mark_ (QUIT: Ping timeout: 258 seconds) (~mjw@gnu.wildebeest.org) | 14:19 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:22 | |
| + bkeys (~Thunderbi@c-67-191-106-248.hsd1.fl.comcast.net) | 14:34 | |
| minute | ok, today i'm gonna work a bit on a311d integration | 14:41 |
| minute | josch: i would like to integrate a311d patches and dts here: https://source.mnt.re/reform/reform-debian-packages/-/tree/main/linux what do you think about creating subdirectories for the SoCs in the patches directory? just for some orderliness | 14:45 |
| minute | also i see a draft MR here https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/6 josch: is this OK to merge or are there problems? | 14:45 |
| minute | for now i will just prefix them | 14:49 |
| josch | minute: MR !6 was something that sigrid suggested and i tested but unfortunately did not work :( | 14:53 |
| minute | ohh ok | 14:54 |
| josch | minute: i have a local branch which creates these subdirectories for the pocket reform | 14:55 |
| josch | minute: i didn't push that one because it depends on linux 6.4 which is only in experimental still | 14:56 |
| minute | ah hmmmmmmmm | 14:56 |
| josch | minute: putting the patches into subdirectories requires some changes of the patching machinery -- would you like me to push the unfinished stuff? | 14:56 |
| minute | i'm not sure how to proceed with making a system image for a311d | 14:56 |
| minute | josch: if i can build on that it would be cool, yes | 14:57 |
| josch | okay, one sec... | 14:57 |
| minute | i'm not sure if it makes sense for me to build on top of the old debian kernel now hmm | 14:57 |
| minute | or if we need to go with a custom kernel until the debian one is new enough | 14:57 |
| minute | (for a311d) | 14:57 |
| minute | why is our hardware always so bleeding edge :D | 14:58 |
| josch | minute: there is now a branch called imx8mplus with two wip commits in it | 14:59 |
| josch | minute: the stuff in that branch already builds and i sent you the result some time ago | 14:59 |
| minute | ah cool | 14:59 |
| josch | so it should be safe to build the a311d stuff on top of that | 14:59 |
| josch | once i finish up the integration once 6.4 is released i can then rebase your a311d changes | 15:00 |
| minute | ahh i see, very nice | 15:00 |
| minute | i have now merged your imx8mplus branch into my new branch a311d | 15:00 |
| + mark_ (~mjw@82-217-174-174.cable.dynamic.v4.ziggo.nl) | 15:00 | |
| minute | now the question is, should we call this meson-g12b or a311d | 15:01 |
| josch | hah :D | 15:01 |
| minute | i think a311d would be easier to identify for most people... | 15:01 |
| minute | ah, lets see what the dts name is that neil submitted | 15:01 |
| minute | https://lore.kernel.org/all/20230512-amlogic-v6-4-upstream-dsi-ccf-vim3-v5-17-56eb7a4d5b8e@linaro.org/ | 15:02 |
| minute | > meson-g12b-bananapi-cm4-mnt-reform2.dts | 15:02 |
| minute | quite the mouthful | 15:02 |
| minute | but i will use that, then | 15:03 |
| josch | the string "MNT Reform 2 with BPI-CM4 Module" also does not match the format you recently suggested, sadly | 15:03 |
| minute | maybe this can still be changed | 15:05 |
| minute | because that huge MR is not yet final | 15:05 |
| minute | does someone know about a tool with which one can auto-download a patch series that from lore, cleanly? | 15:10 |
| minute | it's quite a hassle to do all of this manually | 15:10 |
| minute | hmm https://people.kernel.org/monsieuricon/introducing-b4-and-patch-attestation | 15:10 |
| minute | sudo apt install b4 | 15:11 |
| minute | :3 | 15:11 |
| minute | ha > b4 am 20230512-amlogic-v6-4-upstream-dsi-ccf-vim3-v5-0-56eb7a4d5b8e@linaro.org | 15:12 |
| minute | amazing | 15:12 |
| minute | josch: sorry if this is a lazy question, but where can i see debian's default kernel configuration? | 15:20 |
| minute | josch: i want to check what is already included in terms of CONFIG_MESON_... etc | 15:20 |
| minute | CONFIG_*_MESON | 15:20 |
| - mark_ (QUIT: Ping timeout: 245 seconds) (~mjw@82-217-174-174.cable.dynamic.v4.ziggo.nl) | 15:32 | |
| minute | josch: now checking if the patches will apply on debian's kernel :3 | 15:34 |
| minute | josch: we should perhaps add some requirements for building linux to the README, like python3-jinja and quilt | 15:35 |
| minute | oh > Patch debian/patches/reform/imx8mq-mnt-reform2/0005-imx8mq-import-HDMI-driver-and-make-DCSS-compatible-w.patch does not apply (enforce with -f) | 15:36 |
| minute | i guess i should merge main? | 15:36 |
| josch | minute: /boot/config-*-reform2-arm64 contains the config of the kernel you are running | 15:38 |
| josch | minute: 0005-imx8mq-import-HDMI-driver-and-make-DCSS-compatible-w.patch needing a rebase has to be done all the time... | 15:39 |
| josch | i'd wish somebody who knows how what the patch does could finally get it upstreamed... | 15:39 |
| minute | josch: the patch will never be upstreamed in this form as it is straight from the vendor kernel. i heard that lucas stach has an official hdmi patch brewing for a long time but i haven't seen it | 15:40 |
| josch | yes, lucas told me the same | 15:41 |
| josch | said that there is a way to have a single dtb offering both things | 15:41 |
| minute | not sure what's blocking that | 15:42 |
| minute | TIL "git checkout --theirs" | 15:42 |
| minute | ok, so the HDMI patch for imx8mq from main still works fine. | 15:43 |
| minute | a311d patch reject seem to be small | 15:44 |
| minute | rejects | 15:44 |
| minute | hmm it says > 1 out of 1 hunk FAILED -- rejects in file Documentation/devicetree/bindings/arm/amlogic.yaml | 15:45 |
| minute | normally i know patch saying filename with a .rej ending | 15:46 |
| josch | the patching is done using "patch" but with the tool quilt around | 15:47 |
| minute | ok, so not sure how i can find the actual rejects | 15:47 |
| minute | how did you do it when you had rejects? | 15:48 |
| josch | i applied the patches in a git tree, fixed rejects and then used git-format-patch to create the working patch stack | 15:48 |
| minute | ah hm | 15:48 |
| minute | i'll try to do that then | 15:49 |
| minute | ok, patches apply quite cleanly now | 16:06 |
| josch | oh funny you are in contact with Neil from Linaro -- i know him from Debian all the way back when I was working on Debian integration for the Openmoko Neo1973 and Freerunner :) | 16:11 |
| minute | josch: ah cool! so yeah he did all the mipi-dsi work | 16:11 |
| minute | josch: ah, is fuzz illegal when patching? | 16:12 |
| minute | > dpkg-source: info: the patch has fuzz which is not allowed, or is malformed | 16:12 |
| josch | yup | 16:12 |
| minute | damn | 16:12 |
| minute | sad because so far i didn't have to touch the patches :D | 16:12 |
| josch | it's a lot of fun to do this every time a new kernel comes out ;) | 16:13 |
| josch | 0005-imx8mq-import-HDMI-driver-and-make-DCSS-compatible-w.patch is the worst offender because it's so large and because the drm subsystem changed so much (including file moves and splits) in the past year | 16:13 |
| josch | oh and variable renames... | 16:14 |
| josch | i'm surprised it still works after all this patch has been through with me... | 16:14 |
| minute | yeah :3 | 16:16 |
| minute | actually | 16:17 |
| minute | https://lore.kernel.org/all/cover.1686729444.git.Sandor.yu@nxp.com/ | 16:17 |
| minute | interesting > Fix HDMI scrambling is not enable issue when driver working in 4Kp60 | 16:18 |
| minute | mode | 16:18 |
| minute | so NXP is actually contributing this. so we can test this soon and move to that patch series | 16:19 |
| josch | wow, that makes me very happy :D | 16:20 |
| Boostisbetter | I wonder if NXP recognizes the impact the Reform and Librem 5 have on their chip being seen as a good purchase. | 16:22 |
| Boostisbetter | and the soon to be Pocket Reform as well. | 16:22 |
| Boostisbetter | minute: the amount of support the Pocket crowd funding got was more than the Reform initially right? Were you surprised by this or did you kind of expect it at this point? | 16:23 |
| josch | I fear NXP sells orders of magnitude more of their devices to other industry customers ;) | 16:23 |
| minute | josch: yep | 16:25 |
| minute | but we have a good relationship with laird/boundary devices now | 16:26 |
| minute | so there is some visibility | 16:26 |
| Boostisbetter | I think that is great. Really happy for MNT. | 16:26 |
| Boostisbetter | Do we have any thoughts on Red Hat this past week? I admit at first I thought people were being ridicolous, and I thought RH was in their right for doing what they did and protecting their revenue stream. | 16:28 |
| Boostisbetter | However, I have come to see that people are more upset about all of the software that RHEL uses that they do not send any of their profits back to. | 16:29 |
| Boostisbetter | And from this point of view it makes sense to me the disappointment and anger. | 16:29 |
| Boostisbetter | I think Debian is going to see a nice influx of users in the coming weeks, which is good. | 16:29 |
| minute | Boostisbetter: what do you mean? i haven't been following what RH is doing | 16:36 |
| Boostisbetter | One other thing I'll throw in here this fine Friday is that I think the Reform implements the NXP SoC the best out of the products using it for Linux, simply because it is afforded as many resources as it can handle. 4gb instead of 3. And because of the PCIe lane, it is able have a bigger swap and you barely feel slow downs due to swapping. | 16:37 |
| Boostisbetter | minute: they officially stopped downstreaming the RHEL source to the Centos repo. The only way to get it now is to be a customer. | 16:37 |
| minute | hmm? what about the GPL? | 16:39 |
| Boostisbetter | apparently v2, and I guess they can do that with v2. | 16:39 |
| minute | also, what's interesting in RHEL? the only interesting product that comes to mind from RH is fedora | 16:40 |
| Boostisbetter | in a project I'm developing for work we are using RHEL. | 16:41 |
| Boostisbetter | It matter to most organizations that there is a support structure and certainties for the prducts that are required. | 16:41 |
| Boostisbetter | lets not forget that systemd and pipewire, pulseaudio, etc. all came from there. | 16:41 |
| minute | ah. but then you are a customer? | 16:42 |
| minute | yes, but these are not enterprise only things, right? | 16:43 |
| Boostisbetter | not by choice. That wasn't my decision. I would have gone with Debian or Ubuntu | 16:43 |
| Boostisbetter | right, but what I mean is, that even though this is sucky for many people, RH contributes still via upstreaming to many Linux projects. | 16:43 |
| sknebel | I dont think there is going to be that much that's not upstreamed to CentOS (although I havent seen anything conclusive about that yet) | 16:45 |
| sknebel | so while I dont like the weirdness of selling people GPLed software under expectations they dont share the source (as I understand it they do that, kinda like grsecurity does for their patches), I feel like its kind of overblown? | 16:46 |
| minute | sounds like | 16:46 |
| Boostisbetter | same here. There is a lot of overreacting going on. | 16:46 |
| sknebel | a bunch of projects are going to stop officially testing with RHEL, thats a downside for RedHat | 16:47 |
| sknebel | (*RHEL or compatible distros - clearly the latter are the target here) | 16:47 |
| sknebel | and the entire "we have to stop people building on some aspects of our work for free" squares not so well with "a linux distro thats builds on lots of other peoples work for free"... | 16:49 |
| + mark_ (~mjw@gnu.wildebeest.org) | 16:50 | |
| sknebel | I feel like the noise was predicatable and I dont really see this stopping the compatible distros that effectively, so I'm not sure it makes strategic sense for them, but *shrug* | 16:52 |
| jfred | When the Pocket Reform is open flat, are both sides flush with the table? I.e. are they both the same thickness? | 17:04 |
| minute | jfred: no, the top part is a bit thinner | 17:07 |
| minute | ok, patches sorted out, now > Error creating chroot | 17:07 |
| Boostisbetter | minute: can't wait for the pocket. If you need any help with debugging or working on them, please let me know. Hahaha | 17:07 |
| minute | Boostisbetter: thanks! | 17:08 |
| jfred | minute: Ah, got it - would have been a bit like a Dynabook with a hinge if they were the same, but admittedly the top being thinner does make more sense for a mini laptop haha | 17:09 |
| minute | jfred: you can surely use it at 180 degrees though | 17:09 |
| jfred | And I suppose if I wanted to make the lid thicker that's easier than the reverse :P | 17:10 |
| minute | jfred: yep | 17:10 |
| minute | josch: do i have to do anything special in schroot.conf, like set an architecture or something? is users=rleigh supposed to be my username? | 17:11 |
| minute | hmm > chroot:unstable chroot does not exist | 17:12 |
| minute | trying sudo sbuild-createchroot sid /srv/chroot/sid | 17:15 |
| - mark_ (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 17:15 | |
| josch | minute: as the sbuild maintainer i'd recommend using the unshare backend | 17:17 |
| minute | josch: no idea what that means! | 17:18 |
| josch | minute: the advantage is, that in contrast to the schroot backend, you don't need to be superuser | 17:18 |
| josch | and there is no configuration in /etc | 17:18 |
| josch | first, make sure you have these packages installed: sbuild mmdebstrap uidmap | 17:18 |
| josch | then: | 17:19 |
| josch | mkdir -p ~/.cache/sbuild | 17:19 |
| josch | mmdebstrap --variant=buildd unstable ~/.cache/sbuild/unstable-amd64.tar.xz | 17:19 |
| josch | (assuming you are on amd64, right?) | 17:19 |
| minute | correct! | 17:19 |
| josch | and then put this in your ~/.sbuildrc: | 17:19 |
| josch | $chroot_mode = 'unshare'; | 17:19 |
| minute | ok, it's downloading stuff | 17:19 |
| josch | from that point onward, you are not needing schroot anymore and it should use the tarball in your home directory instead | 17:19 |
| josch | (which you can also create without superuser privileges) | 17:19 |
| minute | cool | 17:19 |
| minute | i will copy this to the README in reform-debian-packages then | 17:20 |
| josch | ACTION is afk now for ~1h -- hope everything works! | 17:20 |
| minute | josch: thank you! | 17:20 |
| minute | sigh > newuidmap: Target process 64220 is owned by a different user: uid:1000 pw_uid:1000 st_uid:1000, gid:131 pw_gid:1000 st_gid:131 | 17:23 |
| minute | ah hm > /usr/bin/fakechroot: 87: Cannot fork | 17:27 |
| minute | hm, i was already in some kind of container or sth :3 | 17:31 |
| minute | > I: success in 120.4871 seconds | 17:32 |
| minute | i did, in a fresh terminal > mmdebstrap --variant=buildd --mode=unshare unstable ~/.cache/sbuild/unstable-amd64.tar.xz | 17:32 |
| minute | ok, sbuild is springing to life | 17:34 |
| minute | > E: Build failure (dpkg-buildpackage died) | 17:36 |
| minute | ah, pocket reform issue | 17:38 |
| minute | Error: /<<PKGBUILDDIR>>/arch/arm64/boot/dts/freescale/imx8mp-mnt-pocket-reform.dts:795.1-8 Label or path lcdif1 not found | 17:38 |
| minute | FATAL ERROR: Syntax error parsing input tree | 17:38 |
| minute | > Error: /<<PKGBUILDDIR>>/arch/arm64/boot/dts/freescale/imx8mp-mnt-pocket-reform.dts:755.1-10 Label or path mipi_dsi not found | 17:38 |
| minute | commenting pocket dts out for now | 17:39 |
| - gnou_liber (QUIT: Read error: Connection reset by peer) (~gnou_libe@223.pool85-50-3.static.orange.es) | 17:47 | |
| + gnou_liber (~gnou_libe@223.pool85-50-3.static.orange.es) | 17:55 | |
| minute | Status: successful | 18:10 |
| minute | Version: 6.3.7-1+reform20230630T153953Z1 | 18:10 |
| minute | nice. i'm just missing the reprepro command | 18:10 |
| - bkeys (QUIT: Read error: Connection reset by peer) (~Thunderbi@c-67-191-106-248.hsd1.fl.comcast.net) | 18:20 | |
| noam | Realllllllly liking the Reform :D | 18:20 |
| Boostisbetter | Reaaaalllly happy for you Noam! | 18:23 |
| + mark_ (~mjw@gnu.wildebeest.org) | 18:45 | |
| - deflated8837 (QUIT: Ping timeout: 240 seconds) (~deflated8@50.53.206.61) | 18:45 | |
| josch | minute: you are building the entire kernel in 20 minutes???? | 18:49 |
| minute | josch: it's a i9-9900k, but also running at half speed (HT turned off) | 18:53 |
| minute | (well, HT probably doesn't really give the same perf as 2x the cores) | 18:53 |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark_!~mjw@gnu.wildebeest.org))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 19:06 | |
| * mark_ -> mjw | 19:06 | |
| + mark_ (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 19:06 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 19:23 | |
| noam | It gets surprisingly close, tbh, at least on my ryzen | 19:27 |
| josch | well... and here i am spending 6+ hours to build the thing on my reform XD | 19:29 |
| noam | Heh, I just don't run Linux and that problem is solved ;) | 19:41 |
| noam | ~121s to build 9front on the reform for me :D [though, in fairness, I removed ghostscript from my branch, or it would be another minute or three probably] | 19:42 |
| Boostisbetter | that is pretty nice! So does Plan 9 work completelty on the Reform? | 19:46 |
| noam | AFAIK, yeah | 19:47 |
| noam | ...except external displays, probably? | 19:47 |
| Boostisbetter | nice, I don't think I have ever tried to use one on Debian either. | 19:51 |
| - ndufresne (QUIT: Quit: The Lounge - https://thelounge.chat) (~ndufresne@gyros.collabora.co.uk) | 20:06 | |
| + ndufresne (~ndufresne@gyros.collabora.co.uk) | 20:07 | |
| minute | i just discovered something | 20:14 |
| minute | a311d and imx8mq ddrc firmware blob(s) are almost identical | 20:14 |
| minute | at least they contain some identical data | 20:14 |
| minute | so they're both synopsys ddrc | 20:14 |
| * cwebber`` -> cwebber | 20:25 | |
| - cwebber (QUIT: Changing host) (~user@2600:6c64:7e3f:fdb0::6db) | 20:25 | |
| + cwebber (~user@user/cwebber) | 20:25 | |
| + bkeys (~Thunderbi@c-67-191-106-248.hsd1.fl.comcast.net) | 20:59 | |
| + deflated8837 (~deflated8@50.53.206.61) | 21:03 | |
| minute | ok, built mainline u-boot (with at-f blobs from amlogic) and that boots on rcm4-bpi | 21:36 |
| minute | hmm, which .deb contains the dtb | 21:41 |
| josch | probably linux-image-6.3.0-0-reform2-arm64 | 21:43 |
| minute | ah yes, but it is in /usr/lib | 21:43 |
| minute | ok, so the new dtb isn't there, hmmmm | 21:44 |
| josch | did you add it to the Makefile? | 21:44 |
| minute | the .dts ended up in the tree at least, and i made a patch to add it to the makefile, lets see | 21:45 |
| josch | they are added using a sed one-liner in build.sh | 21:45 |
| minute | nope, did not end up in the makefile | 21:45 |
| josch | like this for the pocket reform: | 21:45 |
| josch | sed -i '/imx8mq-mnt-reform2.dtb/a dtb-$(CONFIG_ARCH_MXC) += imx8mp-mnt-pocket-reform.dtb' linux/arch/arm64/boot/dts/freescale/Makefile | 21:45 |
| minute | josch: yeah, i've seen that, and... ahhh | 21:46 |
| minute | i made a patch 0002-add-dts-to-makefile | 21:46 |
| minute | but forgot to call it .patch | 21:46 |
| minute | :D | 21:46 |
| josch | ah okay :) | 21:47 |
| minute | before i rebuild all of this i will test with another dtb first | 21:48 |
| minute | ok, > Starting kernel ... | 21:58 |
| minute | and then no more output :D i guess i have to tweak the kernel commandline... | 21:58 |
| minute | ok cool, kernel has booted | 22:04 |
| minute | [ 0.000000] Linux version 6.3.0-1-reform2-arm64 (debian-kernel@lists.debian.org) (aarch64-linux-gnu-gcc-12 (Debian 12.3.0-4) 12.3.0, GNU ld (GNU Binutils for Debian) 2.40.50.20230625) #1 SMP Debian 6.3.7-1+reform20230630T153953Z1 (2023-06-30) | 22:04 |
| minute | [ 0.000000] Machine model: MNT Reform 2 with BPI-CM4 Module | 22:04 |
| vkoskiv | If I want some in-stock bits, as well as the banana pi + adapter board, would it make more sense for me to put in two orders? | 22:05 |
| minute | vkoskiv: which bits are those? | 22:05 |
| vkoskiv | I want the protected bat board, laird antenna, homing caps and the cm4 adapter + A311D | 22:06 |
| vkoskiv | I'm fine with paying shipping twice, if I can get the other stuff a bit sooner | 22:06 |
| minute | yeah, then do two orders i would say | 22:07 |
| minute | apparently, there are still over 600 bpi cm4 in stock at sinovoip | 22:07 |
| minute | last time they took a little to ship, like 2 weeks at least. i will contact them to see if we can do that quicker | 22:08 |
| vkoskiv | And is the A311D just a swap-in replacement? | 22:12 |
| vkoskiv | As in, will my existing sysv3 install boot up on it? | 22:12 |
| noam | Different SoC, IIUC, so probably needs a different kernel build + u-boot firmware, for Linux | 22:12 |
| josch | vkoskiv: in the near future you will not need to swap out the kernel | 22:13 |
| josch | vkoskiv: but you will need a different u-boot and a different dtb | 22:13 |
| minute | the kernel will be unified | 22:14 |
| minute | u-boot goes to offset 512 bytes with this machine btw | 22:14 |
| minute | here is the current status of the kernel integration josch https://source.mnt.re/reform/reform-debian-packages/-/commits/a311d | 22:15 |
| minute | i have to make feierabend now, next week i have to sort out the modules | 22:15 |
| minute | josch: here are my current notes on how the bootloader is built. i wonder how we can integrate this with reform-system-image. i'm thinking to modify the existing image after the fact, i.e. just write this different bootloader to the .img file. but no idea how to integrate the dtb, initrd (?) etc. | 22:19 |
| minute | josch: https://source.mnt.re/-/snippets/3 | 22:19 |
| vkoskiv | josch: Sounds doable. It's not like I'll be swapping between modules on the regular. | 22:22 |
| noam | heck, could just keep two tiny uSD cards :P | 22:22 |
| vkoskiv | All my data is on encrypted lvm on the nvme, so I can keep that and modify the boot stuff when I swap modules | 22:23 |
| noam | Or, wait, there's eMMC isn't there? Yeah | 22:23 |
| vkoskiv | noam: Yeah. I'm using the emmc to boot, so I could just have 2 images I can flash in case I want to swap back for whatever reason | 22:23 |
| noam | Just flash eMMC with u-boot for the given module, and it should JustWork if the kernel's on NVMe :) | 22:23 |
| vkoskiv | Or well, if the emmc is *on* the module, won't that just... work? | 22:23 |
| noam | yeah | 22:33 |
| noam | that's what I was saying ;) | 22:33 |
| josch | no, this will work once but not for long | 22:39 |
| josch | the problem is, that the next kernel upgrade will then destroy your setup | 22:39 |
| noam | ...why's that? | 22:39 |
| josch | because even though your kernel and dtb might be on a different partition, the information which dtb to choose is in your /etc | 22:39 |
| noam | Shouldn't the firmware still be fine? | 22:39 |
| vkoskiv | Ah | 22:39 |
| josch | so even if you have your / on nvme and /boot on emmc, you have to adapt stuff in /etc when you swap modules | 22:40 |
| vagrantc | well, if you copy *all* the dtb files for a given kernel, and use the mechanisms u-boot provides to select a .dtb ... you should be ok? | 22:40 |
| josch | vagrantc: but flash-kernel only selects a single dtb to copy, no? | 22:40 |
| vagrantc | that is true. there are hooks in u-boot-menu to copy all the .dtbs now | 22:41 |
| vagrantc | and also not hard to implement as a one-off script in /etc/kernel*.d | 22:41 |
| josch | yes *if* you switch to u-boot-menu | 22:42 |
| josch | which requires a recent u-boot | 22:42 |
| josch | which we cannot automatically upgrade ;) | 22:42 |
| vagrantc | isn't this talking about systems with the cm4 adapters ... which have not yet even shipped? | 22:42 |
| josch | yes, but we want to have the same userland on all platforms | 22:43 |
| vagrantc | u-boot is not userland | 22:43 |
| josch | because we have a single repo from which all of them upgrade | 22:43 |
| josch | but the config is | 22:43 |
| vagrantc | the config is unchangeable? :P | 22:43 |
| josch | lets discuss this another day -- i have to get up super early tomorrow :) | 22:43 |
| vagrantc | fair :) | 22:44 |
| josch | vagrantc: i'll surely consult you with what i'm coming up with ;) | 22:44 |
| vagrantc | i will be happy to help | 22:44 |
| josch | thank you! :) | 22:44 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!