| + terpri (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 00:04 | |
| - terpri (QUIT: Ping timeout: 260 seconds) (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 00:09 | |
| + terpri (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 00:15 | |
| - erlehmann (QUIT: Ping timeout: 256 seconds) (~erlehmann@46.114.38.193) | 00:19 | |
| + terpri_ (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 00:21 | |
| - terpri (QUIT: Ping timeout: 240 seconds) (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 00:21 | |
| - terpri_ (QUIT: Ping timeout: 240 seconds) (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 00:28 | |
| + erlehmann (~erlehmann@46.114.38.193) | 00:30 | |
| - erlehmann (QUIT: *.net *.split) (~erlehmann@46.114.38.193) | 00:55 | |
| - signop (QUIT: *.net *.split) (~signop@li699-102.members.linode.com) | 00:55 | |
| - technomancy (QUIT: *.net *.split) (~user@170.ip-158-69-211.net) | 00:55 | |
| - ndufresne (QUIT: *.net *.split) (stormer@nat/collabora/x-mskoxdqaszrgmtfo) | 00:55 | |
| - mntmn (QUIT: *.net *.split) (~mntmn@softboy.mntmn.com) | 00:55 | |
| - marlun (QUIT: *.net *.split) (~marlun@unaffiliated/marlun) | 00:55 | |
| + erlehmann (~erlehmann@46.114.38.193) | 00:57 | |
| + signop (~signop@li699-102.members.linode.com) | 00:57 | |
| + technomancy (~user@170.ip-158-69-211.net) | 00:57 | |
| + ndufresne (stormer@nat/collabora/x-mskoxdqaszrgmtfo) | 00:57 | |
| + mntmn (~mntmn@softboy.mntmn.com) | 00:57 | |
| + marlun (~marlun@unaffiliated/marlun) | 00:57 | |
| * hitchcock.freenode.net changed mode (+o, mntmn) | 00:57 | |
| + terpri (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 01:05 | |
| + Albylmannen (~Berof@31-208-217-48.cust.bredband2.com) | 01:09 | |
| + terpri_ (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 01:11 | |
| - terpri (QUIT: Ping timeout: 260 seconds) (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 01:11 | |
| + terpri (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 01:18 | |
| - terpri_ (QUIT: Ping timeout: 260 seconds) (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 01:19 | |
| - rvense (QUIT: Ping timeout: 240 seconds) (~rvense@93-163-17-86-static.dk.customer.tdc.net) | 02:26 | |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@gateway/web/irccloud.com/x-procdvwcozcargvt) | 02:29 | |
| - erlehmann (QUIT: Ping timeout: 256 seconds) (~erlehmann@46.114.38.193) | 02:39 | |
| + erlehmann (~erlehmann@46.114.33.178) | 02:52 | |
| + rvense (~rvense@93-163-17-86-static.dk.customer.tdc.net) | 03:24 | |
| + B[] (~Thunderbi@219-89-228-78-adsl.sparkbb.co.nz) | 03:25 | |
| - Albylmannen (QUIT: Ping timeout: 264 seconds) (~Berof@31-208-217-48.cust.bredband2.com) | 03:38 | |
| - erlehmann (QUIT: Disconnected by services) (~erlehmann@46.114.33.178) | 05:49 | |
| + erlehmann_ (~erlehmann@46.114.37.207) | 05:49 | |
| * erlehmann_ -> erlehmann | 05:49 | |
| - erlehmann (QUIT: Remote host closed the connection) (~erlehmann@46.114.37.207) | 05:55 | |
| + erlehmann (~erlehmann@46.114.37.207) | 05:56 | |
| - B[] (QUIT: Ping timeout: 256 seconds) (~Thunderbi@219-89-228-78-adsl.sparkbb.co.nz) | 06:10 | |
| + B[] (~Thunderbi@219-89-228-78-adsl.sparkbb.co.nz) | 06:12 | |
| + oomono (uid328183@gateway/web/irccloud.com/x-brfzcedtmnhhyjgu) | 09:05 | |
| - dckc (QUIT: Ping timeout: 240 seconds) (~DanC@user2.kumc.edu) | 09:54 | |
| + dckc (~DanC@2605:a601:ac60:1000:c56e:c837:64b4:d70c) | 09:59 | |
| lastebil | mntmn: um let's see (: model name : Intel(R) Core(TM) i3-6100 CPU @ 3.70GHz | 10:09 |
|---|---|---|
| lastebil | so quite a bit less powerful | 10:09 |
| lastebil | it worked, now I am seeing if I can modify it for devuan (: | 10:10 |
| lastebil | (worked last night but then I slept) | 10:10 |
| lastebil | theoretically I can set up a build box for this in my data center but this is a home box | 10:10 |
| lastebil | (yes, I am running xen at home, because I can.) | 10:15 |
| lastebil | hmm, is there a reason you used dd with seek 8000 instead of truncate? (Asking, as I normally don't seek when using dd and when I want sparse, I use truncate) | 10:30 |
| + ehmry (~quassel@2a03:3b40:fe:ab::1) | 10:32 | |
| mntmn | lastebil: no reason. :) | 10:34 |
| swivel | i'm pretty sure truncate(1) is a relatively recent development, the dd method has been around forever | 10:36 |
| lastebil | I'm actually fairly sure it's the opposite; I remember 25 years ago using truncate. | 10:50 |
| lastebil | fairly sure this was the norm for anyone using solaris, and when I worked for a large US telco, we had many folks who had used solaris since ... well, the first version purchased (: | 10:51 |
| lastebil | also that telco "may" have had something to do with unix, and we "may" have had some people who wrote some of the utilities on board. However I never worked with those folks. | 10:51 |
| lastebil | nonetheless I'm not sure. I asked because of habit of using truncate, based on that experience (: | 10:51 |
| mntmn | ha! unfortunately never experienced a version of solaris. i did see some sun fire servers in the flesh though, and some tape robot with them. in ~1999 | 10:51 |
| lastebil | I have some pizzaboxes here (: | 10:52 |
| lastebil | and a copy, boxed, of solaris 10 (: | 10:52 |
| mntmn | nice | 10:52 |
| lastebil | or two. | 10:52 |
| lastebil | anyway messing about with multistrap now (: | 10:52 |
| mntmn | ah yeah! | 10:53 |
| lastebil | I dislike the "huge package list" thing, and ... well. TRIED the bit with | 10:53 |
| lastebil | One alternative is to define your list of packages as multiple groups with packages separated on a functional / dependency basis, e.g. base, Xorg, networking etc. and list each group under 'bootstrap' | 10:53 |
| lastebil | but this would not parse, so - need an example of that. For now, just using a ton of packages= lines | 10:53 |
| mntmn | ah, that's possible? | 10:54 |
| mntmn | the huge list is unwieldy yes, even if it's not so many packages yet. categories would surely be nice | 10:54 |
| swivel | lastebil: truncate was added to gnu coreutils in 2008 | 10:55 |
| swivel | so from a linux perspective, no, truncate is new and dd is the traditional way of doing htis | 10:55 |
| swivel | https://git.savannah.gnu.org/cgit/coreutils.git/commit/?id=75e0047c4b916393b6be3cb985c4c4d4a2d1a836 | 11:01 |
| lastebil | swivel: Linux has always been a bit goofy from the Unix perspective (: | 11:08 |
| lastebil | I'd have to go install a Solaris box to actually check if dd has the seek option on solaris. I suppose I can check the dd manpage for openbsd... | 11:09 |
| lastebil | ok I'd better get moving to my eye doctor appointment, leaving this building... will come back and if it works, try to use devuan as the repo (: | 11:10 |
| lastebil | btw mntmn how are you putting your custom sway portion in, instead of the debian package? maybe it's just that I'm using Beowulf and there IS no sway package, and I build from source - but | 11:11 |
| lastebil | are you actually making a package and sourcing the _local_ package in this multistrap thing? (if so, I don't see the clause for that.) | 11:11 |
| lastebil | (feel free to answer in the future, I have to go to the eye doctor (: ) | 11:12 |
| swivel | what little exposure i had to solaris in the late 90s/early 00s showed me people installing gnu userspaces just to make the systems pleasant to use | 11:12 |
| lastebil | that happened yes; but then there was the "ok, this installer is written by some Linux goofball who doesn't know that bash is not /bin/sh on anything" | 11:14 |
| lastebil | but you wanna talk about pain? | 11:14 |
| lastebil | HPUX (: | 11:15 |
| + Valkyria (~Valkyria@85-76-97-109-nat.elisa-mobile.fi) | 11:23 | |
| mntmn | lastebil: when building from source, the packages just go to the /usr/local/... prefix and take precedence. that's what this relies on. so my sway is /usr/local/bin/sway | 11:28 |
| + terpri_ (~terpri@2607:fb90:cdf:a0f0:605b:7454:c3fe:ad20) | 11:58 | |
| - terpri (QUIT: Ping timeout: 244 seconds) (~terpri@2606-a000-1127-a177-b4ed-3008-1302-e258.inf6.spectrum.com) | 11:59 | |
| - Valkyria (QUIT: Quit: Going offline, see ya! (www.adiirc.com)) (~Valkyria@85-76-97-109-nat.elisa-mobile.fi) | 12:16 | |
| - terpri_ (QUIT: Ping timeout: 240 seconds) (~terpri@2607:fb90:cdf:a0f0:605b:7454:c3fe:ad20) | 12:48 | |
| + Valkyria (~Valkyria@85-76-97-109-nat.elisa-mobile.fi) | 13:36 | |
| - Valkyria (QUIT: Quit: Going offline, see ya! (www.adiirc.com)) (~Valkyria@85-76-97-109-nat.elisa-mobile.fi) | 14:13 | |
| - gcc-addict (QUIT: Remote host closed the connection) (~gcc-addic@gateway/tor-sasl/gcc-addict) | 14:22 | |
| + gcc-addict (~gcc-addic@gateway/tor-sasl/gcc-addict) | 14:23 | |
| lastebil | mntmn: cool. (As I will want to build tmux to be the latest version, rather than the version in Beowulf (: ) | 14:24 |
| mntmn | :3 | 14:25 |
| erlehmann | i actualld told mntmn about the truncate thing some time ago and now i wonder if it has some drawbacks? | 14:34 |
| erlehmann | (does it?) | 14:34 |
| lastebil | I'm not sure. Again, we _used_ to use it. If there ARE drawbacks, it may be 'sparse file' vs 'actual file' rather than the tool that _makes_ the sparse file. | 14:59 |
| lastebil | - first pass at devuan build | 14:59 |
| lastebil | aaand it's not going to work to use chimaera, which, well, that's not stable anyway (: | 14:59 |
| erlehmann | sparse files are a very good idea | 15:00 |
| erlehmann | because that way you can make an image for an 8GB filesystem on an 8GB filesystem | 15:00 |
| erlehmann | as long as it is mostly empty | 15:00 |
| lastebil | yes. they ONLY become an issue when you have a program that does not understand 'holes' | 15:04 |
| erlehmann | lastebil, which would be what? | 15:05 |
| erlehmann | lastebil, i mean it is on a filesystem level | 15:05 |
| lastebil | which SHOULDN'T be a problem here. (I seem to recall "something" with swap space that was a sparse file failing miserably at one time, but I never experienced that myself, so it may just be "idle talk" by someone I picked up) | 15:05 |
| erlehmann | lastebil, so how would a program not understand holes at all? | 15:05 |
| erlehmann | hahaha | 15:05 |
| erlehmann | swap space as sparse file | 15:05 |
| erlehmann | that is a good joke | 15:05 |
| erlehmann | good yolk | 15:05 |
| erlehmann | eggs | 15:05 |
| lastebil | yeah, but you can imagine someone trying it (: | 15:06 |
| erlehmann | i wonder what would go wrong there | 15:06 |
| erlehmann | i mean sparseness is a thing handled by the filesystem | 15:06 |
| mntmn | mmap maybe | 15:29 |
| - dckc (QUIT: Ping timeout: 240 seconds) (~DanC@2605:a601:ac60:1000:c56e:c837:64b4:d70c) | 15:33 | |
| + Valkyria (~Valkyria@85-76-97-109-nat.elisa-mobile.fi) | 15:41 | |
| + dckc (~DanC@user2.kumc.edu) | 15:43 | |
| mntmn | hmno, sparse files can be mmapped. | 15:53 |
| lastebil | well, the devuan expiriment is not going well, problems with multistrap, and as this is a step I don't actually need before setting up void, I'll drop it for now | 16:14 |
| lastebil | but before I do the void portion: remapping the keys at the kernel level was possible, yes? Where was that documented ? | 16:15 |
| mntmn | lastebil: hmm, i'm not sure about how it works on the kernel level. i personally would just change it in the keyboard | 16:24 |
| mntmn | lastebil: but it's probably safer and better to figure out how to do this in software | 16:24 |
| lastebil | ah, ok. Thought I remembered you mentioning it being done at kernel level before when I mentioned remapping via xkb or similar | 16:24 |
| mntmn | lastebil: ah, this might be interesting https://jlk.fjfi.cvut.cz/arch/manpages/man/setkeycodes.8 | 16:24 |
| lastebil | ah yes I believe I have used that with alpine; it uses a binary form of key mapping to do things | 16:25 |
| mntmn | there's also something involving udev | 16:26 |
| lastebil | not sure I want to involve udev. | 16:26 |
| mntmn | the other solution is to edit the firmware of the keyboard and reflashing it, but it involves opening the top bezel of the keyboard and, in this version, unscrewing the oled PCB because the reset button for the keyboard is exactly under that. | 16:26 |
| lastebil | hmm, does it use qwk or similar? | 16:27 |
| mntmn | not yet. | 16:27 |
| mntmn | one second | 16:27 |
| lastebil | right then let's not do that JUST yet. I am familiar with qwk et al for my iris and atreus | 16:27 |
| mntmn | here's the source and the matrix https://source.mntmn.com/MNT/reform/src/branch/master/reform2-keyboard-fw/Keyboard.c#L72 | 16:28 |
| mntmn | so you just need gcc and avr and edit that and do "make" in that folder. and then sudo ./flash.sh but for that the keyboard needs to be reset via its little button. | 16:28 |
| mntmn | this is to prevent unauthorized reflashing. | 16:28 |
| lastebil | yep. makes sense. | 16:29 |
| lastebil | however I don't really want to disassemble it right now, I'll go for the remapping keys in software method. | 16:29 |
| mntmn | yep. | 16:29 |
| mntmn | very understandable! | 16:29 |
| mntmn | also your results would interest me. | 16:29 |
| lastebil | will let you know when I'm finished. | 16:29 |
| mntmn | cool | 16:32 |
| - Valkyria (QUIT: Quit: Going offline, see ya! (www.adiirc.com)) (~Valkyria@85-76-97-109-nat.elisa-mobile.fi) | 16:39 | |
| lastebil | well, dumpkeys output on the reform2 is... significantly different from on other machines I have (: | 17:34 |
| - B[] (QUIT: Ping timeout: 256 seconds) (~Thunderbi@219-89-228-78-adsl.sparkbb.co.nz) | 17:35 | |
| lastebil | hmm. I may actually need to redo the kb firmware to do one thing I had wanted (swap the space and alt keys) in that you have mapped both space keys to produce the same keycode, which actually does make sense, can't think of another way to do it | 17:50 |
| lastebil | unless one was non breaking space or japanese halfspace but yeah we don't really... have those for typing (: | 17:51 |
| lastebil | not in the west... | 17:51 |
| lastebil | but I'll possibly just manage to get my thumb to learn the right spot (: | 17:51 |
| lastebil | think I can fix the others though | 17:51 |
| lastebil | hmm, only one issue with this remapping method; I'm remapping _any_ keyboard at the console level; so this means the usb keyboard I had plugged in suddenly has 'up' on the left shift (: | 18:02 |
| lastebil | of course, 'up' is doing left shift, as I chose to swap those for the reform keyboard (: | 18:02 |
| technomancy | unlabeled keycaps crew represent =D | 18:21 |
| technomancy | I won't say I told you so, but ... | 18:22 |
| lastebil | I actually think I ordered unlabeled , but I don't mind | 18:22 |
| lastebil | because the backlighting isn't really that bright and the keycaps having slight amounts of light on them don't hurt (: | 18:22 |
| technomancy | homing keys > backlight | 18:23 |
| lastebil | homing keys? | 18:23 |
| lastebil | oh | 18:23 |
| lastebil | no that woudln't help | 18:23 |
| technomancy | having your index fingers rest on keycaps that have some kind of nub or indentation on them means you can ensure they're positioned right without looking | 18:23 |
| lastebil | right. that would not help at all. | 18:23 |
| lastebil | ok think with your atreus (if you translate my nick from Norwegian you'll know who I am) | 18:24 |
| lastebil | so for that keymap that you put on my model - I needed to replace the left shift with 'left shift if held, return if tapped' (: | 18:25 |
| lastebil | homing keys don't help that (: | 18:25 |
| lastebil | mntmn: sent you a mail with the loadkeys script I ended up with, rather simple really | 18:29 |
| technomancy | ah I see; yeah | 18:30 |
| lastebil | have not tested with sway; will do that now (: | 18:30 |
| lastebil | technomancy: also haven't tested it with the atreus yet (: | 18:30 |
| lastebil | SHOULD work just fine of course, it works with other usb keyboards (: | 18:30 |
| lastebil | right that just fixes the console, I'll have to do some sway remapping too (: | 18:33 |
| lastebil | but at least now I can type passwords that have shifted characters without as many mistakes (: | 18:33 |
| lastebil | (the actual real issue (: ) | 18:34 |
| mntmn | lastebil: thanks for your mail! i'm thinking to do a docs folder with bits of info like this in the repo, can i include it? | 18:47 |
| mntmn | lastebil: cool that it works btw | 18:47 |
| lastebil | yes. I'll probably put some more documentation together on my sourcehut with various things I do for it, etc. | 18:47 |
| mntmn | lastebil: also while i was just out to go to UPS, loadkeys popped back into my mind, don't know why it didn't occur to me earlier | 18:48 |
| mntmn | lastebil: that's also cool | 18:48 |
| lastebil | because it's the sort of thing you forget (: I was forgetting it also, thinking I needed to use the binary format that loadkeys -b makes | 18:48 |
| lastebil | (which is loaded by busybox... and that's a bit... ugly.) | 18:48 |
| mntmn | technomancy: you are the only one with homing keycaps btw, hence the confusion | 18:49 |
| mntmn | we didn't put them on the engraved once | 18:49 |
| mntmn | ones | 18:49 |
| lastebil | also, loadkeys -b doesn't really _work_ on this... | 18:49 |
| lastebil | oh technomancy got one too? (: | 18:49 |
| technomancy | mntmn: ah, yeah makes sense, because if you want to swap to dvorak or colemak then the homing caps would be wrong. | 18:49 |
| technomancy | lastebil: hell yeah | 18:49 |
| mntmn | not sure, but they said they have blank keycaps and there is only one. | 18:49 |
| technomancy | what! | 18:49 |
| technomancy | only one order for blank caps? what's the world coming to ... | 18:50 |
| mntmn | yes | 18:50 |
| mntmn | oh wait | 18:50 |
| mntmn | maybe i'm the one who's confused. i'm talking about beta reforms. | 18:50 |
| mntmn | technomancy: so probably you don't have your reform _yet_ | 18:50 |
| mntmn | from the 8 beta units, there was one with blank keys and we put homing keycaps on it. | 18:51 |
| technomancy | aha; I see. yeah I will have to wait longer. it's worth it tho! | 18:51 |
| lastebil | it is. (: | 18:52 |
| mntmn | i'm checking now how many blanks there will be... | 18:52 |
| mntmn | almost 20 if i didn't miscount | 18:53 |
| + Albylmannen (Berof@31-208-217-48.cust.bredband2.com) | 19:12 | |
| + terpri (~terpri@cpe-98-26-15-90.nc.res.rr.com) | 19:31 | |
| - terpri (QUIT: Remote host closed the connection) (~terpri@cpe-98-26-15-90.nc.res.rr.com) | 19:38 | |
| + terpri (~terpri@2606-a000-1127-83a8-29bb-d4a9-b6ca-cbc6.inf6.spectrum.com) | 19:38 | |
| lastebil | mntmn: hmm, was there something where we could not build u-boot to actually output to the display? Was it possible to output to an attached hdmi display? | 20:20 |
| lastebil | (I probably should look more closely at your u-boot setup, but what I would actually like to do is have the uboot support for syslinux style boot, which I have on some of my other arm board things... makes things a lot easier for testing (: ) | 20:21 |
| mntmn | lastebil: hdmi is kind of possible but internal display is currently not possible because we don't have mipi-dsi and SN65DSI86 drivers in uboot | 20:21 |
| mntmn | lastebil: also i would like to get rid of u-boot in favor of linuxboot later, like use only the bare minimum of u-boot and immediately launch a linux pre-kernel which would then kexec into a thing you choose | 20:22 |
| lastebil | yeah. and serial would be "open the board and attach the wires" -which I _could_ do but again, easier if I just can press 1 or 2 etc (: | 20:22 |
| mntmn | yeah, this is a flaw. i'm thinking how we can get serial out of the case without opening it for the next board version, too. | 20:22 |
| lastebil | yeah. I would like that too. it's just, well, I'm testing a void tarball rather than trying to redo the reform-system-image to untar the tarball (: | 20:22 |
| lastebil | yeah really the only easy way would be to have a rubber stopper over some pins that went to rx/tx, so you could attach your usb setup to it | 20:23 |
| mntmn | the original boundary vendor kernel does activate the hdmi output in u-boot, so i know that's possible. but you would also need to autoload usb and usbkbd or something to make use of it | 20:24 |
| lastebil | which I'm fine with for testing. | 20:24 |
| mntmn | like: "usb start" and "stdin usbkbd" or something like that | 20:24 |
| lastebil | (a larger uboot isn't that big an issue with the 32 gb card. later, when using emmc/nand, well, that won't be needed, because linuxboot (: | 20:24 |
| mntmn | (not tested by me on reform2, but i did this on reform1) | 20:24 |
| mntmn | lastebil: yeah. | 20:25 |
| lastebil | hmm, my cubietruck I just compiled it in and hit a key... same, actually, for the c64 mini hack I did on Tuesday... | 20:25 |
| lastebil | tho that was over usb. hmm. | 20:25 |
| mntmn | i have a fantasy where i would put a linuxboot thing on eMMC and it would show an amiga-like bootscreen waiting for you to insert the boot SD card or choose to boot from SSD... | 20:25 |
| lastebil | does it have FEL mode at all? | 20:26 |
| lastebil | yes that would be wicked (: | 20:26 |
| lastebil | alternately, it should be Lord Nikon's bootup from Hackers (: | 20:26 |
| mntmn | haha :D | 20:26 |
| mntmn | i quickly googled FEL, so you mean boot via USB yeah? it is possible yes, but i don't remember the dip switch incantation atm | 20:27 |
| lastebil | yeah, FEL mode comes on allwinner. | 20:27 |
| mntmn | you can also compile an environment into uboot to do tftp boot... | 20:27 |
| lastebil | so for example for hacking the c64 mini, you pull out the red light cable, plug it into the exposed FEL thing on the card, and send data across a USB cable connected to the power thing. It's quite odd, but apparently how allwinner expected people to develop (: | 20:28 |
| lastebil | yeah I may do tftp instead. that may be easier. it may also NOT be because of my network setup (: | 20:28 |
| mntmn | fascinating https://github.com/lentinj/u-boot/blob/master/doc/README.NetConsole | 20:28 |
| mntmn | i have never considered that. | 20:28 |
| mntmn | another thing that could be done is use the keyboard's OLED and UART->LPC->UART->IMX | 20:30 |
| mntmn | the LPC has a UART connection to the IMX and another UART connection to the keyboard, but i haven't implemented the LPC<-UART->IMX communication yet. | 20:30 |
| mntmn | but when you press i.e. HYPER-V or HYPER-S, that info comes via LPC<-UART->Keyboard | 20:31 |
| lastebil | oh yes. the oled would work. | 20:43 |
| lastebil | holy crap that netconsole would work fantastic, I wonder if I can get it to run (: | 20:44 |
| lastebil | probably not today tho. I shall instead just... do modifications to the filesystem and test that way (: | 20:46 |
| lastebil | mntmn: hmm this u-boot setup is significantly different from the ones I've dealt with before (either that I've built, or that I've hacked...) - I don't see a script that has been compiled that it is running, and the only indication I have of what it would actually use for the kernel parameters is a mention of "run distro_bootcmd" | 20:56 |
| mntmn | lastebil: one second | 20:57 |
| lastebil | no problem (: | 20:57 |
| mntmn | lastebil: all the magic is here https://source.mntmn.com/MNT/reform-boundary-uboot/src/commit/7d4327e9938f6c3f3d016bbf0a89296a0536053c/board/boundary/nitrogen8m_som/nitrogen8m_som.c#L334 | 20:58 |
| mntmn | (and in that file in general) | 20:58 |
| mntmn | this is just a hacky stopgap until mainline u-boot | 20:58 |
| mntmn | also the indentation in that file is borked, my bad | 20:58 |
| mntmn | and it is like this because i hacked in the ability to boot from SD card (as opposed to emmc-only), but i wrecked the ability to save the environment on the way | 21:00 |
| lastebil | no problem, it just seemed a bit... different, this makes sense tho (: | 21:03 |
| lastebil | now to review the bootargs part for using syslinux type scripts (: | 21:04 |
| lastebil | hmm, do I even need to do this; as there is no initrd, if I were to just throw a different init at it (rather than the debian init, a void setup with runit as init) THEORETICALLY it just would work? | 21:11 |
| mntmn | yeah, that's what i do to boot from SSD | 21:16 |
| mntmn | lastebil: this is my /sbin/init http://dump.mntmn.com/reform-init.txt | 21:16 |
| mntmn | like, on my personal system | 21:16 |
| lastebil | danke, that... will definately help. I'll give it a whirl at least once before bed. | 21:20 |
| mntmn | lastebil: one advantage is that this immediately gives you the drivers for the display and usb keyboard etc | 21:22 |
| lastebil | yeah. just thinking how to get them from the build without actually doing the debian parts (: | 21:29 |
| lastebil | s/looking/reading the scripts/ (: | 21:29 |
| mntmn | ah those modules are not from debian | 21:39 |
| mntmn | they come from the kernel build | 21:40 |
| lastebil | yes. (: Was looking for where you copied them over, so I'd just... use that script (: | 21:54 |
| lastebil | basically, my thought: Build kernel, build the portions of the userland that are not debian; take void tarball, overlay those portions; boot. | 21:55 |
| mntmn | gotcha, but i would rather build on void i think, because you might run into library dependency problems | 21:57 |
| mntmn | i mean if you take userland stuff built on debian straight over to void | 21:57 |
| mntmn | lastebil: my suggestion: don't do mkuserland2.sh as part of your build, but copy that script to your (running) void system and build it on device | 22:00 |
| mntmn | lastebil: (or do similar crosscompile chroot stuff) | 22:00 |
| mntmn | kernel modules are built here together with the kernel Image https://source.mntmn.com/MNT/reform-system-image/src/branch/reform2-redo/reform2-imx8mq/mkkernel.sh#L41 | 22:01 |
| lastebil | ah very much the case, what with openssl being ressl and similar | 22:01 |
| lastebil | also I might want to run musl, and I could do that this way too | 22:01 |
| mntmn | yeah and different versions of everything | 22:01 |
| mntmn | kernel modules are copied into the rootfs here https://source.mntmn.com/MNT/reform-system-image/src/branch/reform2-redo/reform2-imx8mq/mkuserland.sh#L43 | 22:02 |
| mntmn | (it's only 3 modules) | 22:02 |
| mntmn | lastebil: and this is the script that loads those modules https://source.mntmn.com/MNT/reform-system-image/src/branch/reform2-redo/reform2-imx8mq/template-etc/reform-toggle-display.sh | 22:03 |
| lastebil | ok. I'll go to sleep and deal with this tommorrow (: danke | 22:03 |
| mntmn | good night! | 22:04 |
| lastebil | (let's hope the doctor doesn't tell me to not touch a computer for another week tommorrow (: ) | 22:04 |
| mntmn | oh noes :3 | 22:04 |
| ehmry | lastebil: I would recommend trying void with the stock kernel | 22:05 |
| ehmry | and replace it when you have a void system | 22:06 |
| ehmry | unless you are cross-compiling... | 22:06 |
| mntmn | does the stock kernel boot? | 22:06 |
| ehmry | I mean the stock MNT kernel from the SD card | 22:06 |
| mntmn | ah! :D | 22:06 |
| mntmn | haha yes | 22:07 |
| lastebil | yes that was the plan - but as mntmn pointed out I'll also need the modules for the display | 22:13 |
| lastebil | so: void rootfs tarball; kernel + modules; then build the rest on the laptop. | 22:13 |
| mntmn | yep, and some runit (?) script to load the modules probably, so you see something | 22:14 |
| lastebil | yes. which you handed me, basically (: | 22:14 |
| mntmn | yeah | 22:14 |
| lastebil | were I not tired I would just do that; I will sleep and then ... deal with things (: | 22:14 |
| mntmn | that should do the trick. i did similar with arch at some point, just to see if it would boot | 22:14 |
| mntmn | yeah! sleep is important | 22:15 |
| lastebil | yes. I got hung up thinking that I'd need to do some things with the bootloader (: | 22:15 |
| lastebil | talk tommorrow | 22:15 |
| mntmn | ehmry: i think a little guide "quickest way to run $DISTRO on reform" is in order | 22:18 |
| mntmn | ;) | 22:18 |
| ehmry | indeed | 22:28 |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@gateway/web/irccloud.com/x-brfzcedtmnhhyjgu) | 22:52 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!