| - mtm (QUIT: Ping timeout: 252 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 01:04 | |
| - eibachd (QUIT: Ping timeout: 246 seconds) (~eibachd@p200300dcf73d8b00531d08e4b519b9cf.dip0.t-ipconnect.de) | 02:09 | |
| + eibachd (~eibachd@p200300dcf73d8b00ee8d0f249a337332.dip0.t-ipconnect.de) | 02:09 | |
| - Boostisbetter (QUIT: Ping timeout: 255 seconds) (4a410829d7@irc.cheogram.com) | 02:22 | |
| mhoye | josch: fwiw I found the answer - sway has vt switching keysyms hardcoded in. I'd need to patch sway to make my idea work. | 02:31 |
|---|---|---|
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 03:09 | |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 03:40 | |
| - nsc (QUIT: Ping timeout: 264 seconds) (~nicolas@182-121-142-46.pool.kielnet.net) | 03:59 | |
| + nsc (~nicolas@21-98-142-46.pool.kielnet.net) | 04:00 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 05:21 | |
| - sevan (QUIT: Ping timeout: 256 seconds) (~sevan@user/venture37) | 06:07 | |
| josch | mhoye: oh i didn't know -- do you have a link? | 07:19 |
| violet | sorry for asking this again but i forgot what the answer was last time- is there a way to use a311d while staying on debian stable (bookworm), or does it need unstable? | 08:05 |
| + jacobk (~quassel@067-048-061-235.biz.spectrum.com) | 08:06 | |
| josch | violet: a311d works with stable if you use linux 6.5 from backports: https://reform.debian.net/images/ | 08:12 |
| violet | backports!! :D | 08:15 |
| violet | anything special i need to do other than installing that and switching out the hardware? | 08:15 |
| josch | violet: switching out the hardware? | 08:16 |
| violet | taking out the iMX compute module, putting the a311d in | 08:18 |
| josch | violet: you want to migrate your existing system or you want to install fresh? | 08:19 |
| violet | migrate | 08:19 |
| josch | that has been done a few times but no complete write-up has yet been produced | 08:20 |
| violet | *nod* | 08:20 |
| josch | with a fully upgraded sysimage-v4 it is possible to migrate | 08:20 |
| josch | i think vkoskiv has done it? | 08:20 |
| violet | i know i need to move the boot partition since right now it is on eMMC. so i need to move that to SD temporarily | 08:20 |
| violet | but i wonder is there anything in /etc i need to change? | 08:20 |
| violet | relating to device trees maybe or something like that | 08:21 |
| josch | you need to update the contents of /boot | 08:21 |
| josch | because the last time flash-kernel generated your dtbs it still thought you are on imx8mq | 08:21 |
| josch | also you might need to adjust /etc entries for emmc and sd-card because the /dev names are swapped | 08:21 |
| josch | and you cannot (should not) use the emmc of a311d | 08:22 |
| violet | oh? why not | 08:22 |
| josch | a311d will always look on emmc first and if what it finds there is not bootable, then you soft-bricked your device | 08:23 |
| violet | ah. so if i use eMMC i am living very dangerously | 08:23 |
| josch | yes, the reform tools currently forbid writing to it | 08:24 |
| josch | to unbrick you need something like this, for example: https://www.tindie.com/products/superna9999/meson-hdmi-recovery-boot-plug/ | 08:24 |
| violet | interesting. maybe i will try to find one of those, and stick to iMX until then | 08:25 |
| violet | i want to keep my SD slot free because i use it a lot | 08:25 |
| josch | yes, theoretically you could manually umount /boot after the system is booted | 08:26 |
| josch | and then just never forget to re-mount it before upgrading the kernel | 08:26 |
| violet | yeah haha. plus never lose the boot SD when i take it out to put something else in | 08:26 |
| violet | i think i dont trust myself to keep all those in my head :p | 08:27 |
| violet | ill add myself to the waitlist for this tindie and maybe they restock | 08:28 |
| violet | in the mean time, i am safe to install 6.5 from backports to use on iMX right? | 08:30 |
| violet | i have a weird rendering bug in chrome and i dont know if a kernel update will fix it but i might as well try and find out | 08:31 |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 09:15 | |
| - Boostisbetter (QUIT: Ping timeout: 276 seconds) (4a410829d7@irc.cheogram.com) | 09:20 | |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 09:24 | |
| josch | violet: yes, 6.5 from backports should be safe on imx8mq, a311d and ls1028a | 09:43 |
| f_ | violet: yes A311D bootROM looks for eMMC first. If e.g. the u-boot binary installed is b0rked..your reform is officially bricked | 09:50 |
| f_ | But apart from that there's not much risk | 09:50 |
| f_ | I think it's a lot like e.g. your SPI flash on your x86 laptop..if what's installed on it is b0rked, your laptop is bricked | 09:51 |
| f_ | I wish amlogic socs looked on SD first, depending on POC | 09:52 |
| josch | right, i think a big difference is that writing to SPI flash in an x86 laptop is a lot harder than just running dd on /dev/mmcblk1 accidentally. That's what makes it more scary, I think. | 09:55 |
| f_ | writing to SPI is easy depending on the laptop | 09:56 |
| f_ | flashrom -p internal -w whatever.rom | 09:56 |
| josch | oh i had no idea :) | 09:56 |
| f_ | ^ you can run this on a MacBook2,1 and it'll flash to its SPI | 09:56 |
| f_ | But on many other computers the SPI flash is read-only.. | 09:57 |
| - klardotsh (QUIT: Quit: zzz) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 09:58 | |
| f_ | requiring you to open it and use a clip connected to an SPI programmer (such as a raspberrypi) to flash | 09:58 |
| f_ | But yes, running bl2+bl31+SCP_BL2+u-boot from eMMC is risky. | 10:00 |
| f_ | Which is why I like those little removable eMMC modules on librecomputer boards :) | 10:01 |
| - eibachd (QUIT: Ping timeout: 252 seconds) (~eibachd@p200300dcf73d8b00ee8d0f249a337332.dip0.t-ipconnect.de) | 10:20 | |
| + eibachd (~eibachd@2a01:599:310:7433:51c4:a954:9365:87c6) | 10:21 | |
| - eibachd (QUIT: Ping timeout: 255 seconds) (~eibachd@2a01:599:310:7433:51c4:a954:9365:87c6) | 10:43 | |
| + eibachd (~eibachd@2a01:599:303:43bd:bb61:a05:cbf5:4547) | 10:51 | |
| - eibachd (QUIT: Read error: Connection reset by peer) (~eibachd@2a01:599:303:43bd:bb61:a05:cbf5:4547) | 11:04 | |
| + eibachd (~eibachd@p200300dcf73d8b000bfc28dd28ceb794.dip0.t-ipconnect.de) | 11:04 | |
| - eibachd (QUIT: Ping timeout: 268 seconds) (~eibachd@p200300dcf73d8b000bfc28dd28ceb794.dip0.t-ipconnect.de) | 11:09 | |
| + eibachd (~eibachd@2a01:599:303:43bd:bb61:a05:cbf5:4547) | 11:10 | |
| - eibachd (QUIT: Read error: Connection reset by peer) (~eibachd@2a01:599:303:43bd:bb61:a05:cbf5:4547) | 11:15 | |
| + eibachd (~eibachd@p200300dcf73d8b00f1f0db46209c2949.dip0.t-ipconnect.de) | 11:15 | |
| hramrach | violet: you can make your own HDMI dongle follwing the 'documentation' link on the tindie page. As far as I know it is not tested to work on the Reform. It is not documented what *exactly* the ROM does, it may need to be connected to a specific i2c bus, muxed on specific function. | 11:23 |
| hramrach | josch: seems gite is doing server-side rendering (So far) so it won't have this 'waiting for javascript' problem. However, gitlab is special in that it has separate numbering for issues and pull requests which makes migrating from it challenging, would not recommend. | 11:26 |
| hramrach | Also the server-side rendering is problematic with some features for bigger projects (eg Linux size) so I suspect it will need to go at some point, at least for the analytics pages. | 11:28 |
| josch | hramrach: "gite"? Do you mean gitea? | 11:29 |
| hramrach | yes, gitea | 11:30 |
| josch | i have a gitea instance running for my own personal projects and am quite happy with it | 11:32 |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 11:36 | |
| hramrach | yes, it would be ok if it was used from the start. migrating to it from gitlab is a different thing | 11:37 |
| f_ | josch: gitea? I use cgit+gitolite :) | 11:38 |
| f_ | I tell people to send patches to my personal email address. | 11:38 |
| f_ | hramrach: As for what the ROM exactly does, at some point I played around with decompiling BL1 | 11:39 |
| f_ | Succeeded at dumping the bootROM and could look at it in ghidra. | 11:39 |
| f_ | S905 bootROM (dumped from a KII Pro box): https://git.vitali64.duckdns.org/misc/reversing-gxbb-bl2.git/tree/bl1/kii-pro/bl1.bin | 11:40 |
| f_ | S905X bootROM (dumped from a lepotato): https://git.vitali64.duckdns.org/misc/reversing-gxbb-bl2.git/tree/bl1/lepotato/bl1.bin | 11:40 |
| eibachd | 4,99hramrach99,99: Is there something wrong with Gitlab? | 11:42 |
| f_ | eibachd: having issues with your IRC client? | 11:42 |
| eibachd | 13,99f_99,99: why? | 11:43 |
| f_ | <eibachd> 9f_9: why? | 11:43 |
| f_ | ^ how your pings show up in my IRC client | 11:43 |
| f_ | (and it doesn't ping me) | 11:44 |
| f_ | That's why I was wondering if you were having issues with your IRC client. | 11:44 |
| eibachd | Strange, in my client out looks pretty normal. I have no idea how to fix that. | 11:44 |
| f_ | (or it might be just me?) | 11:44 |
| f_ | https://mntre.com/reform-irc-logs/2023-11-25.log.html looks like it's an issue on your side. | 11:45 |
| sknebel | eibachd's client is using color codes for the mentioned nick for whatever reason | 11:46 |
| f_ | But colour codes work fine in my IRC client. | 11:46 |
| sknebel | hm. maybe some variant that gets mangled somehow? | 11:47 |
| eibachd | Hmm, I just used the copy function of the client. Maybe that messes up colors. | 11:47 |
| sknebel | or something that gets mangled to something that my client thinks is a color code | 11:47 |
| f_ | hramrach: now thinking about it, could dump A311D bootROM too :) | 11:49 |
| f_ | (and S905D3 bootROM) | 11:49 |
| - pandora (QUIT: Quit: Connection closed for inactivity) (uid585533@id-585533.ilkley.irccloud.com) | 11:52 | |
| eibachd | hramrach: Anything wrong with Gitlab? (I hope this time nothing is needed up on my message) | 11:54 |
| f_ | At least one thing wrong with GitLab, it b0rks in tor browser. | 11:55 |
| f_ | (even with JS toggled on with NoScript) | 11:56 |
| eibachd | I see | 11:56 |
| f_ | Oh wait a second? | 11:56 |
| f_ | Seems to work better than before? | 11:57 |
| f_ | But it still makes my laptop fan go very fast. | 11:57 |
| f_ | as soon as I enable JS on gitlab | 11:58 |
| eibachd | Wild | 11:58 |
| f_ | https://laboratory.vitali64.duckdns.org/source.mnt.re/reform/reform/ | 11:59 |
| f_ | luckily it works on that flask-powered gitlab frontend I didn't touch in a while. (still uses old CSS!) | 11:59 |
| eibachd | The wonders of this new fangled JS stuff. If everything works, great. If not it is way above my head. | 12:01 |
| f_ | GitLab's API is quite limited unless you authenticate though :/ | 12:02 |
| f_ | So no issue comments, MRs... | 12:02 |
| eibachd | And you dislike authenticating. | 12:04 |
| f_ | Not that, but requiring people to authenticate in order to view MRs in the frontend, could be annoying for many. | 12:04 |
| f_ | And let's not forget that it was meant to be a "private" frontend | 12:05 |
| eibachd | Understood | 12:07 |
| f_ | I do have a GitLab.com account though, mainly for postmarketOS. | 12:09 |
| eibachd | I am simply happy that I can selfhost it without much hassle. And build some nice CI pipelines. | 12:09 |
| f_ | I can also self-host cgit+gitolite without much hassle. | 12:10 |
| eibachd | Understood. CI and pull request handling? | 12:13 |
| hramrach | eibachd: gitlab relis heavily on JavaScript. It sometimes breaks even on powerful client machines because it caches something incorrectly, and it's reportedly slow on the Reform (hav not tried). | 12:14 |
| eibachd | Should be trivial to offer both frontends, right? | 12:16 |
| f_ | eibachd: send PRs via email? | 12:16 |
| hramrach | eibachd: no, there are PRs/issues which should be filed in one place | 12:17 |
| eibachd | Hey, we are not lkml here :) | 12:18 |
| hramrach | The PRs via e-mail are difficult to hook with CI | 12:20 |
| hramrach | als FTR my IRC client has no problem with those color codes. However, IRC protocol (at least the version commonly in use) does not define the content of messages, and properties of messages like encoding and color codes are undefined and subject to arbitrary interpretation. | 12:25 |
| eibachd | Seriously, I have worked long enough with these guys and I know my git send-email. But I really enjoy the simplicity and integration of Gitlab(as long as it works). Maybe I am getting old. | 12:27 |
| hramrach | My takeaway from using gitlab for a while it that it does not really work (reliably enough). | 12:28 |
| hramrach | That said, integration around e-mail is hard, and those forges are in general useful. | 12:29 |
| f_ | integration around email can be easier. | 12:31 |
| f_ | I have no issues with git send-email | 12:31 |
| f_ | >The PRs via e-mail are difficult to hook with CI | 12:31 |
| f_ | Actually not that hard, and builds.sr.ht exists, too. | 12:31 |
| hramrach | that's one side, what would the receiving side integration do? | 12:31 |
| f_ | But I don't use CI anyway. | 12:32 |
| f_ | I used it exactly once, for https://vitali64.gitlab.io/..back when I used gitlab | 12:32 |
| hramrach | The thing is CI has its value if your project happens to recieve random contributions. | 12:34 |
| hramrach | and can be also abused for generating releases | 12:34 |
| f_ | Yeah CI can be very useful. | 12:34 |
| f_ | I just didn't find any use for it, in my infra. | 12:34 |
| hramrach | Sure, it's not always needed. In the MNT infra it is used. | 12:35 |
| f_ | I see. | 12:36 |
| f_ | in the MNT infra Gitea/Forgejo might be better. | 12:36 |
| hramrach | That's my thinking as well, generally. It would be better if it was used from the start, migrating to it is difficult from gitlab. | 12:37 |
| + mjw (~mjw@gnu.wildebeest.org) | 12:40 | |
| eibachd | My reception is, that if you want to lower the bar for onboarding developers these days, it helps a lot to have some GitHub/Gitlab like workflow. But I am not sure whether that is relevant for MNT. | 12:42 |
| minute | we were on gitea before. the merge request and code review functionality was extremely lacking back then, so we migrated to gitlab | 12:43 |
| f_ | And then people ask to migrate back to gitea/forgejo | 12:44 |
| f_ | I think forgejo/gitea matured a lot. | 12:45 |
| hramrach | and maybe gitlab degenerated as well, things change | 12:48 |
| hramrach | Also people started to use the forge on the lower performance Arm devices | 12:48 |
| minute | last night i did that merge and tag on my pocket reform with imx8mp, no issues | 12:50 |
| minute | (in chromium though, unfortunately firefox is slow on imx) | 12:50 |
| hramrach | For me looking at gitea was triggered by gtlab making mutilpe reviews a premium-only feature | 12:51 |
| minute | on a311d i use firefox and it works fine | 12:51 |
| hramrach | I think josh was complaining about the speed of gitlab on Reform | 12:52 |
| f_ | speed of gitlab on hp 8560w :^) | 12:56 |
| f_ | (quite ok, but slows down my entire computer) | 12:56 |
| minute | hramrach: josch is using imx8mq still | 12:56 |
| f_ | A311D is quite powerful for sure. | 12:57 |
| f_ | S905X too. | 12:57 |
| minute | i think most people that don't switch to a311s don't do it because there's no solution for dual display yet | 12:57 |
| f_ | *A311D | 12:58 |
| minute | yes, typo | 12:58 |
| f_ | Or maybe because Amlogic's "trusted"firmware is proprietary. | 12:58 |
| minute | tested rpi cm4s for shipping yesterday and noticed how neat it is to have dual independent displays | 12:58 |
| minute | f_: sure, that's an issue, but not such a bit practical issue | 12:59 |
| minute | big | 12:59 |
| minute | this phone's keyboard is terrible lol (it's an iphone 13 mini) | 12:59 |
| f_ | You could still use amlogic's BSP and get all features, but unless you like sticking on 4.x kernels I can't imagine any reason why you would want to do such a thing. | 12:59 |
| minute | f_: nope. i draw the line at shipping mainline (plus patches) linux only :D | 13:00 |
| f_ | Great. | 13:00 |
| f_ | Amlogic BSP is quite ugly. I only have exactly one Amlogic board using amlogic bsp. | 13:00 |
| f_ | The rest run either mainline (KII Pro, lepotato..) or nothing (lafrite, alta, solitude). | 13:01 |
| minute | at some point i might need to dig in and check how dual display is done, but maybe narmstrong will figure it out before | 13:01 |
| minute | currently i need the weekends to cool off my brain from stress hormones building up during the week while doing the pocket reform crunch ^^ | 13:02 |
| f_ | I'm insane, so I work on amlogic RE in my free time, including weekends. ^^ | 13:03 |
| f_ | As for Pocket Reform, am sure you'll be successful. | 13:03 |
| - mtm (QUIT: Ping timeout: 255 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 13:03 | |
| - Boostisbetter (QUIT: Ping timeout: 255 seconds) (4a410829d7@irc.cheogram.com) | 13:20 | |
| - S0rin (QUIT: Ping timeout: 252 seconds) (~S0rin@user/s0rin) | 13:54 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 14:00 | |
| + S0rin (~S0rin@user/s0rin) | 14:02 | |
| mhoye | josch: https://github.com/swaywm/sway/blob/master/sway/input/keyboard.c#L273 | 14:02 |
| - S0rin (QUIT: Ping timeout: 264 seconds) (~S0rin@user/s0rin) | 14:07 | |
| + S0rin (~S0rin@user/s0rin) | 14:09 | |
| hramrach | it's hadcoded to XKB_KEY_XF86Switch_VT_* so these would have to be in the keymap or am I missing something? | 14:23 |
| hramrach | hm, sourcehat is interesting in that they are not makinf a forge 'product' but rather a sort of a forge ecosystem with many parts that are optional and can be swapped out for parts of alreasy existing infra | 14:34 |
| * jn_ -> jn | 14:34 | |
| hramrach | the downside is that doing things this way is more difficult, and they are in 'open alpha' stage for quite a while. I already saw them once and forgot that thy exist again :) | 14:35 |
| hramrach | An interesting thing is that they have correct message threading because they have e-mail integration. It's not that correct threading cannot be done otherwise but only people doing first class e-mail support seem to care today. | 14:39 |
| - eibachd (QUIT: Ping timeout: 252 seconds) (~eibachd@p200300dcf73d8b00f1f0db46209c2949.dip0.t-ipconnect.de) | 14:50 | |
| + eibachd (~eibachd@p200300dcf73d8b00a470289b5adf1e1d.dip0.t-ipconnect.de) | 14:51 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 15:10 | |
| mhoye | hramrach: what I have is, meta-arrowkey works in tty1-X, and moves you from ttyX to ttyX+1 withone keystroke, which I like. | 16:23 |
| mhoye | what I want is for that to keep working if I open sway from one of those ttys but it looks like I can't have that. | 16:23 |
| mhoye | Because Sway afaict catches everything but is not intrisincally aware of where it comes from, and so has this special bit of code in it to let the traditional control-alt-f# break you out if that's your thing. | 16:25 |
| mhoye | So I think that I will be thwarted in this, and have to escape Sway by using three keys instead of just two, oh the indignity | 16:27 |
| josch | hramrach: yes, gitlab is slow for me because imx8mq is slow. :) I could switch to a311d but i'm hesitant because i also don't want to buy one new SoM for my reform every two years to reduce electronic waste (as that was one of my main motivations for the reform). I do need hdmi output but i do not need both displays at the same time, so in that way, the a311d would probably work for me. | 16:38 |
| josch | But there are only very, very few things i feel that the imx8mq is bad at given my workload so I'll stick around for a while longer. | 16:38 |
| josch | Instead I'll build a second reform-like PC with a real pcie graphics card based on the ls1028a once i can purchase a second motherboard. :) | 16:39 |
| hramrach | the ls1028a should help with gilab performance, the external graphics card probably not so much :) | 17:06 |
| minute | josch: i am thinking we should do a limited buy-back program for imx8mq | 17:06 |
| minute | would be better than to order small quantities from boundary and waiting 24 weeks | 17:07 |
| hramrach | if there is still use for them | 17:08 |
| minute | hramrach: you have ls1028a, or do i misremember? | 17:08 |
| minute | hramrach: yeah, we have a few open orders with imx8mq and mouser has ordered 10 units of reforms with imx8mq | 17:08 |
| hramrach | yes, I have a ls1028a but did not try using it much so far | 17:08 |
| minute | and they do not respond to our offer of upgrading to a311d and other things | 17:08 |
| minute | we have offered our upgrades and peripherals to cs/mouser a few times but they've ghosted me on this | 17:09 |
| hramrach | my thinking is that imx8mq was basically oboslete from the beginning, the gitlab performance is to be expected | 17:10 |
| minute | it was mostly blob free and well documented and supported by mainline, and the module had open schematics | 17:11 |
| minute | that is why we used it | 17:11 |
| hramrach | Apparently there are people who want imx8mq for whatever reason 🤷 | 17:12 |
| minute | remember this was introduced by us 4.5 years ago https://mntre.com/media/reform_md/2019-05-20-reintroducing-reform.html | 17:12 |
| hramrach | Olimex went the route of adding support for Alwinner SoCs, and at the time it worked out | 17:12 |
| hramrach | at this proint the problem is that Allwinner failed to deliver a desktop-grade SoC but that's something that could not be foreseen at the time | 17:13 |
| minute | our strategy from the start was to offer upgrade socs | 17:14 |
| minute | and we delivered on this | 17:14 |
| hramrach | yes, their designs are monolithic, no upgrades there | 17:15 |
| - XYZ_ (QUIT: Remote host closed the connection) (~XYZ@37-48-32-137.nat.epc.tmcz.cz) | 17:16 | |
| hramrach | Also the Alwinner documentation was really bad, and most of the blanks have been filled in by to community by now but again it was not clear it would happen | 17:17 |
| hramrach | Olimex probably contributed some parts but there is a lot that was done by other people | 17:17 |
| + XYZ (~XYZ@37-48-32-137.nat.epc.tmcz.cz) | 17:24 | |
| hramrach | Also the problems with different modules having different interfaces and the mainboard behaving differently because of that is something that future open hardware projects could learn from | 17:30 |
| hramrach | Looks like the modules should be bigger for this to work well | 17:31 |
| minute | i think we managed quite well adapting a system to such diverse socs/soms, i don't know of anything comparable | 17:34 |
| minute | that the a311d doesn't do independent dual display is (afaik) a software issue | 17:34 |
| minute | we could have also put a pcie switch on rcm4 to work around one dead pcie slot, but costs would balloon a bit... still, because rcm4 is oshw anyone experienced in kicad can do that | 17:35 |
| minute | the real estate for the modules is also big | 17:35 |
| minute | and anyone could make alternative motherboards, there's a lot of space in the case | 17:35 |
| minute | so i don't really see why there should be another project doing this differently or sth, instead of doing a replacement part or mod | 17:36 |
| minute | the reform case is big enough to put an amd ryzen embedded in there or whatever people wanted... it's more like, the reality is that there are very few people doing oshw at all | 17:37 |
| minute | an answer could be, well, the reform case is not to the taste of everyone etc etc. next year we'll launch a more compact and more streamlined looking case. if people are still unhappy about that, well, then so be it. | 17:39 |
| f_ | minute: I like that thick case. | 17:40 |
| f_ | ACTION 's laptop is very thick. | 17:40 |
| minute | yeah, it was/is a statement also | 17:41 |
| f_ | I really feel like laptops today are getting thiner, with less and less user servicability too. | 17:41 |
| minute | that was our perspective when we started | 17:42 |
| minute | the new case will be a middle ground. total thickness 2.6cm | 17:42 |
| f_ | Is it still the case? | 17:42 |
| minute | f_: yes, i think we will never make an ultrabook | 17:42 |
| f_ | Well..then the Framework laptop proved some wrong. It was more repairable than other thin laptops. | 17:43 |
| minute | but also we need to offer sth that is a bit thinner. the reform next will be that | 17:43 |
| f_ | Call it Little Reform | 17:43 |
| minute | the framework is not reproducible independently, the manufacturing methods are out of reach for makers | 17:44 |
| minute | so, uninteresting for us | 17:44 |
| f_ | It's also using an intel/amd cpu and proprietary BIOS :/ | 17:44 |
| minute | reform has been cloned independently (by jacqueline) | 17:45 |
| minute | and reform next might be even easier to clone because less parts | 17:45 |
| f_ | Glad to hear that! | 17:45 |
| minute | pocket will also be relatively easy to clone by others | 17:46 |
| minute | we've tried to make it easier and quicker to assemble | 17:46 |
| minute | and no exotic parts | 17:46 |
| minute | i think this is something that most reviewers covering the device etc do not understand at all | 17:46 |
| minute | you will be able to download the parts and make the device yourself if you want | 17:47 |
| minute | haha little saturday rant. i should get going | 17:48 |
| f_ | "You wouldn't download a laptop???" | 17:48 |
| f_ | But yeah nice. | 17:49 |
| minute | f_: https://mastodon.social/@mntmn/110940962003437047 | 17:52 |
| f_ | Oh yeah that was a reference to ^ | 17:54 |
| f_ | But what's "MNT Remix Reform"? | 17:55 |
| Boostisbetter | Isn't that the thinner Reform that is being worked on. | 17:57 |
| minute | no, this is just a logo that you can use if you forked our hardware | 18:00 |
| minute | instead of the mnt logo | 18:00 |
| f_ | Oh ok | 18:00 |
| f_ | Is the MNT logo trademarked? | 18:00 |
| minute | https://source.mnt.re/reform/reform/-/blob/master/symbol-for-derived-works/mnt-based-reform.svg | 18:01 |
| minute | f_: yes | 18:01 |
| f_ | ok | 18:01 |
| minute | f_: if someone else put our logo on electronics that we didn't make, we could get in trouble | 18:01 |
| f_ | Right. | 18:01 |
| minute | because the manufacturer is responsible for recycling, emissions etc etc | 18:01 |
| minute | at least in germany ^^ | 18:02 |
| minute | and is identified by the logo | 18:02 |
| minute | the remix logo is a workaround for this | 18:03 |
| minute | it means, this is compatible/derived, but not manufactured by us | 18:03 |
| minute | Boostisbetter: the thinner reform is codenamed Reform Next | 18:05 |
| jn | oh, that's nice in case i ever get around to my silly project idea of designing my own mnt cpu card and need a logo | 18:06 |
| minute | jn: yes! | 18:06 |
| minute | i think rcm4 is a good starting point for projects like that btw | 18:06 |
| minute | it shows how to adapt a different soc/som | 18:06 |
| jn | (silly because it'll be based on rp2040 and ice40) | 18:06 |
| minute | also if you can output 1080p hdmi (or fake that by upscaling), you can use our hdmi to edp adapter pcb | 18:07 |
| f_ | Run linux on an ESP31. | 18:07 |
| f_ | *ESP32 | 18:07 |
| minute | it would be nice to have ecp5, artix7 or zynq, for cheaper fpga retro emulation (cheaper than kintex-7 which exists) | 18:08 |
| jn | minute: does the adapter and/or panel require a native-resolution signal? | 18:08 |
| jn | oh yeah, ecp5 would be really nice. nextpnr-ecp5 is in debian since recently | 18:08 |
| minute | jn: yes, native resolution and timing. not a problem with an fpga i think, i've done this on spartan6 even. and you could pixel double your input resolution so you don't need a full res buffer | 18:12 |
| f_ | https://mastodon.social/@jacqueline@chaos.social/111186428642185542 | 18:14 |
| f_ | That reform is powered by mana | 18:15 |
| jn | ACTION takes notes | 18:15 |
| josch | minute: uff that's sad regarding how mouser is treating you :( | 18:25 |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@067-048-061-235.biz.spectrum.com) | 18:26 | |
| f_ | oh hi jn | 18:29 |
| josch | minute: uff that's sad regarding how mouser is treating you :( | 18:30 |
| josch | minute: if you are buying back imx8mq boards so that they get used again by others i'd feel much more at ease with buying a new SoM from you | 18:30 |
| - eibachd (QUIT: Ping timeout: 260 seconds) (~eibachd@p200300dcf73d8b00a470289b5adf1e1d.dip0.t-ipconnect.de) | 18:38 | |
| + eibachd (~eibachd@p200300dcf73d8b00537e4b194d65f380.dip0.t-ipconnect.de) | 18:38 | |
| minute | josch: ok, the we'll do that, noting it for monday. it would be cool also to know that you have a bit more performance in your system ^^ | 18:47 |
| hramrach | there is clockwork pi that is adapted to a numebr of SoMs and offers a few form factors but not very familiar with the hardware | 20:13 |
| + sevan (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 20:14 | |
| hramrach | anyway, I think that the different SoCs come with different interfaces - in particular the display/PCI/SATA varies so it would be better if it was on the SoM rather than on the mainboard | 20:14 |
| hramrach | with the big reform splitting tha mainboard so that one part has the IO on one side panel and the other part is SoC-specific with SOC-specific side panel would be viable but it would be difficult to adapt to different form factor | 20:16 |
| minute | hramrach: which processor would you like to use / what is the limiting factor for you right now? | 20:18 |
| hramrach | there isn't limiting factor as such. The problem I see is that the mainboard was designed for one SoM, and the interfaces do not match the other SoMs perfectly. There are some generic interfaces like USB, SPI, i2c, i2s that are almost always available, and some intefaces like PCIe, SATA, HDMI, DSI, DP that vary from SoC to SoC. It would be nice if the slots/connectors matched the interfaces of | 20:23 |
| hramrach | the installed SoC. | 20:24 |
| minute | hramrach: where do you see the biggest mismatch at the moment, i.e. what would be improved? | 20:24 |
| hramrach | That would make it possible to make best use of what the SoC provides and also reduce confusien around what slot/connector does what exactly | 20:24 |
| hramrach | there is one thing that could be improved regardles of SoC: route some USB to the onboard mPCIe/M.2 | 20:26 |
| minute | hramrach: the pocket reform motherboard does this | 20:26 |
| hramrach | the other thing is that the M.2 could be keyed to the interfaces it actually provides which cannot be done when sitting on the mainbpard and not the SoM | 20:26 |
| minute | ok, so your issues are mainly with M.2 functionality and keying | 20:27 |
| hramrach | using the HDMI for external PCIe is also interesting but I can see why you would not want to go for Thunderbolt, and not going there HDMI is one of the more useful phisical connectors for this | 20:28 |
| minute | yeah the other option would have been to output nothing on that connector | 20:29 |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@37-48-32-137.nat.epc.tmcz.cz) | 20:29 | |
| minute | because ls1028a doesnt have hdmi | 20:29 |
| minute | ok gotta run, ttyl | 20:29 |
| hramrach | yes, and that's exactly the thing - the connectors are on the mainboard, and don't match completely to what the SoC has, there is this disconnect | 20:30 |
| + XYZ (~XYZ@37-48-32-137.nat.epc.tmcz.cz) | 20:32 | |
| hramrach | with that said I think the ls1028a SoM works well with the limitations of the existing carrier bord, and at the time the mainboard was designed it would be difficult to do it differently | 20:35 |
| josch | i think when the reform motherboard was designed, since there was just one SoM available (imx) it made sense to design everything around it as it's hard to foresee what the future would bring or if the whole project would even become viable and successful in the first place. Clearly, the motherboard is only really a 100% fit for the imx8mq and a bit less of a fit for other platforms. How much less | 21:18 |
| josch | depends on the platform of course. When there is a new motherboard revision in the future i think there are things that can be learned but as i see it this is already happening with the pocket reform motherboard that does a couple of things differently. And in the past Lukas also pointed out how the Reform Next motherboard will likely be more a successor of the Pocket motherboard than the big Reform | 21:18 |
| josch | motherboard. | 21:18 |
| josch | </walloftext> | 21:18 |
| josch | for example, for the Reform Next there is the idea to have the boards holding the physical ports on the side of the laptop be replacable. Then users can choose which ports they want to have available either depending on their personal use-case or depending on the SoM they use. | 21:20 |
| josch | hramrach: the last bit is in the direction you were thinking about earlier, no? | 21:20 |
| - sevan (QUIT: Changing host) (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 22:00 | |
| + sevan (~sevan@user/venture37) | 22:00 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 22:30 | |
| hramrach | yes, that could be it. | 23:12 |
| eibachd | New blog post about my adventures trying to boot the MNT Reform2 with NXP LS1028A from eMMC. Involves fiddly hardware modifications, writing a disassembler for an exotic architecture and *gasp* lighting a LED. https://zeromips.org/posts/2023-11-24-emmcboot/ | 23:39 |
| mjw | wow, very nice | 23:55 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
