| minute | i will try bluerise's patch | 00:00 |
|---|---|---|
| cinap_lenrek | yeah | 00:00 |
| sigrid | cinap_lenrek: where does it allocate the fb in your patch? | 00:00 |
| cinap_lenrek | i might have filled the struct wrong | 00:00 |
| sigrid | ie to the top of memory | 00:00 |
| cinap_lenrek | sigrid: MAGIC! | 00:00 |
| sigrid | like there is memalign | 00:01 |
| cinap_lenrek | sigrid: with CONFIG_LCD, the fb is allocated by board_f.c | 00:01 |
| cinap_lenrek | very early | 00:01 |
| cinap_lenrek | it just sets gd->fb_base to CONFIG_FB_ADDR | 00:01 |
| sigrid | :( why memalign then? | 00:01 |
| cinap_lenrek | and i arrange CONFIG_SYS_MEM_TOP_HIDE to exclude the framebuffer | 00:01 |
| sigrid | anyway this api is weird af | 00:01 |
| cinap_lenrek | sigrid: lcdif checks if gd->fb_base is NULL | 00:02 |
| cinap_lenrek | if its NULL, it allocates | 00:02 |
| sigrid | ah ok | 00:02 |
| cinap_lenrek | if not, it uses it | 00:02 |
| sigrid | so without config_lcd it will still work. got it | 00:02 |
| cinap_lenrek | note, this is from looking at other drivers | 00:02 |
| cinap_lenrek | to see what they do | 00:02 |
| cinap_lenrek | they do the same | 00:02 |
| cinap_lenrek | so i ASSUMED this is a valid approach | 00:03 |
| cinap_lenrek | sigrid: correct. | 00:03 |
| cinap_lenrek | it is #ifdef mazed. | 00:03 |
| cinap_lenrek | its a bit stupid | 00:03 |
| cinap_lenrek | all i want is the CONFIG_FB_SIMPLE thing | 00:03 |
| cinap_lenrek | but that depends on CONFIG_LCD | 00:03 |
| cinap_lenrek | so i made that happy | 00:03 |
| minute | rebuilding. | 00:04 |
| cinap_lenrek | minute: get a usbsdmux :) | 00:04 |
| cinap_lenrek | https://www.linux-automation.com/en/products/usb-sd-mux.html | 00:06 |
| cinap_lenrek | phantastic device. | 00:06 |
| minute | well, i only have reform atm | 00:10 |
| minute | not true, also pocket | 00:11 |
| cinap_lenrek | you need *TWO* reforms! :) | 00:11 |
| cinap_lenrek | two reforms at the same time! | 00:11 |
| minute | well anyway, openbsd gives me a black screen now. maybe the pcie issue? | 00:11 |
| cinap_lenrek | anything on serial? | 00:11 |
| minute | no serial here on the sofa | 00:11 |
| minute | at home | 00:12 |
| minute | (yet) | 00:12 |
| cinap_lenrek | auf dem sofa zuhause, ja. hab ich alles da, ja | 00:12 |
| minute | bluerise: did you need to disable pcie in the dtb to see anything on screen? can you upload your working dtb? | 00:12 |
| minute | cinap_lenrek: lol | 00:12 |
| minute | das problem ist das system | 00:12 |
| cinap_lenrek | :D | 00:13 |
| cinap_lenrek | bluerise → + return fdt_setup_simplefb_node(blob, off, gd->fb_base, imx_lcdif.winSizeX, | 00:19 |
| cinap_lenrek | bluerise → + imx_lcdif.winSizeY, imx_lcdif.winSizeX * (1 << imx_lcdif.gdfBytesPP) / 8, | 00:19 |
| cinap_lenrek | this is wrong | 00:19 |
| sigrid | bluerise fixed it | 00:20 |
| bluerise | yes, I changed the diff | 00:20 |
| cinap_lenrek | sorry | 00:20 |
| cinap_lenrek | okay! | 00:20 |
| + robin (~robin@user/terpri) | 00:20 | |
| sigrid | see https://mastodon.social/@mntmn/109712632465117425 | 00:20 |
| cinap_lenrek | okay! | 00:20 |
| bluerise | and http://ix.io/4lCn | 00:20 |
| bluerise | name was also wrong | 00:20 |
| bluerise | (the xr8 whatever) | 00:20 |
| cinap_lenrek | fu king hell | 00:20 |
| cinap_lenrek | YES! | 00:20 |
| cinap_lenrek | openbsd aufm sofa! | 00:21 |
| minute | yess | 00:21 |
| bluerise | Told you so :P | 00:21 |
| minute | i also forgot to do > set tty fb0 | 00:21 |
| minute | in the openbsd boot prompt. just for the record | 00:21 |
| bluerise | I rebased and updated my u-boot mainline reform2 branch | 00:22 |
| minute | oh | 00:22 |
| bluerise | but pcie doesn't work, hmhm | 00:22 |
| cinap_lenrek | but usb works? | 00:22 |
| bluerise | scanning bus usb@38100000 for devices... 5 USB Device(s) found | 00:23 |
| bluerise | scanning bus usb@38200000 for devices... cannot reset port 1!? | 00:23 |
| bluerise | 2 USB Device(s) found | 00:23 |
| bluerise | scanning usb for storage devices... 1 Storage Device(s) found | 00:23 |
| bluerise | looks like | 00:23 |
| cinap_lenrek | i mean in obsd | 00:23 |
| cinap_lenrek | it looks like it | 00:24 |
| bluerise | in openbsd? usually, yeah, but my boot disk is on nvme, so without NVMe support I cannot boot from my disk :D | 00:24 |
| bluerise | also, you can see the uhubs | 00:24 |
| cinap_lenrek | yeah | 00:24 |
| bluerise | the trackpad and keyboard | 00:24 |
| cinap_lenrek | very nice | 00:24 |
| cinap_lenrek | bluerise: NFS!!! | 00:24 |
| bluerise | minute: accessing the 'dbi base' on the pcie controller hangs. either the power isn't on, or the clocks are wrong, or they have some MMU setting that stops me from using writel, heh | 00:25 |
| cinap_lenrek | isnt here a patch for uboot to make nvme boot work? | 00:26 |
| minute | bluerise: but it's not the refclk issue? | 00:28 |
| minute | bluerise: linux normally hangs if internal refclk is not configured for the first pcie | 00:29 |
| minute | btw i successfully booted to the OpenBSD installation program | 00:29 |
| bluerise | I have pcie0 disabled, pcie1 enabled | 00:29 |
| minute | oh, shell | 00:30 |
| minute | bluerise: oh ok | 00:30 |
| minute | bluerise: then idk :/ if you can't write, it's normally some AXI thing that doesn't have a clock yeah | 00:30 |
| minute | or power gated yeah | 00:30 |
| cinap_lenrek | man | 00:31 |
| cinap_lenrek | can you use the memory commands in uboot to write the regs? | 00:31 |
| bluerise | minute: pcie1 is not using internal refclk? | 00:31 |
| cinap_lenrek | i added some special commands to uboot on ppc to just do writes in virtual memory | 00:31 |
| minute | bluerise: no | 00:31 |
| cinap_lenrek | we could just flip the bit | 00:31 |
| bluerise | aha! | 00:32 |
| minute | bluerise: it uses an external refclk that comes from a chip on the reform mb | 00:32 |
| bluerise | ah | 00:32 |
| bluerise | wait I'm already doing that right, damnit | 00:32 |
| minute | (which is always enabled) | 00:32 |
| bluerise | yeah it does | 00:32 |
| minute | fun, i have an openbsd shell. not sure what to do in it though lol | 00:33 |
| minute | there's no "top" | 00:33 |
| eery[m] | Well, an upside to breaking everything is I found out the latest sysimage kernel fixes a gpu bug I'd been running into | 00:33 |
| minute | eery[m]: cool | 00:34 |
| eery[m] | I'm using wayfire and by faking GLES 3 compatibility I got the zoom plugin to work | 00:34 |
| eery[m] | So I've been using it to play games at lower resolutions but scaled up full screen | 00:35 |
| eery[m] | Before it would lock up the gpu whenever a window tried to capture the mouse pointer, now it doesn't | 00:35 |
| minute | wow ok | 00:38 |
| eery[m] | Is there a better way to use gles 3 calls like glBlitFramebuffer without setting the version override env variable? I just want that one bit of gles 3 | 00:39 |
| bluerise | huh, I got a ticket for the fusion festival | 00:42 |
| cinap_lenrek | ;_; | 00:42 |
| bluerise | minute: it's a miniroot | 00:42 |
| bluerise | minute: only supposed to be used to install a regular image | 00:42 |
| minute | bluerise: guess i need a real root | 00:45 |
| bluerise | you can install onto a usb stick maybe? | 00:46 |
| bluerise | or on the sd card | 00:46 |
| cinap_lenrek | does the fec work? | 00:47 |
| bluerise | used to :) | 00:47 |
| minute | ok. in any case, this is super cool | 00:49 |
| minute | thank you bluerise cinap_lenrek sigrid | 00:49 |
| cinap_lenrek | no, thank YOU! :) | 00:50 |
| minute | i'm gonna watch elementary and go to sleep and tinker with this more in the next days | 00:50 |
| minute | i also got a riscos binary but it doesn't display anything yet ^^ | 00:50 |
| cinap_lenrek | minute: for sure, some of the safeguards could be thought over now | 00:50 |
| cinap_lenrek | (now that we have a interactive console) | 00:50 |
| minute | in general i think defaulting to emmc in the future would be good | 00:51 |
| sigrid | does riscos support ext4? | 00:52 |
| minute | and it can autoboot sd card, which is great | 00:52 |
| cinap_lenrek | pocoapoco | 00:52 |
| minute | hmm http://splodge.fluff.org/~phil/risc-os/index.html | 00:53 |
| cinap_lenrek | yeah, it might be a good idea to relayout the "rescue" image | 00:53 |
| minute | > IscaFS is a read-only implementation of the ext2 filesystem, as a RISC OS module written in ARM assembly language | 00:53 |
| cinap_lenrek | and have a little fat partition to put kernels for other os and stuff | 00:53 |
| cinap_lenrek | kernels and configs | 00:53 |
| minute | cinap_lenrek: yeah | 00:53 |
| minute | all the kernels | 00:53 |
| - mtm (QUIT: Ping timeout: 272 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 01:03 | |
| - S0rin (QUIT: Ping timeout: 256 seconds) (~S0rin@user/s0rin) | 01:06 | |
| + S0rin (~S0rin@user/s0rin) | 01:17 | |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 01:41 | |
| - bkeys (QUIT: Ping timeout: 256 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 01:48 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 02:44 | |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 02:53 | |
| - bkeys (QUIT: Ping timeout: 272 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 03:05 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 03:08 | |
| + XYZ (~XYZ@89.24.58.41) | 03:21 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@89.24.58.41) | 03:26 | |
| + XYZ (~XYZ@89-24-58-41.nat.epc.tmcz.cz) | 03:27 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@89-24-58-41.nat.epc.tmcz.cz) | 03:33 | |
| - Guest2700 (QUIT: Ping timeout: 260 seconds) (~nicolas@8-49-142-46.pool.kielnet.net) | 03:38 | |
| + nsc (~nicolas@177-48-142-46.pool.kielnet.net) | 03:40 | |
| * nsc -> Guest1048 | 03:41 | |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 03:47 | |
| - S0rin (QUIT: Ping timeout: 256 seconds) (~S0rin@user/s0rin) | 04:34 | |
| + S0rin (~S0rin@user/s0rin) | 04:42 | |
| - S0rin (QUIT: Ping timeout: 256 seconds) (~S0rin@user/s0rin) | 04:48 | |
| + S0rin (~S0rin@user/s0rin) | 05:19 | |
| - S0rin (QUIT: Ping timeout: 256 seconds) (~S0rin@user/s0rin) | 05:24 | |
| + S0rin (~S0rin@user/s0rin) | 05:26 | |
| - S0rin (QUIT: Ping timeout: 256 seconds) (~S0rin@user/s0rin) | 05:32 | |
| + S0rin (~S0rin@user/s0rin) | 05:36 | |
| - robin (QUIT: Remote host closed the connection) (~robin@user/terpri) | 05:44 | |
| + robin (~robin@user/terpri) | 05:58 | |
| - S0rin (QUIT: Ping timeout: 272 seconds) (~S0rin@user/s0rin) | 07:11 | |
| + S0rin (~S0rin@user/s0rin) | 07:13 | |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 07:33 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:46 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:47 | |
| + pandora[m] (~pandorami@2001:470:69fc:105::2:b5e8) | 10:23 | |
| + pandora (uid585533@id-585533.ilkley.irccloud.com) | 10:24 | |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 11:04 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 11:05 | |
| + klardotsh__ (~klardotsh@172.56.105.110) | 11:47 | |
| - klardotsh_ (QUIT: Ping timeout: 272 seconds) (~klardotsh@172.56.105.8) | 11:50 | |
| - mtm (QUIT: Ping timeout: 252 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 13:04 | |
| minute | blurise FYI i've applied your patch to the repo https://source.mnt.re/reform/reform-boundary-uboot/-/commit/f32b8ab1dc2e73ba83f80ca95d4b8c4c64e3a66c | 13:17 |
| minute | bluerise, that is | 13:17 |
| bluerise | :) | 13:21 |
| minute | i collected the instructions here https://community.mnt.re/t/any-news-on-openbsd/646/7 | 13:33 |
| minute | i guess we could easily make a custom image with uboot included | 13:33 |
| minute | for people who have still set their boot switch to SD | 13:33 |
| minute | or for distroboot, we can probably modify the script on the image? | 13:34 |
| vkoskiv | How stable is the new u-boot stuff now? Should I try to flash my bootloader now or wait a bit longer? | 13:40 |
| josch | vkoskiv: i think the current u-boot is a superset of the features of the old u-boot. If your old u-boot was able to boot your system, the new one will be able to as well. Plus, you get more features like distroboot, display and usb-keyboard. | 13:43 |
| vkoskiv | Yeah, mostly just after the display/keyboard stuff :D | 13:44 |
| vkoskiv | I was just observing the discussion as it was being built and wondered if all the bugs were ironed out | 13:44 |
| vkoskiv | Given that lukas just posted and it seems to have been merged to the main reform uboot repo, I'll assume it's good to flash now :D | 13:45 |
| vkoskiv | Will report back once I have it done | 13:45 |
| josch | being able to immediately see that something is happening after pressing the power button is amazing :D | 13:47 |
| * wielaard -> mjw | 13:51 | |
| vkoskiv | Is the bin file spl/u-boot-spl.bin? | 14:01 |
| josch | vkoskiv: no, flash.bin | 14:01 |
| vkoskiv | I didn't get a flash.bin in my build | 14:01 |
| josch | sudo dd if=flash.bin of=/dev/mmcblk1 conv=notrunc bs=1k seek=33 | 14:02 |
| josch | vkoskiv: which commands did you run? | 14:02 |
| vkoskiv | cp mntreform-config .config && time ./build.sh | 14:02 |
| josch | and that exited successfully? because build.sh says "make flash.bin" | 14:03 |
| josch | if make was not able to create flash.bin then you should've gotten an error | 14:03 |
| vkoskiv | Ah, there was an error at some point I bet | 14:03 |
| josch | maybe run without -j to see it better | 14:03 |
| vkoskiv | sh: 1: dtc: not found | 14:03 |
| josch | apt-get --yes install --no-install-recommends gcc-aarch64-linux-gnu build-essential bison flex device-tree-compiler | 14:03 |
| josch | vkoskiv: you can look into .gitlab-ci.yml for the required build steps starting from a clean chroot | 14:04 |
| vkoskiv | Yeah I was just missing the device tree compiler | 14:04 |
| vkoskiv | flash bin got! | 14:05 |
| vkoskiv | Hm, root isn't allowed to toggle the force_ro on mmcblk0boot0 | 14:06 |
| josch | vkoskiv: you are flashing it to emmc? | 14:07 |
| vkoskiv | Hm, sudo is messing it up somehow | 14:07 |
| vkoskiv | If I su to root, then it allows me | 14:07 |
| vkoskiv | Yeah, emmc | 14:07 |
| josch | vkoskiv: what is the command you run? | 14:08 |
| vkoskiv | It worked now | 14:08 |
| vkoskiv | Just echo 0 to the force_ro under sys class | 14:08 |
| josch | vkoskiv: that will just execute echo as root | 14:08 |
| josch | vkoskiv: but the redirect is done by your shell which is not root | 14:08 |
| vkoskiv | Yeah, I figured it was something like that | 14:09 |
| josch | usually you use tee if you do not want to run "su" first | 14:09 |
| josch | echo foobar | sudo /sys/... >/dev/null | 14:09 |
| josch | echo foobar | sudo tee /sys/... >/dev/null | 14:09 |
| vkoskiv | Worked! u-boot shows up now with a cool Reform logo | 14:10 |
| josch | nice!! :D | 14:10 |
| josch | huh... the community.mnt.re forum software does not let me edit my posts?? | 14:17 |
| josch | the "edit this post" button is just missing from some of my posts -- why? | 14:18 |
| bluerise | minute: the DTB you're shipping with linux is not really close to mainline linux, right? | 14:49 |
| bluerise | because my OpenBSD work has always been to follow mainline closely | 14:49 |
| bluerise | for pcie for instance that wasn't always possible because that was never properly upstreamed | 14:49 |
| bluerise | also I'd like to follow U-Boot mainline closely, as they pull their DTBs from linux with only minor changes | 14:50 |
| bluerise | so then you can have a U-Boot that provides a mainline dtb, which is nice | 14:50 |
| bluerise | therefore my eagerness to work on upstreaming U-Boot | 14:51 |
| bluerise | I think I could send out another patch to u-boot later today | 14:51 |
| bluerise | and still need to fix whatever is broken with pcie after the update :( | 14:51 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 15:10 | |
| sknebel | in Germany there's also the option to register as "Selbstverzoller" which means the postal service won't take over at the customs office - and not bill you fees for doing so. means you need to go to the customs office yourself - so only makes sense if thats easy to reach for you | 15:19 |
| sknebel | wops, wrong channel | 16:04 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 17:21 | |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 17:46 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 18:45 | |
| josch | Linux reform 6.1.0-reform2-arm64 #1 SMP Debian 6.1.7-1+reform20230119T115832Z1 (2023-01-19) aarch64 GNU/Linux | 19:20 |
| josch | ACTION just manually re-triggered the reform-debian-packages pipeline -- lets hope things succeed now! | 19:20 |
| + bkeys1 (~Thunderbi@157.201.98.1) | 19:22 | |
| + ajr (~ajr@user/ajr) | 19:26 | |
| - bkeys1 (QUIT: Ping timeout: 256 seconds) (~Thunderbi@157.201.98.1) | 19:27 | |
| + bkeys1 (~Thunderbi@157.201.98.1) | 19:33 | |
| - bkeys1 (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@157.201.98.1) | 19:55 | |
| josch | aaaand another cross-build bug of the linux packaging... | 21:05 |
| sigrid | :( | 21:07 |
| sigrid | maybe qemu-user chroot would be easier | 21:07 |
| - mjw (QUIT: Killed (NickServ (GHOST command used by markw!~wielaard@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 21:10 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 21:11 | |
| - arminweigl (QUIT: Ping timeout: 246 seconds) (~arminweig@sourcehut/user/arminweigl) | 21:14 | |
| + arminweigl (~arminweig@sourcehut/user/arminweigl) | 21:15 | |
| - ajr (QUIT: Ping timeout: 260 seconds) (~ajr@user/ajr) | 21:19 | |
| + jackhill (~jackhill@kalessin.dragonsnail.net) | 21:23 | |
| josch | it builds fine natively on the reform | 21:37 |
| josch | this is about building it on the source.mnt.re ci :) | 21:37 |
| josch | and this worked in the past, so this is just a regression that needs to be fixed | 21:38 |
| josch | if i just use qemu, then others will run into the problem instead | 21:38 |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149-210-16-8.mobile.nym.cosmote.net) | 21:41 | |
| + Ar|stote|is (~linx@149-210-16-8.mobile.nym.cosmote.net) | 21:42 | |
| minute | bluerise: there should be only small differences to mainline | 21:45 |
| josch | minute: i think the problem is, that the order of things is totally different -- if you just "diff" it, the diff will be hundreds of lines long | 21:48 |
| josch | vagrantc tried to minimize the diff and failed | 21:50 |
| josch | that's why in reform-debian-package we just copy over the dtb instead of patching it | 21:50 |
| c-keen[m] | There's a dtb decompiler tool and one can sort the order a bit | 22:00 |
| + ajr (~ajr@user/ajr) | 22:10 | |
| josch | is anybody able to download the job artifacts from source.mnt.re? I get "Connection closed" before it finishes when I try to download this: http://reform.pages.mnt.re/-/reform-debian-packages/-/jobs/1098/artifacts/output.log | 22:15 |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 22:50 | |
| klardotsh__ | josch: failed on the last... presumably few bytes honestly. I thought it had completed and then "Failed" | 23:03 |
| * klardotsh__ -> klardotsh | 23:03 | |
| - ajr (QUIT: Ping timeout: 265 seconds) (~ajr@user/ajr) | 23:08 | |
| minute | josch: ah, adjusting the order first would be quite easy though. | 23:30 |
| Boostisbetter | just out of curiosity, but do we have any EUC riders in here? | 23:32 |
| minute | josch: btw i switched over some repo builds to the big runner earlier. | 23:49 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!