| - Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 00:30 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 00:36 | |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 00:45 | |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@utdpat241106.utdallas.edu) | 01:00 | |
| - mtm (QUIT: Ping timeout: 260 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 01:04 | |
| + jacobk (~quassel@utdpat242094.utdallas.edu) | 01:08 | |
| - mjw (QUIT: Ping timeout: 268 seconds) (~mjw@gnu.wildebeest.org) | 01:29 | |
| - jacobk (QUIT: Ping timeout: 268 seconds) (~quassel@utdpat242094.utdallas.edu) | 01:37 | |
| + jacobk (~quassel@utdpat242094.utdallas.edu) | 01:39 | |
| digitalrayne | does anyone known if the lpc firmware requires a specific version of gcc-arm-none-eabi? | 01:52 |
|---|---|---|
| digitalrayne | *know | 01:52 |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@utdpat242094.utdallas.edu) | 02:05 | |
| jn | digitalrayne: are you hitting a specific error? | 02:13 |
| jn | digitalrayne: perhaps this one? https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1067692 | 02:13 |
| - jn (QUIT: Ping timeout: 272 seconds) (~quassel@user/jn/x-3390946) | 02:50 | |
| + jn (~quassel@user/jn/x-3390946) | 02:50 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 02:56 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 02:57 | |
| digitalrayne | i was trying to compile on OpenBSD, wasn't getting any errors but the built binary didn't seem to be booting - I think it may have been a subtle difference between *BSD's dd and the one shipped with Debian, I was able to reproduce the issue (LPC not booting) with both an alpine-built and OpenBSD-built LPC firmware | 03:00 |
| digitalrayne | if I added bs=1 to the end of the dd command though, it worked OK | 03:01 |
| digitalrayne | (in addition to the documented conv=notrunc) | 03:01 |
| digitalrayne | thank you though - it's good to be aware of potential issues with newer versions | 03:01 |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 03:02 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 03:03 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 03:04 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 03:04 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 03:10 | |
| - nsc_ (QUIT: Ping timeout: 268 seconds) (~nicolas@186-99-142-46.pool.kielnet.net) | 03:40 | |
| + nsc (~nicolas@188-48-142-46.pool.kielnet.net) | 03:42 | |
| + jacobk (~quassel@utdpat242094.utdallas.edu) | 03:44 | |
| + f__ (~AUGESOUND@fases/developer/funderscore) | 03:57 | |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@utdpat242094.utdallas.edu) | 03:58 | |
| - f_ (QUIT: Ping timeout: 260 seconds) (~AUGESOUND@fases/developer/funderscore) | 03:58 | |
| * f__ -> f_ | 04:00 | |
| + jacobk (~quassel@utdpat242068.utdallas.edu) | 06:03 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 06:25 | |
| + hairu (m-uotkmd@user/hairu) | 06:26 | |
| - Gooberpatrol66 (QUIT: Read error: Connection reset by peer) (~Gooberpat@user/gooberpatrol66) | 06:43 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 06:46 | |
| - GNUmoon2 (QUIT: Ping timeout: 260 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 07:36 | |
| - Gooberpatrol66 (QUIT: Remote host closed the connection) (~Gooberpat@user/gooberpatrol66) | 07:48 | |
| + chomwitt (~chomwitt@2a02:587:7a17:f700:1ac0:4dff:fedb:a3f1) | 07:49 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 07:49 | |
| - chomwitt (QUIT: Ping timeout: 260 seconds) (~chomwitt@2a02:587:7a17:f700:1ac0:4dff:fedb:a3f1) | 08:33 | |
| jn | digitalrayne: ah, interesting. i wonder who's at fault here: BSD or alpine for having a POSIX-noncompliant dd, or the firmware installation script for using dd incorrectly | 09:07 |
| + chomwitt (~chomwitt@2a02:587:7a17:f700:1ac0:4dff:fedb:a3f1) | 09:18 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:18 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 09:18 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 09:40 | |
| + hairu (m-uotkmd@user/hairu) | 09:41 | |
| - jacobk (QUIT: Ping timeout: 268 seconds) (~quassel@utdpat242068.utdallas.edu) | 09:49 | |
| - chomwitt (QUIT: Quit: WeeChat 3.8) (~chomwitt@2a02:587:7a17:f700:1ac0:4dff:fedb:a3f1) | 10:16 | |
| + jacobk (~quassel@64.189.201.150) | 10:39 | |
| + colinsane (~colinunin@97-113-159-4.tukw.qwest.net) | 11:17 | |
| digitalrayne | most likely because i was writing the lpc firmware under OpenBSD with different block device behaviour, it could have been dd differences though | 11:17 |
| digitalrayne | the lpc firmware compiled under alpine & openbsd has the exact same checksum though, which was cool to test | 11:18 |
| + mjw (~mjw@gnu.wildebeest.org) | 11:22 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-159-4.tukw.qwest.net) | 11:41 | |
| + colinsane (~colinunin@97-113-159-4.tukw.qwest.net) | 11:44 | |
| jn | hooray for reproducible builds | 11:59 |
| minute | digitalrayne: yeah, the block device implementation in the lpc rom is very peculiar | 12:30 |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 12:30 | |
| * Guest9745 -> mjw | 12:42 | |
| - mtm (QUIT: Ping timeout: 256 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 13:04 | |
| josch | if some versions of dd work, maybe it makes sense to write a tiny C script which does the open() and write() syscalls exactly as required? Or could it be a problem of the vfat implementation? | 14:35 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 15:09 | |
| + mark_ (~mjw@gnu.wildebeest.org) | 15:42 | |
| minute | i'm working now on the pocket reform production install process :0 | 15:45 |
| minute | we're building the first 10 devices... | 15:45 |
| - jacobk (QUIT: Ping timeout: 268 seconds) (~quassel@64.189.201.150) | 16:00 | |
| grimmware | :O | 16:18 |
| grimmware | don't forget to post some pics on the fedi! | 16:18 |
| minute | complicated because we need to bootstrap over ethernet, as the preinstalled uboot can't boot from sd card. | 16:18 |
| + jacobk (~quassel@utdpat242053.utdallas.edu) | 16:21 | |
| josch | digitalrayne: i was able to flash the lpc firmware without mounting the block device by writing the firmware to offset 0x800 on the device it shows to the OS | 16:26 |
| josch | digitalrayne: maybe bypassing the filesystem like that can be a solution for you? | 16:26 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 16:46 | |
| jn | with mentions of block storage and vfat... does the lpc implement UF2? UF2 is pretty cool and easy to use | 17:18 |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@utdpat242053.utdallas.edu) | 17:19 | |
| - mark_ (QUIT: Ping timeout: 272 seconds) (~mjw@gnu.wildebeest.org) | 17:21 | |
| josch | jn: no mention of that in the lpc firmware | 17:28 |
| jn | hm, maybe something for future consideration (although switching from a raw image to UF2 might be difficult, i'm not entirely sure) | 17:29 |
| minute | josch: jn: the filesystem stuff is not part of the firmware. it's implemented in rom | 17:31 |
| josch | i'm inspecting the fat filesystem exposed by the lpc and offset 0x800 there is always a file with 32768 in size -- i guess that's why dd is run with nocreat and notrunc to not mess with any of that and just overwrite the location | 17:31 |
| josch | minute: ahaaa! that explains why the strings i find in the fat filesystem cannot be found anywhere in the firmware :) | 17:32 |
| minute | josch: yep. the disk implementation in the rom is incomplete and they only tested it with windows explorer i think | 17:32 |
| - sevan (QUIT: Ping timeout: 256 seconds) (~sevan@user/venture37) | 17:32 | |
| jn | oh, i see | 17:33 |
| josch | if it's in rom, then the offset of the magic FIRMWARE.BIN file should never be any other place than at offset 0x800 | 17:34 |
| josch | digitalrayne: so on linux, this worked for me: dd if=bin/firmware.bin of=/dev/sda bs=512 seek=4 count=64 | 17:34 |
| jn | where's the firmware source btw? i can't find it under https://source.mnt.re/reform?filter=lpc | 17:36 |
| josch | jn: https://source.mnt.re/reform/reform/-/tree/master/reform2-lpc-fw?ref_type=heads | 17:36 |
| jn | ah thanks | 17:37 |
| jn | the readme also answer the question that has been bugging me for a while: Why is it called LPC? (it's an LPC11U24 µC) | 17:38 |
| jn | for a while i suspected (incorrectly!) that it was an incorrect semantic transfer, i.e.: 1. some laptop ECs are connected via the "low pin count" LPC bus, 2. let's call ECs LPC, 3. let's call this EC "LPC" as well, even tough it's connected by other means | 17:39 |
| minute | jn: LPC means Low Pin Count | 17:40 |
| jn | minute: in the NXP LPC series too? | 17:40 |
| minute | jn: https://en.wikipedia.org/wiki/NXP_LPC | 17:40 |
| jn | terrible naming, in the grand scheme of things :D | 17:40 |
| jn | it's unrelated to the miniaturised ISA bus known as "low pin count", right? | 17:41 |
| minute | i think so | 17:44 |
| jn | reminds me of the two entirely different PowerPC-related processor features called SPE and used in the PowerPC generation of video game consoles (Cell SPE vs. the thing that found its way into debian as powerpcspe) | 17:46 |
| + jacobk (~quassel@64.189.201.150) | 17:47 | |
| minute | hm, our system image doesn't have growpart? | 17:55 |
| minute | (cloud-guest-utils) | 17:55 |
| josch | why not use gparted? | 17:56 |
| minute | ah, parted can do it! | 17:56 |
| josch | parted -s $rootdev resizepart $rootpart_nr 100% | 17:56 |
| josch | it's part of my firstboot MR against reform-system-image | 17:57 |
| minute | > parted /dev/mmcblk2 resizepart 2 100% | 17:57 |
| minute | yeah :3 | 17:57 |
| minute | i've concluded that so far the fastest way of installing our system image to eMMC is to dd it from a USB 3 drive | 17:57 |
| josch | minute: important there is, that the changes are not applied immediately and you might want to apply the udev-waiting logic before you partprobe | 17:57 |
| minute | (during pocket production) | 17:58 |
| + sevan (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 17:59 | |
| minute | the only remaining issue is the slowness of resize2fs | 18:00 |
| minute | well, it's kind of ok i guess | 18:02 |
| josch | minute: maybe instead of first dd and then resize2fs on the device, you could try creating a sparse disk image with fallocate, grow the partition and resize2fs there and then create a bmap file with bmaptool and flash the image with bmaptool instead of dd which will only write those blocks which have content in them | 18:03 |
| - klardotsh_ (QUIT: Quit: zzz) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 18:03 | |
| minute | josch: mhm! | 18:04 |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 18:04 | |
| minute | i will also need to heavily update the imx8mp uboot today | 18:04 |
| minute | it's missing critical variables like kernel_addr_r, fdt_addr_r and ramdisk_addr_r | 18:04 |
| josch | yeah those are kinda important :D | 18:05 |
| minute | but nice, when setting setenv kernel_addr_r $loadaddr; setenv fdt_addr_r $fdt_addr; setenv ramdisk_addr_r 0x44000000; it can boot from the freshly imaged emmc | 18:05 |
| josch | uff awesome!! \o/ | 18:05 |
| minute | but network is broken and wifi is missing, need to fix them today | 18:06 |
| josch | sounds like we soon can generate disk images for the pocket | 18:06 |
| minute | josch: we already do :3 they're just not perfect yet | 18:06 |
| josch | though i wonder if it makes sense to let reform-system-image publish them if the pocket cannot boot from sd-card? | 18:06 |
| minute | josch: pocket will be able to boot them, it only can't load uboot from them | 18:07 |
| josch | ah okay, that's good | 18:07 |
| minute | my factory process, very simplified, is like: 1. flash our uboot to emmc 2. image our system image to emmc | 18:07 |
| minute | in practice it's more complicated because the first boot of the vendor uboot needs to go via ethernet+tftp because everything else is borked | 18:08 |
| josch | so... you will need to do that process 700 times? XD | 18:09 |
| minute | yeah | 18:19 |
| minute | but i'll try to automate it | 18:19 |
| minute | hmhm > [ 17.130879] platform 32ec0000.blk-ctrl: deferred probe pending: imx8m-blk-ctrl: failed to get noc entries | 18:24 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 18:25 | |
| minute | but why was i able to boot the first time around... | 18:26 |
| minute | i wonder if a driver for "fsl,imx8mp-noc", "fsl,imx8m-noc" is missing in the debian kernel | 18:36 |
| minute | > CONFIG_INTERCONNECT_IMX8MP=m | 18:37 |
| minute | that's probably not good | 18:37 |
| minute | hm but on imx8m it was also a module and didn't make such trouble... | 18:38 |
| minute | also, if the system decides to run fsck, all bets are off, because watchdog driver is not loaded and then the system just hard resets before fsck can finish :( | 18:38 |
| minute | i will try to do CONFIG_INTERCONNECT_IMX8MP=y and CONFIG_IMX7ULP_WDT=y and rebuild the kernel package locally | 18:40 |
| - jacobk (QUIT: Ping timeout: 268 seconds) (~quassel@64.189.201.150) | 18:58 | |
| minute | hm, weird error in + env --chdir=linux ../kernel-team/utils/kconfigeditor2/process.py . | 19:02 |
| minute | > KeyError: ('base',) | 19:02 |
| josch | minute: do you have a log? | 19:02 |
| minute | josch: does it log to a file or do i need to copy it out of the terminal?' | 19:03 |
| josch | also, try "rm -rf kernel-team" so that the build.sh script clones a new version | 19:03 |
| minute | ah ok | 19:03 |
| josch | minute: the build log is stored in a file called ${srcpkg}_${version}.build | 19:03 |
| josch | errr... ${srcpkg}-${version}_${arch}.build | 19:04 |
| minute | hmm i believe it did not create such a file yet | 19:04 |
| josch | the debian linux packaging received a very large refactoring with the 6.7 release so it's not surprising that some things break | 19:05 |
| minute | the log is not very long, i will copy it out... and try removing kernel-team | 19:05 |
| minute | josch: http://dump.mntmn.com/linux-buildlog-20240326-1.txt | 19:06 |
| + jacobk (~quassel@utdpat241055.utdallas.edu) | 19:06 | |
| minute | josch: yep, deleting kernel-team did it, many thanks | 19:07 |
| josch | okay, great :) | 19:07 |
| josch | minute: alternatively, you can also comment out that line calling kconfigeditor2 in linux/build.sh. That script is only to "prettify" configs (sorting, structuring, adding header comments etc) and not strictly necessary | 19:08 |
| minute | josch: ah, good to know | 19:10 |
| minute | it is compiling the kernel now. | 19:10 |
| josch | we need compilation times on rk3588 for the reform kernel :D | 19:11 |
| + mark_ (~mjw@gnu.wildebeest.org) | 19:19 | |
| minute | josch: yeah, unfortunately i don't really have time for that at the moment, things are heating up here with pocket :D | 19:21 |
| josch | well, that is also good! :D | 19:23 |
| josch | i ordered my dragonbox pyra in 2016 and still don't have it | 19:24 |
| minute | :0 | 19:25 |
| josch | if you find yourself wondering why i'm so excited contributing to MNT, i'm so happy to finally have found a company which delivers | 19:25 |
| minute | oh, that is very nice to hear :3 | 19:42 |
| f_ | josch: wow haha | 19:43 |
| f_ | I wasn't aware that delivery could take this long :P | 19:43 |
| f_ | In this case it's been almost 10 years and the pyra is still in development | 19:46 |
| kop316 | I think a few folks got them, but it is that one person running the whole shop. I had faith it would come out much quicker due to the open pandora, but...... | 19:47 |
| * mjw -> Guest3473 | 19:50 | |
| - Guest3473 (QUIT: Killed (tungsten.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 19:50 | |
| * mark_ -> mjw | 19:50 | |
| + Guest3473 (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 19:51 | |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@utdpat241055.utdallas.edu) | 19:58 | |
| minute | how big is the pyra batch(es)? | 19:58 |
| minute | ok, a kernel deb fell out. | 20:00 |
| minute | so this took 50 minutes on my i9 oof. | 20:01 |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 20:10 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 20:30 | |
| + hairu (m-uotkmd@user/hairu) | 20:32 | |
| minute | ok, i just had to do some dtb fixes to get usb + ethernet working again | 20:33 |
| minute | next challenge: how to package https://source.mnt.re/reform/qcacld2 | 20:58 |
| minute | i guess similar to the lpc driver, dkms something something | 21:00 |
| minute | reading https://vincent.bernat.ch/en/blog/2018-packaging-driver-debian-dkms | 21:00 |
| grimmware | I looked up the Pyra again thanks to the Pocket and I was astounded it was still in development. | 21:06 |
| josch | minute: for packaging a kernel module with dkms you can just copy what i did for the reform2-lpc module which uses exactly the technique detailed in vincents blog post | 21:10 |
| josch | kop316: they also had a problem with one employee steeling from their warehouse and tons of issues with their case suppliers | 21:14 |
| kop316 | josch: eeesh.....yeah I sort of stop paying attention a few years ago. | 21:15 |
| josch | it's really unfortunate for them -- i think they are very nice people | 21:17 |
| kop316 | yeah they seem like it | 21:27 |
| minute | josch: ok, i have my first qcacld2-dkms_1.0.0-0_all.deb but it's only 1kb in size :D | 21:31 |
| minute | i probably have to make an .install file? | 21:32 |
| josch | minute: if your buildsystem does not install stuff itself, then yes | 21:36 |
| minute | hmm no idea... these are not my sources :D it's makefile based | 21:37 |
| josch | since the module is built upon installation, the package will just ship the Makefile and the sourcecode | 21:37 |
| minute | yeah, but it didn't include the sourcecode in the .deb | 21:37 |
| minute | aha, now that i put all the source paths and the makefile in an .install file, the .deb is 3MB in size | 21:37 |
| josch | cool :) | 21:38 |
| minute | wow, installing the .deb on the pocket literally created the dir 'qcacld2-${env:DEB_VERSION_UPSTREAM}' in /usr/src/ | 21:39 |
| minute | ouch | 21:39 |
| josch | your debian/rules needs to at least include: | 21:39 |
| josch | include /usr/share/dpkg/pkg-info.mk | 21:40 |
| josch | export DEB_VERSION_UPSTREAM | 21:40 |
| josch | this will set that environment variable to the correct value so that it can be used in the .install file | 21:40 |
| minute | it does include that | 21:40 |
| minute | also, when installing the package, no build happens | 21:41 |
| minute | i guess i'm holding it wrong | 21:41 |
| minute | i'll commit my debian folder, maybe you could take a quick look what i'm doing completely wrong? | 21:41 |
| josch | sure thing | 21:41 |
| josch | do you have dh-sequence-dkms in Build-Depends? | 21:42 |
| minute | josch: https://source.mnt.re/reform/qcacld2/-/tree/boundary-CNSS.LEA.NRT_3.1/debian?ref_type=heads | 21:43 |
| minute | let me check | 21:43 |
| minute | josch: no, i have "dkms", probably that tutorial is not current anymore | 21:43 |
| josch | dkms is the package that your qcacld2-dkms needs to depend on | 21:44 |
| josch | it's the package that needs to be installed when *installing* your dkms package not when building it | 21:44 |
| josch | try adding dh-sequence-dkms or in debian/rules add --with=dkms to the dh call | 21:45 |
| josch | (the former method is considered being more elegant) | 21:45 |
| minute | ok, trying | 21:45 |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@utdpat241106.utdallas.edu) | 21:45 | |
| josch | oh wow that MAKE[0] line :D | 21:46 |
| minute | :D | 21:46 |
| minute | not sure if that will work yet | 21:46 |
| josch | oh and instead of debhelper (>= 9) you should use debhelper-compat (= 13) | 21:46 |
| minute | i tried to stuff the vars from my build.sh in there https://source.mnt.re/reform/qcacld2/-/blob/boundary-CNSS.LEA.NRT_3.1/build.sh?ref_type=heads | 21:46 |
| josch | then you can also remove that debian/compat file | 21:46 |
| minute | josch: yeah, seen that and changed that, and deleted compat file | 21:46 |
| minute | aha, this time it wants to build something on the pocket | 21:47 |
| minute | apparently the linux-headers-6.7.9-reform2-arm64 package is not installed but i wonder how i can make the dkms package depend on that dynamically | 21:53 |
| minute | (the build fails with stdarg.h not found) | 21:54 |
| josch | the package cannot know what your currently running kernel is when it gets installed | 21:55 |
| + jacobk (~quassel@129.110.242.224) | 21:55 | |
| josch | otherwise, the moment you install another kernel, the dkms package would have to switch dependencies | 21:55 |
| josch | but the linux-headers-arm64 package is installed by reform-system-image, so it will be available by default | 21:55 |
| josch | and linux-headers-arm64 is the one that then depends on the latest versioned header package | 21:56 |
| minute | hm, i wonder why it doesn't find that header then | 21:56 |
| minute | ah hm | 21:56 |
| minute | the header is in /usr/src/linux-headers-6.7.9-reform2-common | 21:56 |
| minute | and make says > make[1]: Entering directory '/usr/src/linux-headers-6.7.9-reform2-arm64' | 21:57 |
| josch | the directory passed with -C should be /lib/modules/$(shell uname -r)/build | 21:58 |
| minute | ah yeah, and that has a makefile which includes /usr/src/linux-headers-6.7.9-reform2-common/Makefile | 21:59 |
| - jacobk (QUIT: Ping timeout: 268 seconds) (~quassel@129.110.242.224) | 22:00 | |
| josch | what helped me figure this out form the reform2-lpc dkms module was to try out variations of what other dkms modules do: https://codesearch.debian.net/search?q=MAKE%5B0%5D%3D&literal=1 | 22:00 |
| minute | :0 | 22:01 |
| minute | i wonder what causes > make[1]: Entering directory '/usr/src/linux-headers-6.7.9-reform2-arm64' | 22:01 |
| josch | in the dkms file in git you just say: make -j4 -C M=... | 22:03 |
| josch | so there is no path passed to -C | 22:03 |
| josch | others seem to do: -C ${kernel_source_dir} M=${dkms_tree}/${PACKAGE_NAME}/${PACKAGE_VERSION}/build | 22:04 |
| - sevan (QUIT: Changing host) (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 22:08 | |
| + sevan (~sevan@user/venture37) | 22:08 | |
| minute | josch: true | 22:08 |
| minute | looks like a copypaste error | 22:08 |
| minute | it doesn't seem to use the MAKE[0] line anyway... | 22:10 |
| josch | that'd be odd | 22:11 |
| minute | or maybe it does, but the makefile overrides some stuff | 22:11 |
| josch | alternatively, you could write your own makefile | 22:11 |
| minute | ok so the problem is that they include <stdarg.h>, but apparently it has to be <linux/stdarg.h> | 22:15 |
| minute | seems weird though | 22:16 |
| josch | but not uncommon -- the oss4 dkms module in debian also needs to patch this: https://sources.debian.org/src/oss4/4.2-build2020-3/debian/patches/include.patch/?hl=38#L38 | 22:17 |
| minute | josch: ohh ok thanks, that gives me confidence... then i'll patch that in the few driver sources files | 22:17 |
| minute | oof something in the kernel is hogging one core at 100% all the time | 22:29 |
| minute | this driver is so huge... | 22:34 |
| minute | takes forever to compile on the device :D | 22:38 |
| josch | it will taken even longer when it gets compiled on the gitlab CI because there the build gets emulated :D | 22:40 |
| minute | gaaah | 22:40 |
| minute | ok, build finished (i have to say that 1 core is 100% hogged... and i don't know why) | 22:43 |
| minute | ah, usb related | 22:43 |
| minute | ah yes, goes away upon `rmmod dwc3` | 22:43 |
| minute | probably i still got one of the fsl,over-current-active-low wrong | 22:44 |
| minute | > INSTALL /lib/modules/6.7.9-reform2-arm64/updates/wlan.ko | 22:46 |
| minute | funny that it's just called "wlan.ko" sigh | 22:46 |
| josch | minute: reading man page of dkms(8) it seems you can use DEST_MODULE_NAME to rename modules | 22:49 |
| minute | oh nice | 22:50 |
| minute | > ar6k_wlan: probe of mmc1:0001:1 failed with error -1 | 22:59 |
| minute | :3 | 22:59 |
| minute | at least i fixed the usb issue... | 23:00 |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-159-4.tukw.qwest.net) | 23:08 | |
| + colinsane (~colinunin@97-113-159-4.tukw.qwest.net) | 23:11 | |
| minute | ah, i managed to revive the wifi. i just had the wrong collection of firmware blobs | 23:32 |
| + jacobk (~quassel@64.189.201.150) | 23:43 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!