#archlinux-ports | Logs for 2026-11-04
Back
[00:40:32] <alkersan> so, the standard kernel can't be loaded by rpi own bootloader - because it is a zstd compressed kernel image, and rpi bootloader only knows gzip
[02:24:57] -!- linkmauve has quit [Remote host closed the connection]
[02:33:37] -!- linkmauve has joined #archlinux-ports
[02:45:10] -!- jclsn has quit [Ping timeout: 251 seconds]
[02:47:14] -!- jclsn has joined #archlinux-ports
[02:55:05] -!- linkmauve has quit [Remote host closed the connection]
[03:03:50] -!- linkmauve has joined #archlinux-ports
[03:58:54] -!- drathir_tor has quit [Remote host closed the connection]
[04:21:52] -!- drathir_tor has joined #archlinux-ports
[04:38:07] -!- hcmb_ has joined #archlinux-ports
[04:38:07] -!- hcmb has quit [Killed (erbium.libera.chat (Nickname regained by services))]
[04:38:07] hcmb_ is now known as hcmb
[04:50:59] <alkersan> with few install hooks it's possible to make it work https://github.com , but i can't call that stable.
[04:51:00] <phrik> Title: arch-rpi5/mkosi.profiles/linux/mkosi.skeleton/etc at main · alkersan/arch-rpi5 · GitHub (at github.com)
[04:52:26] -!- drathir_tor has quit [Ping timeout: 270 seconds]
[05:00:55] -!- drathir_tor has joined #archlinux-ports
[05:43:50] -!- ismailarilik has joined #archlinux-ports
[06:04:39] -!- hch1290722 has joined #archlinux-ports
[06:07:20] -!- hch129072 has quit [Ping timeout: 262 seconds]
[06:07:20] hch1290722 is now known as hch129072
[06:58:40] -!- ismailarilik has quit [Ping timeout: 245 seconds]
[07:01:16] -!- alkersan has quit [Quit: Connection closed for inactivity]
[07:40:25] -!- ismailarilik has joined #archlinux-ports
[08:27:18] -!- ismailarilik has quit [Ping timeout: 264 seconds]
[08:41:08] <solskogen|M> I think SuSE makes uboot boot grub, which in turn can load zstd
[08:56:48] -!- drathir_tor has quit [Remote host closed the connection]
[08:57:19] -!- drathir_tor has joined #archlinux-ports
[09:17:49] -!- ismailarilik has joined #archlinux-ports
[09:20:15] -!- Danct12 has quit [Quit: ZNC 1.10.3 - https://znc.in]
[09:20:53] -!- Danct12 has joined #archlinux-ports
[09:58:13] -!- filmroellchen has joined #archlinux-ports
[10:49:21] -!- ismailarilik has quit [Ping timeout: 240 seconds]
[11:50:39] <fermino> I'm working in having linux-rpi install the kernel with kernel-install instead of having the package manage it. It's still a bit of a mess and for the pi4 kernel and earlier BUT it would be possible to ship the uncompressed one and compress it on the install, or the opposite
[11:57:32] -!- ismailarilik has joined #archlinux-ports
[12:33:05] <solskogen|M> Why would you want it uncompressed?
[12:33:43] <solskogen|M> (I'm not asking because I think I know better, I'm just genuinely curious)
[12:37:09] -!- iwuvmika has quit [Ping timeout: 256 seconds]
[12:37:43] -!- Danct12 has quit [Quit: ZNC 1.10.3 - https://znc.in]
[12:38:23] -!- Danct12 has joined #archlinux-ports
[12:42:46] -!- iwuvmika has joined #archlinux-ports
[12:53:03] -!- alkersan has joined #archlinux-ports
[12:58:30] -!- Danct12 has quit [Quit: ZNC 1.10.3 - https://znc.in]
[12:59:09] -!- Danct12 has joined #archlinux-ports
[13:06:36] -!- alkersan has quit []
[13:06:38] -!- iwuvmika has quit [Ping timeout: 267 seconds]
[13:23:43] -!- alkersan has joined #archlinux-ports
[13:27:28] <alkersan> fermino I used an mkinitcpio hook that repackage the mainline kernel into a gzip compressed Image. There is a handy helper function for that "decompress_cat"
[13:28:25] <alkersan> check the "linux" profile in the mkosi repo I've linked above
[13:30:15] -!- cjc7373 has quit [Ping timeout: 244 seconds]
[13:31:07] -!- cjc7373 has joined #archlinux-ports
[13:49:13] -!- alkersan has quit [Quit: The Lounge - https://thelounge.chat]
[13:49:35] -!- alkersan has joined #archlinux-ports
[14:36:03] -!- ismailarilik has quit [Ping timeout: 247 seconds]
[14:45:56] -!- ismailarilik has joined #archlinux-ports
[15:05:54] -!- jelle|M has quit [Ping timeout: 249 seconds]
[15:06:09] -!- artafinde|M has quit [Ping timeout: 270 seconds]
[15:06:09] -!- bgyorgy|M has quit [Ping timeout: 260 seconds]
[15:06:19] -!- prawn has quit [Ping timeout: 250 seconds]
[15:06:19] -!- tippfehlr|M has quit [Ping timeout: 250 seconds]
[15:06:33] -!- tpkessler|M has quit [Ping timeout: 251 seconds]
[15:06:33] -!- heftig has quit [Ping timeout: 251 seconds]
[15:06:38] -!- yuvadm has quit [Ping timeout: 259 seconds]
[15:09:32] -!- yuvadm has joined #archlinux-ports
[15:19:20] -!- tippfehlr|M has joined #archlinux-ports
[15:19:24] -!- prawn has joined #archlinux-ports
[15:20:02] -!- artafinde|M has joined #archlinux-ports
[15:20:45] -!- jelle|M has joined #archlinux-ports
[15:20:47] -!- bgyorgy|M has joined #archlinux-ports
[15:21:12] -!- heftig has joined #archlinux-ports
[15:21:21] -!- tpkessler|M has joined #archlinux-ports
[15:40:04] -!- cbix has parted #archlinux-ports
[15:49:58] -!- cbix has joined #archlinux-ports
[15:54:59] -!- cbix has parted #archlinux-ports
[15:55:23] -!- cbix has joined #archlinux-ports
[16:04:16] -!- ismailarilik has quit [Ping timeout: 252 seconds]
[16:44:30] <linkmauve> Hi, I’m running Dolphin on my board, and I see the lroundf32() function from libm.so taking 2.7% of the CPU runtime, even though this function only contains a single useful instruction (besides handling the stack): fcvtas x0, s0
[16:44:57] <linkmauve> I would expect inlining it to improve things a bunch, given how often it gets called.
[16:45:30] <linkmauve> Where could I modify things so that it becomes inlined, if possible in all programs?
[16:47:46] <linkmauve> https://codebrowser.dev
[16:47:47] <phrik> Title: s_lroundf.c source code [glibc/sysdeps/aarch64/fpu/s_lroundf.c] - Codebrowser (at codebrowser.dev)
[16:49:19] -!- ismailarilik has joined #archlinux-ports
[16:54:31] -!- cbix has parted #archlinux-ports
[16:54:33] -!- cbix has joined #archlinux-ports
[17:03:37] -!- ismailarilik has quit [Ping timeout: 269 seconds]
[17:19:01] -!- cbix has parted #archlinux-ports
[17:19:04] -!- cbix has joined #archlinux-ports
[18:11:38] -!- iwuvmika has joined #archlinux-ports
[19:35:53] -!- iwuvmika has quit [Ping timeout: 247 seconds]
[19:55:29] -!- ismailarilik has joined #archlinux-ports
[20:08:34] -!- hcmb has quit [Remote host closed the connection]
[20:09:37] -!- hcmb has joined #archlinux-ports
[20:11:11] -!- iwuvmika has joined #archlinux-ports
[20:20:45] -!- ismailarilik has quit [Ping timeout: 243 seconds]
[21:17:22] -!- ismailarilik has joined #archlinux-ports
[21:24:35] <bertptrs> linkmauve: as long as libm is being dynamically linked, you're not going to achieve that, and there is no desire from a distro perspective to statically link libm
[22:12:30] <fermino> @linkmauve, you can always patch glibc and install it system wide, but you'll have to keep track of updates pretty closely or things will break
[22:12:31] <fermino> https://gitlab.archlinux.org
[22:12:32] <phrik> Title: Arch Linux / Packaging / Packages / glibc · GitLab (at gitlab.archlinux.org)
[22:52:22] -!- leanguedes has quit [Remote host closed the connection]
[22:52:46] -!- leanguedes has joined #archlinux-ports
[22:55:00] <fermino> !paste
[22:55:01] <phrik> Paste short snippets (up to 3 lines) in channel. Please do NOT use pastebin.com or paste.ubuntu.com -- some good, sane paste services are https://paste.rs, https://x0.at, https://bpaste.net and https://gist.github.com. See !rs, !x0 or !gist for an easy way to paste from the cli.
[23:00:52] <fermino> I'm gonna ask a maybe stupid question that I've had for a while @Solskogen you are the owner of drzee.net, because if not I've been directing a ton of questions to you that were meant to go there :)
[23:02:17] <fermino> Anyways, besides that an in case you are :p Any chance that the gitlab-runner package from the repos is being passed through from x86_64? https://bpa.st
[23:02:18] <phrik> Title: View paste WIEKC (at bpa.st)
[23:02:40] -!- filmroellchen has quit [Quit: filmroellchen]
[23:04:38] <fermino> Ohhhh, I think it's just being built wrong `make out/binaries/gitlab-runner-linux-amd64`
[23:10:14] -!- titus_livius has joined #archlinux-ports
[23:55:43] <bschnei> well that's interesting. wonder why it doesn't fail to build...