Linux CacULE (feedback & testing)

5.12.12-1 Changes:

  • Kernel got updated :smiley:
5 Likes

Hey,
over that whole thread we get told to set new "sysctl options" in order to make everything work even faster / better. Now there are a lot of replys and changes of the parameters.

here_1
here_2
here_3
and so on. I get lost there...

Would it be an idea to put the "actual perfect sysctl-cacule.conf" at some public place and keep that always updated? Then you just have to link that once for everyone.

5.13-1 Changes:

  • Patches got updated for 5.13 release
  • new futex2 patch
  • updated cacule patchset
  • winesync, futex also got updated

In general ive tested the 5.13 Kernel so far and everything runs smooth and i didnt faced any bugs so far. Seems in my testcase a stable running kernel, i dont faced any dramatic bugs.

Kernel got pushed to the AUR, Generic v3 Kernel will be pushed soon to the chaotic-aur.
@pamalosebi

In general the sysctl-settings a forced to the default settings. You can just play around with them, maybe you face any better values then the default ones.
You also find them the "files" with the containing value in /proc/sys/kernel

There you can find for example:

ptr1337@archpc /p/s/kernel> cat sched_interactivity_factor
32768

So i know the value is right now "32768" for the sched_interactivity_factor.

I have changed two default values to another one, maybe placebo if it runs better or not. Didnt benchmarked it.

My running Config placed in /etc/sysctl.d/99-cacule-settings.conf :

    # Cacule-Settings # 
    kernel.sched_interactivity_factor= 65536
    kernel.sched_max_lifetime_ms=11000
   ### Some general Network-Settings ##

    net.core.netdev_max_backlog = 16384
    net.core.somaxconn = 8192
    net.core.rmem_default = 1048576
    net.core.rmem_max = 16777216
    net.core.wmem_default = 1048576
    net.core.wmem_max = 16777216
    net.core.optmem_max = 65536
    net.ipv4.tcp_rmem = 4096 1048576 2097152
    net.ipv4.tcp_wmem = 4096 65536 16777216
    net.ipv4.udp_rmem_min = 8192
    net.ipv4.udp_wmem_min = 8192
    net.ipv4.tcp_fastopen = 3
    net.ipv4.tcp_keepalive_time = 60
    net.ipv4.tcp_keepalive_intvl = 10
    net.ipv4.tcp_keepalive_probes = 6
    net.ipv4.conf.default.log_martians = 1
    net.ipv4.conf.all.log_martians = 1
    net.ipv4.tcp_mtu_probing = 1
    net.ipv4.tcp_syncookies = 1
    net.core.default_qdisc = cake
    net.ipv4.tcp_congestion_control = bbr2

You can set these settings with following command:

ptr1337@archpc /p/s/kernel> sudo sysctl --system
6 Likes

5.13-2 Changes

  • NO_HZ_FULL patch which is right now used in linux-next is now included, no_hz_full didnt work since some time correctly, now it should be fixed
  • some security patches
5 Likes

Same here, running cacule-rdb, no issues for now.

2 Likes

The only thing what i found is when using systemd-boot you see the complete dmesg prompt until you can type in your luks password. probally some configurations things, need to watch these days to fix that.

But all in all it is running really good.

2 Likes

5.13-4 Changes:

  • Security patches included
  • lrng patches included
  • futex2 patch updated
  • pf-patches updated
  • nohzfull-fix.patch from linux-next included

GENv3 and the normal kernel will be soon updated in the chaotic-aur.

5 Likes

Many changes last days, Kernel will be updated as soon as 5.13.2 gets released, i think that will be tomorrow.
Alsa got some patches inputlag with usbb is finally away.
some rcu fixes. and so on.

regards.

5.13.2 Changes:

  • added lrng patches
  • added toggle for: CONFIG_LSM lockdown,yama,brute
  • updated bfq, btrfs, pf, lru, mm patches
  • included alsa patches for lower delay with usb audio
4 Likes

on linux-cacule-rdb 5.13.4-4 it fails to build bec of md5 check

rdb-5.13.patch … FEHLGESCHLAGEN

Make sure to clean the cache of paru prior to trying to build the package. The patches are statically linked and content gets updated without the file name changing which results in wrong checksums as the cached file is used instead of being redownloaded.

2 Likes

i Have try now

Paru -Scc
sudo paccache -r
sudo pacman -Scc
yay -Scc

paru -Scc

Zwischenspeicher-Verzeichnis: /var/cache/pacman/pkg/
:: Möchten Sie ALLE Dateien aus dem Zwischenspeicher entfernen? [j/N] j
Alle Dateien aus dem Zwischenspeicher entfernt …

Datenbank-Verzeichnis: /var/lib/pacman/
:: Möchten Sie ungenutzte Repositorien entfernen? [J/n] j
Ungenutzte Sync-Repositorien werden entfernt …

Clone Directory: /home/smoky/.cache/paru/clone
:: Do you want to remove all other AUR packages from cache? [Y/n]: y

Diff Directory: /home/smoky/.cache/paru/diff
:: Do you want to remove all saved diffs? [Y/n]: y

I also delete all files from /.cache/paru/clone/ and from /.cache/yay/
Also use Garuda assistant and clear caches and clear packed cache

sudo paccache -r

==> no candidate packages found for pruning 
sudo ls /var/cache/pacman/pkg/ | wc -l
0

After all this same error

rdb-5.13.patch … FEHLGESCHLAGEN

Have i forget something or do wrong ?

Alright, if its really a PKGBUILD issue I would suggest fixing it yourself by downloading it and running updpkgsums:

paru -G linux-cacule-rdb
cd linux-cacule-rdb
updpkgsums
makepkg -Si
2 Likes

I'm very intrigued by the look of this kernel. Been running Zen but with some changes as detailed in the optimise Garuda for laptop guide and something to do with 7th generation and so on changes. So I was just wondering how people are finding this? Like how's the battery life and performance compared to Zen when left as defaults?

This works fine ...but if i will build then after it same .. i think i need to wait till next version


╰─λ paru -G linux-cacule-rdb
:: Querying AUR...

:: Downloading PKGBUILDs...
(1/1) linux-cacule-rdb-5.13.4-4                     [------------------------------------------------------]

╭─smoky@smoky in ~ took 609ms
╰─λ cd linux-cacule-rdb

╭─smoky@smoky in repo: linux-cacule-rdb on  master [?] 
╰─λ updpkgsums
==> Empfange Quellen...
-> linux-5.13.tar.xz gefunden
-> patch-5.13.4.xz gefunden
-> config gefunden
-> 0001-arch-patches.patch gefunden
-> cacule-5.13.patch gefunden
-> rdb-5.13.patch gefunden
-> 0001-cpu-patches.patch gefunden
-> 0001-futex-resync-from-gitlab.collabora.com.patch gefunden
-> 0007-v5.13-futex2_interface.patch gefunden
-> 5.13-winesync.patch gefunden
-> 0001-zen-patches.patch gefunden
-> 0001-zen-Allow-MSR-writes-by-default.patch gefunden
-> 0001-bfq-patches.patch gefunden
-> 0001-block-patches.patch gefunden
-> 0001-fixes-miscellaneous.patch gefunden
-> 0001-bbr2-patches.patch gefunden
-> 0001-btrfs-patches.patch gefunden
-> 0001-android-export-symbold-and-enable-building-ashmem-an.patch gefunden
-> 0001-pf-patches.patch gefunden
-> 0001-lru-patches.patch gefunden
-> 0001-ntfs3-patches.patch gefunden
-> 0001-lrng-patches.patch gefunden
-> 0001-security-patches.patch gefunden
-> nohzfull.patch gefunden
-> 0001-alsa-patches.patch gefunden
-> 0001-zstd-upstream-patches.patch gefunden
-> 0001-clearlinux-patches.patch gefunden
-> 0001-v4l2loopback-patches.patch gefunden
==> Erstelle Prüfsummen für Quell-Dateien...

╭─smoky@smoky in repo: linux-cacule-rdb on  master [!?] took 1s
╰─λ makepkg -Si
==> Erstelle Paket: linux-cacule-rdb 5.13.4-4 (Sa 24 Jul 2021 19:42:31 CEST)
==> Empfange Quellen...
-> linux-5.13.tar.xz gefunden
-> patch-5.13.4.xz gefunden
-> config gefunden
-> 0001-arch-patches.patch gefunden
-> cacule-5.13.patch gefunden
-> rdb-5.13.patch gefunden
-> 0001-cpu-patches.patch gefunden
-> 0001-futex-resync-from-gitlab.collabora.com.patch gefunden
-> 0007-v5.13-futex2_interface.patch gefunden
-> 5.13-winesync.patch gefunden
-> 0001-zen-patches.patch gefunden
-> 0001-zen-Allow-MSR-writes-by-default.patch gefunden
-> 0001-bfq-patches.patch gefunden
-> 0001-block-patches.patch gefunden
-> 0001-fixes-miscellaneous.patch gefunden
-> 0001-bbr2-patches.patch gefunden
-> 0001-btrfs-patches.patch gefunden
-> 0001-android-export-symbold-and-enable-building-ashmem-an.patch gefunden
-> 0001-pf-patches.patch gefunden
-> 0001-lru-patches.patch gefunden
-> 0001-ntfs3-patches.patch gefunden
-> 0001-lrng-patches.patch gefunden
-> 0001-security-patches.patch gefunden
-> nohzfull.patch gefunden
-> 0001-alsa-patches.patch gefunden
-> 0001-zstd-upstream-patches.patch gefunden
-> 0001-clearlinux-patches.patch gefunden
-> 0001-v4l2loopback-patches.patch gefunden
==> Überprüfe source Dateien mit md5sums...
linux-5.13.tar.xz ... Durchgelaufen
patch-5.13.4.xz ... Durchgelaufen
config ... Durchgelaufen
0001-arch-patches.patch ... Durchgelaufen
cacule-5.13.patch ... Durchgelaufen
rdb-5.13.patch ... Durchgelaufen
0001-cpu-patches.patch ... Durchgelaufen
0001-futex-resync-from-gitlab.collabora.com.patch ... Durchgelaufen
0007-v5.13-futex2_interface.patch ... Durchgelaufen
5.13-winesync.patch ... Durchgelaufen
0001-zen-patches.patch ... Durchgelaufen
0001-zen-Allow-MSR-writes-by-default.patch ... Durchgelaufen
0001-bfq-patches.patch ... Durchgelaufen
0001-block-patches.patch ... Durchgelaufen
0001-fixes-miscellaneous.patch ... Durchgelaufen
0001-bbr2-patches.patch ... Durchgelaufen
0001-btrfs-patches.patch ... Durchgelaufen
0001-android-export-symbold-and-enable-building-ashmem-an.patch ... Durchgelaufen
0001-pf-patches.patch ... Durchgelaufen
0001-lru-patches.patch ... Durchgelaufen
0001-ntfs3-patches.patch ... Durchgelaufen
0001-lrng-patches.patch ... Durchgelaufen
0001-security-patches.patch ... Durchgelaufen
nohzfull.patch ... Durchgelaufen
0001-alsa-patches.patch ... Durchgelaufen
0001-zstd-upstream-patches.patch ... Durchgelaufen
0001-clearlinux-patches.patch ... Durchgelaufen
0001-v4l2loopback-patches.patch ... Durchgelaufen
==> Betrete fakeroot Umgebung...
==> Erstelle Quell-Paket...
-> Füge PKGBUILD hinzu...
-> Erstelle .SRCINFO Datei...
-> Füge config hinzu...
-> Komprimiere Quell-Paket...
==> Verlasse fakeroot Umgebung.
==> Quell-Paket erstellt: linux-cacule-rdb (Sa 24 Jul 2021 19:42:33 CEST)

Now with the new linux-cacule-rdb 5.13.5-1 it works agin without errors :slight_smile:

1 Like

Hey there i download cacUle and built it on my pc and went to sleep so it built it not installing the packages it timed out asking for passwords do i only have to install linux-cacule-5.13.5-1-x86_64.pkg.tar.zst and linux-cacule-headers-5.13.5-1-x86_64.pkg.tar.zst using pacman -U

Edit - Installed and THANK YOU it really minimized my freezes

2 Likes

Work fine on my AMD PC :slight_smile:

--  85M root 29 Jul 19:17 initramfs-linux-zen-fallback.img
--  80M root 29 Jul 19:17 initramfs-linux-cacule-fallback.img

--  61M root 29 Jul 19:17 initramfs-linux-zen.img
--  60M root 29 Jul 19:17 initramfs-linux-cacule.img
4 Likes

Some warnings not in function,
and ==> WARNING: package contains a reference to $srcdir
I use

 62) MNATIVE_AMD : AMD-Native optimizations autodetected by GCC.

Don't know if this is helpful?

In function "memcpy",
    inserted by "load_and_flush" at drivers/staging/media/atomisp/pci/hmm/hmm.c:404:3,
    inserted by "hmm_flush" at drivers/staging/media/atomisp/pci/hmm/hmm.c:431:9:
./include/linux/fortify-string.h:20:33: Warning: argument 1 is null, but non-null is expected [-Wnonnull].
   20 | #define __underlying_memcpy __builtin_memcpy
      | ^
./include/linux/fortify-string.h:191:16: Note: when substituting the macro "__underlying_memcpy".
  191 | return __underlying_memcpy(p, q, size);
      | ^~~~~~~~~~~~~~~~~~~
drivers/staging/media/atomisp/pci/hmm/hmm.c: In function "hmm_flush":
./include/linux/fortify-string.h:20:33: Note: in a call to the built-in function "__builtin_memcpy".
   20 | #define __underlying_memcpy __builtin_memcpy
      | ^
./include/linux/fortify-string.h:191:16: Note: if macro "__underlying_memcpy" is substituted
  191 | return __underlying_memcpy(p, q, size);
      | ^~~~~~~~~~~~~~~~~~~
In file included by drivers/staging/wlan-ng/p80211netdev.c:92:
drivers/staging/wlan-ng/cfg80211.c: In function "prism2_connect":
drivers/staging/wlan-ng/cfg80211.c:538:1: Warning: frame size of 2728 bytes exceeds 2048 bytes [-Wframe-larger-than=]
  538 | }
      | ^
drivers/mtd/nand/raw/nand_base.c: In function "nand_ecc_choose_conf":
drivers/mtd/nand/raw/nand_base.c:5546:13: Warning: "best_step" is not initialised, but is used [-Wuninitialized].
 5546 | int best_step, best_strength, best_ecc_bytes;
      | ^~~~~~~~~
drivers/mtd/nand/raw/nand_base.c:5546:39: Warning: "best_ecc_bytes" is not initialised but is used [-Wuninitialized]
 5546 | int best_step, best_strength, best_ecc_bytes;
      | ^~~~~~~~~~~~~~
drivers/staging/wfx/queue.o: warning: objtool: .text.unlikely: unexpected end of section
==> WARNING: package contains a reference to $srcdir
usr/lib/modules/5.13.7-1-cacule/build/scripts/gcc-plugins/structleak_plugin.so
usr/lib/modules/5.13.7-1-cacule/build/tools/objtool/objtool

 ╭─sgs@i3 in repo: linux-cacule on master [?] took 8m49s


Translated with www.DeepL.com/Translator (free version)

Boot perfect :slight_smile:

Summary
 ╭─sgs@i3 in ~ 
 ╰─λ systemd-analyze
Startup finished in 9.722s (firmware) + 1.948s (loader) + 2.091s (kernel) + 2.655s (userspace) = 16.418s 
graphical.target reached after 2.411s in userspace

 ╭─sgs@i3 in ~ took 3ms
 ╰─λ systemd-analyze blame
693ms udisks2.service
491ms mnt-sda3.mount
322ms linux-modules-cleanup.service
230ms mnt-sda1.mount
222ms mnt-sda2.mount
170ms dev-nvme0n1p2.device
 87ms cups.service
 83ms systemd-udev-trigger.service
 81ms blueman-mechanism.service
 79ms lm_sensors.service
 76ms user@1000.service
 69ms systemd-modules-load.service
 67ms systemd-journal-flush.service
 59ms systemd-zram-setup@zram0.service
 55ms systemd-remount-fs.service
 41ms sys-kernel-config.mount
 40ms systemd-journald.service
 39ms sys-fs-fuse-connections.mount
 38ms accounts-daemon.service
 35ms boot-efi.mount
 35ms systemd-hostnamed.service
 31ms systemd-timesyncd.service
 31ms systemd-vconsole-setup.service
 28ms polkit.service
 28ms systemd-logind.service
 27ms upower.service
 26ms systemd-udevd.service
 25ms mnt-sdb2.mount
 23ms systemd-fsck@dev-disk-by\x2duuid-B3B3\x2dD44D.service
 22ms NetworkManager.service
 22ms lightdm-plymouth.service
 20ms plymouth-quit-wait.service
 19ms mnt-nvme1.mount
 19ms dev-zram0.swap
 16ms systemd-rfkill.service
 14ms plymouth-start.service
 14ms systemd-random-seed.service
 13ms avahi-daemon.service
 12ms bluetooth.service
 11ms plymouth-read-write.service
 10ms systemd-tmpfiles-setup.service
  9ms systemd-sysctl.service
  7ms udisks2-zram-setup@zram0.service
  6ms colord.service
  5ms dev-hugepages.mount
  5ms systemd-tmpfiles-setup-dev.service
  5ms dev-mqueue.mount
  4ms modprobe@fuse.service
  4ms sys-kernel-debug.mount
  4ms user-runtime-dir@1000.service
  4ms sys-kernel-tracing.mount
  4ms root.mount
  4ms kmod-static-nodes.service
  3ms home.mount
  3ms modprobe@configfs.service
  3ms srv.mount
  3ms var-cache.mount
  3ms systemd-update-utmp.service
  3ms systemd-fsck@dev-disk-by\x2duuid-35075ddb\x2dbcf7\x2d47aa\x2da43a\x2de759a082af57.service
  3ms var-log.mount
  2ms var-tmp.mount
  2ms systemd-user-sessions.service
  2ms rtkit-daemon.service
  1ms tmp.mount
  1ms modprobe@drm.service
1 Like

5.13.7 Changes and some other informations:

I was some days in the hospital and couldnt update the kernel or anything else.

First the Changes:

  • Kernel got updated
  • NTFS driver got updated
  • Cacule/RDB patches got updated
  • NTFS patches got updated

Now i also provide linux-cacule-llvm and linux-cacule-rdb-llvm, which are compiled with lto then.

Both are in the aur.

Next suggestion, as long you dont use the normal RDB version, i can really suggest to use ananicy-cpp, he got finally all his features mostly done and supports autogroup now. the normal ananicy disabled it by default.

RDB does not use autgroup, but @Hamadmarri is working on a solution and my latest tests seems good so far!

Also i provide Generic3 and Generic Builds in my own repo.

https://build.cachyos.org

Check it out.

For mirrorlist and pacman.conf if interest is there just hit me up.

4 Likes

This could be a GCC Error or maybe from other modules.

Check if the problem still exist. I can also sent you a GCC 11.2 Version.

2 Likes
 pacman -Qi gcc
Name                     : gcc
Version                  : 11.1.0-1

Let me see :slight_smile:

2 Likes