Package Details: linux-pf 6.12.pf1-1

Git Clone URL: https://aur.archlinux.org/linux-pf.git (read-only, click to copy)
Package Base: linux-pf
Description: The pf-kernel and modules
Upstream URL: https://pfkernel.natalenko.name
Keywords: bbr bbr3 kernel ksm linux linux-pf pf-kernel uksm uksmd v4l2loopback zstd
Licenses: GPL-2.0-only
Provides: KSMBD-MODULE, linux-pf, NTFS3-MODULE, UKSMD-BUILTIN, V4L2LOOPBACK-MODULE, VIRTUALBOX-GUEST-MODULES, WIREGUARD-MODULE
Replaces: virtualbox-guest-modules-arch, wireguard-arch
Submitter: nous
Maintainer: post-factum
Last Packager: post-factum
Votes: 209
Popularity: 0.21
First Submitted: 2011-07-24 12:01 (UTC)
Last Updated: 2024-11-27 09:08 (UTC)

Dependencies (18)

Required by (15)

Sources (2)

Pinned Comments

post-factum commented on 2023-09-25 20:30 (UTC) (edited on 2024-10-08 14:21 (UTC) by post-factum)

Official binary builds for various x86_64 ISA levels are available here.

Latest Comments

« First ‹ Previous 1 .. 45 46 47 48 49 50 51 52 53 54 55 .. 106 Next › Last »

nous commented on 2012-12-21 13:39 (UTC)

Thaodan: Good to know, as I've included in the PKGBUILD the option to follow the 3-number versioning in 3.7 (right after the CPU-naming selection); it leaves SUBLEVEL at the original upstream number with no further tweaking.

nous commented on 2012-12-21 13:33 (UTC)

About the proposed optimization schemes and the benchmarks graysky provided: the slowest tested CPU vs GENERIC gives a echo '(191.9256-191.8459)/191.8459*100'|bc -l .04154375986143045000 WOW! 0.0415% gain in performance! 79.7 timeless and torturing milliseconds after 191 seconds of wait. And, mind you, that gain is against the GENERIC kernel. Now, if graysky is kind enough to compare a core2 kernel vs an avx-i, I bet my sideburns that the measured gain won't be greater than 1 millisecond, if that. It's just nonsensical and futile; optimize your userspace, not the kernel. Again, see http://tinyurl.com/cp7wjv5 and read the comments [0]. Thaodan: thanks for your work but I won't spend 20 hours (40 minutes * 15 CPUs * 2 arches) compiling 30 different kernels which run barely 1/10th of a second faster in benchmarks than linux-pf (and 1/1000th second faster than linux-pf-[core2|k8]. Not to mention the nvidia-pf mess that'll ensue. graysky: I've queued the watchdog patch for inclusion, thanks. [0] Read http://tinyurl.com/cp7wjv5. It was the last thing I ever did on Gentoo before moving to Arch.

Thaodan commented on 2012-12-21 12:38 (UTC)

nous: i fixed it by setting SUBLEVEL in /usr/src/linux-pf-3.7/Makefile to "1"

nous commented on 2012-12-21 11:59 (UTC)

Thaodan: I do. % cat /usr/lib/modules/3.6-pf/build/include/linux/version.h #define LINUX_VERSION_CODE 198144 #define KERNEL_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c)) % cat /usr/lib/modules/3.7-pf/build/include/linux/version.h cat: /usr/lib/modules/3.7-pf/build/include/linux/version.h: No such file or directory % cat /usr/lib/modules/3.7.0-1-ARCH/build/include/linux/version.h cat: /usr/lib/modules/3.7.0-1-ARCH/build/include/linux/version.h: No such file or directory It would seem version.h is missing from 3.7. A simple copy from a *-headers-3.6 will temporarily resolve this.

Thaodan commented on 2012-12-21 05:25 (UTC)

Has someone an idea why drivers like nvidia don't build if SUBLEVEL in /usr/src/linux-pf-3.7 is empty?

Thaodan commented on 2012-12-21 03:40 (UTC)

My modifed PKGBUILD with proper pkgnames and description for optimized builds, I addded grayskys docs and disable numa (if not enabled) and I the upstream fix for watchdog regression as suggest: https://docs.google.com/open?id=0B_FQtIHEHNn8UTBQempFcl9jb1E

graysky commented on 2012-12-20 22:32 (UTC)

/type on/typo/

« First ‹ Previous 1 .. 45 46 47 48 49 50 51 52 53 54 55 .. 106 Next › Last »