Bug#1070083: linux: Do we still need sched-autogroup-disabled.patch in 2024?
On Tue, 2024-04-30 at 12:42 +0200, Salvatore Bonaccorso wrote: > Hi, > > On Tue, Apr 30, 2024 at 02:10:20AM +0500, Alex Volkov wrote: > > Source: linux > > Severity: normal > > > > Dear Maintainer, > > > > I can't see why something which can be done with a kernel boot parameter or > > a > > sysctl variable > > needs to be forced in the source since 2011. Also, the very existence of > > this > > new default is non-transparent for anyone relying on official kernel > > documentation. Is it even mentioned in any README or something? Because I > > couldn't find anything. > > It was set as such a long time ago when we enabled the option: > > https://salsa.debian.org/kernel-team/linux/-/commit/25e71bd23902b424ec9457244db24faad8b4359f > > I suspect the idea was to make it possible for users to use this new > feature, but only as a opt-in, while it was still very fresh. > > Maybe Ben remembers. Yes that was exactly it. > But I guess we could drop the Debian specific patch. I had a look at other distribution configurations to see if CONFIG_SCHED_AUTOGROUP is sensible to enable by default: - Fedora enables it except on RT configurations - SUSE disables it - Ubuntu enables it Neither Fedora nor Ubuntu has a similar patch, though it's possible they change the default via a sysctl.conf file. So I think we should: - Drop the patch - Disable CONFIG_SCHED_AUTOGRUP on RT Ben. -- Ben Hutchings - Debian developer, member of kernel, installer and LTS teams signature.asc Description: This is a digitally signed message part
Bug#1070083: linux: Do we still need sched-autogroup-disabled.patch in 2024?
Hi, On Tue, Apr 30, 2024 at 02:10:20AM +0500, Alex Volkov wrote: > Source: linux > Severity: normal > > Dear Maintainer, > > I can't see why something which can be done with a kernel boot parameter or a > sysctl variable > needs to be forced in the source since 2011. Also, the very existence of this > new default is non-transparent for anyone relying on official kernel > documentation. Is it even mentioned in any README or something? Because I > couldn't find anything. It was set as such a long time ago when we enabled the option: https://salsa.debian.org/kernel-team/linux/-/commit/25e71bd23902b424ec9457244db24faad8b4359f I suspect the idea was to make it possible for users to use this new feature, but only as a opt-in, while it was still very fresh. Maybe Ben remembers. But I guess we could drop the Debian specific patch. Regards, Salvatore
Bug#1070083: linux: Do we still need sched-autogroup-disabled.patch in 2024?
Source: linux Severity: normal Dear Maintainer, I can't see why something which can be done with a kernel boot parameter or a sysctl variable needs to be forced in the source since 2011. Also, the very existence of this new default is non-transparent for anyone relying on official kernel documentation. Is it even mentioned in any README or something? Because I couldn't find anything. -- System Information: Debian Release: 12.5 APT prefers stable-updates APT policy: (991, 'stable-updates'), (991, 'stable-security'), (991, 'stable'), (99, 'testing'), (90, 'unstable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 6.6.13-bootes0-p-1000 (SMP w/16 CPU threads; PREEMPT) Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE Locale: LANG=en_DK.UTF-8, LC_CTYPE=en_DK.UTF-8 (charmap=UTF-8), LANGUAGE=en_GB:en_US Shell: /bin/sh linked to /usr/bin/dash Init: sysvinit (via /sbin/init) LSM: AppArmor: enabled
