Running ./ps.test/ps_personality.exp ... Running ./ps.test/ps_sched_batch.exp ... FAIL: ps SCHED_BATCH scheduler === ps Summary === ------------------------------------------------------------------- This is an unstable amd64 chroot image at a tinderbox (==build bot) name: 17.1_no-multilib-test-20200203-215003 ------------------------------------------------------------------- gcc-config -l: [1] x86_64-pc-linux-gnu-9.2.0 * Available Python interpreters, in order of preference: [1] python3.6 [2] python3.8 (fallback) [3] python3.7 (fallback) [4] python2.7 (fallback) repository: ==> /var/db/repos/gentoo/metadata/timestamp.chk <== Tue, 04 Feb 2020 07:33:32 +0000 emerge -qpvO sys-process/procps [ebuild U ] sys-process/procps-3.3.16 [3.3.15-r1] USE="kill ncurses nls (split-usr) test* unicode -elogind -modern-top (-selinux) -static-libs -systemd"
Created attachment 611570 [details] emerge-info.txt
Created attachment 611572 [details] emerge-history.txt
Created attachment 611574 [details] environment
Created attachment 611576 [details] etc.portage.tbz2
Created attachment 611578 [details] logs.tbz2
Created attachment 611580 [details] sys-process:procps-3.3.16:20200204-083947.log
Created attachment 611582 [details] temp.tbz2
I only get this error when the package is emerged with SCHED_IDLE scheduling policy, i.e. via PORTAGE_IONICE_COMMAND="chrt -i -p 0 \${PID}" commenting that out / creating a seperate package.env fixes it - looks to me like the sched_batch test fails if it's not started as SCHED_RR tested with both 3.3.15-r1 and 3.3.16-r2
Update summary. [something] is used when something refers to an overlay
Is there any way to restrict PORTAGE_SCHEDULING_POLICY on a per-ebuild basis? Even if it's just a bail-out in pkg_setup etc?
