Re: [PATCH v2 0/4] Powerpc: Better preemption for shared processor

From: Michael Ellerman
Date: Thu Dec 10 2020 - 06:31:35 EST


On Wed, 2 Dec 2020 10:34:52 +0530, Srikar Dronamraju wrote:
> Currently, vcpu_is_preempted will return the yield_count for
> shared_processor. On a PowerVM LPAR, Phyp schedules at SMT8 core boundary
> i.e all CPUs belonging to a core are either group scheduled in or group
> scheduled out. This can be used to better predict non-preempted CPUs on
> PowerVM shared LPARs.
>
> perf stat -r 5 -a perf bench sched pipe -l 10000000 (lesser time is better)
>
> [...]

Applied to powerpc/next.

[1/4] powerpc: Refactor is_kvm_guest() declaration to new header
https://git.kernel.org/powerpc/c/92cc6bf01c7f4c5cfefd1963985c0064687ebeda
[2/4] powerpc: Rename is_kvm_guest() to check_kvm_guest()
https://git.kernel.org/powerpc/c/16520a858a995742c2d2248e86a6026bd0316562
[3/4] powerpc: Reintroduce is_kvm_guest() as a fast-path check
https://git.kernel.org/powerpc/c/a21d1becaa3f17a97b933ffa677b526afc514ec5
[4/4] powerpc/paravirt: Use is_kvm_guest() in vcpu_is_preempted()
https://git.kernel.org/powerpc/c/ca3f969dcb111d35674b66bdcb72beb2c426b9b5

cheers