| Message ID | 20210709043713.887098-4-senozhatsky@chromium.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | arm64:kvm: teach guest sched that VCPUs can be preempted | expand |
On Fri, 09 Jul 2021 05:37:12 +0100, Sergey Senozhatsky <senozhatsky@chromium.org> wrote: > > vcpu_is_preempted() now can represent the actual state of > the VCPU, so the scheduler can make better decisions when > it picks the idle CPU to enqueue a task on. > > Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org> > --- > arch/arm64/include/asm/spinlock.h | 18 ++++++++++-------- > 1 file changed, 10 insertions(+), 8 deletions(-) > > diff --git a/arch/arm64/include/asm/spinlock.h b/arch/arm64/include/asm/spinlock.h > index 0525c0b089ed..1d579497e1b8 100644 > --- a/arch/arm64/include/asm/spinlock.h > +++ b/arch/arm64/include/asm/spinlock.h > @@ -7,21 +7,23 @@ > > #include <asm/qspinlock.h> > #include <asm/qrwlock.h> > +#include <asm/paravirt.h> > > /* See include/linux/spinlock.h */ > #define smp_mb__after_spinlock() smp_mb() > > -/* > - * Changing this will break osq_lock() thanks to the call inside > - * smp_cond_load_relaxed(). > - * > - * See: > - * https://lore.kernel.org/lkml/20200110100612.GC2827@hirez.programming.kicks-ass.net > - */ Why are you deleting this? Please explain your reasoning in the commit message. It seems to me that it still makes complete sense when CONFIG_PARAVIRT is not defined. > #define vcpu_is_preempted vcpu_is_preempted > -static inline bool vcpu_is_preempted(int cpu) > + > +#ifdef CONFIG_PARAVIRT > +static inline bool vcpu_is_preempted(unsigned int cpu) > +{ > + return paravirt_vcpu_is_preempted(cpu); > +} > +#else > +static inline bool vcpu_is_preempted(unsigned int cpu) > { > return false; > } > +#endif /* CONFIG_PARAVIRT */ > > #endif /* __ASM_SPINLOCK_H */ Thanks, M.
On (21/07/12 16:47), Marc Zyngier wrote: > > #include <asm/qspinlock.h> > > #include <asm/qrwlock.h> > > +#include <asm/paravirt.h> > > > > /* See include/linux/spinlock.h */ > > #define smp_mb__after_spinlock() smp_mb() > > > > -/* > > - * Changing this will break osq_lock() thanks to the call inside > > - * smp_cond_load_relaxed(). > > - * > > - * See: > > - * https://lore.kernel.org/lkml/20200110100612.GC2827@hirez.programming.kicks-ass.net > > - */ > > Why are you deleting this? Please explain your reasoning in the commit > message. It seems to me that it still makes complete sense when > CONFIG_PARAVIRT is not defined. You are right. I'll move it to !PARAVIRT #else-branch.
diff --git a/arch/arm64/include/asm/spinlock.h b/arch/arm64/include/asm/spinlock.h index 0525c0b089ed..1d579497e1b8 100644 --- a/arch/arm64/include/asm/spinlock.h +++ b/arch/arm64/include/asm/spinlock.h @@ -7,21 +7,23 @@ #include <asm/qspinlock.h> #include <asm/qrwlock.h> +#include <asm/paravirt.h> /* See include/linux/spinlock.h */ #define smp_mb__after_spinlock() smp_mb() -/* - * Changing this will break osq_lock() thanks to the call inside - * smp_cond_load_relaxed(). - * - * See: - * https://lore.kernel.org/lkml/20200110100612.GC2827@hirez.programming.kicks-ass.net - */ #define vcpu_is_preempted vcpu_is_preempted -static inline bool vcpu_is_preempted(int cpu) + +#ifdef CONFIG_PARAVIRT +static inline bool vcpu_is_preempted(unsigned int cpu) +{ + return paravirt_vcpu_is_preempted(cpu); +} +#else +static inline bool vcpu_is_preempted(unsigned int cpu) { return false; } +#endif /* CONFIG_PARAVIRT */ #endif /* __ASM_SPINLOCK_H */
vcpu_is_preempted() now can represent the actual state of the VCPU, so the scheduler can make better decisions when it picks the idle CPU to enqueue a task on. Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org> --- arch/arm64/include/asm/spinlock.h | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-)