[RFC PATCH 1/7] arm64: Use physical counter for in-kernel reads

[Christoffer Dall]

Hi Marc,

On Thu, Jan 05, 2017 at 06:11:14PM +0000, Marc Zyngier wrote:
> [adding the arm64 maintainers, plus Mark as arch timer maintainer]

Right, sorry, I should have done that already.

> 
> On 10/12/16 20:47, Christoffer Dall wrote:
> > Using the physical counter allows KVM to retain the offset between the
> > virtual and physical counter as long as it is actively running a VCPU.
> > 
> > As soon as a VCPU is released, another thread is scheduled or we start
> > running userspace applications, we reset the offset to 0, so that VDSO
> > operations can still read the virtual counter and get the same view of
> > time as the kernel.
> > 
> > This opens up potential improvements for KVM performance.
> > 
> > VHE kernels or kernels using the virtual timer are unaffected by this.
> > 
> > Signed-off-by: Christoffer Dall <christoffer.dall at linaro.org>
> > ---
> >  arch/arm64/include/asm/arch_timer.h  | 6 ++++--
> >  drivers/clocksource/arm_arch_timer.c | 2 +-
> >  2 files changed, 5 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/arm64/include/asm/arch_timer.h b/arch/arm64/include/asm/arch_timer.h
> > index eaa5bbe..cec2549 100644
> > --- a/arch/arm64/include/asm/arch_timer.h
> > +++ b/arch/arm64/include/asm/arch_timer.h
> > @@ -139,11 +139,13 @@ static inline void arch_timer_set_cntkctl(u32 cntkctl)
> >  
> >  static inline u64 arch_counter_get_cntpct(void)
> >  {
> > +	u64 cval;
> >  	/*
> >  	 * AArch64 kernel and user space mandate the use of CNTVCT.
> >  	 */
> > -	BUG();
> > -	return 0;
> > +	isb();
> > +	asm volatile("mrs %0, cntpct_el0" : "=r" (cval));
> > +	return cval;
> >  }
> >  
> >  static inline u64 arch_counter_get_cntvct(void)
> > diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
> > index 73c487d..a5b0789 100644
> > --- a/drivers/clocksource/arm_arch_timer.c
> > +++ b/drivers/clocksource/arm_arch_timer.c
> > @@ -597,7 +597,7 @@ static void __init arch_counter_register(unsigned type)
> >  
> >  	/* Register the CP15 based counter if we have one */
> >  	if (type & ARCH_CP15_TIMER) {
> > -		if (IS_ENABLED(CONFIG_ARM64) || arch_timer_uses_ppi == VIRT_PPI)
> > +		if (arch_timer_uses_ppi == VIRT_PPI || is_kernel_in_hyp_mode())
> 
> Why do we have this is_kernel_in_hyp_mode clause? I can't think of any
> reason for a VHE kernel to use the virtual counter at all...
> 

Good question.  I think I just didn't want to change behavior from the
existing functionality mre than necessary.

Note that on a VHE kernel this will be the EL2 virtual counter, not the
EL1 virtual counter, due to the register redirection.  Are the virtual
and physical EL2 counters always equivalent on a VHE system?


> >  			arch_timer_read_counter = arch_counter_get_cntvct;
> >  		else
> >  			arch_timer_read_counter = arch_counter_get_cntpct;
> > 
> 

Thanks,
-Christoffer