Re: [PATCH] x86/hyperv: Pass on the lpj value from host to guest

[Wei Liu <wei.liu@kernel.org>]

On Tue, Feb 07, 2023 at 03:24:47PM -0800, Nuno Das Neves wrote:
> On 2/6/2023 12:49 PM, Stanislav Kinsburskii wrote:
> > From: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> > 
> > And have it preset.

In the future please add a blank line between two paragraphs.

> > This change allows to significantly reduce time to bring up guest SMP
> > configuration as well as make sure the guest won't get inaccurate
> > calibration results due to "noisy neighbour" situation.
> > 

This looks like a good idea. 0293615f3fb9 was committed in 2008, so
we're very late to the party. Better late than never though.

If I hear no objections in a few days' time I will apply this to
hyperv-next with Nuno's Rb tag.

Thanks,
Wei.

> > Below are the numbers for 16 VCPU guest before the patch (~1300 msec)
> > 
> > [    0.562938] x86: Booting SMP configuration:
> > ...
> > [    1.859447] smp: Brought up 1 node, 16 CPUs
> > 
> > and after the patch (~130 msec):
> > 
> > [    0.445079] x86: Booting SMP configuration:
> > ...
> > [    0.575035] smp: Brought up 1 node, 16 CPUs
> > 
> > This change is inspired by commit 0293615f3fb9 ("x86: KVM guest: use
> > paravirt function to calculate cpu khz").
> > 
> > Signed-off-by: Stanislav Kinsburskiy <stanislav.kinsburskiy@gmail.com>
> > CC: "K. Y. Srinivasan" <kys@microsoft.com>
> > CC: Haiyang Zhang <haiyangz@microsoft.com>
> > CC: Wei Liu <wei.liu@kernel.org>
> > CC: Dexuan Cui <decui@microsoft.com>
> > CC: Thomas Gleixner <tglx@linutronix.de>
> > CC: Ingo Molnar <mingo@redhat.com>
> > CC: Borislav Petkov <bp@alien8.de>
> > CC: Dave Hansen <dave.hansen@linux.intel.com>
> > CC: x86@kernel.org
> > CC: "H. Peter Anvin" <hpa@zytor.com>
> > CC: linux-hyperv@vger.kernel.org
> > CC: linux-kernel@vger.kernel.org
> > ---
> >  arch/x86/kernel/cpu/mshyperv.c |   21 +++++++++++++++++++++
> >  1 file changed, 21 insertions(+)
> > 
> > diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> > index dedec2f23ad1..0282b2e96cc2 100644
> > --- a/arch/x86/kernel/cpu/mshyperv.c
> > +++ b/arch/x86/kernel/cpu/mshyperv.c
> > @@ -320,6 +320,21 @@ static void __init hv_smp_prepare_cpus(unsigned int max_cpus)
> >  }
> >  #endif
> >  
> > +static void __init __maybe_unused hv_preset_lpj(void)
> > +{
> > +	unsigned long khz;
> > +	u64 lpj;
> > +
> > +	if (!x86_platform.calibrate_tsc)
> > +		return;
> > +
> > +	khz = x86_platform.calibrate_tsc();
> > +
> > +	lpj = ((u64)khz * 1000);
> > +	do_div(lpj, HZ);
> > +	preset_lpj = lpj;
> > +}
> > +
> >  static void __init ms_hyperv_init_platform(void)
> >  {
> >  	int hv_max_functions_eax;
> > @@ -521,6 +536,12 @@ static void __init ms_hyperv_init_platform(void)
> >  
> >  	/* Register Hyper-V specific clocksource */
> >  	hv_init_clocksource();
> > +
> > +	/*
> > +	 * Preset lpj to make calibrate_delay a no-op, which is turn helps to
> > +	 * speed up secondary cores initialization.
> > +	 */
> > +	hv_preset_lpj();
> >  #endif
> >  	/*
> >  	 * TSC should be marked as unstable only after Hyper-V
> > 
> 
> Reviewed-by: Nuno Das Neves <nunodasneves@linux.microsoft.com>