Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
From:   Michael Ellerman <mpe@ellerman.id.au>
To:     Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc:     linuxppc-dev <linuxppc-dev@lists.ozlabs.org>,
        Srikar Dronamraju <srikar@linux.vnet.ibm.com>,
        Aboorva Devarajan <aboorvad@linux.vnet.ibm.com>,
        Shrikanth Hegde <sshegde@linux.vnet.ibm.com>,
        Ajay Kaher <akaher@vmware.com>,
        Alexey Makhalov <amakhalov@vmware.com>,
        VMware PV-Drivers Reviewers <pv-drivers@vmware.com>,
        Nicholas Piggin <npiggin@gmail.com>,
        Christophe Leroy <christophe.leroy@csgroup.eu>,
        virtualization@lists.linux-foundation.org, x86@kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] powerpc/paravirt: Improve vcpu_is_preempted
In-Reply-To: <20231018155838.2332822-1-srikar@linux.vnet.ibm.com>
References: <20231018155838.2332822-1-srikar@linux.vnet.ibm.com>
Date:   Thu, 19 Oct 2023 12:09:38 +1100
Message-ID: <877cnj76zx.fsf@mail.lhotse>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

Hi Srikar,

Srikar Dronamraju <srikar@linux.vnet.ibm.com> writes:
> PowerVM Hypervisor dispatches on a whole core basis. In a shared LPAR, a
> CPU from a core that is CEDED or preempted may have a larger latency. In
> such a scenario, its preferable to choose a different CPU to run.
>
> If one of the CPUs in the core is active, i.e neither CEDED nor
> preempted, then consider this CPU as not preempted.
>
> Also if any of the CPUs in the core has yielded but OS has not requested
> CEDE or CONFER, then consider this CPU to be preempted.

I think the change is OK, but the change log and comments are slightly
confusing IMHO.

In several places you use "this CPU", but that usually means "the CPU
the code is currently executing on".

I think it would be clearer if you used eg. "target CPU" or something to
make it clear that you're not talking about the currently executing CPU.

cheers

> Correct detection of preempted CPUs is important for detecting idle
> CPUs/cores in task scheduler.
>
> Changelog:
> v1 -> v2: Handle lppaca_of(cpu) in !PPC_SPLPAR case.
> v1: https://lore.kernel.org/r/20231009051740.17683-1-srikar%40linux.vnet.ibm.com
> 1. Fixed some compilation issues reported by kernelbot
> a. https://lore.kernel.org/oe-kbuild-all/202310102341.K0sgoqQL-lkp@intel.com/
> b.  https://lore.kernel.org/oe-kbuild-all/202310091636.lElmJkYV-lkp@intel.com/
> 2. Resolved comments from Shrikanth

That change log should appear below the break "---".

> Tested-by: Aboorva Devarajan <aboorvad@linux.vnet.ibm.com>
> Reviewed-by: Shrikanth Hegde <sshegde@linux.vnet.ibm.com>
> Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
> ---
>  arch/powerpc/include/asm/paravirt.h | 42 ++++++++++++++++++++++++++---
>  1 file changed, 39 insertions(+), 3 deletions(-)
>
> diff --git a/arch/powerpc/include/asm/paravirt.h b/arch/powerpc/include/asm/paravirt.h
> index e08513d73119..0372b0093f72 100644
> --- a/arch/powerpc/include/asm/paravirt.h
> +++ b/arch/powerpc/include/asm/paravirt.h
> @@ -71,6 +71,11 @@ static inline void yield_to_any(void)
>  {
>  	plpar_hcall_norets_notrace(H_CONFER, -1, 0);
>  }
> +
> +static inline bool is_vcpu_idle(int vcpu)
> +{
> +	return lppaca_of(vcpu).idle;
> +}
>  #else
>  static inline bool is_shared_processor(void)
>  {
> @@ -100,6 +105,10 @@ static inline void prod_cpu(int cpu)
>  	___bad_prod_cpu(); /* This would be a bug */
>  }
>  
> +static inline bool is_vcpu_idle(int vcpu)
> +{
> +	return false;
> +}
>  #endif
>  
>  #define vcpu_is_preempted vcpu_is_preempted
> @@ -121,9 +130,19 @@ static inline bool vcpu_is_preempted(int cpu)
>  	if (!is_shared_processor())
>  		return false;
>  
> +	if (!(yield_count_of(cpu) & 1))
> +		return false;

Would be nice for that to have a short comment too.

> +
> +	/*
> +	 * If CPU has yielded to Hypervisor but OS has not requested idle
> +	 * then this CPU is definitely preempted.

eg. If the target CPU has yielded to the Hypervisor, but the OS has not
requested idle then the target CPU has definitely been preempted.

> +	 */
> +	if (!is_vcpu_idle(cpu))
> +		return true;
> +
>  #ifdef CONFIG_PPC_SPLPAR
>  	if (!is_kvm_guest()) {
> -		int first_cpu;
> +		int first_cpu, i;
>  
>  		/*
>  		 * The result of vcpu_is_preempted() is used in a
> @@ -149,11 +168,28 @@ static inline bool vcpu_is_preempted(int cpu)
>  		 */
>  		if (cpu_first_thread_sibling(cpu) == first_cpu)
>  			return false;
> +
> +		/*
> +		 * If any of the threads of this core is not preempted or
> +		 * ceded, then consider this CPU to be non-preempted
> +		 */

eg. If any of the threads of the target CPU's core are not preempted or
ceded, then consider that the target CPU is also not preempted.

> +		first_cpu = cpu_first_thread_sibling(cpu);
> +		for (i = first_cpu; i < first_cpu + threads_per_core; i++) {
> +			if (i == cpu)
> +				continue;
> +			if (!(yield_count_of(i) & 1))
> +				return false;
> +			if (!is_vcpu_idle(i))
> +				return true;
> +		}
>  	}
>  #endif
>  
> -	if (yield_count_of(cpu) & 1)
> -		return true;
> +	/*
> +	 * None of the threads in this core are running but none of
> +	 * them were preempted too. Hence assume the thread to be
> +	 * non-preempted.
> +	 */
>  	return false;
>  }
>  
>
> base-commit: eddc90ea2af5933249ea1a78119f2c8ef8d07156
> -- 
> 2.31.1