Re: [PATCH] x86/rapl: Do not load in a guest

From: "H. Peter Anvin" <hpa@zytor.com>
To: Peter Zijlstra <peterz@infradead.org>,
	Jacob Pan <jacob.jun.pan@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>, Borislav Petkov <bp@alien8.de>,
	Amy Wiles <amy.l.wiles@intel.com>,
	"Rafael J. Wysocki" <rafael.j.wysocki@intel.com>,
	LKML <linux-kernel@vger.kernel.org>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	Ingo Molnar <mingo@redhat.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: Re: [PATCH] x86/rapl: Do not load in a guest
Date: Fri, 04 Dec 2015 14:39:26 -0800	[thread overview]
Message-ID: <1142766D-212C-47CF-8400-89616715BCF8@zytor.com> (raw)
In-Reply-To: <20151204221446.GK17308@twins.programming.kicks-ass.net>

On December 4, 2015 2:14:46 PM PST, Peter Zijlstra <peterz@infradead.org> wrote:
>On Fri, Dec 04, 2015 at 09:51:02AM -0800, Jacob Pan wrote:
>> On Fri, 4 Dec 2015 09:22:56 +0100
>> Peter Zijlstra <peterz@infradead.org> wrote:
>> 
>> > Also, yuck @ powercap/intel_rapl.c for doing rdmsr_on_cpu() +
>> > wrmsr_on_cpu() all over the place.
>> Can you please be more specific? is the concern related to the
>> overhead of IPI? I am doing these calls based on MSR CPU scope and
>> consider the fact that access is less frequent.
>
>Yeah, its just offensive to do an IPI to read a value, then twiddle a
>few bits on the value and then IPI again to store the value.
>
>I know its low freq, and that MSR access is slow, but *groan*.

Yes, for that it would be better to invoke a common routine to do all the accesses on the target CPU.  MSR accesses may be slow, but IPIs are way slower.
-- 
Sent from my Android device with K-9 Mail. Please excuse brevity and formatting.