From mboxrd@z Thu Jan  1 00:00:00 1970
From: mark.rutland@arm.com (Mark Rutland)
Date: Thu, 3 Dec 2015 18:22:27 +0000
Subject: [PATCH] arm64: calculate the various pages number to show
In-Reply-To: <5657B750.6080908@huawei.com>
References: <1448458872-39897-1-git-send-email-zhongjiang@huawei.com>
 <20151125150448.GD12434@leverpostej> <56571FBC.2020300@huawei.com>
 <20151126154926.GG32343@leverpostej> <5657B750.6080908@huawei.com>
Message-ID: <20151203182227.GC3527@leverpostej>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Fri, Nov 27, 2015 at 09:52:16AM +0800, Xishi Qiu wrote:
> On 2015/11/26 23:49, Mark Rutland wrote:
> 
> > On Thu, Nov 26, 2015 at 11:05:32PM +0800, zhong jiang wrote:
> >> On 2015/11/25 23:04, Mark Rutland wrote:
> >>> On Wed, Nov 25, 2015 at 09:41:12PM +0800, zhongjiang wrote:
> >>>> This patch add the interface to show the number of 4KB or 64KB page,
> >>>> aims to statistics the number of different types of pages.
> >>>
> >>> What is this useful for? Why do we want it?
> >>>
> >>> What does it account for, just the swapper?
> >>>
> >>
> >> The patch is wirtten when I was in backport set_memory_ro. It can be used to
> >> detect whether there is a large page spliting and merging. large page will
> >> significantly reduce the TLB miss, and improve the system performance.
> > 
> > Ok, but typically the user isn't going to be able to do much with this
> > information. It feels more like something that should be in the page
> > table dump code (where we can calculate the values as we walk the
> > tables).
> > 
> > What is it intended to account for?
> > 
> > The entire swapper?
> > 
> > Just the linear mapping?
> 
> Hi Mark,
> 
> x86 has this information when cat /proc/meminfo, so how about just
> like x86 to show it?

The fact that another architecture has some implementation doesn't
necessarily mean it's a good idea. In this case there are concerns that
don't apply to x86, in that we support a number of page sizes, and
anything reading this needs to handle that fact.

If there's a sensible use-case, then I am not opposed to this. I don't
see the point in adding it just because we can.

A prerequisite for adding it is knowing precisely what it is intended to
describe. Otherwise it's impossible to review.

Thanks,
Mark.