Re: [RFC/PATCHSET 0/6] perf kmem: Implement page allocation analysis (v1)

[Ingo Molnar <mingo@kernel.org>]

* Namhyung Kim <namhyung@kernel.org> wrote:

> Hi Ingo,
> 
> On Thu, Mar 12, 2015 at 04:54:22PM +0100, Ingo Molnar wrote:
> > 
> > * Namhyung Kim <namhyung@kernel.org> wrote:
> > 
> > > > I.e. something like this (mockup) output:
> > > > 
> > > >    SUMMARY (page allocator)
> > > >    ========================
> > > > 
> > > >    Pages allocated+freed:       12,593   [     51,630,080 bytes ]
> > > > 
> > > >    Pages allocated-only:         2,342   [      1,235,010 bytes ]
> > > >    Pages freed-only:                67   [        135,311 bytes ]
> > > > 
> > > >    Page allocation failures :        0
> > > 
> > > Looks a lot better!
> > > 
> > > One thing I need to tell you is that the numbers are not pages but 
> > > requests.
> > 
> > Yes, but in the MM code we tend to call larger order allocations 
> > 'pages' as well: higher order pages, such as a 2MB hugepage, or a 8K 
> > order-1 page. So at least in MM-speak it should be OK to call them 
> > 'pages'.
> > 
> > But your call!
> 
> How about this?
> 
>   SUMMARY (page allocator)
>   ========================
>   Total allocation requests   :        9,015  [     37,200 Kbytes ]  (A)
>   Total free requests         :        8,093  [     33,176 Kbytes ]  (B)
> 
>   Total alloc+freed requests  :        7,985  [     32,732 Kbytes ]  (C)
>   Total alloc-only requests   :        1,030  [      4,468 Kbytes ]  (D)
>   Total free-only requests    :          108  [        444 Kbytes ]  (E)
> 
>   Total allocation failure    :            0  [          0 Kbytes ]

s/failure/failures
s/Kbytes/KB

I'd leave a bit more space for the numbers, for up into billions of 
requests and terabytes of data. Other than that, sounds good to me!

Thanks,

	Ingo