From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
To: Dave Hansen <dave@sr71.net>, linux-kernel@vger.kernel.org
Cc: Dave Hansen <dave@sr71.net>, hch@lst.de, akpm@linux-foundation.org,
        dan.j.williams@intel.com, khandual@linux.vnet.ibm.com, vbabka@suse.cz,
        linux-mm@kvack.org, linux-arch@vger.kernel.org
Subject: Re: [PATCH] proc: mm: export PTE sizes directly in smaps (v3)
In-Reply-To: <20161129201703.CE9D5054@viggo.jf.intel.com>
References: <20161129201703.CE9D5054@viggo.jf.intel.com>
Date: Fri, 02 Dec 2016 21:13:19 +0530
MIME-Version: 1.0
Content-Type: text/plain
Message-Id: <877f7ie3qg.fsf@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 979
Lines: 32

Dave Hansen <dave@sr71.net> writes:

> Andrew, you can drop proc-mm-export-pte-sizes-directly-in-smaps-v2.patch,
> and replace it with this.
>
.....

> diff -puN mm/hugetlb.c~smaps-pte-sizes mm/hugetlb.c
> --- a/mm/hugetlb.c~smaps-pte-sizes	2016-11-28 14:21:37.555519365 -0800
> +++ b/mm/hugetlb.c	2016-11-28 14:28:49.186234688 -0800
> @@ -2763,6 +2763,17 @@ void __init hugetlb_add_hstate(unsigned
>  					huge_page_size(h)/1024);
>  
>  	parsed_hstate = h;
> +
> +	/*
> +	 * PGD_SIZE isn't widely made available by architecures,
> +	 * so use PUD_SIZE*PTRS_PER_PUD as a substitute.
> +	 *
> +	 * Check for sizes that might be mapped by a PGD.  There
> +	 * are none of these known today, but be on the lookout.
> +	 * If this trips, we will need to update the mss->rss_*
> +	 * code in fs/proc/task_mmu.c.
> +	 */
> +	WARN_ON_ONCE((PAGE_SIZE << order) >= PUD_SIZE * PTRS_PER_PUD);
>  }

This will trip for ppc64 16GB hugepage.

For ppc64 we have the 16G at pgd level.

-aneesh