From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id AD7122C00A7 for ; Fri, 22 Nov 2013 10:46:35 +1100 (EST) Message-ID: <1385077581.4882.30.camel@pasglop> Subject: Re: [PATCH 2/2] powerpc: mm: change pgtable index size for 64K page From: Benjamin Herrenschmidt To: Liu Ping Fan Date: Fri, 22 Nov 2013 10:46:21 +1100 In-Reply-To: <20131121221150.GB26359@iris.ozlabs.ibm.com> References: <1385000275-5988-1-git-send-email-pingfank@linux.vnet.ibm.com> <1385000275-5988-2-git-send-email-pingfank@linux.vnet.ibm.com> <20131121221150.GB26359@iris.ozlabs.ibm.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Cc: Paul Mackerras , linuxppc-dev@lists.ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Fri, 2013-11-22 at 09:11 +1100, Paul Mackerras wrote: > On Thu, Nov 21, 2013 at 10:17:55AM +0800, Liu Ping Fan wrote: > > For 64K page, we waste half of the pte_t page. With this patch, after > > changing PGD_INDEX_SIZE from 12 to 11, PTE_INDEX_SIZE from 8 to 9, > > we can improve the usage of pte_t page and shrink the continuous phys > > size for pgd_t. Also you did you miss that we use the second half to store the per-subpage hash info when using 64k on top of HW 4k ? Cheers, Ben. > > Signed-off-by: Liu Ping Fan > > --- > > arch/powerpc/include/asm/pgtable-ppc64-64k.h | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/arch/powerpc/include/asm/pgtable-ppc64-64k.h b/arch/powerpc/include/asm/pgtable-ppc64-64k.h > > index a56b82f..f6955ff 100644 > > --- a/arch/powerpc/include/asm/pgtable-ppc64-64k.h > > +++ b/arch/powerpc/include/asm/pgtable-ppc64-64k.h > > @@ -4,10 +4,10 @@ > > #include > > > > > > -#define PTE_INDEX_SIZE 8 > > +#define PTE_INDEX_SIZE 9 > > #define PMD_INDEX_SIZE 10 > > #define PUD_INDEX_SIZE 0 > > -#define PGD_INDEX_SIZE 12 > > +#define PGD_INDEX_SIZE 11 > > Nack. Those definitions are the way they are in order to have the PMD > map 16MB, which is our large page size, so that transparent huge pages > work. > > Paul.