From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from tx2outboundpool.messaging.microsoft.com (tx2ehsobe005.messaging.microsoft.com [65.55.88.15]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (Client CN "mail.global.frontbridge.com", Issuer "MSIT Machine Auth CA 2" (not verified)) by ozlabs.org (Postfix) with ESMTPS id DCE532C008A for ; Fri, 7 Jun 2013 08:42:48 +1000 (EST) Date: Thu, 6 Jun 2013 17:42:39 -0500 From: Scott Wood Subject: Re: [PATCH -V7 09/18] powerpc: Switch 16GB and 16MB explicit hugepages to a different page table format To: Aneesh Kumar K.V References: <1367177859-7893-1-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1367177859-7893-10-git-send-email-aneesh.kumar@linux.vnet.ibm.com> In-Reply-To: <1367177859-7893-10-git-send-email-aneesh.kumar@linux.vnet.ibm.com> (from aneesh.kumar@linux.vnet.ibm.com on Sun Apr 28 14:37:30 2013) Message-ID: <1370558559.32518.4@snotra> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; delsp=Yes; format=Flowed Cc: dwg@au1.ibm.com, linux-mm@kvack.org, paulus@samba.org, "Aneesh Kumar K.V" , linuxppc-dev@lists.ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 04/28/2013 02:37:30 PM, Aneesh Kumar K.V wrote: > From: "Aneesh Kumar K.V" >=20 > We will be switching PMD_SHIFT to 24 bits to facilitate THP =20 > impmenetation. > With PMD_SHIFT set to 24, we now have 16MB huge pages allocated at =20 > PGD level. > That means with 32 bit process we cannot allocate normal pages at > all, because we cover the entire address space with one pgd entry. =20 > Fix this > by switching to a new page table format for hugepages. With the new =20 > page table > format for 16GB and 16MB hugepages we won't allocate hugepage =20 > directory. Instead > we encode the PTE information directly at the directory level. This =20 > forces 16MB > hugepage at PMD level. This will also make the page take walk much =20 > simpler later > when we add the THP support. >=20 > With the new table format we have 4 cases for pgds and pmds: > (1) invalid (all zeroes) > (2) pointer to next table, as normal; bottom 6 bits =3D=3D 0 > (3) leaf pte for huge page, bottom two bits !=3D 00 > (4) hugepd pointer, bottom two bits =3D=3D 00, next 4 bits indicate size = =20 > of table >=20 > Signed-off-by: Aneesh Kumar K.V > --- > arch/powerpc/include/asm/page.h | 2 + > arch/powerpc/include/asm/pgtable.h | 2 + > arch/powerpc/mm/gup.c | 18 +++- > arch/powerpc/mm/hugetlbpage.c | 176 =20 > +++++++++++++++++++++++++++++++------ > 4 files changed, 168 insertions(+), 30 deletions(-) After this patch, on 64-bit book3e (e5500, and thus 4K pages), I see =20 messages like this after exiting a program that uses hugepages =20 (specifically, qemu): /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221516. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221516. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc2214d6. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc2214d6. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221916. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221916. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc2218d6. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc2218d6. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221496. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221496. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221856. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221856. /home/scott/fsl/git/linux/upstream/mm/memory.c:407: bad pmd =20 40000001fc221816. -Scott=