From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964939Ab2KEXYU (ORCPT ); Mon, 5 Nov 2012 18:24:20 -0500 Received: from mga14.intel.com ([143.182.124.37]:15558 "EHLO mga14.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964927Ab2KEXYN (ORCPT ); Mon, 5 Nov 2012 18:24:13 -0500 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.80,718,1344236400"; d="scan'208";a="213808715" From: Andi Kleen To: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org, akpm@linux-foundation.org, mtk.manpages@gmail.com Subject: Updated MMAP/SHMGET 1GB patchkit Date: Mon, 5 Nov 2012 15:24:07 -0800 Message-Id: <1352157848-29473-1-git-send-email-andi@firstfloor.org> X-Mailer: git-send-email 1.7.7.6 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org This is the updated patchkit to allow to use 1GB pages for mmap/shmget directly. This is needed for various software and makes it far more convenient to use 1GB pages. There were a lot of discussions on the ABI of v6 due to a conflict of the new bits with the NOMMU only feature MAP_UNINITIALIZED. However I don't think this matters at all because this feature only works with HUGETLBFS and HUGETLBFS and NOMMU are mutually exclusive. So any architecture which suppors HUGETLBFS will never do MAP_UNINITIALIZED and vice versa. So there's no actual conflict in these flags. Based on that I don't think any changes are needed. So I decided to just repost with a rebase. Please consider applying, Thanks, -Andi