Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932066AbZLRQFK (ORCPT ); Fri, 18 Dec 2009 11:05:10 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754899AbZLRQFC (ORCPT ); Fri, 18 Dec 2009 11:05:02 -0500 Received: from vpn.id2.novell.com ([195.33.99.129]:40067 "EHLO vpn.id2.novell.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751563AbZLRQE7 convert rfc822-to-8bit (ORCPT ); Fri, 18 Dec 2009 11:04:59 -0500 Message-Id: <4B2BB66F0200007800026AD6@vpn.id2.novell.com> X-Mailer: Novell GroupWise Internet Agent 8.0.1 Date: Fri, 18 Dec 2009 16:05:51 +0000 From: "Jan Beulich" To: , , Cc: Subject: [PATCH] x86: lift restriction on the location of FIX_BTMAP_* Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 8BIT Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2854 Lines: 65 The early ioremap fixmap entries cover half (or for 32-bit non-PAE, a quarter) of a page table, yet they got uncondtitionally aligned so far to a 256-entry boundary. This is not necessary if the range of page table entries anyway falls into a single page table. This buys back, for (theoretically) 50% of all configurations (25% of all non-PAE ones), at least some of the lowmem necessarily lost with commit e621bd18958ef5dbace3129ebe17a0a475e127d9. Signed-off-by: Jan Beulich --- arch/x86/include/asm/fixmap.h | 16 +++++++++++----- arch/x86/mm/ioremap.c | 4 ++++ 2 files changed, 15 insertions(+), 5 deletions(-) --- linux-2.6.33-rc1/arch/x86/include/asm/fixmap.h 2009-12-03 04:51:21.000000000 +0100 +++ 2.6.33-rc1-x86-fix-btmap-reduce/arch/x86/include/asm/fixmap.h 2009-12-10 17:24:58.000000000 +0100 @@ -118,14 +118,20 @@ enum fixed_addresses { * 256 temporary boot-time mappings, used by early_ioremap(), * before ioremap() is functional. * - * We round it up to the next 256 pages boundary so that we - * can have a single pgd entry and a single pte table: + * If necessary we round it up to the next 256 pages boundary so + * that we can have a single pgd entry and a single pte table: */ #define NR_FIX_BTMAPS 64 #define FIX_BTMAPS_SLOTS 4 - FIX_BTMAP_END = __end_of_permanent_fixed_addresses + 256 - - (__end_of_permanent_fixed_addresses & 255), - FIX_BTMAP_BEGIN = FIX_BTMAP_END + NR_FIX_BTMAPS*FIX_BTMAPS_SLOTS - 1, +#define TOTAL_FIX_BTMAPS (NR_FIX_BTMAPS * FIX_BTMAPS_SLOTS) + FIX_BTMAP_END = + (__end_of_permanent_fixed_addresses ^ + (__end_of_permanent_fixed_addresses + TOTAL_FIX_BTMAPS - 1)) & + -PTRS_PER_PTE + ? __end_of_permanent_fixed_addresses + TOTAL_FIX_BTMAPS - + (__end_of_permanent_fixed_addresses & (TOTAL_FIX_BTMAPS - 1)) + : __end_of_permanent_fixed_addresses, + FIX_BTMAP_BEGIN = FIX_BTMAP_END + TOTAL_FIX_BTMAPS - 1, #ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT FIX_OHCI1394_BASE, #endif --- linux-2.6.33-rc1/arch/x86/mm/ioremap.c 2009-12-18 16:05:40.000000000 +0100 +++ 2.6.33-rc1-x86-fix-btmap-reduce/arch/x86/mm/ioremap.c 2009-12-10 17:30:57.000000000 +0100 @@ -422,6 +422,10 @@ void __init early_ioremap_init(void) * The boot-ioremap range spans multiple pmds, for which * we are not prepared: */ +#define __FIXADDR_TOP (-PAGE_SIZE) + BUILD_BUG_ON((__fix_to_virt(FIX_BTMAP_BEGIN) >> PMD_SHIFT) + != (__fix_to_virt(FIX_BTMAP_END) >> PMD_SHIFT)); +#undef __FIXADDR_TOP if (pmd != early_ioremap_pmd(fix_to_virt(FIX_BTMAP_END))) { WARN_ON(1); printk(KERN_WARNING "pmd %p != %p\n", -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/