From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 6 Feb 2002 21:34:31 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 6 Feb 2002 21:34:22 -0500 Received: from mailout09.sul.t-online.com ([194.25.134.84]:21722 "EHLO mailout09.sul.t-online.com") by vger.kernel.org with ESMTP id ; Wed, 6 Feb 2002 21:34:04 -0500 Date: Thu, 7 Feb 2002 03:33:42 +0100 From: Andi Kleen To: torvalds@transmeta.com Cc: Alexander Viro , linux-kernel@vger.kernel.org Subject: [PATCH] Fix page cache limit wrapping in filesystems Message-ID: <20020207033342.A2032@averell> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.22.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Several file systems in tree that nominally support files >2GB set their s_maxbytes value to ~0ULL. This has the nasty side effect on 32bit machines that when a file write reaches the page cache limit (e.g. 2^43) it'll silently wrap and destroy data at the beginning of the file. This patch changes the file systems in question to fill in a proper limit. I also have an alternate patch that adds a check for this generically in super.c, but preliminary comments from Al suggested that he prefered to do it in the file systems, so it is done this way way. Patch for 2.5.4pre1. Please consider applying. -Andi --- linux-2.5.4pre1-work/fs/nfs/inode.c-MAXLFS Thu Feb 7 01:44:22 2002 +++ linux-2.5.4pre1-work/fs/nfs/inode.c Thu Feb 7 02:11:24 2002 @@ -455,6 +455,8 @@ server->namelen = maxlen; sb->s_maxbytes = fsinfo.maxfilesize; + if (sb->s_maxbytes > MAX_LFS_FILESIZE) + sb->s_maxbytes = MAX_LFS_FILESIZE; /* Fire up the writeback cache */ if (nfs_reqlist_alloc(server) < 0) { --- linux-2.5.4pre1-work/fs/ntfs/fs.c-MAXLFS Wed Jan 30 22:38:09 2002 +++ linux-2.5.4pre1-work/fs/ntfs/fs.c Thu Feb 7 02:11:24 2002 @@ -1130,7 +1130,7 @@ /* Inform the kernel about which super operations are available. */ sb->s_op = &ntfs_super_operations; sb->s_magic = NTFS_SUPER_MAGIC; - sb->s_maxbytes = ~0ULL >> 1; + sb->s_maxbytes = MAX_LFS_FILESIZE; ntfs_debug(DEBUG_OTHER, "Reading special files\n"); if (ntfs_load_special_files(vol)) { ntfs_error("Error loading special files\n"); --- linux-2.5.4pre1-work/fs/udf/super.c-MAXLFS Wed Jan 30 22:38:09 2002 +++ linux-2.5.4pre1-work/fs/udf/super.c Thu Feb 7 02:11:24 2002 @@ -1544,7 +1544,7 @@ iput(inode); goto error_out; } - sb->s_maxbytes = ~0ULL; + sb->s_maxbytes = MAX_LFS_FILESIZE; return sb; error_out: --- linux-2.5.4pre1-work/include/linux/fs.h-MAXLFS Thu Feb 7 02:09:02 2002 +++ linux-2.5.4pre1-work/include/linux/fs.h Thu Feb 7 02:11:24 2002 @@ -508,6 +508,14 @@ #define MAX_NON_LFS ((1UL<<31) - 1) +/* Page cache limit. The filesystems should put that into their s_maxbytes + limits, otherwise bad things can happen in VM. */ +#if BITS_PER_LONG==32 +#define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1) +#elif BITS_PER_LONG==64 +#define MAX_LFS_FILESIZE 0x7fffffffffffffff +#endif + #define FL_POSIX 1 #define FL_FLOCK 2 #define FL_BROKEN 4 /* broken flock() emulation */