Fix page cache limit wrapping in filesystems

PATCH In review Andi Kleen 2 participants 2 messages Feb 7, 2002

Subsystems (4)

VFS NFS client NTFS filesystem UDF filesystem

lore

Share & tools

lore

Andi Kleen

author

2002-02-07 02:34 UTC #0 lore raw

PATCH 1/1 Fix page cache limit wrapping in filesystems

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

fs/nfs/inode.c renamed from fs/nfs/inode.c-MAXLFS +2 -0

@@ -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) {

fs/ntfs/fs.c renamed from fs/ntfs/fs.c-MAXLFS +1 -1

@@ -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");

fs/udf/super.c renamed from fs/udf/super.c-MAXLFS +1 -1

@@ -1544,7 +1544,7 @@
 		iput(inode);
 		goto error_out;
 	}
-	sb->s_maxbytes = ~0ULL;
+	sb->s_maxbytes = MAX_LFS_FILESIZE;
 	return sb;
 
 error_out:

include/linux/fs.h renamed from include/linux/fs.h-MAXLFS +8 -0

@@ -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 */

Alexander Viro

2002-02-07 02:40 UTC #1 lore raw

Quoted 6 lines · Andi Kleen

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.

Quoted 7 lines

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.

Looks OK.