Re: [PATCH] FAT: Support the large partition (> 128GB) for 2.4

From: Erik Andersen
Date: Wed Dec 31 2003 - 04:16:54 EST

Next message: Andrew Morton: "Re: [PATCH linux-2.6.0-test10-mm1] filemap_fdatawait.patch"
Previous message: Nigel Cunningham: "SMP testers wanted for Software Suspend 2"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue Dec 23, 2003 at 05:02:01AM +0900, OGAWA Hirofumi wrote:
> Hello,
>
> Some peoples reported the problem of fatfs on large partion (> 128GB),
> and fixed by the following patch.
>
>
> The "directory entry pointer" is position of the directory entry in
> the partition. like the following,
>
> in fs/fat/misc.c:fat__get_entry()
>
> offset &= sb->s_blocksize - 1;
> *de = (struct msdos_dir_entry *) ((*bh)->b_data + offset);
> *ino = (sector << sbi->dir_per_block_bits) + (offset >> MSDOS_DIR_BITS);
> ^^^
>
> This is used for updates (not create) of the directory entry, and
> overflowed by large partition (> 128GB).

I think this additional fat32 patch would be a good idea for
2.4.x. Could you review these changes and perhaps fold them into
your patch for inclusion into 2.4.x. This patch fixes support
for the full 4GB (-1 bytes) allowable fat32 file size, and should
be added onto of your previous patch for large fat32 filesystems.

diff -urN linux/fs.orig/buffer.c linux/fs/buffer.c
--- linux/fs.orig/buffer.c 2003-12-31 00:34:57.000000000 -0700
+++ linux/fs/buffer.c 2003-12-31 00:45:40.000000000 -0700
@@ -1940,7 +1940,7 @@
* We may have to extend the file.
*/

-int cont_prepare_write(struct page *page, unsigned offset, unsigned to, get_block_t *get_block, unsigned long *bytes)
+int cont_prepare_write(struct page *page, unsigned offset, unsigned to, get_block_t *get_block, loff_t *bytes)
{
struct address_space *mapping = page->mapping;
struct inode *inode = mapping->host;
diff -urN linux/fs.orig/fat/dir.c linux/fs/fat/dir.c
--- linux/fs.orig/fat/dir.c 2003-12-31 00:36:59.000000000 -0700
+++ linux/fs/fat/dir.c 2003-12-31 01:22:34.000000000 -0700
@@ -362,15 +362,14 @@
unsigned char long_slots;
wchar_t *unicode = NULL;
char c, work[8], bufname[56], *ptname = bufname;
- unsigned long lpos, dummy, *furrfu = &lpos;
+ loff_t lpos, dummy, *furrfu = &lpos;
int uni_xlate = MSDOS_SB(sb)->options.unicode_xlate;
int isvfat = MSDOS_SB(sb)->options.isvfat;
int utf8 = MSDOS_SB(sb)->options.utf8;
int nocase = MSDOS_SB(sb)->options.nocase;
unsigned short opt_shortname = MSDOS_SB(sb)->options.shortname;
- unsigned long inum;
int chi, chl, i, i2, j, last, last_u, dotoffset = 0;
- loff_t i_pos, cpos;
+ loff_t i_pos, inum, cpos;

cpos = filp->f_pos;
/* Fake . and .. for the root directory. */
diff -urN linux/fs.orig/fat/file.c linux/fs/fat/file.c
--- linux/fs.orig/fat/file.c 2001-08-12 11:56:56.000000000 -0600
+++ linux/fs/fat/file.c 2003-12-31 01:39:20.000000000 -0700
@@ -62,7 +62,7 @@
}
if (!create)
return 0;
- if (iblock << sb->s_blocksize_bits != MSDOS_I(inode)->mmu_private) {
+ if (iblock != MSDOS_I(inode)->mmu_private >> sb->s_blocksize_bits) {
BUG();
return -EIO;
}
diff -urN linux/fs.orig/fat/inode.c linux/fs/fat/inode.c
--- linux/fs.orig/fat/inode.c 2003-12-31 00:36:59.000000000 -0700
+++ linux/fs/fat/inode.c 2003-12-31 01:53:02.000000000 -0700
@@ -406,7 +406,7 @@
}
inode->i_blksize = 1 << sbi->cluster_bits;
inode->i_blocks = ((inode->i_size + inode->i_blksize - 1)
- & ~(inode->i_blksize - 1)) >> 9;
+ & ~((loff_t)inode->i_blksize - 1)) >> 9;
MSDOS_I(inode)->i_logstart = 0;
MSDOS_I(inode)->mmu_private = inode->i_size;

@@ -556,8 +556,8 @@
struct fat_boot_sector *b;
struct msdos_sb_info *sbi = MSDOS_SB(sb);
char *p;
- int logical_sector_size, hard_blksize, fat_clusters = 0;
- unsigned int total_sectors, rootdir_sectors;
+ int logical_sector_size, hard_blksize;
+ unsigned long total_sectors, rootdir_sectors, fat_clusters = 0;
int fat32, debug, error, fat, cp;
struct fat_mount_options opts;
char buf[50];
@@ -651,6 +651,7 @@
fat32 = 1;
sbi->fat_length = CF_LE_L(b->fat32_length);
sbi->root_cluster = CF_LE_L(b->root_cluster);
+ sb->s_maxbytes = 0xffffffff;

sbi->fsinfo_sector = CF_LE_W(b->info_sector);
/* MC - if info_sector is 0, don't multiply by 0 */
@@ -957,7 +958,7 @@
/* this is as close to the truth as we can get ... */
inode->i_blksize = 1 << sbi->cluster_bits;
inode->i_blocks = ((inode->i_size + inode->i_blksize - 1)
- & ~(inode->i_blksize - 1)) >> 9;
+ & ~((loff_t)inode->i_blksize - 1)) >> 9;
inode->i_mtime = inode->i_atime =
date_dos2unix(CF_LE_W(de->time),CF_LE_W(de->date));
inode->i_ctime =
diff -urN linux/include/linux.orig/adfs_fs_i.h linux/include/linux/adfs_fs_i.h
--- linux/include/linux.orig/adfs_fs_i.h 2000-02-13 11:45:13.000000000 -0700
+++ linux/include/linux/adfs_fs_i.h 2003-12-31 00:45:40.000000000 -0700
@@ -11,7 +11,7 @@
* adfs file system inode data in memory
*/
struct adfs_inode_info {
- unsigned long mmu_private;
+ loff_t mmu_private;
unsigned long parent_id; /* object id of parent */
__u32 loadaddr; /* RISC OS load address */
__u32 execaddr; /* RISC OS exec address */
diff -urN linux/include/linux.orig/affs_fs_i.h linux/include/linux/affs_fs_i.h
--- linux/include/linux.orig/affs_fs_i.h 2001-11-22 12:46:19.000000000 -0700
+++ linux/include/linux/affs_fs_i.h 2003-12-31 00:45:40.000000000 -0700
@@ -36,7 +36,7 @@
struct affs_ext_key *i_ac; /* associative cache of extended blocks */
u32 i_ext_last; /* last accessed extended block */
struct buffer_head *i_ext_bh; /* bh of last extended block */
- unsigned long mmu_private;
+ loff_t mmu_private;
u32 i_protect; /* unused attribute bits */
u32 i_lastalloc; /* last allocated block */
int i_pa_cnt; /* number of preallocated blocks */
diff -urN linux/include/linux.orig/fs.h linux/include/linux/fs.h
--- linux/include/linux.orig/fs.h 2003-12-31 00:34:59.000000000 -0700
+++ linux/include/linux/fs.h 2003-12-31 00:45:40.000000000 -0700
@@ -1505,7 +1505,7 @@
extern int block_read_full_page(struct page*, get_block_t*);
extern int block_prepare_write(struct page*, unsigned, unsigned, get_block_t*);
extern int cont_prepare_write(struct page*, unsigned, unsigned, get_block_t*,
- unsigned long *);
+ loff_t *);
extern int generic_cont_expand(struct inode *inode, loff_t size) ;
extern int block_commit_write(struct page *page, unsigned from, unsigned to);
extern int block_sync_page(struct page *);
diff -urN linux/include/linux.orig/hfs_fs_i.h linux/include/linux/hfs_fs_i.h
--- linux/include/linux.orig/hfs_fs_i.h 2001-02-13 15:13:44.000000000 -0700
+++ linux/include/linux/hfs_fs_i.h 2003-12-31 00:45:40.000000000 -0700
@@ -19,7 +19,7 @@
struct hfs_inode_info {
int magic; /* A magic number */

- unsigned long mmu_private;
+ loff_t mmu_private;
struct hfs_cat_entry *entry;

/* For a regular or header file */
diff -urN linux/include/linux.orig/hpfs_fs_i.h linux/include/linux/hpfs_fs_i.h
--- linux/include/linux.orig/hpfs_fs_i.h 2000-02-10 13:17:00.000000000 -0700
+++ linux/include/linux/hpfs_fs_i.h 2003-12-31 00:45:40.000000000 -0700
@@ -2,7 +2,7 @@
#define _HPFS_FS_I

struct hpfs_inode_info {
- unsigned long mmu_private;
+ loff_t mmu_private;
ino_t i_parent_dir; /* (directories) gives fnode of parent dir */
unsigned i_dno; /* (directories) root dnode */
unsigned i_dpos; /* (directories) temp for readdir */
diff -urN linux/include/linux.orig/msdos_fs_i.h linux/include/linux/msdos_fs_i.h
--- linux/include/linux.orig/msdos_fs_i.h 2003-12-31 00:36:59.000000000 -0700
+++ linux/include/linux/msdos_fs_i.h 2003-12-31 00:46:21.000000000 -0700
@@ -6,7 +6,7 @@
*/

struct msdos_inode_info {
- unsigned long mmu_private;
+ loff_t mmu_private;
int i_start; /* first cluster or 0 */
int i_logstart; /* logical first cluster */
int i_attrs; /* unused attribute bits */

-Erik

--
Erik B. Andersen http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Andrew Morton: "Re: [PATCH linux-2.6.0-test10-mm1] filemap_fdatawait.patch"
Previous message: Nigel Cunningham: "SMP testers wanted for Software Suspend 2"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]