--=_courier-11923-1056046783-0001-2
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 7bit
This adds i_seqcnt to the inode structure and then uses i_size_read()
and i_size_write() to provide atomic access to i_size. This is a port
of Andrea Arcangeli's i_size atomic access patch from 2.4. This only
uses the generic reader/writer consistent mechanism.
re-diff against 2.5.72.
This patch has been tested on 1 proc, 2 proc, 4 proc and 8 proc
machines.
This patch is available for download from OSDL's patch lifecycle
manager (PLM):
http://www.osdl.org/cgi-bin/plm?module=patch_info&patch_id=1940
Daniel McNeil <daniel@osdl.org>
--=_courier-11923-1056046783-0001-2
Content-Type: text/x-patch; name="patch.2.5.72-isize.2"; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment; filename=patch.2.5.72-isize.2
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/drivers/block/loop.c l=
inux-2.5.72-isize/drivers/block/loop.c
--- linux-2.5.72/drivers/block/loop.c 2003-06-16 21:20:05.000000000 -0700
+++ linux-2.5.72-isize/drivers/block/loop.c 2003-06-17 17:07:49.130734182 -=
0700
@@ -146,7 +146,7 @@ struct loop_func_table *xfer_funcs[MAX_L
=20
static int figure_loop_size(struct loop_device *lo)
{
- loff_t size =3D lo->lo_backing_file->f_dentry->d_inode->i_mapping->host->=
i_size;
+ loff_t size =3D i_size_read(lo->lo_backing_file->f_dentry->d_inode->i_map=
ping->host);
sector_t x;
/*
* Unfortunately, if we want to do I/O on the device,
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/attr.c linux-2.5.72=
-isize/fs/attr.c
--- linux-2.5.72/fs/attr.c 2003-06-16 21:19:41.000000000 -0700
+++ linux-2.5.72-isize/fs/attr.c 2003-06-17 17:24:16.817360469 -0700
@@ -68,7 +68,7 @@ int inode_setattr(struct inode * inode,=20
int error =3D 0;
=20
if (ia_valid & ATTR_SIZE) {
- if (attr->ia_size !=3D inode->i_size) {
+ if (attr->ia_size !=3D i_size_read(inode)) {
error =3D vmtruncate(inode, attr->ia_size);
if (error || (ia_valid =3D=3D ATTR_SIZE))
goto out;
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/binfmt_aout.c linux=
-2.5.72-isize/fs/binfmt_aout.c
--- linux-2.5.72/fs/binfmt_aout.c 2003-06-16 21:19:42.000000000 -0700
+++ linux-2.5.72-isize/fs/binfmt_aout.c 2003-06-17 17:07:49.140734047 -0700
@@ -269,7 +269,7 @@ static int load_aout_binary(struct linux
if ((N_MAGIC(ex) !=3D ZMAGIC && N_MAGIC(ex) !=3D OMAGIC &&
N_MAGIC(ex) !=3D QMAGIC && N_MAGIC(ex) !=3D NMAGIC) ||
N_TRSIZE(ex) || N_DRSIZE(ex) ||
- bprm->file->f_dentry->d_inode->i_size < ex.a_text+ex.a_data+N_SYMSIZE=
(ex)+N_TXTOFF(ex)) {
+ i_size_read(bprm->file->f_dentry->d_inode) < ex.a_text+ex.a_data+N_SY=
MSIZE(ex)+N_TXTOFF(ex)) {
return -ENOEXEC;
}
=20
@@ -454,7 +454,7 @@ static int load_aout_library(struct file
/* We come in here for the regular a.out style of shared libraries */
if ((N_MAGIC(ex) !=3D ZMAGIC && N_MAGIC(ex) !=3D QMAGIC) || N_TRSIZE(ex) =
||
N_DRSIZE(ex) || ((ex.a_entry & 0xfff) && N_MAGIC(ex) =3D=3D ZMAGIC) |=
|
- inode->i_size < ex.a_text+ex.a_data+N_SYMSIZE(ex)+N_TXTOFF(ex)) {
+ i_size_read(inode) < ex.a_text+ex.a_data+N_SYMSIZE(ex)+N_TXTOFF(ex)) =
{
goto out;
}
=20
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/block_dev.c linux-2=
.5.72-isize/fs/block_dev.c
--- linux-2.5.72/fs/block_dev.c 2003-06-16 21:20:28.000000000 -0700
+++ linux-2.5.72-isize/fs/block_dev.c 2003-06-17 17:07:49.140734047 -0700
@@ -29,7 +29,7 @@
static sector_t max_block(struct block_device *bdev)
{
sector_t retval =3D ~((sector_t)0);
- loff_t sz =3D bdev->bd_inode->i_size;
+ loff_t sz =3D i_size_read(bdev->bd_inode);
=20
if (sz) {
unsigned int size =3D block_size(bdev);
@@ -156,7 +156,7 @@ static int blkdev_commit_write(struct fi
static loff_t block_llseek(struct file *file, loff_t offset, int origin)
{
/* ewww */
- loff_t size =3D file->f_dentry->d_inode->i_bdev->bd_inode->i_size;
+ loff_t size =3D i_size_read(file->f_dentry->d_inode->i_bdev->bd_inode);
loff_t retval;
=20
lock_kernel();
@@ -485,7 +485,7 @@ int check_disk_change(struct block_devic
static void bd_set_size(struct block_device *bdev, loff_t size)
{
unsigned bsize =3D bdev_hardsect_size(bdev);
- bdev->bd_inode->i_size =3D size;
+ i_size_write(bdev->bd_inode, size);
while (bsize < PAGE_CACHE_SIZE) {
if (size & bsize)
break;
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/buffer.c linux-2.5.=
72-isize/fs/buffer.c
--- linux-2.5.72/fs/buffer.c 2003-06-16 21:20:05.000000000 -0700
+++ linux-2.5.72-isize/fs/buffer.c 2003-06-17 17:07:49.140734047 -0700
@@ -1707,7 +1707,7 @@ static int __block_write_full_page(struc
=20
BUG_ON(!PageLocked(page));
=20
- last_block =3D (inode->i_size - 1) >> inode->i_blkbits;
+ last_block =3D (i_size_read(inode) - 1) >> inode->i_blkbits;
=20
if (!page_has_buffers(page)) {
if (!PageUptodate(page))
@@ -2043,7 +2043,7 @@ int block_read_full_page(struct page *pa
head =3D page_buffers(page);
=20
iblock =3D (sector_t)page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits)=
;
- lblock =3D (inode->i_size+blocksize-1) >> inode->i_blkbits;
+ lblock =3D (i_size_read(inode)+blocksize-1) >> inode->i_blkbits;
bh =3D head;
nr =3D 0;
i =3D 0;
@@ -2268,8 +2268,12 @@ int generic_commit_write(struct file *fi
struct inode *inode =3D page->mapping->host;
loff_t pos =3D ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
__block_commit_write(inode,page,from,to);
+ /*
+ * No need to use i_size_read() here, the i_size
+ * cannot change under us because we hold i_sem.
+ */
if (pos > inode->i_size) {
- inode->i_size =3D pos;
+ i_size_write(inode, pos);
mark_inode_dirty(inode);
}
return 0;
@@ -2421,7 +2425,7 @@ int nobh_commit_write(struct file *file,
=20
set_page_dirty(page);
if (pos > inode->i_size) {
- inode->i_size =3D pos;
+ i_size_write(inode, pos);
mark_inode_dirty(inode);
}
return 0;
@@ -2551,7 +2555,8 @@ int block_write_full_page(struct page *p
struct writeback_control *wbc)
{
struct inode * const inode =3D page->mapping->host;
- const unsigned long end_index =3D inode->i_size >> PAGE_CACHE_SHIFT;
+ loff_t i_size =3D i_size_read(inode);
+ const unsigned long end_index =3D i_size >> PAGE_CACHE_SHIFT;
unsigned offset;
void *kaddr;
=20
@@ -2560,7 +2565,7 @@ int block_write_full_page(struct page *p
return __block_write_full_page(inode, page, get_block, wbc);
=20
/* Is the page fully outside i_size? (truncate in progress) */
- offset =3D inode->i_size & (PAGE_CACHE_SIZE-1);
+ offset =3D i_size & (PAGE_CACHE_SIZE-1);
if (page->index >=3D end_index+1 || !offset) {
/*
* The page may have dirty, unmapped buffers. For example,
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/ext3/inode.c linux-=
2.5.72-isize/fs/ext3/inode.c
--- linux-2.5.72/fs/ext3/inode.c 2003-06-16 21:20:07.000000000 -0700
+++ linux-2.5.72-isize/fs/ext3/inode.c 2003-06-17 17:07:49.150733912 -0700
@@ -1163,7 +1163,7 @@ static int ext3_commit_write(struct file
if (!partial)
SetPageUptodate(page);
if (pos > inode->i_size)
- inode->i_size =3D pos;
+ i_size_write(inode, pos);
EXT3_I(inode)->i_state |=3D EXT3_STATE_JDATA;
if (inode->i_size > EXT3_I(inode)->i_disksize) {
EXT3_I(inode)->i_disksize =3D inode->i_size;
@@ -1495,7 +1495,7 @@ out_stop:
loff_t end =3D offset + ret;
if (end > inode->i_size) {
ei->i_disksize =3D end;
- inode->i_size =3D end;
+ i_size_write(inode, end);
err =3D ext3_mark_inode_dirty(handle, inode);
if (!ret)=20
ret =3D err;
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/inode.c linux-2.5.7=
2-isize/fs/inode.c
--- linux-2.5.72/fs/inode.c 2003-06-16 21:20:27.000000000 -0700
+++ linux-2.5.72-isize/fs/inode.c 2003-06-17 17:07:49.150733912 -0700
@@ -189,6 +189,7 @@ void inode_init_once(struct inode *inode
INIT_LIST_HEAD(&inode->i_data.i_mmap);
INIT_LIST_HEAD(&inode->i_data.i_mmap_shared);
spin_lock_init(&inode->i_lock);
+ i_size_ordered_init(inode);
}
=20
static void init_once(void * foo, kmem_cache_t * cachep, unsigned long fla=
gs)
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/ioctl.c linux-2.5.7=
2-isize/fs/ioctl.c
--- linux-2.5.72/fs/ioctl.c 2003-06-16 21:19:42.000000000 -0700
+++ linux-2.5.72-isize/fs/ioctl.c 2003-06-17 17:07:49.150733912 -0700
@@ -40,7 +40,7 @@ static int file_ioctl(struct file *filp,
return -EBADF;
return put_user(inode->i_sb->s_blocksize, (int *) arg);
case FIONREAD:
- return put_user(inode->i_size - filp->f_pos, (int *) arg);
+ return put_user(i_size_read(inode) - filp->f_pos, (int *) arg);
}
if (filp->f_op && filp->f_op->ioctl)
return filp->f_op->ioctl(inode, filp, cmd, arg);
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/libfs.c linux-2.5.7=
2-isize/fs/libfs.c
--- linux-2.5.72/fs/libfs.c 2003-06-16 21:19:39.000000000 -0700
+++ linux-2.5.72-isize/fs/libfs.c 2003-06-17 17:07:49.150733912 -0700
@@ -328,8 +328,12 @@ int simple_commit_write(struct file *fil
struct inode *inode =3D page->mapping->host;
loff_t pos =3D ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
=20
+ /*
+ * No need to use i_size_read() here, the i_size
+ * cannot change under us because we hold the i_sem.
+ */
if (pos > inode->i_size)
- inode->i_size =3D pos;
+ i_size_write(inode, pos);
set_page_dirty(page);
return 0;
}
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/locks.c linux-2.5.7=
2-isize/fs/locks.c
--- linux-2.5.72/fs/locks.c 2003-06-16 21:20:23.000000000 -0700
+++ linux-2.5.72-isize/fs/locks.c 2003-06-17 17:07:49.150733912 -0700
@@ -285,7 +285,7 @@ static int flock_to_posix_lock(struct fi
start =3D filp->f_pos;
break;
case 2: /*SEEK_END*/
- start =3D filp->f_dentry->d_inode->i_size;
+ start =3D i_size_read(filp->f_dentry->d_inode);
break;
default:
return -EINVAL;
@@ -335,7 +335,7 @@ static int flock64_to_posix_lock(struct=20
start =3D filp->f_pos;
break;
case 2: /*SEEK_END*/
- start =3D filp->f_dentry->d_inode->i_size;
+ start =3D i_size_read(filp->f_dentry->d_inode);
break;
default:
return -EINVAL;
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/mpage.c linux-2.5.7=
2-isize/fs/mpage.c
--- linux-2.5.72/fs/mpage.c 2003-06-16 21:19:44.000000000 -0700
+++ linux-2.5.72-isize/fs/mpage.c 2003-06-17 17:07:49.160733777 -0700
@@ -227,7 +227,7 @@ do_mpage_readpage(struct bio *bio, struc
goto confused;
=20
block_in_file =3D page->index << (PAGE_CACHE_SHIFT - blkbits);
- last_block =3D (inode->i_size + blocksize - 1) >> blkbits;
+ last_block =3D (i_size_read(inode) + blocksize - 1) >> blkbits;
=20
bh.b_page =3D page;
for (page_block =3D 0; page_block < blocks_per_page;
@@ -459,7 +459,7 @@ mpage_writepage(struct bio *bio, struct=20
*/
BUG_ON(!PageUptodate(page));
block_in_file =3D page->index << (PAGE_CACHE_SHIFT - blkbits);
- last_block =3D (inode->i_size - 1) >> blkbits;
+ last_block =3D (i_size_read(inode) - 1) >> blkbits;
map_bh.b_page =3D page;
for (page_block =3D 0; page_block < blocks_per_page; ) {
=20
@@ -489,9 +489,9 @@ mpage_writepage(struct bio *bio, struct=20
=20
first_unmapped =3D page_block;
=20
- end_index =3D inode->i_size >> PAGE_CACHE_SHIFT;
+ end_index =3D i_size_read(inode) >> PAGE_CACHE_SHIFT;
if (page->index >=3D end_index) {
- unsigned offset =3D inode->i_size & (PAGE_CACHE_SIZE - 1);
+ unsigned offset =3D i_size_read(inode) & (PAGE_CACHE_SIZE - 1);
char *kaddr;
=20
if (page->index > end_index || !offset)
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/open.c linux-2.5.72=
-isize/fs/open.c
--- linux-2.5.72/fs/open.c 2003-06-16 21:19:40.000000000 -0700
+++ linux-2.5.72-isize/fs/open.c 2003-06-17 17:07:49.160733777 -0700
@@ -908,7 +908,7 @@ asmlinkage long sys_vhangup(void)
*/
int generic_file_open(struct inode * inode, struct file * filp)
{
- if (!(filp->f_flags & O_LARGEFILE) && inode->i_size > MAX_NON_LFS)
+ if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
return -EFBIG;
return 0;
}
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/quota_v1.c linux-2.=
5.72-isize/fs/quota_v1.c
--- linux-2.5.72/fs/quota_v1.c 2003-06-16 21:20:19.000000000 -0700
+++ linux-2.5.72-isize/fs/quota_v1.c 2003-06-17 17:07:49.160733777 -0700
@@ -132,12 +132,14 @@ static int v1_check_quota_file(struct su
mm_segment_t fs;
ssize_t size;
loff_t offset =3D 0;
+ loff_t isize;
static const uint quota_magics[] =3D V2_INITQMAGICS;
=20
- if (!inode->i_size)
+ isize =3D i_size_read(inode);
+ if (!isize)
return 0;
- blocks =3D inode->i_size >> BLOCK_SIZE_BITS;
- off =3D inode->i_size & (BLOCK_SIZE - 1);
+ blocks =3D isize >> BLOCK_SIZE_BITS;
+ off =3D isize & (BLOCK_SIZE - 1);
if ((blocks % sizeof(struct v1_disk_dqblk) * BLOCK_SIZE + off) % sizeof(s=
truct v1_disk_dqblk))
return 0;
/* Doublecheck whether we didn't get file with new format - with old quot=
actl() this could happen */
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/fs/stat.c linux-2.5.72=
-isize/fs/stat.c
--- linux-2.5.72/fs/stat.c 2003-06-16 21:19:42.000000000 -0700
+++ linux-2.5.72-isize/fs/stat.c 2003-06-17 17:07:49.160733777 -0700
@@ -28,7 +28,7 @@ void generic_fillattr(struct inode *inod
stat->atime =3D inode->i_atime;
stat->mtime =3D inode->i_mtime;
stat->ctime =3D inode->i_ctime;
- stat->size =3D inode->i_size;
+ stat->size =3D i_size_read(inode);
stat->blocks =3D inode->i_blocks;
stat->blksize =3D inode->i_blksize;
}
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/include/linux/fs.h lin=
ux-2.5.72-isize/include/linux/fs.h
--- linux-2.5.72/include/linux/fs.h 2003-06-16 21:19:46.000000000 -0700
+++ linux-2.5.72-isize/include/linux/fs.h 2003-06-17 17:07:49.180733507 -07=
00
@@ -349,6 +349,17 @@ struct block_device {
struct gendisk * bd_disk;
};
=20
+/*
+ * Use sequence counter to get consistent i_size on 32-bit processors.
+ */
+#if BITS_PER_LONG=3D=3D32 && defined(CONFIG_SMP)
+#include <linux/seqlock.h>
+#define __NEED_I_SIZE_ORDERED
+#define i_size_ordered_init(inode) seqcnt_init(&inode->i_size_seqcnt)
+#else
+#define i_size_ordered_init(inode) do { } while (0)
+#endif
+
struct inode {
struct hlist_node i_hash;
struct list_head i_list;
@@ -399,8 +410,60 @@ struct inode {
union {
void *generic_ip;
} u;
+#ifdef __NEED_I_SIZE_ORDERED
+ seqcnt_t i_size_seqcnt;
+#endif
};
=20
+/*
+ * NOTE: in a 32bit arch with a preemptable kernel and
+ * an UP compile the i_size_read/write must be atomic
+ * with respect to the local cpu (unlike with preempt disabled),
+ * but they don't need to be atomic with respect to other cpus like in
+ * true SMP (so they need either to either locally disable irq around
+ * the read or for example on x86 they can be still implemented as a
+ * cmpxchg8b without the need of the lock prefix). For SMP compiles
+ * and 64bit archs it makes no difference if preempt is enabled or not.
+ */
+static inline loff_t i_size_read(struct inode * inode)
+{
+#if BITS_PER_LONG=3D=3D32 && defined(CONFIG_SMP)
+ loff_t i_size;
+ unsigned int seq;
+
+ do {
+ seq =3D read_seqcntbegin(&inode->i_size_seqcnt);
+ i_size =3D inode->i_size;
+ } while (read_seqcntretry(&inode->i_size_seqcnt, seq));
+ return i_size;
+#elif BITS_PER_LONG=3D=3D32 && defined(CONFIG_PREEMPT)
+ loff_t i_size;
+
+ preempt_disable();
+ i_size =3D inode->i_size;
+ preempt_enable();
+ return i_size;
+#else
+ return inode->i_size;
+#endif
+}
+
+
+static inline void i_size_write(struct inode * inode, loff_t i_size)
+{
+#if BITS_PER_LONG=3D=3D32 && defined(CONFIG_SMP)
+ write_seqcntbegin(&inode->i_size_seqcnt);
+ inode->i_size =3D i_size;
+ write_seqcntend(&inode->i_size_seqcnt);
+#elif BITS_PER_LONG=3D=3D32 && defined(CONFIG_PREEMPT)
+ preempt_disable();
+ inode->i_size =3D i_size;
+ preempt_enable();
+#else
+ inode->i_size =3D i_size;
+#endif
+}
+
struct fown_struct {
rwlock_t lock; /* protects pid, uid, euid fields */
int pid; /* pid or -pgrp where SIGIO should be sent */
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/ipc/shm.c linux-2.5.72=
-isize/ipc/shm.c
--- linux-2.5.72/ipc/shm.c 2003-06-16 21:20:01.000000000 -0700
+++ linux-2.5.72-isize/ipc/shm.c 2003-06-17 17:07:49.180733507 -0700
@@ -703,7 +703,7 @@ long sys_shmat(int shmid, char __user *s
}
=09
file =3D shp->shm_file;
- size =3D file->f_dentry->d_inode->i_size;
+ size =3D i_size_read(file->f_dentry->d_inode);
shp->shm_nattch++;
shm_unlock(shp);
=20
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/mm/filemap.c linux-2.5=
.72-isize/mm/filemap.c
--- linux-2.5.72/mm/filemap.c 2003-06-16 21:19:49.000000000 -0700
+++ linux-2.5.72-isize/mm/filemap.c 2003-06-17 17:07:49.180733507 -0700
@@ -555,14 +555,15 @@ void do_generic_mapping_read(struct addr
for (;;) {
struct page *page;
unsigned long end_index, nr, ret;
+ loff_t isize =3D i_size_read(inode);
=20
- end_index =3D inode->i_size >> PAGE_CACHE_SHIFT;
+ end_index =3D isize >> PAGE_CACHE_SHIFT;
=09
if (index > end_index)
break;
nr =3D PAGE_CACHE_SIZE;
if (index =3D=3D end_index) {
- nr =3D inode->i_size & ~PAGE_CACHE_MASK;
+ nr =3D isize & ~PAGE_CACHE_MASK;
if (nr <=3D offset)
break;
}
@@ -763,7 +764,7 @@ __generic_file_aio_read(struct kiocb *io
retval =3D 0;
if (!count)
goto out; /* skip atime */
- size =3D inode->i_size;
+ size =3D i_size_read(inode);
if (pos < size) {
retval =3D generic_file_direct_IO(READ, iocb,
iov, pos, nr_segs);
@@ -952,7 +953,7 @@ retry_all:
* An external ptracer can access pages that normally aren't
* accessible..
*/
- size =3D (inode->i_size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+ size =3D (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
if ((pgoff >=3D size) && (area->vm_mm =3D=3D current->mm))
return NULL;
=20
@@ -1219,7 +1220,7 @@ static int filemap_populate(struct vm_ar
pgoff, len >> PAGE_CACHE_SHIFT);
=20
repeat:
- size =3D (inode->i_size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+ size =3D (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
if (pgoff + (len >> PAGE_CACHE_SHIFT) > size)
return -EINVAL;
=20
@@ -1530,7 +1531,7 @@ inline int generic_write_checks(struct i
if (!isblk) {
/* FIXME: this is for backwards compatibility with 2.4 */
if (file->f_flags & O_APPEND)
- *pos =3D inode->i_size;
+ *pos =3D i_size_read(inode);
=20
if (limit !=3D RLIM_INFINITY) {
if (*pos >=3D limit) {
@@ -1576,15 +1577,17 @@ inline int generic_write_checks(struct i
if (unlikely(*pos + *count > inode->i_sb->s_maxbytes))
*count =3D inode->i_sb->s_maxbytes - *pos;
} else {
+ loff_t isize;
if (bdev_read_only(inode->i_bdev))
return -EPERM;
- if (*pos >=3D inode->i_size) {
- if (*count || *pos > inode->i_size)
+ isize =3D i_size_read(inode);
+ if (*pos >=3D isize) {
+ if (*count || *pos > isize)
return -ENOSPC;
}
=20
- if (*pos + *count > inode->i_size)
- *count =3D inode->i_size - *pos;
+ if (*pos + *count > isize)
+ *count =3D isize - *pos;
}
return 0;
}
@@ -1671,8 +1674,8 @@ generic_file_aio_write_nolock(struct kio
iov, pos, nr_segs);
if (written > 0) {
loff_t end =3D pos + written;
- if (end > inode->i_size && !isblk) {
- inode->i_size =3D end;
+ if (end > i_size_read(inode) && !isblk) {
+ i_size_write(inode, end);
mark_inode_dirty(inode);
}
*ppos =3D end;
@@ -1716,14 +1719,15 @@ generic_file_aio_write_nolock(struct kio
=20
status =3D a_ops->prepare_write(file, page, offset, offset+bytes);
if (unlikely(status)) {
+ loff_t isize =3D i_size_read(inode);
/*
* prepare_write() may have instantiated a few blocks
* outside i_size. Trim these off again.
*/
unlock_page(page);
page_cache_release(page);
- if (pos + bytes > inode->i_size)
- vmtruncate(inode, inode->i_size);
+ if (pos + bytes > isize)
+ vmtruncate(inode, isize);
break;
}
if (likely(nr_segs =3D=3D 1))
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/mm/memory.c linux-2.5.=
72-isize/mm/memory.c
--- linux-2.5.72/mm/memory.c 2003-06-16 21:20:00.000000000 -0700
+++ linux-2.5.72-isize/mm/memory.c 2003-06-17 17:07:49.180733507 -0700
@@ -1109,7 +1109,7 @@ int vmtruncate(struct inode * inode, lof
=20
if (inode->i_size < offset)
goto do_expand;
- inode->i_size =3D offset;
+ i_size_write(inode, offset);
pgoff =3D (offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
down(&mapping->i_shared_sem);
if (unlikely(!list_empty(&mapping->i_mmap)))
@@ -1126,7 +1126,7 @@ do_expand:
goto out_sig;
if (offset > inode->i_sb->s_maxbytes)
goto out;
- inode->i_size =3D offset;
+ i_size_write(inode, offset);
=20
out_truncate:
if (inode->i_op && inode->i_op->truncate)
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/mm/nommu.c linux-2.5.7=
2-isize/mm/nommu.c
--- linux-2.5.72/mm/nommu.c 2003-06-16 21:20:21.000000000 -0700
+++ linux-2.5.72-isize/mm/nommu.c 2003-06-17 17:07:49.190733372 -0700
@@ -48,7 +48,7 @@ int vmtruncate(struct inode *inode, loff
=20
if (inode->i_size < offset)
goto do_expand;
- inode->i_size =3D offset;
+ i_size_write(inode, offset);
=20
truncate_inode_pages(mapping, offset);
goto out_truncate;
@@ -59,7 +59,7 @@ do_expand:
goto out_sig;
if (offset > inode->i_sb->s_maxbytes)
goto out;
- inode->i_size =3D offset;
+ i_size_write(inode, offset);
=20
out_truncate:
if (inode->i_op && inode->i_op->truncate) {
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/mm/readahead.c linux-2=
.5.72-isize/mm/readahead.c
--- linux-2.5.72/mm/readahead.c 2003-06-16 21:19:40.000000000 -0700
+++ linux-2.5.72-isize/mm/readahead.c 2003-06-17 17:07:49.190733372 -0700
@@ -208,11 +208,12 @@ __do_page_cache_readahead(struct address
LIST_HEAD(page_pool);
int page_idx;
int ret =3D 0;
+ loff_t isize =3D i_size_read(inode);
=20
- if (inode->i_size =3D=3D 0)
+ if (isize =3D=3D 0)
goto out;
=20
- end_index =3D ((inode->i_size - 1) >> PAGE_CACHE_SHIFT);
+ end_index =3D ((isize - 1) >> PAGE_CACHE_SHIFT);
=20
/*
* Preallocate as many pages as we will need.
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/mm/shmem.c linux-2.5.7=
2-isize/mm/shmem.c
--- linux-2.5.72/mm/shmem.c 2003-06-16 21:20:00.000000000 -0700
+++ linux-2.5.72-isize/mm/shmem.c 2003-06-17 17:07:49.190733372 -0700
@@ -298,7 +298,7 @@ static swp_entry_t *shmem_swp_alloc(stru
static const swp_entry_t unswapped =3D {0};
=20
if (sgp !=3D SGP_WRITE &&
- ((loff_t) index << PAGE_CACHE_SHIFT) >=3D inode->i_size)
+ ((loff_t) index << PAGE_CACHE_SHIFT) >=3D i_size_read(inode))
return ERR_PTR(-EINVAL);
=20
while (!(entry =3D shmem_swp_entry(info, index, &page))) {
@@ -331,7 +331,7 @@ static swp_entry_t *shmem_swp_alloc(stru
return ERR_PTR(-ENOMEM);
}
if (sgp !=3D SGP_WRITE &&
- ((loff_t) index << PAGE_CACHE_SHIFT) >=3D inode->i_size) {
+ ((loff_t) index << PAGE_CACHE_SHIFT) >=3D i_size_read(inode)) {
entry =3D ERR_PTR(-EINVAL);
break;
}
diff -rupN -X /home/daniel_nfs/dontdiff linux-2.5.72/mm/swapfile.c linux-2.=
5.72-isize/mm/swapfile.c
--- linux-2.5.72/mm/swapfile.c 2003-06-16 21:19:42.000000000 -0700
+++ linux-2.5.72-isize/mm/swapfile.c 2003-06-17 17:07:49.190733372 -0700
@@ -922,7 +922,7 @@ static int setup_swap_extents(struct swa
*/
probe_block =3D 0;
page_no =3D 0;
- last_block =3D inode->i_size >> blkbits;
+ last_block =3D i_size_read(inode) >> blkbits;
while ((probe_block + blocks_per_page) <=3D last_block &&
page_no < sis->max) {
unsigned block_in_page;
@@ -1308,7 +1308,7 @@ asmlinkage long sys_swapon(const char __
goto bad_swap;
}
=20
- swapfilesize =3D mapping->host->i_size >> PAGE_SHIFT;
+ swapfilesize =3D i_size_read(mapping->host) >> PAGE_SHIFT;
=20
/*
* Read the swap header.
--=_courier-11923-1056046783-0001-2--