2008-05-22 08:50:20

by Takashi Sato

[permalink] [raw]
Subject: [RFC PATCH 1/3] Implement generic freeze feature

I have modified to set the suitable error number (EOPNOTSUPP)
in case the filesystem doesn't support the freeze feature.
This was pointed out by Andreas Dilger.

The ioctls for the generic freeze feature are below.
o Freeze the filesystem
int ioctl(int fd, int FIFREEZE, arg)
fd: The file descriptor of the mountpoint
FIFREEZE: request code for the freeze
arg: Ignored
Return value: 0 if the operation succeeds. Otherwise, -1

o Unfreeze the filesystem
int ioctl(int fd, int FITHAW, arg)
fd: The file descriptor of the mountpoint
FITHAW: request code for unfreeze
arg: Ignored
Return value: 0 if the operation succeeds. Otherwise, -1

Signed-off-by: Takashi Sato <[email protected]>
Signed-off-by: Masayuki Hamaguchi <[email protected]>
---
fs/buffer.c | 30 ++++++++++++++++++++++++
fs/ioctl.c | 53 ++++++++++++++++++++++++++++++++++++++++++++
fs/super.c | 32 +++++++++++++++++++++++++-
include/linux/buffer_head.h | 2 -
include/linux/fs.h | 14 +++++++++++
5 files changed, 128 insertions(+), 3 deletions(-)

diff -uprN -X /home/sho/pub/MC/freeze-set/dontdiff linux-2.6.26-rc3.org/fs/buffer.c linux-2.6.26-rc3-freeze/fs/buffer.c
--- linux-2.6.26-rc3.org/fs/buffer.c 2008-05-19 06:36:41.000000000 +0900
+++ linux-2.6.26-rc3-freeze/fs/buffer.c 2008-05-19 21:03:29.000000000 +0900
@@ -201,6 +201,21 @@ struct super_block *freeze_bdev(struct b
{
struct super_block *sb;

+ if (test_and_set_bit(BD_FREEZE_OP, &bdev->bd_state))
+ return ERR_PTR(-EBUSY);
+
+ sb = get_super_without_lock(bdev);
+
+ /* If super_block has been already frozen, return. */
+ if (sb && sb->s_frozen != SB_UNFROZEN) {
+ put_super(sb);
+ clear_bit(BD_FREEZE_OP, &bdev->bd_state);
+ return sb;
+ }
+
+ if (sb)
+ put_super(sb);
+
down(&bdev->bd_mount_sem);
sb = get_super(bdev);
if (sb && !(sb->s_flags & MS_RDONLY)) {
@@ -219,6 +234,8 @@ struct super_block *freeze_bdev(struct b
}

sync_blockdev(bdev);
+ clear_bit(BD_FREEZE_OP, &bdev->bd_state);
+
return sb; /* thaw_bdev releases s->s_umount and bd_mount_sem */
}
EXPORT_SYMBOL(freeze_bdev);
@@ -230,8 +247,17 @@ EXPORT_SYMBOL(freeze_bdev);
*
* Unlocks the filesystem and marks it writeable again after freeze_bdev().
*/
-void thaw_bdev(struct block_device *bdev, struct super_block *sb)
+int thaw_bdev(struct block_device *bdev, struct super_block *sb)
{
+
+ if (test_and_set_bit(BD_FREEZE_OP, &bdev->bd_state))
+ return -EBUSY;
+
+ if (sb && sb->s_frozen == SB_UNFROZEN) {
+ clear_bit(BD_FREEZE_OP, &bdev->bd_state);
+ return 0;
+ }
+
if (sb) {
BUG_ON(sb->s_bdev != bdev);

@@ -244,6 +270,8 @@ void thaw_bdev(struct block_device *bdev
}

up(&bdev->bd_mount_sem);
+ clear_bit(BD_FREEZE_OP, &bdev->bd_state);
+ return 0;
}
EXPORT_SYMBOL(thaw_bdev);

diff -uprN -X /home/sho/pub/MC/freeze-set/dontdiff linux-2.6.26-rc3.org/fs/ioctl.c linux-2.6.26-rc3-freeze/fs/ioctl.c
--- linux-2.6.26-rc3.org/fs/ioctl.c 2008-05-19 06:36:41.000000000 +0900
+++ linux-2.6.26-rc3-freeze/fs/ioctl.c 2008-05-21 11:43:39.000000000 +0900
@@ -13,6 +13,7 @@
#include <linux/security.h>
#include <linux/module.h>
#include <linux/uaccess.h>
+#include <linux/buffer_head.h>

#include <asm/ioctls.h>

@@ -141,6 +142,49 @@ static int ioctl_fioasync(unsigned int f
}

/*
+ * ioctl_freeze - Freeze the filesystem.
+ *
+ * @filp: target file
+ *
+ * Call freeze_bdev() to freeze the filesystem.
+ */
+static int ioctl_freeze(struct file *filp)
+{
+ struct super_block *sb = filp->f_path.dentry->d_inode->i_sb;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ /* If filesystem doesn't support freeze feature, return. */
+ if (sb->s_op->write_super_lockfs == NULL)
+ return -EOPNOTSUPP;
+
+ /* Freeze */
+ sb = freeze_bdev(sb->s_bdev);
+ if (IS_ERR(sb))
+ return PTR_ERR(sb);
+ return 0;
+}
+
+/*
+ * ioctl_thaw - Thaw the filesystem.
+ *
+ * @filp: target file
+ *
+ * Call thaw_bdev() to thaw the filesystem.
+ */
+static int ioctl_thaw(struct file *filp)
+{
+ struct super_block *sb = filp->f_path.dentry->d_inode->i_sb;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ /* Thaw */
+ return thaw_bdev(sb->s_bdev, sb);
+}
+
+/*
* When you add any new common ioctls to the switches above and below
* please update compat_sys_ioctl() too.
*
@@ -181,6 +225,15 @@ int do_vfs_ioctl(struct file *filp, unsi
} else
error = -ENOTTY;
break;
+
+ case FIFREEZE:
+ error = ioctl_freeze(filp);
+ break;
+
+ case FITHAW:
+ error = ioctl_thaw(filp);
+ break;
+
default:
if (S_ISREG(filp->f_path.dentry->d_inode->i_mode))
error = file_ioctl(filp, cmd, arg);
diff -uprN -X /home/sho/pub/MC/freeze-set/dontdiff linux-2.6.26-rc3.org/fs/super.c linux-2.6.26-rc3-freeze/fs/super.c
--- linux-2.6.26-rc3.org/fs/super.c 2008-05-19 06:36:41.000000000 +0900
+++ linux-2.6.26-rc3-freeze/fs/super.c 2008-05-19 21:03:29.000000000 +0900
@@ -156,7 +156,7 @@ int __put_super_and_need_restart(struct
* Drops a temporary reference, frees superblock if there's no
* references left.
*/
-static void put_super(struct super_block *sb)
+void put_super(struct super_block *sb)
{
spin_lock(&sb_lock);
__put_super(sb);
@@ -509,6 +509,36 @@ rescan:

EXPORT_SYMBOL(get_super);

+/*
+ * get_super_without_lock - Get super_block from block_device without lock.
+ * @bdev: block device struct
+ *
+ * Scan the superblock list and finds the superblock of the file system
+ * mounted on the block device given. This doesn't lock anyone.
+ * %NULL is returned if no match is found.
+ */
+struct super_block *get_super_without_lock(struct block_device *bdev)
+{
+ struct super_block *sb;
+
+ if (!bdev)
+ return NULL;
+
+ spin_lock(&sb_lock);
+ list_for_each_entry(sb, &super_blocks, s_list) {
+ if (sb->s_bdev == bdev) {
+ if (sb->s_root) {
+ sb->s_count++;
+ spin_unlock(&sb_lock);
+ return sb;
+ }
+ }
+ }
+ spin_unlock(&sb_lock);
+ return NULL;
+}
+EXPORT_SYMBOL(get_super_without_lock);
+
struct super_block * user_get_super(dev_t dev)
{
struct super_block *sb;
diff -uprN -X /home/sho/pub/MC/freeze-set/dontdiff linux-2.6.26-rc3.org/include/linux/buffer_head.h linux-2.6.26-rc3-fre
eze/include/linux/buffer_head.h
--- linux-2.6.26-rc3.org/include/linux/buffer_head.h 2008-05-19 06:36:41.000000000 +0900
+++ linux-2.6.26-rc3-freeze/include/linux/buffer_head.h 2008-05-19 21:03:29.000000000 +0900
@@ -171,7 +171,7 @@ void __wait_on_buffer(struct buffer_head
wait_queue_head_t *bh_waitq_head(struct buffer_head *bh);
int fsync_bdev(struct block_device *);
struct super_block *freeze_bdev(struct block_device *);
-void thaw_bdev(struct block_device *, struct super_block *);
+int thaw_bdev(struct block_device *, struct super_block *);
int fsync_super(struct super_block *);
int fsync_no_super(struct block_device *);
struct buffer_head *__find_get_block(struct block_device *bdev, sector_t block,
diff -uprN -X /home/sho/pub/MC/freeze-set/dontdiff linux-2.6.26-rc3.org/include/linux/fs.h linux-2.6.26-rc3-freeze/inclu
de/linux/fs.h
--- linux-2.6.26-rc3.org/include/linux/fs.h 2008-05-19 06:36:41.000000000 +0900
+++ linux-2.6.26-rc3-freeze/include/linux/fs.h 2008-05-19 21:03:29.000000000 +0900
@@ -223,6 +223,8 @@ extern int dir_notify_enable;
#define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
#define FIBMAP _IO(0x00,1) /* bmap access */
#define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
+#define FIFREEZE _IOWR('X', 119, int) /* Freeze */
+#define FITHAW _IOWR('X', 120, int) /* Thaw */

#define FS_IOC_GETFLAGS _IOR('f', 1, long)
#define FS_IOC_SETFLAGS _IOW('f', 2, long)
@@ -494,6 +496,13 @@ int pagecache_write_end(struct file *, s
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata);

+/*
+ * Bits in block_device.bd_state.
+ */
+enum bd_state {
+ BD_FREEZE_OP /* In freeze operation */
+};
+
struct backing_dev_info;
struct address_space {
struct inode *host; /* owner: inode, block_device */
@@ -547,6 +556,9 @@ struct block_device {
* care to not mess up bd_private for that case.
*/
unsigned long bd_private;
+
+ /* State of the block device. (Used by freeze feature) */
+ unsigned long bd_state;
};

/*
@@ -1964,7 +1976,9 @@ extern int do_vfs_ioctl(struct file *fil
extern void get_filesystem(struct file_system_type *fs);
extern void put_filesystem(struct file_system_type *fs);
extern struct file_system_type *get_fs_type(const char *name);
+extern void put_super(struct super_block *sb);
extern struct super_block *get_super(struct block_device *);
+extern struct super_block *get_super_without_lock(struct block_device *);
extern struct super_block *user_get_super(dev_t);
extern void drop_super(struct super_block *sb);



2008-05-25 19:32:11

by Christoph Hellwig

[permalink] [raw]
Subject: Re: [RFC PATCH 1/3] Implement generic freeze feature

> + if (test_and_set_bit(BD_FREEZE_OP, &bdev->bd_state))
> + return ERR_PTR(-EBUSY);
> +
> + sb = get_super_without_lock(bdev);
> +
> + /* If super_block has been already frozen, return. */
> + if (sb && sb->s_frozen != SB_UNFROZEN) {
> + put_super(sb);
> + clear_bit(BD_FREEZE_OP, &bdev->bd_state);
> + return sb;
> + }

The BD_FREEZE_OP flag in the block_device already prevents multiple
freezes for a singe block device, so there is no need for this
additional check and the get_super_without_lock helper.

> down(&bdev->bd_mount_sem);

And with that flag bd_mount_sem is also obsolete for preventing the
multiple freeze operations. We still need investigate what
synchronization we need vs unmount which also takes bd_mount_sem without
every having document what it exactly protects.


2008-05-27 10:20:11

by Takashi Sato

[permalink] [raw]
Subject: Re: [RFC PATCH 1/3] Implement generic freeze feature

Hi,

>> + if (test_and_set_bit(BD_FREEZE_OP, &bdev->bd_state))
>> + return ERR_PTR(-EBUSY);
>> +
>> + sb = get_super_without_lock(bdev);
>> +
>> + /* If super_block has been already frozen, return. */
>> + if (sb && sb->s_frozen != SB_UNFROZEN) {
>> + put_super(sb);
>> + clear_bit(BD_FREEZE_OP, &bdev->bd_state);
>> + return sb;
>> + }
>
> The BD_FREEZE_OP flag in the block_device already prevents multiple
> freezes for a singe block device, so there is no need for this
> additional check and the get_super_without_lock helper.

I think the above check is needed because BD_FREEZE_OP is used to
guarantee that the following two operations are done atomically.
1. Check if the filesystem has already been frozen.
(sb->s_frozen != SB_UNFROZEN)
2. Do the filesystem specific freeze operation
(write_super_lockfs)

The freeze operation can be done again for the frozen filesystem
unless the filesystem is checked.
And if write_super_lockfs doesn't consider that the filesystem
has already been frozen, a problem might occur.

>> down(&bdev->bd_mount_sem);
>
> And with that flag bd_mount_sem is also obsolete for preventing the
> multiple freeze operations.

bd_mount_sem is needed because it is used to guarantee that the freeze
operation never run while mounting (get_sb_bdev()).

> We still need investigate what
> synchronization we need vs unmount which also takes bd_mount_sem without
> every having document what it exactly protects.

And s_umount (semaphore) is acquired in get_user() (called by freeze_bdev())
to guarantee that unmount never run while the filesystem is frozen.

Cheers, Takashi