2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* linux/fs/block_dev.c
|
|
|
|
*
|
|
|
|
* Copyright (C) 1991, 1992 Linus Torvalds
|
|
|
|
* Copyright (C) 2001 Andrea Arcangeli <andrea@suse.de> SuSE
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/fcntl.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/kmod.h>
|
|
|
|
#include <linux/major.h>
|
|
|
|
#include <linux/smp_lock.h>
|
2008-06-06 07:46:27 +02:00
|
|
|
#include <linux/device_cgroup.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <linux/highmem.h>
|
|
|
|
#include <linux/blkdev.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/blkpg.h>
|
|
|
|
#include <linux/buffer_head.h>
|
2009-02-25 10:44:19 +01:00
|
|
|
#include <linux/pagevec.h>
|
2006-08-29 20:06:09 +02:00
|
|
|
#include <linux/writeback.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <linux/mpage.h>
|
|
|
|
#include <linux/mount.h>
|
|
|
|
#include <linux/uio.h>
|
|
|
|
#include <linux/namei.h>
|
2007-05-08 09:24:32 +02:00
|
|
|
#include <linux/log2.h>
|
2009-06-11 14:24:13 +02:00
|
|
|
#include <linux/kmemleak.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
#include <asm/uaccess.h>
|
2006-09-30 20:52:18 +02:00
|
|
|
#include "internal.h"
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
struct bdev_inode {
|
|
|
|
struct block_device bdev;
|
|
|
|
struct inode vfs_inode;
|
|
|
|
};
|
|
|
|
|
2008-02-18 13:48:31 +01:00
|
|
|
static const struct address_space_operations def_blk_aops;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
static inline struct bdev_inode *BDEV_I(struct inode *inode)
|
|
|
|
{
|
|
|
|
return container_of(inode, struct bdev_inode, vfs_inode);
|
|
|
|
}
|
|
|
|
|
|
|
|
inline struct block_device *I_BDEV(struct inode *inode)
|
|
|
|
{
|
|
|
|
return &BDEV_I(inode)->bdev;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(I_BDEV);
|
|
|
|
|
|
|
|
static sector_t max_block(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
sector_t retval = ~((sector_t)0);
|
|
|
|
loff_t sz = i_size_read(bdev->bd_inode);
|
|
|
|
|
|
|
|
if (sz) {
|
|
|
|
unsigned int size = block_size(bdev);
|
|
|
|
unsigned int sizebits = blksize_bits(size);
|
|
|
|
retval = (sz >> sizebits);
|
|
|
|
}
|
|
|
|
return retval;
|
|
|
|
}
|
|
|
|
|
2007-05-06 23:49:55 +02:00
|
|
|
/* Kill _all_ buffers and pagecache , dirty or not.. */
|
2005-04-17 00:20:36 +02:00
|
|
|
static void kill_bdev(struct block_device *bdev)
|
|
|
|
{
|
2007-05-06 23:49:55 +02:00
|
|
|
if (bdev->bd_inode->i_mapping->nrpages == 0)
|
|
|
|
return;
|
|
|
|
invalidate_bh_lrus();
|
2005-04-17 00:20:36 +02:00
|
|
|
truncate_inode_pages(bdev->bd_inode->i_mapping, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
int set_blocksize(struct block_device *bdev, int size)
|
|
|
|
{
|
|
|
|
/* Size must be a power of two, and between 512 and PAGE_SIZE */
|
2007-05-08 09:24:32 +02:00
|
|
|
if (size > PAGE_SIZE || size < 512 || !is_power_of_2(size))
|
2005-04-17 00:20:36 +02:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/* Size cannot be smaller than the size supported by the device */
|
2009-05-22 23:17:49 +02:00
|
|
|
if (size < bdev_logical_block_size(bdev))
|
2005-04-17 00:20:36 +02:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/* Don't change the size if it is same as current */
|
|
|
|
if (bdev->bd_block_size != size) {
|
|
|
|
sync_blockdev(bdev);
|
|
|
|
bdev->bd_block_size = size;
|
|
|
|
bdev->bd_inode->i_blkbits = blksize_bits(size);
|
|
|
|
kill_bdev(bdev);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(set_blocksize);
|
|
|
|
|
|
|
|
int sb_set_blocksize(struct super_block *sb, int size)
|
|
|
|
{
|
|
|
|
if (set_blocksize(sb->s_bdev, size))
|
|
|
|
return 0;
|
|
|
|
/* If we get here, we know size is power of two
|
|
|
|
* and it's value is between 512 and PAGE_SIZE */
|
|
|
|
sb->s_blocksize = size;
|
2006-03-24 12:18:05 +01:00
|
|
|
sb->s_blocksize_bits = blksize_bits(size);
|
2005-04-17 00:20:36 +02:00
|
|
|
return sb->s_blocksize;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(sb_set_blocksize);
|
|
|
|
|
|
|
|
int sb_min_blocksize(struct super_block *sb, int size)
|
|
|
|
{
|
2009-05-22 23:17:49 +02:00
|
|
|
int minsize = bdev_logical_block_size(sb->s_bdev);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (size < minsize)
|
|
|
|
size = minsize;
|
|
|
|
return sb_set_blocksize(sb, size);
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(sb_min_blocksize);
|
|
|
|
|
|
|
|
static int
|
|
|
|
blkdev_get_block(struct inode *inode, sector_t iblock,
|
|
|
|
struct buffer_head *bh, int create)
|
|
|
|
{
|
|
|
|
if (iblock >= max_block(I_BDEV(inode))) {
|
|
|
|
if (create)
|
|
|
|
return -EIO;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* for reads, we're just trying to fill a partial page.
|
|
|
|
* return a hole, they will have to call get_block again
|
|
|
|
* before they can fill it, and they will get -EIO at that
|
|
|
|
* time
|
|
|
|
*/
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
bh->b_bdev = I_BDEV(inode);
|
|
|
|
bh->b_blocknr = iblock;
|
|
|
|
set_buffer_mapped(bh);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-02-03 10:14:01 +01:00
|
|
|
static int
|
|
|
|
blkdev_get_blocks(struct inode *inode, sector_t iblock,
|
|
|
|
struct buffer_head *bh, int create)
|
|
|
|
{
|
|
|
|
sector_t end_block = max_block(I_BDEV(inode));
|
|
|
|
unsigned long max_blocks = bh->b_size >> inode->i_blkbits;
|
|
|
|
|
|
|
|
if ((iblock + max_blocks) > end_block) {
|
|
|
|
max_blocks = end_block - iblock;
|
|
|
|
if ((long)max_blocks <= 0) {
|
|
|
|
if (create)
|
|
|
|
return -EIO; /* write fully beyond EOF */
|
|
|
|
/*
|
|
|
|
* It is a read which is fully beyond EOF. We return
|
|
|
|
* a !buffer_mapped buffer
|
|
|
|
*/
|
|
|
|
max_blocks = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
bh->b_bdev = I_BDEV(inode);
|
|
|
|
bh->b_blocknr = iblock;
|
|
|
|
bh->b_size = max_blocks << inode->i_blkbits;
|
|
|
|
if (max_blocks)
|
|
|
|
set_buffer_mapped(bh);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static ssize_t
|
|
|
|
blkdev_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
|
|
|
|
loff_t offset, unsigned long nr_segs)
|
|
|
|
{
|
|
|
|
struct file *file = iocb->ki_filp;
|
|
|
|
struct inode *inode = file->f_mapping->host;
|
|
|
|
|
|
|
|
return blockdev_direct_IO_no_locking(rw, iocb, inode, I_BDEV(inode),
|
|
|
|
iov, offset, nr_segs, blkdev_get_blocks, NULL);
|
|
|
|
}
|
|
|
|
|
2009-02-25 10:44:19 +01:00
|
|
|
/*
|
|
|
|
* Write out and wait upon all the dirty data associated with a block
|
|
|
|
* device via its mapping. Does not take the superblock lock.
|
|
|
|
*/
|
|
|
|
int sync_blockdev(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (bdev)
|
|
|
|
ret = filemap_write_and_wait(bdev->bd_inode->i_mapping);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(sync_blockdev);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Write out and wait upon all dirty data associated with this
|
|
|
|
* device. Filesystem data as well as the underlying block
|
|
|
|
* device. Takes the superblock lock.
|
|
|
|
*/
|
|
|
|
int fsync_bdev(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
struct super_block *sb = get_super(bdev);
|
|
|
|
if (sb) {
|
|
|
|
int res = fsync_super(sb);
|
|
|
|
drop_super(sb);
|
|
|
|
return res;
|
|
|
|
}
|
|
|
|
return sync_blockdev(bdev);
|
|
|
|
}
|
2009-04-01 13:07:16 +02:00
|
|
|
EXPORT_SYMBOL(fsync_bdev);
|
2009-02-25 10:44:19 +01:00
|
|
|
|
|
|
|
/**
|
|
|
|
* freeze_bdev -- lock a filesystem and force it into a consistent state
|
|
|
|
* @bdev: blockdevice to lock
|
|
|
|
*
|
|
|
|
* This takes the block device bd_mount_sem to make sure no new mounts
|
|
|
|
* happen on bdev until thaw_bdev() is called.
|
|
|
|
* If a superblock is found on this device, we take the s_umount semaphore
|
|
|
|
* on it to make sure nobody unmounts until the snapshot creation is done.
|
|
|
|
* The reference counter (bd_fsfreeze_count) guarantees that only the last
|
|
|
|
* unfreeze process can unfreeze the frozen filesystem actually when multiple
|
|
|
|
* freeze requests arrive simultaneously. It counts up in freeze_bdev() and
|
|
|
|
* count down in thaw_bdev(). When it becomes 0, thaw_bdev() will unfreeze
|
|
|
|
* actually.
|
|
|
|
*/
|
|
|
|
struct super_block *freeze_bdev(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
struct super_block *sb;
|
|
|
|
int error = 0;
|
|
|
|
|
|
|
|
mutex_lock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
if (bdev->bd_fsfreeze_count > 0) {
|
|
|
|
bdev->bd_fsfreeze_count++;
|
|
|
|
sb = get_super(bdev);
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
return sb;
|
|
|
|
}
|
|
|
|
bdev->bd_fsfreeze_count++;
|
|
|
|
|
|
|
|
down(&bdev->bd_mount_sem);
|
|
|
|
sb = get_super(bdev);
|
|
|
|
if (sb && !(sb->s_flags & MS_RDONLY)) {
|
|
|
|
sb->s_frozen = SB_FREEZE_WRITE;
|
|
|
|
smp_wmb();
|
|
|
|
|
|
|
|
__fsync_super(sb);
|
|
|
|
|
|
|
|
sb->s_frozen = SB_FREEZE_TRANS;
|
|
|
|
smp_wmb();
|
|
|
|
|
|
|
|
sync_blockdev(sb->s_bdev);
|
|
|
|
|
|
|
|
if (sb->s_op->freeze_fs) {
|
|
|
|
error = sb->s_op->freeze_fs(sb);
|
|
|
|
if (error) {
|
|
|
|
printk(KERN_ERR
|
|
|
|
"VFS:Filesystem freeze failed\n");
|
|
|
|
sb->s_frozen = SB_UNFROZEN;
|
|
|
|
drop_super(sb);
|
|
|
|
up(&bdev->bd_mount_sem);
|
|
|
|
bdev->bd_fsfreeze_count--;
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
return ERR_PTR(error);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
sync_blockdev(bdev);
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
|
|
|
|
return sb; /* thaw_bdev releases s->s_umount and bd_mount_sem */
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(freeze_bdev);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* thaw_bdev -- unlock filesystem
|
|
|
|
* @bdev: blockdevice to unlock
|
|
|
|
* @sb: associated superblock
|
|
|
|
*
|
|
|
|
* Unlocks the filesystem and marks it writeable again after freeze_bdev().
|
|
|
|
*/
|
|
|
|
int thaw_bdev(struct block_device *bdev, struct super_block *sb)
|
|
|
|
{
|
|
|
|
int error = 0;
|
|
|
|
|
|
|
|
mutex_lock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
if (!bdev->bd_fsfreeze_count) {
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
bdev->bd_fsfreeze_count--;
|
|
|
|
if (bdev->bd_fsfreeze_count > 0) {
|
|
|
|
if (sb)
|
|
|
|
drop_super(sb);
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (sb) {
|
|
|
|
BUG_ON(sb->s_bdev != bdev);
|
|
|
|
if (!(sb->s_flags & MS_RDONLY)) {
|
|
|
|
if (sb->s_op->unfreeze_fs) {
|
|
|
|
error = sb->s_op->unfreeze_fs(sb);
|
|
|
|
if (error) {
|
|
|
|
printk(KERN_ERR
|
|
|
|
"VFS:Filesystem thaw failed\n");
|
|
|
|
sb->s_frozen = SB_FREEZE_TRANS;
|
|
|
|
bdev->bd_fsfreeze_count++;
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sb->s_frozen = SB_UNFROZEN;
|
|
|
|
smp_wmb();
|
|
|
|
wake_up(&sb->s_wait_unfrozen);
|
|
|
|
}
|
|
|
|
drop_super(sb);
|
|
|
|
}
|
|
|
|
|
|
|
|
up(&bdev->bd_mount_sem);
|
|
|
|
mutex_unlock(&bdev->bd_fsfreeze_mutex);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(thaw_bdev);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
static int blkdev_writepage(struct page *page, struct writeback_control *wbc)
|
|
|
|
{
|
|
|
|
return block_write_full_page(page, blkdev_get_block, wbc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int blkdev_readpage(struct file * file, struct page * page)
|
|
|
|
{
|
|
|
|
return block_read_full_page(page, blkdev_get_block);
|
|
|
|
}
|
|
|
|
|
2007-10-16 10:25:04 +02:00
|
|
|
static int blkdev_write_begin(struct file *file, struct address_space *mapping,
|
|
|
|
loff_t pos, unsigned len, unsigned flags,
|
|
|
|
struct page **pagep, void **fsdata)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-10-16 10:25:04 +02:00
|
|
|
*pagep = NULL;
|
|
|
|
return block_write_begin(file, mapping, pos, len, flags, pagep, fsdata,
|
|
|
|
blkdev_get_block);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2007-10-16 10:25:04 +02:00
|
|
|
static int blkdev_write_end(struct file *file, struct address_space *mapping,
|
|
|
|
loff_t pos, unsigned len, unsigned copied,
|
|
|
|
struct page *page, void *fsdata)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-10-16 10:25:04 +02:00
|
|
|
int ret;
|
|
|
|
ret = block_write_end(file, mapping, pos, len, copied, page, fsdata);
|
|
|
|
|
|
|
|
unlock_page(page);
|
|
|
|
page_cache_release(page);
|
|
|
|
|
|
|
|
return ret;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* private llseek:
|
2006-12-08 11:36:35 +01:00
|
|
|
* for a block special file file->f_path.dentry->d_inode->i_size is zero
|
2005-04-17 00:20:36 +02:00
|
|
|
* so we compute the size by hand (just as in block_read/write above)
|
|
|
|
*/
|
|
|
|
static loff_t block_llseek(struct file *file, loff_t offset, int origin)
|
|
|
|
{
|
|
|
|
struct inode *bd_inode = file->f_mapping->host;
|
|
|
|
loff_t size;
|
|
|
|
loff_t retval;
|
|
|
|
|
2006-01-10 00:59:24 +01:00
|
|
|
mutex_lock(&bd_inode->i_mutex);
|
2005-04-17 00:20:36 +02:00
|
|
|
size = i_size_read(bd_inode);
|
|
|
|
|
|
|
|
switch (origin) {
|
|
|
|
case 2:
|
|
|
|
offset += size;
|
|
|
|
break;
|
|
|
|
case 1:
|
|
|
|
offset += file->f_pos;
|
|
|
|
}
|
|
|
|
retval = -EINVAL;
|
|
|
|
if (offset >= 0 && offset <= size) {
|
|
|
|
if (offset != file->f_pos) {
|
|
|
|
file->f_pos = offset;
|
|
|
|
}
|
|
|
|
retval = offset;
|
|
|
|
}
|
2006-01-10 00:59:24 +01:00
|
|
|
mutex_unlock(&bd_inode->i_mutex);
|
2005-04-17 00:20:36 +02:00
|
|
|
return retval;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Filp is never NULL; the only case when ->fsync() is called with
|
|
|
|
* NULL first argument is nfsd_sync_dir() and that's not a directory.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int block_fsync(struct file *filp, struct dentry *dentry, int datasync)
|
|
|
|
{
|
|
|
|
return sync_blockdev(I_BDEV(filp->f_mapping->host));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* pseudo-fs
|
|
|
|
*/
|
|
|
|
|
|
|
|
static __cacheline_aligned_in_smp DEFINE_SPINLOCK(bdev_lock);
|
2006-12-07 05:33:20 +01:00
|
|
|
static struct kmem_cache * bdev_cachep __read_mostly;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
static struct inode *bdev_alloc_inode(struct super_block *sb)
|
|
|
|
{
|
2006-12-07 05:33:17 +01:00
|
|
|
struct bdev_inode *ei = kmem_cache_alloc(bdev_cachep, GFP_KERNEL);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (!ei)
|
|
|
|
return NULL;
|
|
|
|
return &ei->vfs_inode;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void bdev_destroy_inode(struct inode *inode)
|
|
|
|
{
|
|
|
|
struct bdev_inode *bdi = BDEV_I(inode);
|
|
|
|
|
|
|
|
bdi->bdev.bd_inode_backing_dev_info = NULL;
|
|
|
|
kmem_cache_free(bdev_cachep, bdi);
|
|
|
|
}
|
|
|
|
|
2008-07-26 04:45:34 +02:00
|
|
|
static void init_once(void *foo)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct bdev_inode *ei = (struct bdev_inode *) foo;
|
|
|
|
struct block_device *bdev = &ei->bdev;
|
|
|
|
|
2007-05-17 07:10:57 +02:00
|
|
|
memset(bdev, 0, sizeof(*bdev));
|
|
|
|
mutex_init(&bdev->bd_mutex);
|
|
|
|
sema_init(&bdev->bd_mount_sem, 1);
|
|
|
|
INIT_LIST_HEAD(&bdev->bd_inodes);
|
|
|
|
INIT_LIST_HEAD(&bdev->bd_list);
|
2006-03-27 11:17:57 +02:00
|
|
|
#ifdef CONFIG_SYSFS
|
2007-05-17 07:10:57 +02:00
|
|
|
INIT_LIST_HEAD(&bdev->bd_holder_list);
|
2006-03-27 11:17:57 +02:00
|
|
|
#endif
|
2007-05-17 07:10:57 +02:00
|
|
|
inode_init_once(&ei->vfs_inode);
|
2009-01-10 01:40:59 +01:00
|
|
|
/* Initialize mutex for freeze. */
|
|
|
|
mutex_init(&bdev->bd_fsfreeze_mutex);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void __bd_forget(struct inode *inode)
|
|
|
|
{
|
|
|
|
list_del_init(&inode->i_devices);
|
|
|
|
inode->i_bdev = NULL;
|
|
|
|
inode->i_mapping = &inode->i_data;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void bdev_clear_inode(struct inode *inode)
|
|
|
|
{
|
|
|
|
struct block_device *bdev = &BDEV_I(inode)->bdev;
|
|
|
|
struct list_head *p;
|
|
|
|
spin_lock(&bdev_lock);
|
|
|
|
while ( (p = bdev->bd_inodes.next) != &bdev->bd_inodes ) {
|
|
|
|
__bd_forget(list_entry(p, struct inode, i_devices));
|
|
|
|
}
|
|
|
|
list_del_init(&bdev->bd_list);
|
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
}
|
|
|
|
|
2007-02-12 09:55:41 +01:00
|
|
|
static const struct super_operations bdev_sops = {
|
2005-04-17 00:20:36 +02:00
|
|
|
.statfs = simple_statfs,
|
|
|
|
.alloc_inode = bdev_alloc_inode,
|
|
|
|
.destroy_inode = bdev_destroy_inode,
|
|
|
|
.drop_inode = generic_delete_inode,
|
|
|
|
.clear_inode = bdev_clear_inode,
|
|
|
|
};
|
|
|
|
|
[PATCH] VFS: Permit filesystem to override root dentry on mount
Extend the get_sb() filesystem operation to take an extra argument that
permits the VFS to pass in the target vfsmount that defines the mountpoint.
The filesystem is then required to manually set the superblock and root dentry
pointers. For most filesystems, this should be done with simple_set_mnt()
which will set the superblock pointer and then set the root dentry to the
superblock's s_root (as per the old default behaviour).
The get_sb() op now returns an integer as there's now no need to return the
superblock pointer.
This patch permits a superblock to be implicitly shared amongst several mount
points, such as can be done with NFS to avoid potential inode aliasing. In
such a case, simple_set_mnt() would not be called, and instead the mnt_root
and mnt_sb would be set directly.
The patch also makes the following changes:
(*) the get_sb_*() convenience functions in the core kernel now take a vfsmount
pointer argument and return an integer, so most filesystems have to change
very little.
(*) If one of the convenience function is not used, then get_sb() should
normally call simple_set_mnt() to instantiate the vfsmount. This will
always return 0, and so can be tail-called from get_sb().
(*) generic_shutdown_super() now calls shrink_dcache_sb() to clean up the
dcache upon superblock destruction rather than shrink_dcache_anon().
This is required because the superblock may now have multiple trees that
aren't actually bound to s_root, but that still need to be cleaned up. The
currently called functions assume that the whole tree is rooted at s_root,
and that anonymous dentries are not the roots of trees which results in
dentries being left unculled.
However, with the way NFS superblock sharing are currently set to be
implemented, these assumptions are violated: the root of the filesystem is
simply a dummy dentry and inode (the real inode for '/' may well be
inaccessible), and all the vfsmounts are rooted on anonymous[*] dentries
with child trees.
[*] Anonymous until discovered from another tree.
(*) The documentation has been adjusted, including the additional bit of
changing ext2_* into foo_* in the documentation.
[akpm@osdl.org: convert ipath_fs, do other stuff]
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Nathan Scott <nathans@sgi.com>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 11:02:57 +02:00
|
|
|
static int bd_get_sb(struct file_system_type *fs_type,
|
|
|
|
int flags, const char *dev_name, void *data, struct vfsmount *mnt)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
[PATCH] VFS: Permit filesystem to override root dentry on mount
Extend the get_sb() filesystem operation to take an extra argument that
permits the VFS to pass in the target vfsmount that defines the mountpoint.
The filesystem is then required to manually set the superblock and root dentry
pointers. For most filesystems, this should be done with simple_set_mnt()
which will set the superblock pointer and then set the root dentry to the
superblock's s_root (as per the old default behaviour).
The get_sb() op now returns an integer as there's now no need to return the
superblock pointer.
This patch permits a superblock to be implicitly shared amongst several mount
points, such as can be done with NFS to avoid potential inode aliasing. In
such a case, simple_set_mnt() would not be called, and instead the mnt_root
and mnt_sb would be set directly.
The patch also makes the following changes:
(*) the get_sb_*() convenience functions in the core kernel now take a vfsmount
pointer argument and return an integer, so most filesystems have to change
very little.
(*) If one of the convenience function is not used, then get_sb() should
normally call simple_set_mnt() to instantiate the vfsmount. This will
always return 0, and so can be tail-called from get_sb().
(*) generic_shutdown_super() now calls shrink_dcache_sb() to clean up the
dcache upon superblock destruction rather than shrink_dcache_anon().
This is required because the superblock may now have multiple trees that
aren't actually bound to s_root, but that still need to be cleaned up. The
currently called functions assume that the whole tree is rooted at s_root,
and that anonymous dentries are not the roots of trees which results in
dentries being left unculled.
However, with the way NFS superblock sharing are currently set to be
implemented, these assumptions are violated: the root of the filesystem is
simply a dummy dentry and inode (the real inode for '/' may well be
inaccessible), and all the vfsmounts are rooted on anonymous[*] dentries
with child trees.
[*] Anonymous until discovered from another tree.
(*) The documentation has been adjusted, including the additional bit of
changing ext2_* into foo_* in the documentation.
[akpm@osdl.org: convert ipath_fs, do other stuff]
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Nathan Scott <nathans@sgi.com>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 11:02:57 +02:00
|
|
|
return get_sb_pseudo(fs_type, "bdev:", &bdev_sops, 0x62646576, mnt);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static struct file_system_type bd_type = {
|
|
|
|
.name = "bdev",
|
|
|
|
.get_sb = bd_get_sb,
|
|
|
|
.kill_sb = kill_anon_super,
|
|
|
|
};
|
|
|
|
|
2008-12-01 23:34:56 +01:00
|
|
|
struct super_block *blockdev_superblock __read_mostly;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
void __init bdev_cache_init(void)
|
|
|
|
{
|
|
|
|
int err;
|
2008-12-01 23:34:56 +01:00
|
|
|
struct vfsmount *bd_mnt;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev_cachep = kmem_cache_create("bdev_cache", sizeof(struct bdev_inode),
|
2006-03-24 12:16:06 +01:00
|
|
|
0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
|
|
|
|
SLAB_MEM_SPREAD|SLAB_PANIC),
|
2007-07-20 03:11:58 +02:00
|
|
|
init_once);
|
2005-04-17 00:20:36 +02:00
|
|
|
err = register_filesystem(&bd_type);
|
|
|
|
if (err)
|
|
|
|
panic("Cannot register bdev pseudo-fs");
|
|
|
|
bd_mnt = kern_mount(&bd_type);
|
|
|
|
if (IS_ERR(bd_mnt))
|
|
|
|
panic("Cannot create bdev pseudo-fs");
|
2009-06-11 14:24:13 +02:00
|
|
|
/*
|
|
|
|
* This vfsmount structure is only used to obtain the
|
|
|
|
* blockdev_superblock, so tell kmemleak not to report it.
|
|
|
|
*/
|
|
|
|
kmemleak_not_leak(bd_mnt);
|
2005-04-17 00:20:36 +02:00
|
|
|
blockdev_superblock = bd_mnt->mnt_sb; /* For writeback */
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Most likely _very_ bad one - but then it's hardly critical for small
|
|
|
|
* /dev and can be fixed when somebody will need really large one.
|
|
|
|
* Keep in mind that it will be fed through icache hash function too.
|
|
|
|
*/
|
|
|
|
static inline unsigned long hash(dev_t dev)
|
|
|
|
{
|
|
|
|
return MAJOR(dev)+MINOR(dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int bdev_test(struct inode *inode, void *data)
|
|
|
|
{
|
|
|
|
return BDEV_I(inode)->bdev.bd_dev == *(dev_t *)data;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int bdev_set(struct inode *inode, void *data)
|
|
|
|
{
|
|
|
|
BDEV_I(inode)->bdev.bd_dev = *(dev_t *)data;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static LIST_HEAD(all_bdevs);
|
|
|
|
|
|
|
|
struct block_device *bdget(dev_t dev)
|
|
|
|
{
|
|
|
|
struct block_device *bdev;
|
|
|
|
struct inode *inode;
|
|
|
|
|
2008-12-01 23:34:56 +01:00
|
|
|
inode = iget5_locked(blockdev_superblock, hash(dev),
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev_test, bdev_set, &dev);
|
|
|
|
|
|
|
|
if (!inode)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
bdev = &BDEV_I(inode)->bdev;
|
|
|
|
|
|
|
|
if (inode->i_state & I_NEW) {
|
|
|
|
bdev->bd_contains = NULL;
|
|
|
|
bdev->bd_inode = inode;
|
|
|
|
bdev->bd_block_size = (1 << inode->i_blkbits);
|
|
|
|
bdev->bd_part_count = 0;
|
|
|
|
bdev->bd_invalidated = 0;
|
|
|
|
inode->i_mode = S_IFBLK;
|
|
|
|
inode->i_rdev = dev;
|
|
|
|
inode->i_bdev = bdev;
|
|
|
|
inode->i_data.a_ops = &def_blk_aops;
|
|
|
|
mapping_set_gfp_mask(&inode->i_data, GFP_USER);
|
|
|
|
inode->i_data.backing_dev_info = &default_backing_dev_info;
|
|
|
|
spin_lock(&bdev_lock);
|
|
|
|
list_add(&bdev->bd_list, &all_bdevs);
|
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
unlock_new_inode(inode);
|
|
|
|
}
|
|
|
|
return bdev;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(bdget);
|
|
|
|
|
|
|
|
long nr_blockdev_pages(void)
|
|
|
|
{
|
2007-07-16 08:40:23 +02:00
|
|
|
struct block_device *bdev;
|
2005-04-17 00:20:36 +02:00
|
|
|
long ret = 0;
|
|
|
|
spin_lock(&bdev_lock);
|
2007-07-16 08:40:23 +02:00
|
|
|
list_for_each_entry(bdev, &all_bdevs, bd_list) {
|
2005-04-17 00:20:36 +02:00
|
|
|
ret += bdev->bd_inode->i_mapping->nrpages;
|
|
|
|
}
|
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
void bdput(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
iput(bdev->bd_inode);
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(bdput);
|
|
|
|
|
|
|
|
static struct block_device *bd_acquire(struct inode *inode)
|
|
|
|
{
|
|
|
|
struct block_device *bdev;
|
2006-06-22 23:47:21 +02:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_lock(&bdev_lock);
|
|
|
|
bdev = inode->i_bdev;
|
2006-06-22 23:47:21 +02:00
|
|
|
if (bdev) {
|
|
|
|
atomic_inc(&bdev->bd_inode->i_count);
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
return bdev;
|
|
|
|
}
|
|
|
|
spin_unlock(&bdev_lock);
|
2006-06-22 23:47:21 +02:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev = bdget(inode->i_rdev);
|
|
|
|
if (bdev) {
|
|
|
|
spin_lock(&bdev_lock);
|
2006-06-22 23:47:21 +02:00
|
|
|
if (!inode->i_bdev) {
|
|
|
|
/*
|
|
|
|
* We take an additional bd_inode->i_count for inode,
|
|
|
|
* and it's released in clear_inode() of inode.
|
|
|
|
* So, we can access it via ->i_mapping always
|
|
|
|
* without igrab().
|
|
|
|
*/
|
|
|
|
atomic_inc(&bdev->bd_inode->i_count);
|
|
|
|
inode->i_bdev = bdev;
|
|
|
|
inode->i_mapping = bdev->bd_inode->i_mapping;
|
|
|
|
list_add(&inode->i_devices, &bdev->bd_inodes);
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
}
|
|
|
|
return bdev;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Call when you free inode */
|
|
|
|
|
|
|
|
void bd_forget(struct inode *inode)
|
|
|
|
{
|
2006-06-22 23:47:21 +02:00
|
|
|
struct block_device *bdev = NULL;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_lock(&bdev_lock);
|
2006-06-22 23:47:21 +02:00
|
|
|
if (inode->i_bdev) {
|
2008-12-01 23:34:56 +01:00
|
|
|
if (!sb_is_blkdev_sb(inode->i_sb))
|
2006-06-22 23:47:21 +02:00
|
|
|
bdev = inode->i_bdev;
|
2005-04-17 00:20:36 +02:00
|
|
|
__bd_forget(inode);
|
2006-06-22 23:47:21 +02:00
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_unlock(&bdev_lock);
|
2006-06-22 23:47:21 +02:00
|
|
|
|
|
|
|
if (bdev)
|
|
|
|
iput(bdev->bd_inode);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
int bd_claim(struct block_device *bdev, void *holder)
|
|
|
|
{
|
|
|
|
int res;
|
|
|
|
spin_lock(&bdev_lock);
|
|
|
|
|
|
|
|
/* first decide result */
|
|
|
|
if (bdev->bd_holder == holder)
|
|
|
|
res = 0; /* already a holder */
|
|
|
|
else if (bdev->bd_holder != NULL)
|
|
|
|
res = -EBUSY; /* held by someone else */
|
|
|
|
else if (bdev->bd_contains == bdev)
|
|
|
|
res = 0; /* is a whole device which isn't held */
|
|
|
|
|
|
|
|
else if (bdev->bd_contains->bd_holder == bd_claim)
|
|
|
|
res = 0; /* is a partition of a device that is being partitioned */
|
|
|
|
else if (bdev->bd_contains->bd_holder != NULL)
|
|
|
|
res = -EBUSY; /* is a partition of a held device */
|
|
|
|
else
|
|
|
|
res = 0; /* is a partition of an un-held device */
|
|
|
|
|
|
|
|
/* now impose change */
|
|
|
|
if (res==0) {
|
|
|
|
/* note that for a whole device bd_holders
|
|
|
|
* will be incremented twice, and bd_holder will
|
|
|
|
* be set to bd_claim before being set to holder
|
|
|
|
*/
|
|
|
|
bdev->bd_contains->bd_holders ++;
|
|
|
|
bdev->bd_contains->bd_holder = bd_claim;
|
|
|
|
bdev->bd_holders++;
|
|
|
|
bdev->bd_holder = holder;
|
|
|
|
}
|
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
return res;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(bd_claim);
|
|
|
|
|
|
|
|
void bd_release(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
spin_lock(&bdev_lock);
|
|
|
|
if (!--bdev->bd_contains->bd_holders)
|
|
|
|
bdev->bd_contains->bd_holder = NULL;
|
|
|
|
if (!--bdev->bd_holders)
|
|
|
|
bdev->bd_holder = NULL;
|
|
|
|
spin_unlock(&bdev_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(bd_release);
|
|
|
|
|
2006-03-27 11:17:57 +02:00
|
|
|
#ifdef CONFIG_SYSFS
|
|
|
|
/*
|
|
|
|
* Functions for bd_claim_by_kobject / bd_release_from_kobject
|
|
|
|
*
|
|
|
|
* If a kobject is passed to bd_claim_by_kobject()
|
|
|
|
* and the kobject has a parent directory,
|
|
|
|
* following symlinks are created:
|
|
|
|
* o from the kobject to the claimed bdev
|
|
|
|
* o from "holders" directory of the bdev to the parent of the kobject
|
|
|
|
* bd_release_from_kobject() removes these symlinks.
|
|
|
|
*
|
|
|
|
* Example:
|
|
|
|
* If /dev/dm-0 maps to /dev/sda, kobject corresponding to
|
|
|
|
* /sys/block/dm-0/slaves is passed to bd_claim_by_kobject(), then:
|
|
|
|
* /sys/block/dm-0/slaves/sda --> /sys/block/sda
|
|
|
|
* /sys/block/sda/holders/dm-0 --> /sys/block/dm-0
|
|
|
|
*/
|
|
|
|
|
[PATCH] blockdev.c: check driver layer errors
Check driver layer errors.
Fix from: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
In blockdevc-check-errors.patch, add_bd_holder() is modified to return error
values when some of its operation failed. Among them, it returns -EEXIST when
a given bd_holder object already exists in the list.
However, in this case, the function completed its work successfully and need
no action by its caller other than freeing unused bd_holder object. So I
think it's better to return success after freeing by itself.
Otherwise, bd_claim-ing with same claim pointer will fail.
Typically, lvresize will fails with following message:
device-mapper: reload ioctl failed: Invalid argument
and you'll see messages like below in kernel log:
device-mapper: table: 254:13: linear: dm-linear: Device lookup failed
device-mapper: ioctl: error adding target to table
Similarly, it should not add bd_holder to the list if either one of symlinking
fails. I don't have a test case for this to happen but it should cause
dereference of freed pointer.
If a matching bd_holder is found in bd_holder_list, add_bd_holder() completes
its job by just incrementing the reference count. In this case, it should be
considered as success but it used to return 'fail' to let the caller free
temporary bd_holder. Fixed it to return success and free given object by
itself.
Also, if either one of symlinking fails, the bd_holder should not be added to
the list so that it can be discarded later. Otherwise, the caller will free
bd_holder which is in the list.
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 10:58:56 +02:00
|
|
|
static int add_symlink(struct kobject *from, struct kobject *to)
|
2006-03-27 11:17:57 +02:00
|
|
|
{
|
|
|
|
if (!from || !to)
|
[PATCH] blockdev.c: check driver layer errors
Check driver layer errors.
Fix from: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
In blockdevc-check-errors.patch, add_bd_holder() is modified to return error
values when some of its operation failed. Among them, it returns -EEXIST when
a given bd_holder object already exists in the list.
However, in this case, the function completed its work successfully and need
no action by its caller other than freeing unused bd_holder object. So I
think it's better to return success after freeing by itself.
Otherwise, bd_claim-ing with same claim pointer will fail.
Typically, lvresize will fails with following message:
device-mapper: reload ioctl failed: Invalid argument
and you'll see messages like below in kernel log:
device-mapper: table: 254:13: linear: dm-linear: Device lookup failed
device-mapper: ioctl: error adding target to table
Similarly, it should not add bd_holder to the list if either one of symlinking
fails. I don't have a test case for this to happen but it should cause
dereference of freed pointer.
If a matching bd_holder is found in bd_holder_list, add_bd_holder() completes
its job by just incrementing the reference count. In this case, it should be
considered as success but it used to return 'fail' to let the caller free
temporary bd_holder. Fixed it to return success and free given object by
itself.
Also, if either one of symlinking fails, the bd_holder should not be added to
the list so that it can be discarded later. Otherwise, the caller will free
bd_holder which is in the list.
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 10:58:56 +02:00
|
|
|
return 0;
|
|
|
|
return sysfs_create_link(from, to, kobject_name(to));
|
2006-03-27 11:17:57 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static void del_symlink(struct kobject *from, struct kobject *to)
|
|
|
|
{
|
|
|
|
if (!from || !to)
|
|
|
|
return;
|
|
|
|
sysfs_remove_link(from, kobject_name(to));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* 'struct bd_holder' contains pointers to kobjects symlinked by
|
|
|
|
* bd_claim_by_kobject.
|
|
|
|
* It's connected to bd_holder_list which is protected by bdev->bd_sem.
|
|
|
|
*/
|
|
|
|
struct bd_holder {
|
|
|
|
struct list_head list; /* chain of holders of the bdev */
|
|
|
|
int count; /* references from the holder */
|
|
|
|
struct kobject *sdir; /* holder object, e.g. "/block/dm-0/slaves" */
|
|
|
|
struct kobject *hdev; /* e.g. "/block/dm-0" */
|
|
|
|
struct kobject *hdir; /* e.g. "/block/sda/holders" */
|
|
|
|
struct kobject *sdev; /* e.g. "/block/sda" */
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Get references of related kobjects at once.
|
|
|
|
* Returns 1 on success. 0 on failure.
|
|
|
|
*
|
|
|
|
* Should call bd_holder_release_dirs() after successful use.
|
|
|
|
*/
|
|
|
|
static int bd_holder_grab_dirs(struct block_device *bdev,
|
|
|
|
struct bd_holder *bo)
|
|
|
|
{
|
|
|
|
if (!bdev || !bo)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
bo->sdir = kobject_get(bo->sdir);
|
|
|
|
if (!bo->sdir)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
bo->hdev = kobject_get(bo->sdir->parent);
|
|
|
|
if (!bo->hdev)
|
|
|
|
goto fail_put_sdir;
|
|
|
|
|
2008-08-25 12:56:12 +02:00
|
|
|
bo->sdev = kobject_get(&part_to_dev(bdev->bd_part)->kobj);
|
2006-03-27 11:17:57 +02:00
|
|
|
if (!bo->sdev)
|
|
|
|
goto fail_put_hdev;
|
|
|
|
|
2008-08-25 12:56:11 +02:00
|
|
|
bo->hdir = kobject_get(bdev->bd_part->holder_dir);
|
2006-03-27 11:17:57 +02:00
|
|
|
if (!bo->hdir)
|
|
|
|
goto fail_put_sdev;
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
fail_put_sdev:
|
|
|
|
kobject_put(bo->sdev);
|
|
|
|
fail_put_hdev:
|
|
|
|
kobject_put(bo->hdev);
|
|
|
|
fail_put_sdir:
|
|
|
|
kobject_put(bo->sdir);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Put references of related kobjects at once. */
|
|
|
|
static void bd_holder_release_dirs(struct bd_holder *bo)
|
|
|
|
{
|
|
|
|
kobject_put(bo->hdir);
|
|
|
|
kobject_put(bo->sdev);
|
|
|
|
kobject_put(bo->hdev);
|
|
|
|
kobject_put(bo->sdir);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct bd_holder *alloc_bd_holder(struct kobject *kobj)
|
|
|
|
{
|
|
|
|
struct bd_holder *bo;
|
|
|
|
|
|
|
|
bo = kzalloc(sizeof(*bo), GFP_KERNEL);
|
|
|
|
if (!bo)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
bo->count = 1;
|
|
|
|
bo->sdir = kobj;
|
|
|
|
|
|
|
|
return bo;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void free_bd_holder(struct bd_holder *bo)
|
|
|
|
{
|
|
|
|
kfree(bo);
|
|
|
|
}
|
|
|
|
|
2006-10-30 22:23:56 +01:00
|
|
|
/**
|
|
|
|
* find_bd_holder - find matching struct bd_holder from the block device
|
|
|
|
*
|
|
|
|
* @bdev: struct block device to be searched
|
|
|
|
* @bo: target struct bd_holder
|
|
|
|
*
|
|
|
|
* Returns matching entry with @bo in @bdev->bd_holder_list.
|
|
|
|
* If found, increment the reference count and return the pointer.
|
|
|
|
* If not found, returns NULL.
|
|
|
|
*/
|
2006-10-31 07:07:03 +01:00
|
|
|
static struct bd_holder *find_bd_holder(struct block_device *bdev,
|
|
|
|
struct bd_holder *bo)
|
2006-10-30 22:23:56 +01:00
|
|
|
{
|
|
|
|
struct bd_holder *tmp;
|
|
|
|
|
|
|
|
list_for_each_entry(tmp, &bdev->bd_holder_list, list)
|
|
|
|
if (tmp->sdir == bo->sdir) {
|
|
|
|
tmp->count++;
|
|
|
|
return tmp;
|
|
|
|
}
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2006-03-27 11:17:57 +02:00
|
|
|
/**
|
|
|
|
* add_bd_holder - create sysfs symlinks for bd_claim() relationship
|
|
|
|
*
|
|
|
|
* @bdev: block device to be bd_claimed
|
|
|
|
* @bo: preallocated and initialized by alloc_bd_holder()
|
|
|
|
*
|
2006-10-30 22:23:56 +01:00
|
|
|
* Add @bo to @bdev->bd_holder_list, create symlinks.
|
2006-03-27 11:17:57 +02:00
|
|
|
*
|
2006-10-30 22:23:56 +01:00
|
|
|
* Returns 0 if symlinks are created.
|
|
|
|
* Returns -ve if something fails.
|
2006-03-27 11:17:57 +02:00
|
|
|
*/
|
|
|
|
static int add_bd_holder(struct block_device *bdev, struct bd_holder *bo)
|
|
|
|
{
|
2007-07-16 08:41:25 +02:00
|
|
|
int err;
|
2006-03-27 11:17:57 +02:00
|
|
|
|
|
|
|
if (!bo)
|
[PATCH] blockdev.c: check driver layer errors
Check driver layer errors.
Fix from: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
In blockdevc-check-errors.patch, add_bd_holder() is modified to return error
values when some of its operation failed. Among them, it returns -EEXIST when
a given bd_holder object already exists in the list.
However, in this case, the function completed its work successfully and need
no action by its caller other than freeing unused bd_holder object. So I
think it's better to return success after freeing by itself.
Otherwise, bd_claim-ing with same claim pointer will fail.
Typically, lvresize will fails with following message:
device-mapper: reload ioctl failed: Invalid argument
and you'll see messages like below in kernel log:
device-mapper: table: 254:13: linear: dm-linear: Device lookup failed
device-mapper: ioctl: error adding target to table
Similarly, it should not add bd_holder to the list if either one of symlinking
fails. I don't have a test case for this to happen but it should cause
dereference of freed pointer.
If a matching bd_holder is found in bd_holder_list, add_bd_holder() completes
its job by just incrementing the reference count. In this case, it should be
considered as success but it used to return 'fail' to let the caller free
temporary bd_holder. Fixed it to return success and free given object by
itself.
Also, if either one of symlinking fails, the bd_holder should not be added to
the list so that it can be discarded later. Otherwise, the caller will free
bd_holder which is in the list.
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 10:58:56 +02:00
|
|
|
return -EINVAL;
|
2006-03-27 11:17:57 +02:00
|
|
|
|
|
|
|
if (!bd_holder_grab_dirs(bdev, bo))
|
[PATCH] blockdev.c: check driver layer errors
Check driver layer errors.
Fix from: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
In blockdevc-check-errors.patch, add_bd_holder() is modified to return error
values when some of its operation failed. Among them, it returns -EEXIST when
a given bd_holder object already exists in the list.
However, in this case, the function completed its work successfully and need
no action by its caller other than freeing unused bd_holder object. So I
think it's better to return success after freeing by itself.
Otherwise, bd_claim-ing with same claim pointer will fail.
Typically, lvresize will fails with following message:
device-mapper: reload ioctl failed: Invalid argument
and you'll see messages like below in kernel log:
device-mapper: table: 254:13: linear: dm-linear: Device lookup failed
device-mapper: ioctl: error adding target to table
Similarly, it should not add bd_holder to the list if either one of symlinking
fails. I don't have a test case for this to happen but it should cause
dereference of freed pointer.
If a matching bd_holder is found in bd_holder_list, add_bd_holder() completes
its job by just incrementing the reference count. In this case, it should be
considered as success but it used to return 'fail' to let the caller free
temporary bd_holder. Fixed it to return success and free given object by
itself.
Also, if either one of symlinking fails, the bd_holder should not be added to
the list so that it can be discarded later. Otherwise, the caller will free
bd_holder which is in the list.
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 10:58:56 +02:00
|
|
|
return -EBUSY;
|
2006-03-27 11:17:57 +02:00
|
|
|
|
2007-07-16 08:41:25 +02:00
|
|
|
err = add_symlink(bo->sdir, bo->sdev);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
err = add_symlink(bo->hdir, bo->hdev);
|
|
|
|
if (err) {
|
|
|
|
del_symlink(bo->sdir, bo->sdev);
|
|
|
|
return err;
|
[PATCH] blockdev.c: check driver layer errors
Check driver layer errors.
Fix from: "Jun'ichi Nomura" <j-nomura@ce.jp.nec.com>
In blockdevc-check-errors.patch, add_bd_holder() is modified to return error
values when some of its operation failed. Among them, it returns -EEXIST when
a given bd_holder object already exists in the list.
However, in this case, the function completed its work successfully and need
no action by its caller other than freeing unused bd_holder object. So I
think it's better to return success after freeing by itself.
Otherwise, bd_claim-ing with same claim pointer will fail.
Typically, lvresize will fails with following message:
device-mapper: reload ioctl failed: Invalid argument
and you'll see messages like below in kernel log:
device-mapper: table: 254:13: linear: dm-linear: Device lookup failed
device-mapper: ioctl: error adding target to table
Similarly, it should not add bd_holder to the list if either one of symlinking
fails. I don't have a test case for this to happen but it should cause
dereference of freed pointer.
If a matching bd_holder is found in bd_holder_list, add_bd_holder() completes
its job by just incrementing the reference count. In this case, it should be
considered as success but it used to return 'fail' to let the caller free
temporary bd_holder. Fixed it to return success and free given object by
itself.
Also, if either one of symlinking fails, the bd_holder should not be added to
the list so that it can be discarded later. Otherwise, the caller will free
bd_holder which is in the list.
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-29 10:58:56 +02:00
|
|
|
}
|
2007-07-16 08:41:25 +02:00
|
|
|
|
|
|
|
list_add_tail(&bo->list, &bdev->bd_holder_list);
|
|
|
|
return 0;
|
2006-03-27 11:17:57 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* del_bd_holder - delete sysfs symlinks for bd_claim() relationship
|
|
|
|
*
|
|
|
|
* @bdev: block device to be bd_claimed
|
|
|
|
* @kobj: holder's kobject
|
|
|
|
*
|
|
|
|
* If there is matching entry with @kobj in @bdev->bd_holder_list
|
|
|
|
* and no other bd_claim() from the same kobject,
|
|
|
|
* remove the struct bd_holder from the list, delete symlinks for it.
|
|
|
|
*
|
|
|
|
* Returns a pointer to the struct bd_holder when it's removed from the list
|
|
|
|
* and ready to be freed.
|
|
|
|
* Returns NULL if matching claim isn't found or there is other bd_claim()
|
|
|
|
* by the same kobject.
|
|
|
|
*/
|
|
|
|
static struct bd_holder *del_bd_holder(struct block_device *bdev,
|
|
|
|
struct kobject *kobj)
|
|
|
|
{
|
|
|
|
struct bd_holder *bo;
|
|
|
|
|
|
|
|
list_for_each_entry(bo, &bdev->bd_holder_list, list) {
|
|
|
|
if (bo->sdir == kobj) {
|
|
|
|
bo->count--;
|
|
|
|
BUG_ON(bo->count < 0);
|
|
|
|
if (!bo->count) {
|
|
|
|
list_del(&bo->list);
|
|
|
|
del_symlink(bo->sdir, bo->sdev);
|
|
|
|
del_symlink(bo->hdir, bo->hdev);
|
|
|
|
bd_holder_release_dirs(bo);
|
|
|
|
return bo;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bd_claim_by_kobject - bd_claim() with additional kobject signature
|
|
|
|
*
|
|
|
|
* @bdev: block device to be claimed
|
|
|
|
* @holder: holder's signature
|
|
|
|
* @kobj: holder's kobject
|
|
|
|
*
|
|
|
|
* Do bd_claim() and if it succeeds, create sysfs symlinks between
|
|
|
|
* the bdev and the holder's kobject.
|
|
|
|
* Use bd_release_from_kobject() when relesing the claimed bdev.
|
|
|
|
*
|
|
|
|
* Returns 0 on success. (same as bd_claim())
|
|
|
|
* Returns errno on failure.
|
|
|
|
*/
|
|
|
|
static int bd_claim_by_kobject(struct block_device *bdev, void *holder,
|
|
|
|
struct kobject *kobj)
|
|
|
|
{
|
2007-07-16 08:41:25 +02:00
|
|
|
int err;
|
2006-10-30 22:23:56 +01:00
|
|
|
struct bd_holder *bo, *found;
|
2006-03-27 11:17:57 +02:00
|
|
|
|
|
|
|
if (!kobj)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
bo = alloc_bd_holder(kobj);
|
|
|
|
if (!bo)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2006-12-08 11:36:13 +01:00
|
|
|
mutex_lock(&bdev->bd_mutex);
|
2006-10-30 22:23:56 +01:00
|
|
|
|
2007-07-16 08:41:25 +02:00
|
|
|
err = bd_claim(bdev, holder);
|
|
|
|
if (err)
|
2007-07-16 08:41:28 +02:00
|
|
|
goto fail;
|
2007-07-16 08:41:25 +02:00
|
|
|
|
|
|
|
found = find_bd_holder(bdev, bo);
|
|
|
|
if (found)
|
2007-07-16 08:41:28 +02:00
|
|
|
goto fail;
|
2007-07-16 08:41:25 +02:00
|
|
|
|
|
|
|
err = add_bd_holder(bdev, bo);
|
|
|
|
if (err)
|
|
|
|
bd_release(bdev);
|
2007-07-16 08:41:28 +02:00
|
|
|
else
|
|
|
|
bo = NULL;
|
|
|
|
fail:
|
2006-03-27 11:18:00 +02:00
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
2007-07-16 08:41:28 +02:00
|
|
|
free_bd_holder(bo);
|
2007-07-16 08:41:25 +02:00
|
|
|
return err;
|
2006-03-27 11:17:57 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bd_release_from_kobject - bd_release() with additional kobject signature
|
|
|
|
*
|
|
|
|
* @bdev: block device to be released
|
|
|
|
* @kobj: holder's kobject
|
|
|
|
*
|
|
|
|
* Do bd_release() and remove sysfs symlinks created by bd_claim_by_kobject().
|
|
|
|
*/
|
|
|
|
static void bd_release_from_kobject(struct block_device *bdev,
|
|
|
|
struct kobject *kobj)
|
|
|
|
{
|
|
|
|
if (!kobj)
|
|
|
|
return;
|
|
|
|
|
2006-12-08 11:36:13 +01:00
|
|
|
mutex_lock(&bdev->bd_mutex);
|
2006-03-27 11:17:57 +02:00
|
|
|
bd_release(bdev);
|
2007-07-16 08:41:28 +02:00
|
|
|
free_bd_holder(del_bd_holder(bdev, kobj));
|
2006-03-27 11:18:00 +02:00
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
2006-03-27 11:17:57 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bd_claim_by_disk - wrapper function for bd_claim_by_kobject()
|
|
|
|
*
|
|
|
|
* @bdev: block device to be claimed
|
|
|
|
* @holder: holder's signature
|
|
|
|
* @disk: holder's gendisk
|
|
|
|
*
|
|
|
|
* Call bd_claim_by_kobject() with getting @disk->slave_dir.
|
|
|
|
*/
|
|
|
|
int bd_claim_by_disk(struct block_device *bdev, void *holder,
|
|
|
|
struct gendisk *disk)
|
|
|
|
{
|
|
|
|
return bd_claim_by_kobject(bdev, holder, kobject_get(disk->slave_dir));
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(bd_claim_by_disk);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* bd_release_from_disk - wrapper function for bd_release_from_kobject()
|
|
|
|
*
|
|
|
|
* @bdev: block device to be claimed
|
|
|
|
* @disk: holder's gendisk
|
|
|
|
*
|
|
|
|
* Call bd_release_from_kobject() and put @disk->slave_dir.
|
|
|
|
*/
|
|
|
|
void bd_release_from_disk(struct block_device *bdev, struct gendisk *disk)
|
|
|
|
{
|
|
|
|
bd_release_from_kobject(bdev, disk->slave_dir);
|
|
|
|
kobject_put(disk->slave_dir);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(bd_release_from_disk);
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* Tries to open block device by device number. Use it ONLY if you
|
|
|
|
* really do not have anything better - i.e. when you are behind a
|
|
|
|
* truly sucky interface and all you are given is a device number. _Never_
|
|
|
|
* to be used for internal purposes. If you ever need it - reconsider
|
|
|
|
* your API.
|
|
|
|
*/
|
2008-09-02 21:28:45 +02:00
|
|
|
struct block_device *open_by_devnum(dev_t dev, fmode_t mode)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct block_device *bdev = bdget(dev);
|
|
|
|
int err = -ENOMEM;
|
|
|
|
if (bdev)
|
2007-10-08 19:24:05 +02:00
|
|
|
err = blkdev_get(bdev, mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
return err ? ERR_PTR(err) : bdev;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(open_by_devnum);
|
|
|
|
|
2008-09-04 22:27:40 +02:00
|
|
|
/**
|
|
|
|
* flush_disk - invalidates all buffer-cache entries on a disk
|
|
|
|
*
|
|
|
|
* @bdev: struct block device to be flushed
|
|
|
|
*
|
|
|
|
* Invalidates all buffer-cache entries on a disk. It should be called
|
|
|
|
* when a disk has been changed -- either by a media change or online
|
|
|
|
* resize.
|
|
|
|
*/
|
|
|
|
static void flush_disk(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
if (__invalidate_device(bdev)) {
|
|
|
|
char name[BDEVNAME_SIZE] = "";
|
|
|
|
|
|
|
|
if (bdev->bd_disk)
|
|
|
|
disk_name(bdev->bd_disk, 0, name);
|
|
|
|
printk(KERN_WARNING "VFS: busy inodes on changed media or "
|
|
|
|
"resized disk %s\n", name);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!bdev->bd_disk)
|
|
|
|
return;
|
|
|
|
if (disk_partitionable(bdev->bd_disk))
|
|
|
|
bdev->bd_invalidated = 1;
|
|
|
|
}
|
|
|
|
|
2008-09-04 22:27:25 +02:00
|
|
|
/**
|
2008-10-09 10:42:38 +02:00
|
|
|
* check_disk_size_change - checks for disk size change and adjusts bdev size.
|
2008-09-04 22:27:25 +02:00
|
|
|
* @disk: struct gendisk to check
|
|
|
|
* @bdev: struct bdev to adjust.
|
|
|
|
*
|
|
|
|
* This routine checks to see if the bdev size does not match the disk size
|
|
|
|
* and adjusts it if it differs.
|
|
|
|
*/
|
|
|
|
void check_disk_size_change(struct gendisk *disk, struct block_device *bdev)
|
|
|
|
{
|
|
|
|
loff_t disk_size, bdev_size;
|
|
|
|
|
|
|
|
disk_size = (loff_t)get_capacity(disk) << 9;
|
|
|
|
bdev_size = i_size_read(bdev->bd_inode);
|
|
|
|
if (disk_size != bdev_size) {
|
|
|
|
char name[BDEVNAME_SIZE];
|
|
|
|
|
|
|
|
disk_name(disk, 0, name);
|
|
|
|
printk(KERN_INFO
|
|
|
|
"%s: detected capacity change from %lld to %lld\n",
|
|
|
|
name, bdev_size, disk_size);
|
|
|
|
i_size_write(bdev->bd_inode, disk_size);
|
2008-09-04 22:27:45 +02:00
|
|
|
flush_disk(bdev);
|
2008-09-04 22:27:25 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(check_disk_size_change);
|
|
|
|
|
2008-09-04 22:27:20 +02:00
|
|
|
/**
|
2008-10-09 10:42:38 +02:00
|
|
|
* revalidate_disk - wrapper for lower-level driver's revalidate_disk call-back
|
2008-09-04 22:27:20 +02:00
|
|
|
* @disk: struct gendisk to be revalidated
|
|
|
|
*
|
|
|
|
* This routine is a wrapper for lower-level driver's revalidate_disk
|
|
|
|
* call-backs. It is used to do common pre and post operations needed
|
|
|
|
* for all revalidate_disk operations.
|
|
|
|
*/
|
|
|
|
int revalidate_disk(struct gendisk *disk)
|
|
|
|
{
|
2008-09-04 22:27:25 +02:00
|
|
|
struct block_device *bdev;
|
2008-09-04 22:27:20 +02:00
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (disk->fops->revalidate_disk)
|
|
|
|
ret = disk->fops->revalidate_disk(disk);
|
|
|
|
|
2008-09-04 22:27:25 +02:00
|
|
|
bdev = bdget_disk(disk, 0);
|
|
|
|
if (!bdev)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
mutex_lock(&bdev->bd_mutex);
|
|
|
|
check_disk_size_change(disk, bdev);
|
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
|
|
|
bdput(bdev);
|
2008-09-04 22:27:20 +02:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(revalidate_disk);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* This routine checks whether a removable media has been changed,
|
|
|
|
* and invalidates all buffer-cache-entries in that case. This
|
|
|
|
* is a relatively slow routine, so we have to try to minimize using
|
|
|
|
* it. Thus it is called only upon a 'mount' or 'open'. This
|
|
|
|
* is the best way of combining speed and utility, I think.
|
|
|
|
* People changing diskettes in the middle of an operation deserve
|
|
|
|
* to lose :-)
|
|
|
|
*/
|
|
|
|
int check_disk_change(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
struct gendisk *disk = bdev->bd_disk;
|
|
|
|
struct block_device_operations * bdops = disk->fops;
|
|
|
|
|
|
|
|
if (!bdops->media_changed)
|
|
|
|
return 0;
|
|
|
|
if (!bdops->media_changed(bdev->bd_disk))
|
|
|
|
return 0;
|
|
|
|
|
2008-09-04 22:27:40 +02:00
|
|
|
flush_disk(bdev);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (bdops->revalidate_disk)
|
|
|
|
bdops->revalidate_disk(bdev->bd_disk);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(check_disk_change);
|
|
|
|
|
|
|
|
void bd_set_size(struct block_device *bdev, loff_t size)
|
|
|
|
{
|
2009-05-22 23:17:49 +02:00
|
|
|
unsigned bsize = bdev_logical_block_size(bdev);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
bdev->bd_inode->i_size = size;
|
|
|
|
while (bsize < PAGE_CACHE_SIZE) {
|
|
|
|
if (size & bsize)
|
|
|
|
break;
|
|
|
|
bsize <<= 1;
|
|
|
|
}
|
|
|
|
bdev->bd_block_size = bsize;
|
|
|
|
bdev->bd_inode->i_blkbits = blksize_bits(bsize);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(bd_set_size);
|
|
|
|
|
2008-02-23 02:40:24 +01:00
|
|
|
static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part);
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
|
2007-02-20 22:58:18 +01:00
|
|
|
/*
|
|
|
|
* bd_mutex locking:
|
|
|
|
*
|
|
|
|
* mutex_lock(part->bd_mutex)
|
|
|
|
* mutex_lock_nested(whole->bd_mutex, 1)
|
|
|
|
*/
|
|
|
|
|
2007-10-08 19:24:05 +02:00
|
|
|
static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct gendisk *disk;
|
2008-06-06 07:46:27 +02:00
|
|
|
int ret;
|
2008-09-03 09:01:09 +02:00
|
|
|
int partno;
|
2008-06-23 14:30:55 +02:00
|
|
|
int perm = 0;
|
|
|
|
|
2007-10-08 19:24:05 +02:00
|
|
|
if (mode & FMODE_READ)
|
2008-06-23 14:30:55 +02:00
|
|
|
perm |= MAY_READ;
|
2007-10-08 19:24:05 +02:00
|
|
|
if (mode & FMODE_WRITE)
|
2008-06-23 14:30:55 +02:00
|
|
|
perm |= MAY_WRITE;
|
|
|
|
/*
|
|
|
|
* hooks: /n/, see "layering violations".
|
|
|
|
*/
|
|
|
|
ret = devcgroup_inode_permission(bdev->bd_inode, perm);
|
2008-08-01 11:32:04 +02:00
|
|
|
if (ret != 0) {
|
|
|
|
bdput(bdev);
|
2008-06-06 07:46:27 +02:00
|
|
|
return ret;
|
2008-08-01 11:32:04 +02:00
|
|
|
}
|
2008-06-06 07:46:27 +02:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
lock_kernel();
|
md: make devices disappear when they are no longer needed.
Currently md devices, once created, never disappear until the module
is unloaded. This is essentially because the gendisk holds a
reference to the mddev, and the mddev holds a reference to the
gendisk, this a circular reference.
If we drop the reference from mddev to gendisk, then we need to ensure
that the mddev is destroyed when the gendisk is destroyed. However it
is not possible to hook into the gendisk destruction process to enable
this.
So we drop the reference from the gendisk to the mddev and destroy the
gendisk when the mddev gets destroyed. However this has a
complication.
Between the call
__blkdev_get->get_gendisk->kobj_lookup->md_probe
and the call
__blkdev_get->md_open
there is no obvious way to hold a reference on the mddev any more, so
unless something is done, it will disappear and gendisk will be
destroyed prematurely.
Also, once we decide to destroy the mddev, there will be an unlockable
moment before the gendisk is unlinked (blk_unregister_region) during
which a new reference to the gendisk can be created. We need to
ensure that this reference can not be used. i.e. the ->open must
fail.
So:
1/ in md_probe we set a flag in the mddev (hold_active) which
indicates that the array should be treated as active, even
though there are no references, and no appearance of activity.
This is cleared by md_release when the device is closed if it
is no longer needed.
This ensures that the gendisk will survive between md_probe and
md_open.
2/ In md_open we check if the mddev we expect to open matches
the gendisk that we did open.
If there is a mismatch we return -ERESTARTSYS and modify
__blkdev_get to retry from the top in that case.
In the -ERESTARTSYS sys case we make sure to wait until
the old gendisk (that we succeeded in opening) is really gone so
we loop at most once.
Some udev configurations will always open an md device when it first
appears. If we allow an md device that was just created by an open
to disappear on an immediate close, then this can race with such udev
configurations and result in an infinite loop the device being opened
and closed, then re-open due to the 'ADD' even from the first open,
and then close and so on.
So we make sure an md device, once created by an open, remains active
at least until some md 'ioctl' has been made on it. This means that
all normal usage of md devices will allow them to disappear promptly
when not needed, but the worst that an incorrect usage will do it
cause an inactive md device to be left in existence (it can easily be
removed).
As an array can be stopped by writing to a sysfs attribute
echo clear > /sys/block/mdXXX/md/array_state
we need to use scheduled work for deleting the gendisk and other
kobjects. This allows us to wait for any pending gendisk deletion to
complete by simply calling flush_scheduled_work().
Signed-off-by: NeilBrown <neilb@suse.de>
2009-01-08 22:31:10 +01:00
|
|
|
restart:
|
2008-08-25 12:56:12 +02:00
|
|
|
|
2008-11-05 10:21:06 +01:00
|
|
|
ret = -ENXIO;
|
2008-09-03 09:01:09 +02:00
|
|
|
disk = get_gendisk(bdev->bd_dev, &partno);
|
2008-08-25 12:56:12 +02:00
|
|
|
if (!disk)
|
|
|
|
goto out_unlock_kernel;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2006-12-08 11:36:16 +01:00
|
|
|
mutex_lock_nested(&bdev->bd_mutex, for_part);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (!bdev->bd_openers) {
|
|
|
|
bdev->bd_disk = disk;
|
|
|
|
bdev->bd_contains = bdev;
|
2008-09-03 09:01:09 +02:00
|
|
|
if (!partno) {
|
2005-04-17 00:20:36 +02:00
|
|
|
struct backing_dev_info *bdi;
|
2008-11-05 10:21:06 +01:00
|
|
|
|
|
|
|
ret = -ENXIO;
|
|
|
|
bdev->bd_part = disk_get_part(disk, partno);
|
|
|
|
if (!bdev->bd_part)
|
|
|
|
goto out_clear;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
if (disk->fops->open) {
|
2007-10-08 19:24:05 +02:00
|
|
|
ret = disk->fops->open(bdev, mode);
|
md: make devices disappear when they are no longer needed.
Currently md devices, once created, never disappear until the module
is unloaded. This is essentially because the gendisk holds a
reference to the mddev, and the mddev holds a reference to the
gendisk, this a circular reference.
If we drop the reference from mddev to gendisk, then we need to ensure
that the mddev is destroyed when the gendisk is destroyed. However it
is not possible to hook into the gendisk destruction process to enable
this.
So we drop the reference from the gendisk to the mddev and destroy the
gendisk when the mddev gets destroyed. However this has a
complication.
Between the call
__blkdev_get->get_gendisk->kobj_lookup->md_probe
and the call
__blkdev_get->md_open
there is no obvious way to hold a reference on the mddev any more, so
unless something is done, it will disappear and gendisk will be
destroyed prematurely.
Also, once we decide to destroy the mddev, there will be an unlockable
moment before the gendisk is unlinked (blk_unregister_region) during
which a new reference to the gendisk can be created. We need to
ensure that this reference can not be used. i.e. the ->open must
fail.
So:
1/ in md_probe we set a flag in the mddev (hold_active) which
indicates that the array should be treated as active, even
though there are no references, and no appearance of activity.
This is cleared by md_release when the device is closed if it
is no longer needed.
This ensures that the gendisk will survive between md_probe and
md_open.
2/ In md_open we check if the mddev we expect to open matches
the gendisk that we did open.
If there is a mismatch we return -ERESTARTSYS and modify
__blkdev_get to retry from the top in that case.
In the -ERESTARTSYS sys case we make sure to wait until
the old gendisk (that we succeeded in opening) is really gone so
we loop at most once.
Some udev configurations will always open an md device when it first
appears. If we allow an md device that was just created by an open
to disappear on an immediate close, then this can race with such udev
configurations and result in an infinite loop the device being opened
and closed, then re-open due to the 'ADD' even from the first open,
and then close and so on.
So we make sure an md device, once created by an open, remains active
at least until some md 'ioctl' has been made on it. This means that
all normal usage of md devices will allow them to disappear promptly
when not needed, but the worst that an incorrect usage will do it
cause an inactive md device to be left in existence (it can easily be
removed).
As an array can be stopped by writing to a sysfs attribute
echo clear > /sys/block/mdXXX/md/array_state
we need to use scheduled work for deleting the gendisk and other
kobjects. This allows us to wait for any pending gendisk deletion to
complete by simply calling flush_scheduled_work().
Signed-off-by: NeilBrown <neilb@suse.de>
2009-01-08 22:31:10 +01:00
|
|
|
if (ret == -ERESTARTSYS) {
|
|
|
|
/* Lost a race with 'disk' being
|
|
|
|
* deleted, try again.
|
|
|
|
* See md.c
|
|
|
|
*/
|
|
|
|
disk_put_part(bdev->bd_part);
|
|
|
|
bdev->bd_part = NULL;
|
|
|
|
module_put(disk->fops->owner);
|
|
|
|
put_disk(disk);
|
|
|
|
bdev->bd_disk = NULL;
|
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
|
|
|
goto restart;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
if (ret)
|
2008-08-25 12:56:12 +02:00
|
|
|
goto out_clear;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
if (!bdev->bd_openers) {
|
|
|
|
bd_set_size(bdev,(loff_t)get_capacity(disk)<<9);
|
|
|
|
bdi = blk_get_backing_dev_info(bdev);
|
|
|
|
if (bdi == NULL)
|
|
|
|
bdi = &default_backing_dev_info;
|
|
|
|
bdev->bd_inode->i_data.backing_dev_info = bdi;
|
|
|
|
}
|
|
|
|
if (bdev->bd_invalidated)
|
|
|
|
rescan_partitions(disk, bdev);
|
|
|
|
} else {
|
|
|
|
struct block_device *whole;
|
|
|
|
whole = bdget_disk(disk, 0);
|
|
|
|
ret = -ENOMEM;
|
|
|
|
if (!whole)
|
2008-08-25 12:56:12 +02:00
|
|
|
goto out_clear;
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
BUG_ON(for_part);
|
2007-10-08 19:24:05 +02:00
|
|
|
ret = __blkdev_get(whole, mode, 1);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (ret)
|
2008-08-25 12:56:12 +02:00
|
|
|
goto out_clear;
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev->bd_contains = whole;
|
|
|
|
bdev->bd_inode->i_data.backing_dev_info =
|
|
|
|
whole->bd_inode->i_data.backing_dev_info;
|
2008-11-05 10:21:06 +01:00
|
|
|
bdev->bd_part = disk_get_part(disk, partno);
|
2008-09-03 09:03:02 +02:00
|
|
|
if (!(disk->flags & GENHD_FL_UP) ||
|
2008-11-05 10:21:06 +01:00
|
|
|
!bdev->bd_part || !bdev->bd_part->nr_sects) {
|
2005-04-17 00:20:36 +02:00
|
|
|
ret = -ENXIO;
|
2008-08-25 12:56:12 +02:00
|
|
|
goto out_clear;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
2008-11-05 10:21:06 +01:00
|
|
|
bd_set_size(bdev, (loff_t)bdev->bd_part->nr_sects << 9);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
put_disk(disk);
|
2008-08-25 12:56:12 +02:00
|
|
|
module_put(disk->fops->owner);
|
|
|
|
disk = NULL;
|
2005-04-17 00:20:36 +02:00
|
|
|
if (bdev->bd_contains == bdev) {
|
|
|
|
if (bdev->bd_disk->fops->open) {
|
2007-10-08 19:24:05 +02:00
|
|
|
ret = bdev->bd_disk->fops->open(bdev, mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (ret)
|
2008-08-25 12:56:12 +02:00
|
|
|
goto out_unlock_bdev;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
if (bdev->bd_invalidated)
|
|
|
|
rescan_partitions(bdev->bd_disk, bdev);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
bdev->bd_openers++;
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
if (for_part)
|
|
|
|
bdev->bd_part_count++;
|
2006-03-23 12:00:28 +01:00
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
2005-04-17 00:20:36 +02:00
|
|
|
unlock_kernel();
|
|
|
|
return 0;
|
|
|
|
|
2008-08-25 12:56:12 +02:00
|
|
|
out_clear:
|
2008-11-05 10:21:06 +01:00
|
|
|
disk_put_part(bdev->bd_part);
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev->bd_disk = NULL;
|
2008-08-25 12:56:12 +02:00
|
|
|
bdev->bd_part = NULL;
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev->bd_inode->i_data.backing_dev_info = &default_backing_dev_info;
|
|
|
|
if (bdev != bdev->bd_contains)
|
2007-10-08 19:24:05 +02:00
|
|
|
__blkdev_put(bdev->bd_contains, mode, 1);
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev->bd_contains = NULL;
|
2008-08-25 12:56:12 +02:00
|
|
|
out_unlock_bdev:
|
2006-03-23 12:00:28 +01:00
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
2008-08-25 12:56:12 +02:00
|
|
|
out_unlock_kernel:
|
2005-04-17 00:20:36 +02:00
|
|
|
unlock_kernel();
|
2008-08-25 12:56:12 +02:00
|
|
|
|
|
|
|
if (disk)
|
|
|
|
module_put(disk->fops->owner);
|
|
|
|
put_disk(disk);
|
|
|
|
bdput(bdev);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2007-10-08 19:24:05 +02:00
|
|
|
int blkdev_get(struct block_device *bdev, fmode_t mode)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-10-08 19:24:05 +02:00
|
|
|
return __blkdev_get(bdev, mode, 0);
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
EXPORT_SYMBOL(blkdev_get);
|
|
|
|
|
|
|
|
static int blkdev_open(struct inode * inode, struct file * filp)
|
|
|
|
{
|
|
|
|
struct block_device *bdev;
|
|
|
|
int res;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Preserve backwards compatibility and allow large file access
|
|
|
|
* even if userspace doesn't ask for it explicitly. Some mkfs
|
|
|
|
* binary needs it. We might want to drop this workaround
|
|
|
|
* during an unstable branch.
|
|
|
|
*/
|
|
|
|
filp->f_flags |= O_LARGEFILE;
|
|
|
|
|
2007-10-08 19:24:05 +02:00
|
|
|
if (filp->f_flags & O_NDELAY)
|
|
|
|
filp->f_mode |= FMODE_NDELAY;
|
|
|
|
if (filp->f_flags & O_EXCL)
|
|
|
|
filp->f_mode |= FMODE_EXCL;
|
|
|
|
if ((filp->f_flags & O_ACCMODE) == 3)
|
|
|
|
filp->f_mode |= FMODE_WRITE_IOCTL;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
bdev = bd_acquire(inode);
|
2006-10-28 19:38:33 +02:00
|
|
|
if (bdev == NULL)
|
|
|
|
return -ENOMEM;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-10-08 19:24:05 +02:00
|
|
|
filp->f_mapping = bdev->bd_inode->i_mapping;
|
|
|
|
|
|
|
|
res = blkdev_get(bdev, filp->f_mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (res)
|
|
|
|
return res;
|
|
|
|
|
2008-11-05 14:54:41 +01:00
|
|
|
if (filp->f_mode & FMODE_EXCL) {
|
|
|
|
res = bd_claim(bdev, filp);
|
|
|
|
if (res)
|
|
|
|
goto out_blkdev_put;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-11-05 14:54:41 +01:00
|
|
|
return 0;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-11-05 14:54:41 +01:00
|
|
|
out_blkdev_put:
|
2008-02-23 02:40:24 +01:00
|
|
|
blkdev_put(bdev, filp->f_mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
return res;
|
|
|
|
}
|
|
|
|
|
2008-02-23 02:40:24 +01:00
|
|
|
static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part)
|
2006-12-08 11:36:13 +01:00
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
struct gendisk *disk = bdev->bd_disk;
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
struct block_device *victim = NULL;
|
2006-12-08 11:36:13 +01:00
|
|
|
|
2006-12-08 11:36:16 +01:00
|
|
|
mutex_lock_nested(&bdev->bd_mutex, for_part);
|
2006-12-08 11:36:13 +01:00
|
|
|
lock_kernel();
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
if (for_part)
|
|
|
|
bdev->bd_part_count--;
|
|
|
|
|
2006-12-08 11:36:13 +01:00
|
|
|
if (!--bdev->bd_openers) {
|
|
|
|
sync_blockdev(bdev);
|
|
|
|
kill_bdev(bdev);
|
|
|
|
}
|
|
|
|
if (bdev->bd_contains == bdev) {
|
|
|
|
if (disk->fops->release)
|
2008-02-23 02:40:24 +01:00
|
|
|
ret = disk->fops->release(disk, mode);
|
2006-12-08 11:36:13 +01:00
|
|
|
}
|
|
|
|
if (!bdev->bd_openers) {
|
|
|
|
struct module *owner = disk->fops->owner;
|
|
|
|
|
|
|
|
put_disk(disk);
|
|
|
|
module_put(owner);
|
2008-08-25 12:56:12 +02:00
|
|
|
disk_put_part(bdev->bd_part);
|
|
|
|
bdev->bd_part = NULL;
|
2006-12-08 11:36:13 +01:00
|
|
|
bdev->bd_disk = NULL;
|
|
|
|
bdev->bd_inode->i_data.backing_dev_info = &default_backing_dev_info;
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
if (bdev != bdev->bd_contains)
|
|
|
|
victim = bdev->bd_contains;
|
2006-12-08 11:36:13 +01:00
|
|
|
bdev->bd_contains = NULL;
|
|
|
|
}
|
|
|
|
unlock_kernel();
|
|
|
|
mutex_unlock(&bdev->bd_mutex);
|
|
|
|
bdput(bdev);
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
if (victim)
|
2008-02-23 02:40:24 +01:00
|
|
|
__blkdev_put(victim, mode, 1);
|
2006-12-08 11:36:13 +01:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-02-23 02:40:24 +01:00
|
|
|
int blkdev_put(struct block_device *bdev, fmode_t mode)
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
{
|
2008-02-23 02:40:24 +01:00
|
|
|
return __blkdev_put(bdev, mode, 0);
|
[PATCH] lockdep: simplify some aspects of bd_mutex nesting
When we open (actually blkdev_get) a partition we need to also open (get) the
whole device that holds the partition. The involves some limited recursion.
This patch tries to simplify some aspects of this.
As well as opening the whole device, we need to increment ->bd_part_count when
a partition is opened (this is used by rescan_partitions to avoid a rescan if
any partition is active, as that would be confusing).
The main change this patch makes is to move the inc/dec of bd_part_count into
blkdev_{get,put} for the whole rather than doing it in blkdev_{get,put} for
the partition.
More specifically, we introduce __blkdev_get and __blkdev_put which do exactly
what blkdev_{get,put} did, only with an extra "for_part" argument
(blkget_{get,put} then call the __ version with a '0' for the extra argument).
If for_part is 1, then the blkdev is being get(put) because a partition is
being opened(closed) for the first(last) time, and so bd_part_count should be
updated (on success). The particular advantage of pushing this function down
is that the bd_mutex lock (which is needed to update bd_part_count) is already
held at the lower level.
Note that this slightly changes the semantics of bd_part_count. Instead of
updating it whenever a partition is opened or released, it is now only updated
on the first open or last release. This is an adequate semantic as it is only
ever tested for "== 0".
Having introduced these functions we remove the current bd_part_count updates
from do_open (which is really the body of blkdev_get) and call
__blkdev_get(... 1). Similarly in blkget_put we remove the old bd_part_count
updates and call __blkget_put(..., 1). This call is moved to the end of
__blkdev_put to avoid nested locks of bd_mutex.
Finally the mutex_lock on whole->bd_mutex in do_open can be removed. It was
only really needed to protect bd_part_count, and that is now managed (and
protected) within the recursive call.
The observation that bd_part_count is central to the locking issues, and the
modifications to create __blkdev_put are from Peter Zijlstra.
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 11:36:16 +01:00
|
|
|
}
|
2006-12-08 11:36:13 +01:00
|
|
|
EXPORT_SYMBOL(blkdev_put);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
static int blkdev_close(struct inode * inode, struct file * filp)
|
|
|
|
{
|
|
|
|
struct block_device *bdev = I_BDEV(filp->f_mapping->host);
|
|
|
|
if (bdev->bd_holder == filp)
|
|
|
|
bd_release(bdev);
|
2008-02-23 02:40:24 +01:00
|
|
|
return blkdev_put(bdev, filp->f_mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2005-06-23 09:10:15 +02:00
|
|
|
static long block_ioctl(struct file *file, unsigned cmd, unsigned long arg)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2008-09-19 09:17:36 +02:00
|
|
|
struct block_device *bdev = I_BDEV(file->f_mapping->host);
|
|
|
|
fmode_t mode = file->f_mode;
|
2008-11-05 14:58:42 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
* O_NDELAY can be altered using fcntl(.., F_SETFL, ..), so we have
|
|
|
|
* to updated it before every ioctl.
|
|
|
|
*/
|
2008-09-19 09:17:36 +02:00
|
|
|
if (file->f_flags & O_NDELAY)
|
2008-11-05 14:58:42 +01:00
|
|
|
mode |= FMODE_NDELAY;
|
|
|
|
else
|
|
|
|
mode &= ~FMODE_NDELAY;
|
|
|
|
|
2008-09-19 09:17:36 +02:00
|
|
|
return blkdev_ioctl(bdev, mode, cmd, arg);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2009-01-03 15:47:09 +01:00
|
|
|
/*
|
|
|
|
* Try to release a page associated with block device when the system
|
|
|
|
* is under memory pressure.
|
|
|
|
*/
|
|
|
|
static int blkdev_releasepage(struct page *page, gfp_t wait)
|
|
|
|
{
|
|
|
|
struct super_block *super = BDEV_I(page->mapping->host)->bdev.bd_super;
|
|
|
|
|
|
|
|
if (super && super->s_op->bdev_try_to_free_page)
|
|
|
|
return super->s_op->bdev_try_to_free_page(super, page, wait);
|
|
|
|
|
|
|
|
return try_to_free_buffers(page);
|
|
|
|
}
|
|
|
|
|
2008-02-18 13:48:31 +01:00
|
|
|
static const struct address_space_operations def_blk_aops = {
|
2005-04-17 00:20:36 +02:00
|
|
|
.readpage = blkdev_readpage,
|
|
|
|
.writepage = blkdev_writepage,
|
|
|
|
.sync_page = block_sync_page,
|
2007-10-16 10:25:04 +02:00
|
|
|
.write_begin = blkdev_write_begin,
|
|
|
|
.write_end = blkdev_write_end,
|
2005-04-17 00:20:36 +02:00
|
|
|
.writepages = generic_writepages,
|
2009-01-03 15:47:09 +01:00
|
|
|
.releasepage = blkdev_releasepage,
|
2005-04-17 00:20:36 +02:00
|
|
|
.direct_IO = blkdev_direct_IO,
|
|
|
|
};
|
|
|
|
|
2006-03-28 11:56:42 +02:00
|
|
|
const struct file_operations def_blk_fops = {
|
2005-04-17 00:20:36 +02:00
|
|
|
.open = blkdev_open,
|
|
|
|
.release = blkdev_close,
|
|
|
|
.llseek = block_llseek,
|
2006-10-01 08:28:48 +02:00
|
|
|
.read = do_sync_read,
|
|
|
|
.write = do_sync_write,
|
2005-04-17 00:20:36 +02:00
|
|
|
.aio_read = generic_file_aio_read,
|
2006-10-01 08:28:46 +02:00
|
|
|
.aio_write = generic_file_aio_write_nolock,
|
2005-04-17 00:20:36 +02:00
|
|
|
.mmap = generic_file_mmap,
|
|
|
|
.fsync = block_fsync,
|
2005-06-23 09:10:15 +02:00
|
|
|
.unlocked_ioctl = block_ioctl,
|
2005-04-17 00:20:36 +02:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
.compat_ioctl = compat_blkdev_ioctl,
|
|
|
|
#endif
|
2006-05-01 19:59:32 +02:00
|
|
|
.splice_read = generic_file_splice_read,
|
|
|
|
.splice_write = generic_file_splice_write,
|
2005-04-17 00:20:36 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
int ioctl_by_bdev(struct block_device *bdev, unsigned cmd, unsigned long arg)
|
|
|
|
{
|
|
|
|
int res;
|
|
|
|
mm_segment_t old_fs = get_fs();
|
|
|
|
set_fs(KERNEL_DS);
|
2008-09-19 09:17:36 +02:00
|
|
|
res = blkdev_ioctl(bdev, 0, cmd, arg);
|
2005-04-17 00:20:36 +02:00
|
|
|
set_fs(old_fs);
|
|
|
|
return res;
|
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(ioctl_by_bdev);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* lookup_bdev - lookup a struct block_device by name
|
2009-01-06 23:41:15 +01:00
|
|
|
* @pathname: special file representing the block device
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
2008-10-09 10:42:38 +02:00
|
|
|
* Get a reference to the blockdevice at @pathname in the current
|
2005-04-17 00:20:36 +02:00
|
|
|
* namespace if possible and return it. Return ERR_PTR(error)
|
|
|
|
* otherwise.
|
|
|
|
*/
|
2008-08-02 07:04:36 +02:00
|
|
|
struct block_device *lookup_bdev(const char *pathname)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct block_device *bdev;
|
|
|
|
struct inode *inode;
|
2008-08-02 07:04:36 +02:00
|
|
|
struct path path;
|
2005-04-17 00:20:36 +02:00
|
|
|
int error;
|
|
|
|
|
2008-08-02 07:04:36 +02:00
|
|
|
if (!pathname || !*pathname)
|
2005-04-17 00:20:36 +02:00
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
|
2008-08-02 07:04:36 +02:00
|
|
|
error = kern_path(pathname, LOOKUP_FOLLOW, &path);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (error)
|
|
|
|
return ERR_PTR(error);
|
|
|
|
|
2008-08-02 07:04:36 +02:00
|
|
|
inode = path.dentry->d_inode;
|
2005-04-17 00:20:36 +02:00
|
|
|
error = -ENOTBLK;
|
|
|
|
if (!S_ISBLK(inode->i_mode))
|
|
|
|
goto fail;
|
|
|
|
error = -EACCES;
|
2008-08-02 07:04:36 +02:00
|
|
|
if (path.mnt->mnt_flags & MNT_NODEV)
|
2005-04-17 00:20:36 +02:00
|
|
|
goto fail;
|
|
|
|
error = -ENOMEM;
|
|
|
|
bdev = bd_acquire(inode);
|
|
|
|
if (!bdev)
|
|
|
|
goto fail;
|
|
|
|
out:
|
2008-08-02 07:04:36 +02:00
|
|
|
path_put(&path);
|
2005-04-17 00:20:36 +02:00
|
|
|
return bdev;
|
|
|
|
fail:
|
|
|
|
bdev = ERR_PTR(error);
|
|
|
|
goto out;
|
|
|
|
}
|
2008-08-01 11:00:11 +02:00
|
|
|
EXPORT_SYMBOL(lookup_bdev);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/**
|
2008-02-23 01:50:45 +01:00
|
|
|
* open_bdev_exclusive - open a block device by name and set it up for use
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
* @path: special file representing the block device
|
2008-02-23 01:50:45 +01:00
|
|
|
* @mode: FMODE_... combination to pass be used
|
2005-04-17 00:20:36 +02:00
|
|
|
* @holder: owner for exclusion
|
|
|
|
*
|
|
|
|
* Open the blockdevice described by the special file at @path, claim it
|
|
|
|
* for the @holder.
|
|
|
|
*/
|
2008-02-23 01:50:45 +01:00
|
|
|
struct block_device *open_bdev_exclusive(const char *path, fmode_t mode, void *holder)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct block_device *bdev;
|
|
|
|
int error = 0;
|
|
|
|
|
|
|
|
bdev = lookup_bdev(path);
|
|
|
|
if (IS_ERR(bdev))
|
|
|
|
return bdev;
|
|
|
|
|
2007-10-08 19:24:05 +02:00
|
|
|
error = blkdev_get(bdev, mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (error)
|
|
|
|
return ERR_PTR(error);
|
|
|
|
error = -EACCES;
|
2008-02-23 01:50:45 +01:00
|
|
|
if ((mode & FMODE_WRITE) && bdev_read_only(bdev))
|
2005-04-17 00:20:36 +02:00
|
|
|
goto blkdev_put;
|
|
|
|
error = bd_claim(bdev, holder);
|
|
|
|
if (error)
|
|
|
|
goto blkdev_put;
|
|
|
|
|
|
|
|
return bdev;
|
|
|
|
|
|
|
|
blkdev_put:
|
2008-02-23 02:40:24 +01:00
|
|
|
blkdev_put(bdev, mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
return ERR_PTR(error);
|
|
|
|
}
|
|
|
|
|
2008-02-23 01:50:45 +01:00
|
|
|
EXPORT_SYMBOL(open_bdev_exclusive);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/**
|
2008-02-23 01:50:45 +01:00
|
|
|
* close_bdev_exclusive - close a blockdevice opened by open_bdev_exclusive()
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
* @bdev: blockdevice to close
|
2008-02-23 01:50:45 +01:00
|
|
|
* @mode: mode, must match that used to open.
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
2008-02-23 01:50:45 +01:00
|
|
|
* This is the counterpart to open_bdev_exclusive().
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
2008-02-23 01:50:45 +01:00
|
|
|
void close_bdev_exclusive(struct block_device *bdev, fmode_t mode)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
bd_release(bdev);
|
2008-02-23 01:50:45 +01:00
|
|
|
blkdev_put(bdev, mode);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2008-02-23 01:50:45 +01:00
|
|
|
EXPORT_SYMBOL(close_bdev_exclusive);
|
2006-08-29 20:06:11 +02:00
|
|
|
|
|
|
|
int __invalidate_device(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
struct super_block *sb = get_super(bdev);
|
|
|
|
int res = 0;
|
|
|
|
|
|
|
|
if (sb) {
|
|
|
|
/*
|
|
|
|
* no need to lock the super, get_super holds the
|
|
|
|
* read mutex so the filesystem cannot go away
|
|
|
|
* under us (->put_super runs with the write lock
|
|
|
|
* hold).
|
|
|
|
*/
|
|
|
|
shrink_dcache_sb(sb);
|
|
|
|
res = invalidate_inodes(sb);
|
|
|
|
drop_super(sb);
|
|
|
|
}
|
2007-05-06 23:49:54 +02:00
|
|
|
invalidate_bdev(bdev);
|
2006-08-29 20:06:11 +02:00
|
|
|
return res;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(__invalidate_device);
|