2008-03-02 07:10:45

by Peter Teoh

[permalink] [raw]
Subject: Unmatched semaphore usage in fs/super.c

Please pardon me if wrong. In fs/super.c, I saw a down_read()
without a matching (when sb->s_root is non-zero) up_read() as shown
below:

/**
* get_super - get the superblock of a device
* @bdev: device to get the superblock for
*
* Scans the superblock list and finds the superblock of the file system
* mounted on the device given. %NULL is returned if no match is found.
*/

struct super_block * get_super(struct block_device *bdev)
{
struct super_block *sb;

if (!bdev)
return NULL;

spin_lock(&sb_lock);
rescan:
list_for_each_entry(sb, &super_blocks, s_list) {
if (sb->s_bdev == bdev) {
sb->s_count++;
spin_unlock(&sb_lock);
down_read(&sb->s_umount);
if (sb->s_root)
return sb;
up_read(&sb->s_umount);
/* restart only when sb is no longer on the list */
spin_lock(&sb_lock);
if (__put_super_and_need_restart(sb))
goto rescan;
}
}
spin_unlock(&sb_lock);
return NULL;
}

Neither did the get_super() caller remember and correct the semaphore
either. Is this a problem?


2008-03-02 07:58:30

by Al Viro

[permalink] [raw]
Subject: Re: Unmatched semaphore usage in fs/super.c

On Sun, Mar 02, 2008 at 03:10:35PM +0800, Peter Teoh wrote:
> Please pardon me if wrong. In fs/super.c, I saw a down_read()
> without a matching (when sb->s_root is non-zero) up_read() as shown
> below:

Correct. get_super() and user_get_super() return locked superblock.

> Neither did the get_super() caller remember and correct the semaphore
> either. Is this a problem?

It would be, if it would be true. Take a look at drop_super().