Re: [PATCH] ext3 [linux-2.6.2.]: accessing already freed inodes when under memory pressure

[Andrew Morton <akpm@osdl.org>]

Carsten Otte <cotte@freenet.de> wrote:
>
> Am Samstag 21 Februar 2004 01:43 schrieb Andrew Morton:
> > Could you chage it so we generate a backtrace in iput() if the inode has an
> > i_count of one (the dentry) and it is still on the orphan list?
>
> I did change iput() to the following:
> void iput(struct inode *inode)
> {
>         if (inode) {
>                 struct super_operations *op = inode->i_sb->s_op;
> 
>                 if (inode->i_state == I_CLEAR)
>                         BUG();
> 
>                 if (op && op->put_inode)
>                         op->put_inode(inode);
> 
>                 if (atomic_dec_and_lock(&inode->i_count, &inode_lock))
>                         iput_final(inode);
>                 if (atomic_read(&inode->i_count) == 1)
>                         if (!strcmp(inode->i_sb->s_type->name, "ext3")) {
>                         struct ext3_inode_info *ext3_i;
>                                ext3_i = container_of(inode,
>                                                struct ext3_inode_info,
>                                                vfs_inode);
>                                if (!list_empty(&ext3_i->i_orphan))
>                                        WARN_ON(1);
>                         }
>         }
> }

atomic_dec_and_lock() decrements i_count, so this warning will be triggered
by people calling iput() with an orphan inode which has a refcount of 2,
not of 1.   Please change the

	if (atomic_read(&inode->i_count) == 1)

to

	if (atomic_read(&inode->i_count) == 0)

Thanks.