Re: [patch 5/5] ext2: convert to use the new truncate convention.

[Christoph Hellwig <hch@infradead.org>]

On Sun, Aug 16, 2009 at 08:25:38PM +1000, npiggin@suse.de wrote:
> @@ -66,9 +68,10 @@ void ext2_delete_inode (struct inode * i
>  	mark_inode_dirty(inode);
>  	ext2_write_inode(inode, inode_needs_sync(inode));
>  
> +	/* XXX: where is truncate_inode_pages? */
>  	inode->i_size = 0;
>  	if (inode->i_blocks)
> -		ext2_truncate (inode);
> +		ext2_truncate_blocks(inode, 0);
>  	ext2_free_inode (inode);

At the beginning of the function, just before the diff window.  Because
this is ->delete_inode we truncate away all pages, down to offset 0.

> +static void ext2_truncate_blocks(struct inode *inode, loff_t offset)
> +{
> +	/*
> +	 * XXX: it seems like a bug here that we don't allow
> +	 * IS_APPEND inode to have blocks-past-i_size trimmed off.
> +	 * review and fix this.
> +	 */
> +	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
> +	    S_ISLNK(inode->i_mode)))
> +		return -EINVAL;
> +	if (ext2_inode_is_fast_symlink(inode))
> +		return -EINVAL;
> +	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
> +		return -EPERM;
> +	__ext2_truncate_blocks(inode, offset);

Yes, I think the IS_APPEND(inode) || IS_IMMUTABLE(inode) checks should move
into ext2_setsize.  But let's leave that for a separate patch.

Btw, the above code gives me warnings like this:

/home/hch/work/linux-2.6/fs/ext2/inode.c: In function
'ext2_truncate_blocks':
/home/hch/work/linux-2.6/fs/ext2/inode.c:1158: warning: 'return' with a
value, in function returning void
/home/hch/work/linux-2.6/fs/ext2/inode.c:1160: warning: 'return' with a
value, in function returning void
/home/hch/work/linux-2.6/fs/ext2/inode.c:1162: warning: 'return' with a
value, in function returning void

because you try to return errors from a function delcared as void.