Re: [PATCH ver3] exofs: New truncate sequence

[Nick Piggin <npiggin@suse.de>]

On Tue, Jun 01, 2010 at 02:31:30PM +0300, Boaz Harrosh wrote:
> 
> These changes are crafted based on the similar
> conversion done to ext2 by Nick Piggin.
> 
> * Remove the deprecated ->truncate vector. Let exofs_setattr
>   take care of on-disk size updates.
> * Call truncate_pagecache on the unused pages if
>   write_begin/end fails.
> * Cleanup exofs_delete_inode that did stupid inode
>   writes and updates on an inode that will be
>   removed.
> * And finally get rid of exofs_get_block. We never
>   had any blocks it was all for calling nobh_truncate_page.
>   nobh_truncate_page is not actually needed in exofs since
>   the last page is complete and gone, just like all the other
>   pages. There is no partial blocks in exofs.
>   [OK do I need a partial read here upto i_size ???]

You shouldn't need to. We do have some interesting code trying to deal
with mmap data past i_size in core/buffer code. I hope Jan's patches
can improve that.

But anyway if you are not doing byte based IO, you have to remember
that if you mark a page uptodate, it should not contain uninitialized
data past i_size. Also, in the writeback path, junk data can get
written past i_size too.

The latter case is difficult to avoid (due to concurrent mmap writes).
It is not technically a leak perhaps, but it's a bit nasty if you allow
data to be hidden past i_size. The best course of action IMO is to
zero this out at read-in and truncate-extend time.

Doesn't seem like you have to worry about any of that if you have
byte granularity , although you do have EXOFS_BLKSHIFT defined?