Re: [PATCH v8 4/9] dax: support dirty DAX entries in radix tree

[Ross Zwisler]

On Fri, Jan 15, 2016 at 02:22:49PM +0100, Jan Kara wrote:
> On Wed 13-01-16 11:48:32, Ross Zwisler wrote:
> > On Wed, Jan 13, 2016 at 10:44:11AM +0100, Jan Kara wrote:
> > > On Thu 07-01-16 22:27:54, Ross Zwisler wrote:
> > > > Add support for tracking dirty DAX entries in the struct address_space
> > > > radix tree.  This tree is already used for dirty page writeback, and it
> > > > already supports the use of exceptional (non struct page*) entries.
> > > > 
> > > > In order to properly track dirty DAX pages we will insert new exceptional
> > > > entries into the radix tree that represent dirty DAX PTE or PMD pages.
> > > > These exceptional entries will also contain the writeback sectors for the
> > > > PTE or PMD faults that we can use at fsync/msync time.
> > > > 
> > > > There are currently two types of exceptional entries (shmem and shadow)
> > > > that can be placed into the radix tree, and this adds a third.  We rely on
> > > > the fact that only one type of exceptional entry can be found in a given
> > > > radix tree based on its usage.  This happens for free with DAX vs shmem but
> > > > we explicitly prevent shadow entries from being added to radix trees for
> > > > DAX mappings.
> > > > 
> > > > The only shadow entries that would be generated for DAX radix trees would
> > > > be to track zero page mappings that were created for holes.  These pages
> > > > would receive minimal benefit from having shadow entries, and the choice
> > > > to have only one type of exceptional entry in a given radix tree makes the
> > > > logic simpler both in clear_exceptional_entry() and in the rest of DAX.
> > > > 
> > > > Signed-off-by: Ross Zwisler <ross.zwisler@xxxxxxxxxxxxxxx>
> > > > Reviewed-by: Jan Kara <jack@xxxxxxx>
> > > 
> > > I have realized there's one issue with this code. See below:
> > > 
> > > > @@ -34,31 +35,39 @@ static void clear_exceptional_entry(struct address_space *mapping,
> > > >  		return;
> > > >  
> > > >  	spin_lock_irq(&mapping->tree_lock);
> > > > -	/*
> > > > -	 * Regular page slots are stabilized by the page lock even
> > > > -	 * without the tree itself locked.  These unlocked entries
> > > > -	 * need verification under the tree lock.
> > > > -	 */
> > > > -	if (!__radix_tree_lookup(&mapping->page_tree, index, &node, &slot))
> > > > -		goto unlock;
> > > > -	if (*slot != entry)
> > > > -		goto unlock;
> > > > -	radix_tree_replace_slot(slot, NULL);
> > > > -	mapping->nrshadows--;
> > > > -	if (!node)
> > > > -		goto unlock;
> > > > -	workingset_node_shadows_dec(node);
> > > > -	/*
> > > > -	 * Don't track node without shadow entries.
> > > > -	 *
> > > > -	 * Avoid acquiring the list_lru lock if already untracked.
> > > > -	 * The list_empty() test is safe as node->private_list is
> > > > -	 * protected by mapping->tree_lock.
> > > > -	 */
> > > > -	if (!workingset_node_shadows(node) &&
> > > > -	    !list_empty(&node->private_list))
> > > > -		list_lru_del(&workingset_shadow_nodes, &node->private_list);
> > > > -	__radix_tree_delete_node(&mapping->page_tree, node);
> > > > +
> > > > +	if (dax_mapping(mapping)) {
> > > > +		if (radix_tree_delete_item(&mapping->page_tree, index, entry))
> > > > +			mapping->nrexceptional--;
> > > 
> > > So when you punch hole in a file, you can delete a PMD entry from a radix
> > > tree which covers part of the file which still stays. So in this case you
> > > have to split the PMD entry into PTE entries (probably that needs to happen
> > > up in truncate_inode_pages_range()) or something similar...
> > 
> > I think (and will verify) that the DAX code just unmaps the entire PMD range
> > when we receive a hole punch request inside of the PMD.  If this is true then
> > I think the radix tree code should behave the same way and just remove the PMD
> > entry in the radix tree.
> 
> But you cannot just remove it if it is dirty... You have to keep somewhere
> information that part of the PMD range is still dirty (or write that range
> out before removing the radix tree entry).

Yep, agreed.