On Wed, 15 Jun 2011 17:37:13 +0200
Jan Kara <jack@xxxxxxx> wrote:
> Under heavy memory and filesystem load, users observe the assertion
> mapping->nrpages == 0 in end_writeback() trigger. This can be caused
> by page reclaim reclaiming the last page from a mapping in the following
> race:
> CPU0 CPU1
> ...
> shrink_page_list()
> __remove_mapping()
> __delete_from_page_cache()
> radix_tree_delete()
> evict_inode()
> truncate_inode_pages()
> truncate_inode_pages_range()
> pagevec_lookup() - finds nothing
> end_writeback()
> mapping->nrpages != 0 -> BUG
> page->mapping = NULL
> mapping->nrpages--
>
> Fix the problem by doing a reliable check of mapping->nrpages under
> mapping->tree_lock in end_writeback().
>
> Analyzed by Jay <jinshan.xiong@xxxxxxxxxxxxx>, lost in LKML, and dug
> out by Miklos Szeredi <mszeredi@xxxxxxx>.
>
> CC: Jay <jinshan.xiong@xxxxxxxxxxxxx>
> CC: Miklos Szeredi <mszeredi@xxxxxxx>
> Signed-off-by: Jan Kara <jack@xxxxxxx>
> ---
> fs/inode.c | 7 +++++++
> include/linux/fs.h | 1 +
> mm/truncate.c | 5 +++++
> 3 files changed, 13 insertions(+), 0 deletions(-)
>
> Andrew, does this look better?
spose so.
> diff --git a/fs/inode.c b/fs/inode.c
> index 33c963d..1133cb0 100644
> --- a/fs/inode.c
> +++ b/fs/inode.c
> @@ -467,7 +467,14 @@ EXPORT_SYMBOL(remove_inode_hash);
> void end_writeback(struct inode *inode)
> {
> might_sleep();
> + /*
> + * We have to cycle tree_lock here because reclaim can be still in the
> + * process of removing the last page (in __delete_from_page_cache())
> + * and we must not free mapping under it.
> + */
> + spin_lock(&inode->i_data.tree_lock);
> BUG_ON(inode->i_data.nrpages);
> + spin_unlock(&inode->i_data.tree_lock);
That's an expensive assertion. We might want to wrap all this in
CONFIG_DEBUG_VM.
Or we could do
if (unlikely(inode->i_data.nrpages)) {
/* comment goes here */
spin_lock(&inode->i_data.tree_lock);
BUG_ON(inode->i_data.nrpages);
spin_unlock(&inode->i_data.tree_lock);
}
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>