On Tue, 5 Dec 2000, Linus Torvalds wrote:
> And this is not just a "it happens to be like this" kind of thing. It
> _has_ to be like this, because every time we call clear_inode() we are
> going to physically free the memory associated with the inode very soon
> afterwards. Which means that _any_ use of the inode had better be long
> gone. Dirty buffers included.
Urgh. Linus, AFAICS we _all_ agree on that. The only real question is
whether we consider calling clear_inode() with droppable dirty buffers
to be OK. It can't happen on the dispose_list() path and I'ld rather
see it _not_ happening on the delete_inode() one. It's a policy question,
not the correctness one.
IOW, I would prefer to have BUG() instead of invalidate_inode_buffers()
and let the ->delete_inode() make sure that list is empty. I'm not saying
that current code doesn't work. However, "let's clean after the
truncate_inode_page()/foo_delete_inode(), they might leave some junk
on the list" looks like a wrong thing.
Notice that policy wrt pages is already of that kind - clear_inode()
expects the callers to make sure that ->i_data.nr_pages is zero instead of
trying to clean after them. I think that we will be better off with
similar rules wrt dirty buffers list.
Comments?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Dec 07 2000 - 21:00:13 EST