G'day!

On Tue, 12 Oct 1999, Stephen C. Tweedie wrote:
...
> Andrea, you are just trying to relax carefully designed buffer cache
> semantics which are relied upon by the current filesystems.  Saying it
> is a trick doesn't help matters much.

Andrea's right in that the semantics make it far too easy to introduce
bugs.  How about making bforget set a flag that the buffer head is to be
destroyed as soon as it's released?  If a filesystem re-uses the buffer,
mark_buffer_dirty can clear this destroyed flag (to handle the case where
a buffer head becomes legitimately reused).

                -ben

Reply via email to