PATCH: Forcible delaying of UFS (soft)updates

[Alfred Perlstein]

* Michael Sierchio <kudzu at tenebras.com> [030412 11:38] wrote:
> Marko Zec wrote:
> 
> >- fsync() no longer flushes the buffers to disk, but returns immediately
> >instead;
> 
> Any system that does this should be flushed down the toilet.  Softupdates
> already breaks transaction semantics of FFS by breaking 
> link()/unlink()/rename()
> etc.

How does it do that?