Re: Review of Thomas's >2GB ext2fs proposal

[Ognyan Kulev]

Thomas Bushnell BSG wrote:
> Ognyan Kulev <ogi@fmi.uni-sofia.bg> writes:
> > Just a small techical correction: we remove mapping only when we are
> > remapping page to contain another block.  So when page is evicted, it
> > continues to map the same disk block until ext2fs decides that we want
> > this page to hold another block (only if page is still evicted, of
> > course).
>
> What do you do if the page is evicted while it is still in use?

Eviction in itself does nothing except to raise the flag that page is 
evicted.  When ext2fs searches for a page that can be remapped[1], only 
those that are marked as evicted _and_ with reference count of zero are 
considered for remapping.  Locking and other stuff take care of the 
tricky remapping.

[1] "remapping page" means that this page held one disk block, and we 
change it to contain another block -- future pager_read_page will read 
this other block.  This change is only in ext2fs data structures and 
it's important that the page is evicted and unreferenced while we do it.

Regards,
ogi