Re: working of balance_classzone()

[Hugh Dickins <hugh@veritas.com>]

On Sun, 19 May 2002, Abhishek Nayani wrote:
> 
> 	So the code in balance_classzone() looks very suspicious as it
> is acting as if there were many blocks of free pages of different orders
> on the list and we are trying to get the block of the correct order and
> then freeing the rest in reverse order.... 
> 
> 	Since there is only one block, we can cut the code to just check
> the order of that block, if its greater than our requirement, call
> rmqueue() else return NULL. 

The code hereabouts has been marooned for many releases, somewhere
in between what Andrea intended and what Linus wanted.  You're right
that there's a (harmless) mismatch between what __free_pages_ok is
actually doing, and what balance_classzone is expecting it to do.
But that should get sorted out when Andrea's further VM patches are
merged into 2.4.20... (always seems to be around the next corner).

Hugh

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/