drm/ttm: fix uncached page deallocation to properly fill page pool v3.

Current code never allowed the page pool to actualy fill in anyway.
This fix it, so that we only start freeing page from the pool when
we go over the pool size.

Changed since v1:
  - Move the page batching optimization to its separate patch.

Changed since v2:
  - Do not remove code part of the batching optimization with
    this patch.
  - Better commit message.

Signed-off-by: Jérôme Glisse <[email protected]>
Reviewed-by: Mario Kleiner <[email protected]>
Reviewed-and-Tested-by: Michel Dänzer <[email protected]>
Reviewed-by: Konrad Rzeszutek Wilk <[email protected]>
Cc: Thomas Hellstrom <[email protected]>
Reviewed-by: Alex Deucher <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>

Author: Jérôme Glisse

drivers/gpu/drm/ttm/ttm_page_alloc_dma.c

@@ -963,7 +963,6 @@ void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev)
 	} else {
 		pool->npages_free += count;
 		list_splice(&ttm_dma->pages_list, &pool->free_list);
-		npages = count;
 		if (pool->npages_free > _manager->options.max_size) {
 			npages = pool->npages_free - _manager->options.max_size;
 			/* free at least NUM_PAGES_TO_ALLOC number of pages