On 12/03/15 03:30, Roger Pau Monné wrote:
>
> @@ -343,20 +353,19 @@ void xen_blkbk_unmap_purged_grants(struct work_struct
> *work)
>
> pages[segs_to_unmap] = persistent_gnt->page;
>
> - if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST) {
> - ret = gnttab_unmap_refs(unmap, NULL, pages,
> - segs_to_unmap);
> - BUG_ON(ret);
> + if (++segs_to_unmap == BLKIF_MAX_SEGMENTS_PER_REQUEST ||
> + list_empty(&blkif->persistent_purge_list)) {
> +
> + unmap_data.count = segs_to_unmap;
> + gnttab_unmap_refs_async(&unmap_data);
> + wait_for_completion(&unmap_completion);
> +
We now have a number of places that use the async unmap and wait for
completion. Could this be moved into common code?
David
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
