virtio-mem: retry fake-offlining via alloc_contig_range() on ZONE_MOVABLE (f2d799d5) · Commits · 戴 / test

ZONE_MOVABLE is supposed to give some guarantees, yet, alloc_contig_range() isn't prepared to properly deal with some racy cases properly (e.g., temporary page pinning when exiting processed, PCP). Retry 5 times for now. There is certainly room for improvement in the future. Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Jason Wang <jasowang@redhat.com> Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com> Signed-off-by:

David Hildenbrand <david@redhat.com> Link: https://lore.kernel.org/r/20201112133815.13332-11-david@redhat.com Signed-off-by:

Michael S. Tsirkin <mst@redhat.com>

drivers/virtio/virtio_mem.c

+26 −11

Original line number	Diff line number	Diff line
		@@ -784,21 +784,36 @@ static void virtio_mem_fake_online(unsigned long pfn, unsigned long nr_pages)
		*/
		static int virtio_mem_fake_offline(unsigned long pfn, unsigned long nr_pages)
		{
		int rc;
		const bool is_movable = zone_idx(page_zone(pfn_to_page(pfn))) ==
		ZONE_MOVABLE;
		int rc, retry_count;

		/*
		* TODO: We want an alloc_contig_range() mode that tries to allocate
		* harder (e.g., dealing with temporarily pinned pages, PCP), especially
		* with ZONE_MOVABLE. So for now, retry a couple of times with
		* ZONE_MOVABLE before giving up - because that zone is supposed to give
		* some guarantees.
		*/
		for (retry_count = 0; retry_count < 5; retry_count++) {
		rc = alloc_contig_range(pfn, pfn + nr_pages, MIGRATE_MOVABLE,
		GFP_KERNEL);
		if (rc == -ENOMEM)
		/* whoops, out of memory */
		return rc;
		if (rc)
		return -EBUSY;
		else if (rc && !is_movable)
		break;
		else if (rc)
		continue;

		virtio_mem_set_fake_offline(pfn, nr_pages, true);
		adjust_managed_page_count(pfn_to_page(pfn), -nr_pages);
		return 0;
		}

		return -EBUSY;
		}

		/*
		* Handle fake-offline pages when memory is going offline - such that the
		* pages can be skipped by mm-core when offlining.

Admin message