xen/swiotlb: don't initialize swiotlb twice on arm64

On arm64 swiotlb is often (not always) already initialized by mem_init.
We don't want to initialize it twice, which would trigger a second
memory allocation. Moreover, the second memory pool is typically made of
high pages and ends up replacing the original memory pool of low pages.
As a side effect of this change, it is possible to have low pages in
swiotlb-xen on arm64.

Signed-off-by: Stefano Stabellini <stefanos@xilinx.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

1 files changed, 11 insertions, 1 deletions

diff --git a/drivers/xen/swiotlb-xen.c b/drivers/xen/swiotlb-xen.c
index 64cb94dfedd4..342762b1aefb 100644
--- a/drivers/xen/swiotlb-xen.c
+++ b/drivers/xen/swiotlb-xen.c
@@ -211,6 +211,15 @@ int __ref xen_swiotlb_init(int verbose, bool early)
 retry:
 	bytes = xen_set_nslabs(xen_io_tlb_nslabs);
 	order = get_order(xen_io_tlb_nslabs << IO_TLB_SHIFT);
+
+	/*
+	 * IO TLB memory already allocated. Just use it.
+	 */
+	if (io_tlb_start != 0) {
+		xen_io_tlb_start = phys_to_virt(io_tlb_start);
+		goto end;
+	}
+
 	/*
 	 * Get IO TLB memory from any location.
 	 */
@@ -237,7 +246,6 @@ retry:
 		m_ret = XEN_SWIOTLB_ENOMEM;
 		goto error;
 	}
-	xen_io_tlb_end = xen_io_tlb_start + bytes;
 	/*
 	 * And replace that memory with pages under 4GB.
 	 */
@@ -264,6 +272,8 @@ retry:
 	} else
 		rc = swiotlb_late_init_with_tbl(xen_io_tlb_start, xen_io_tlb_nslabs);
 
+end:
+	xen_io_tlb_end = xen_io_tlb_start + bytes;
 	if (!rc)
 		swiotlb_set_max_segment(PAGE_SIZE);