Although Xen's documentation states that the address_bits field is not used

by XENMEM_decrease_reservation, it is checked by the hypervisor. In certain
circumstances (stack leak), the field could have an improper value, leading
to a fail of the hypercall.

Set it to 0 ("no addressing restriction") to avoid that.

Patch tested by Sam Fourman and haad@.

This should fix the rare "failed allocating DMA memory" encountered
under NetBSD dom0. Will ask for a pull-up.
This commit is contained in:
jym 2010-03-09 23:12:06 +00:00
parent af1ec92289
commit d0a75a07db

View File

@ -1,4 +1,4 @@
/* $NetBSD: xen_bus_dma.c,v 1.19 2010/03/02 00:13:50 jym Exp $ */
/* $NetBSD: xen_bus_dma.c,v 1.20 2010/03/09 23:12:06 jym Exp $ */
/* NetBSD bus_dma.c,v 1.21 2005/04/16 07:53:35 yamt Exp */
/*-
@ -32,7 +32,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: xen_bus_dma.c,v 1.19 2010/03/02 00:13:50 jym Exp $");
__KERNEL_RCSID(0, "$NetBSD: xen_bus_dma.c,v 1.20 2010/03/09 23:12:06 jym Exp $");
#include <sys/param.h>
#include <sys/systm.h>
@ -95,6 +95,7 @@ _xen_alloc_contig(bus_size_t size, bus_size_t alignment, bus_size_t boundary,
xenguest_handle(res.extent_start) = &mfn;
res.nr_extents = 1;
res.extent_order = 0;
res.address_bits = 0;
res.domid = DOMID_SELF;
error = HYPERVISOR_memory_op(XENMEM_decrease_reservation, &res);
if (error != 1) {