hw/arm: xenpvh: Reverse virtio-mmio creation order

We've been creating the virtio-mmio devices in forwards order
but since the qbus lists prepend (rather than append) entries,
the virtio busses end up with decreasing base address order.

Xen enables virtio-mmio nodes in forwards order so there's been
a missmatch. So far, we've been working around this with an
out-of-tree patch to Xen.

This reverses the order making sure the virtio busses end up
ordered with increasing base addresses avoiding the need to
patch Xen.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Acked-by: Stefano Stabellini <sstabellini@kernel.org>

1 files changed, 12 insertions, 1 deletions

diff --git a/hw/xen/xen-pvh-common.c b/hw/xen/xen-pvh-common.c
index 880e8143d7..295f920442 100644
--- a/hw/xen/xen-pvh-common.c
+++ b/hw/xen/xen-pvh-common.c
@@ -75,7 +75,18 @@ static void xen_create_virtio_mmio_devices(XenPVHMachineState *s)
 {
     int i;
 
-    for (i = 0; i < s->cfg.virtio_mmio_num; i++) {
+    /*
+     * We create the transports in reverse order. Since qbus_realize()
+     * prepends (not appends) new child buses, the decrementing loop below will
+     * create a list of virtio-mmio buses with increasing base addresses.
+     *
+     * When a -device option is processed from the command line,
+     * qbus_find_recursive() picks the next free virtio-mmio bus in forwards
+     * order.
+     *
+     * This is what the Xen tools expect.
+     */
+    for (i = s->cfg.virtio_mmio_num - 1; i >= 0; i--) {
         hwaddr base = s->cfg.virtio_mmio.base + i * s->cfg.virtio_mmio.size;
         qemu_irq irq = qemu_allocate_irq(xen_set_irq, NULL,
                                          s->cfg.virtio_mmio_irq_base + i);