virtio: force vring descriptors to be allocated from lowmem

Virtio devices may attempt to add descriptors to a virtqueue from atomic
context using GFP_ATOMIC allocation. This is problematic because such
allocations can fall outside of the lowmem mapping, causing virt_to_phys
to report bogus physical addresses which are subsequently passed to
userspace via the buffers for the virtual device.

This patch masks out __GFP_HIGH and __GFP_HIGHMEM from the requested
flags when allocating descriptors for a virtqueue. If an atomic
allocation is requested and later fails, we will return -ENOSPC which
will be handled by the driver.

Cc: [email protected]
Cc: Sasha Levin <[email protected]>
Signed-off-by: Will Deacon <[email protected]>
Signed-off-by: Rusty Russell <[email protected]>

diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index e639584b2dbd1c24e36017db1978009b73f8e33f..286c30cb393d245dda6068d8df07efc0aa3699d8 100644 (file)
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -135,6 +135,13 @@ static int vring_add_indirect(struct vring_virtqueue *vq,
        unsigned head;
        int i;
 
+       /*
+        * We require lowmem mappings for the descriptors because
+        * otherwise virt_to_phys will give us bogus addresses in the
+        * virtqueue.
+        */
+       gfp &= ~(__GFP_HIGHMEM | __GFP_HIGH);
+
        desc = kmalloc((out + in) * sizeof(struct vring_desc), gfp);
        if (!desc)
                return -ENOMEM;