On Mon, Mar 19, 2018 at 4:17 AM, Bjorn Andersson
<bjorn.andersson@xxxxxxxxxx> wrote:
> On Wed 10 Jan 05:16 PST 2018, Anup Patel wrote:
>> diff --git a/drivers/rpmsg/virtio_rpmsg_bus.c b/drivers/rpmsg/virtio_rpmsg_bus.c
> [..]
>> @@ -924,9 +925,16 @@ static int rpmsg_probe(struct virtio_device *vdev)
>> total_buf_space, &vrp->bufs_dma,
>> GFP_KERNEL);
>> if (!bufs_va) {
>> - err = -ENOMEM;
>> - goto vqs_del;
>> - }
>> + bufs_va = dma_alloc_coherent(vdev->dev.parent,
>> + total_buf_space, &vrp->bufs_dma,
>> + GFP_KERNEL);
>> + if (!bufs_va) {
>> + err = -ENOMEM;
>> + goto vqs_del;
>> + } else
>> + vrp->bufs_dev = vdev->dev.parent;
>> + } else
>> + vrp->bufs_dev = vdev->dev.parent->parent;
>
> I really don't fancy the idea of us allocating on behalf of our
> grandparent here, as you show it's not certain that our grandparent is
> what someone originally expected it to be.
>
> With the purpose of being able to control these allocations there is an
> ongoing discussion related to this, which I believe will result in this
> being changed to at least vdev->dev.parent..
>
>
> I do expect that this discussion will be brought up during Linaro
> Connect the coming week.
>
Currently, rpmsg_probe() is broken for virtio-mmio transport
hence I send this patch as a stable fix.
In general, I am fine if we are eventually going towards
vdev->dev.parent usage.
Regards,
Anup