RE: [PATCH 1/2] virtio_pmem: initialize provider_data through nd_region_desc

[Dan Williams <dan.j.williams@intel.com>]

Jason Wang wrote:
> We used to initialize the provider_data manually after
> nvdimm_pemm_region_create(). This seems to be racy if the flush is

It would be nice to include the actual backtrace / bug signature that
this fixes if it is available.

> issued before the initialization of provider_data. Fixing this by
> initialize the provider_data through nd_region_desc to make sure the
> provider_data is ready after the pmem is created.
> 
> Fixes 6e84200c0a29 ("virtio-pmem: Add virtio pmem driver")
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
>  drivers/nvdimm/virtio_pmem.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/nvdimm/virtio_pmem.c b/drivers/nvdimm/virtio_pmem.c
> index 995b6cdc67ed..48f8327d0431 100644
> --- a/drivers/nvdimm/virtio_pmem.c
> +++ b/drivers/nvdimm/virtio_pmem.c
> @@ -81,6 +81,7 @@ static int virtio_pmem_probe(struct virtio_device *vdev)
>  	ndr_desc.res = &res;
>  	ndr_desc.numa_node = nid;
>  	ndr_desc.flush = async_pmem_flush;
> +	ndr_desc.provider_data = vdev;

For my untrained eye, why not
"dev_to_virtio(nd_region->dev.parent->parent)"? If that is indeed
equivalent "vdev" then you can do a follow-on cleanup patch to reduce
that syntax. Otherwise, if by chance they are not equivalent, then this
conversion is introducing a new problem.

Outside of that you can add:

Reviewed-by: Dan Williams <dan.j.williams@intel.com>