1:20 p.m.
On Wed, 2017-06-28 at 13:14 -0700, Dan Williams wrote:
On Wed, Jun 28, 2017 at 1:07 PM, Vishal Verma
<vishal.l.verma(a)intel.co
m> wrote:
> The UEFI 2.7 specification defines an updated BTT metadata format,
> bumping the revision to 2.0. Add support for the new format, while
> retaining compatibility for the old 1.1 format.
>
> Cc: Toshi Kani <toshi.kani(a)hpe.com>
> Cc: Linda Knippers <linda.knippers(a)hpe.com>
> Cc: Dan Williams <dan.j.williams(a)intel.com>
> Signed-off-by: Vishal Verma <vishal.l.verma(a)intel.com>
[..]
> diff --git a/drivers/nvdimm/namespace_devs.c
> b/drivers/nvdimm/namespace_devs.c
> index f05d9b0..09e7e76 100644
> --- a/drivers/nvdimm/namespace_devs.c
> +++ b/drivers/nvdimm/namespace_devs.c
> @@ -1411,6 +1411,58 @@ static ssize_t dpa_extents_show(struct device
> *dev,
> }
> static DEVICE_ATTR_RO(dpa_extents);
>
> +static int btt_claim_class(struct device *dev)
> +{
> + struct nd_region *nd_region = to_nd_region(dev->parent);
> + int i, loop_bitmask = 0;
> +
> + for (i = 0; i < nd_region->ndr_mappings; i++) {
> + struct nd_mapping *nd_mapping = &nd_region-
> >mapping[i];
> + struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
> + struct nd_namespace_index *nsindex;
> +
> + nsindex = to_namespace_index(ndd, ndd->ns_current);
> + if (nsindex == NULL)
> + loop_bitmask |= 1;
> + else {
> + /* check whether existing labels are v1.1 or
> v1.2 */
> + if (__le16_to_cpu(nsindex->major) == 1
> + && __le16_to_cpu(nsindex-
> >minor) == 1)
> + loop_bitmask |= 2;
> + else
> + loop_bitmask |= 4;
> + }
> + }
> + /*
> + * If nsindex is null loop_bitmask's bit 0 will be set, and
> if an index
> + * block is found, a v1.1 label for any mapping will set bit
> 1, and a
> + * v1.2 label will set bit 2.
> + *
> + * At the end of the loop, only one of the three bits must
> be set.
> + * If multiple bits were set, it means the different
> mappings disagree
> + * about their labels, and this must be cleaned up first.
> + *
> + * If all the label index blocks are found to agree, nsindex
> of NULL
> + * implies labels haven't been initialized yet, and when
> they will,
> + * they will be of the 1.2 format, so we can assume BTT2.0
> + *
> + * If 1.1 labels are found, we enforce BTT1.1, and if 1.2
> labels are
> + * found, we enforce BTT2.0
> + *
> + * If the loop was never entered, default to BTT1.1 (legacy
> namespaces)
> + */
> + switch (loop_bitmask) {
> + case 0:
> + case 2:
> + return NVDIMM_CCLASS_BTT;
> + case 1:
> + case 4:
> + return NVDIMM_CCLASS_BTT2;
> + default:
> + return -ENXIO;
> + }
Clever, looks good to me. I'll get this queued.
Err Hold on, I just noticed I forgot to properly propagate the ENXIO in
__holder_class_store.
v5 incoming.