When user unbinds a BTT disk and binds again with a different
sector size without wiping out the disk, a BTT disk is created
with a wrong size.
This is because the bind operation keeps the previous metadata,
which leads nd_btt->lbasize inconsistent with internal_lbasize
and external_lbasize in the arena. A reboot also reattaches
the BTT from the previous metadata.
Change nd_btt_arena_is_valid() to check if nd_btt->uuid matches
with super->uuid when a new UUID is set for binding. This
assures the bind operation writes the metadata with the values
specified by user.
Cc: Vishal Verma <vishal.l.ve...@intel.com>
Cc: Dan Williams <dan.j.willi...@intel.com>
Reported-by: Micah Parrish <micah.parr...@hpe.com>
Signed-off-by: Toshi Kani <toshi.k...@hpe.com>
---
drivers/nvdimm/btt_devs.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/nvdimm/btt_devs.c b/drivers/nvdimm/btt_devs.c
index cb47751..176ea25 100644
--- a/drivers/nvdimm/btt_devs.c
+++ b/drivers/nvdimm/btt_devs.c
@@ -218,6 +218,8 @@ static bool uuid_is_null(u8 *uuid)
* Check consistency of the btt info block with itself by validating
* the checksum, and with the parent namespace by verifying the
* parent_uuid contained in the info block with the one supplied in.
+ * When nd_btt->uuid is set for binding, verify if the metadata is
+ * stale.
*
* Returns:
* false for an invalid info block, true for a valid one
@@ -234,6 +236,10 @@ bool nd_btt_arena_is_valid(struct nd_btt *nd_btt, struct
btt_sb *super)
if (memcmp(super->parent_uuid, parent_uuid, 16) != 0)
return false;
+ if (nd_btt->uuid)
+ if (memcmp(super->uuid, nd_btt->uuid, 16) != 0)
+ return false;
+
checksum = le64_to_cpu(super->checksum);
super->checksum = 0;
if (checksum != nd_sb_checksum((struct nd_gen_sb *) super))
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
