libnvdimm, pmem: fix up max_hw_sectors
authorDan Williams <dan.j.williams@intel.com>
Sat, 16 May 2015 16:28:50 +0000 (12:28 -0400)
committerDan Williams <dan.j.williams@intel.com>
Fri, 26 Jun 2015 15:23:38 +0000 (11:23 -0400)
There is no hardware limit to enforce on the size of the i/o that can be passed
to an nvdimm block device, so set it to UINT_MAX.

Reviewed-by: Vishal Verma <vishal.l.verma@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
drivers/nvdimm/pmem.c

index 7346054bccbb6228fd362c432524b5797290e443..d29a42adb95a8fcfb776be449b2634bdf36dad10 100644 (file)
@@ -167,7 +167,7 @@ static int pmem_attach_disk(struct nd_namespace_common *ndns,
                return -ENOMEM;
 
        blk_queue_make_request(pmem->pmem_queue, pmem_make_request);
-       blk_queue_max_hw_sectors(pmem->pmem_queue, 1024);
+       blk_queue_max_hw_sectors(pmem->pmem_queue, UINT_MAX);
        blk_queue_bounce_limit(pmem->pmem_queue, BLK_BOUNCE_ANY);
 
        disk = alloc_disk(0);