Remove potentially_dirty counter from bottom volume
This counter is not accurate (missing required memory barrier to avoid unwanted behavior due to processor optimizations) and performance gain is not clear - generally global atomic variables are something we would like to avoid going forward. Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
This commit is contained in:
@@ -14,18 +14,8 @@ struct casdsk_disk;
|
||||
|
||||
struct bd_object {
|
||||
struct casdsk_disk *dsk;
|
||||
|
||||
struct block_device *btm_bd;
|
||||
/**
|
||||
* This denotes state of volatile write cache of the device.
|
||||
* This is set to true when:
|
||||
* - opening the device
|
||||
* - when writing to a device without FUA/FLUSH flags
|
||||
* This is set to false when:
|
||||
* - FLUSH request is completed on device.
|
||||
* When it is false
|
||||
* - FLUSH requests from upper layer are NOT passed to the device.
|
||||
*/
|
||||
atomic_t potentially_dirty;
|
||||
|
||||
uint32_t expobj_valid : 1;
|
||||
/*!< Bit indicates that exported object was created */
|
||||
|
||||
Reference in New Issue
Block a user