block: remove BLK_DEV_DAX config option
authorRoss Zwisler <[email protected]>
Wed, 3 Aug 2016 20:46:15 +0000 (13:46 -0700)
committerLinus Torvalds <[email protected]>
Thu, 4 Aug 2016 12:50:07 +0000 (08:50 -0400)
The functionality for block device DAX was already removed with commit
acc93d30d7d4 ("Revert "block: enable dax for raw block devices"")

However, we still had a config option hanging around that was always
disabled because it depended on CONFIG_BROKEN.  This config option was
introduced in commit 03cdadb04077 ("block: disable block device DAX by
default")

This change reverts that commit, removing the dead config option.

Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ross Zwisler <[email protected]>
Cc: Dave Hansen <[email protected]>
Acked-by: Dan Williams <[email protected]>
Cc: Jens Axboe <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
block/Kconfig
fs/block_dev.c

index 0363cd731320d8d9d93cdcd8e3fb5351682e6464..161491d0a879ed11ce2276a1923a60bc1d92479e 100644 (file)
@@ -88,19 +88,6 @@ config BLK_DEV_INTEGRITY
        T10/SCSI Data Integrity Field or the T13/ATA External Path
        Protection.  If in doubt, say N.
 
-config BLK_DEV_DAX
-       bool "Block device DAX support"
-       depends on FS_DAX
-       depends on BROKEN
-       help
-         When DAX support is available (CONFIG_FS_DAX) raw block
-         devices can also support direct userspace access to the
-         storage capacity via MMAP(2) similar to a file on a
-         DAX-enabled filesystem.  However, the DAX I/O-path disables
-         some standard I/O-statistics, and the MMAP(2) path has some
-         operational differences due to bypassing the page
-         cache.  If in doubt, say N.
-
 config BLK_DEV_THROTTLING
        bool "Block layer bio throttling support"
        depends on BLK_CGROUP=y
index ada42cf42d0667a5d41be90daa203766854e3705..2033a3f91d58a13b6b38334850f2edd3520fd996 100644 (file)
@@ -1275,11 +1275,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
                bdev->bd_disk = disk;
                bdev->bd_queue = disk->queue;
                bdev->bd_contains = bdev;
-               if (IS_ENABLED(CONFIG_BLK_DEV_DAX) &&
-                   blk_queue_dax(disk->queue))
-                       bdev->bd_inode->i_flags = S_DAX;
-               else
-                       bdev->bd_inode->i_flags = 0;
+               bdev->bd_inode->i_flags = 0;
 
                if (!partno) {
                        ret = -ENXIO;