Btrfs: fix free space tree bitmaps on big-endian systems
In convert_free_space_to_{bitmaps,extents}(), we buffer the free space
bitmaps in memory and copy them directly to/from the extent buffers with
{read,write}_extent_buffer(). The extent buffer bitmap helpers use byte
granularity, which is equivalent to a little-endian bitmap. This means
that on big-endian systems, the in-memory bitmaps will be written to
disk byte-swapped. To fix this, use byte-granularity for the bitmaps in
memory.
Fixes: a5ed91828518 ("Btrfs: implement the free space B-tree")
Cc: stable@vger.kernel.org # 4.5+
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Tested-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index 28cd88f..1cf4e42 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -59,6 +59,28 @@
*/
#define EXTENT_PAGE_PRIVATE 1
+/*
+ * The extent buffer bitmap operations are done with byte granularity instead of
+ * word granularity for two reasons:
+ * 1. The bitmaps must be little-endian on disk.
+ * 2. Bitmap items are not guaranteed to be aligned to a word and therefore a
+ * single word in a bitmap may straddle two pages in the extent buffer.
+ */
+#define BIT_BYTE(nr) ((nr) / BITS_PER_BYTE)
+#define BYTE_MASK ((1 << BITS_PER_BYTE) - 1)
+#define BITMAP_FIRST_BYTE_MASK(start) \
+ ((BYTE_MASK << ((start) & (BITS_PER_BYTE - 1))) & BYTE_MASK)
+#define BITMAP_LAST_BYTE_MASK(nbits) \
+ (BYTE_MASK >> (-(nbits) & (BITS_PER_BYTE - 1)))
+
+static inline int le_test_bit(int nr, const u8 *addr)
+{
+ return 1U & (addr[BIT_BYTE(nr)] >> (nr & (BITS_PER_BYTE-1)));
+}
+
+extern void le_bitmap_set(u8 *map, unsigned int start, int len);
+extern void le_bitmap_clear(u8 *map, unsigned int start, int len);
+
struct extent_state;
struct btrfs_root;
struct btrfs_io_bio;