2008-06-09 20:42:57

by Eric Sandeen

Subject: __set_bit for BH flags in ext4_ext_get_blocks?

I was wondering why ext4_ext_get_blocks calls things like:

__set_bit(BH_Unwritten, &bh_result->b_state);
__set_bit(BH_New, &bh_result->b_state);
__set_bit(BH_Mapped, &bh_result->b_state);

instead of set_buffer_unwritten, etc.

Is there any reason it's calling __set_bit instead of set_bit (via the
BUFFER_FNS macros)?

Thanks,
-Eric

2008-06-09 20:59:43

by Eric Sandeen

Subject: [PATCH] use atomic functions to set bh_state

use the BUFFER_FNS functions (set_buffer_foo) to set buffer
head state atomically, not the nonatomic __set_bit.

Alex, please speak up if you had a good reason for __set_bit() :)

Signed-off-by: Eric Sandeen <[email protected]>
---


Index: linux-2.6/fs/ext4/extents.c
===================================================================
--- linux-2.6.orig/fs/ext4/extents.c 2008-06-05 13:44:20.000000000 -0500
+++ linux-2.6/fs/ext4/extents.c 2008-06-09 15:53:11.060045685 -0500
@@ -2616,8 +2616,7 @@ int ext4_ext_get_blocks(handle_t *handle
*/
if (allocated > max_blocks)
allocated = max_blocks;
- /* mark the buffer unwritten */
- __set_bit(BH_Unwritten, &bh_result->b_state);
+ set_buffer_unwritten(bh);
goto out2;
}

@@ -2723,7 +2722,7 @@ int ext4_ext_get_blocks(handle_t *handle
newblock = ext_pblock(&newex);
allocated = ext4_ext_get_actual_len(&newex);
outnew:
- __set_bit(BH_New, &bh_result->b_state);
+ set_buffer_new(bh_result);

/* Cache only when it is _not_ an uninitialized extent */
if (create != EXT4_CREATE_UNINITIALIZED_EXT)
@@ -2733,7 +2732,7 @@ out:
if (allocated > max_blocks)
allocated = max_blocks;
ext4_ext_show_leaf(inode, path);
- __set_bit(BH_Mapped, &bh_result->b_state);
+ set_buffer_mapped(bh_result);
bh_result->b_bdev = inode->i_sb->s_bdev;
bh_result->b_blocknr = newblock;
out2: