Patchwork [1/2] ext4: get free blocks count of BLOCK_UNINIT groups from group desc directly

login
register
mail settings
Submitter Yongqiang Yang
Date July 20, 2011, 1:32 p.m.
Message ID <1311168774-26378-2-git-send-email-xiaoqiangnk@gmail.com>
Download mbox | patch
Permalink /patch/105691/
State New
Headers show

Comments

Yongqiang Yang - July 20, 2011, 1:32 p.m.
This patch teaches ext4 to get free blocks count of BLOCK_UNINIT groups from
group desc directly.

Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
---
 fs/ext4/ialloc.c  |    2 --
 fs/ext4/mballoc.c |   14 ++------------
 2 files changed, 2 insertions(+), 14 deletions(-)
Theodore Ts'o - Aug. 1, 2011, 11:20 a.m.
On Wed, Jul 20, 2011 at 09:32:53PM +0800, Yongqiang Yang wrote:
> This patch teaches ext4 to get free blocks count of BLOCK_UNINIT groups from
> group desc directly.
> 
> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>

The reason why we weren't doing this is because older versions of
e2fsprogs I don't think set the free blocks count correctly.  Or at
least, I have a vague memory that this was the case.  I'll have to do
some code archeology before can be 100% sure this is safe to do.

Regards,                      	              	

					- Ted
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andreas Dilger - Aug. 2, 2011, 3:15 a.m.
On 2011-08-01, at 5:20 AM, Ted Ts'o <tytso@mit.edu> wrote:
> On Wed, Jul 20, 2011 at 09:32:53PM +0800, Yongqiang Yang wrote:
>> This patch teaches ext4 to get free blocks count of BLOCK_UNINIT groups from
>> group desc directly.
>> 
>> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
> 
> The reason why we weren't doing this is because older versions of
> e2fsprogs I don't think set the free blocks count correctly.  Or at
> least, I have a vague memory that this was the case.  I'll have to do
> some code archeology before can be 100% sure this is safe to do.

I think the other reason is that this gives the kernel code some flexibility in how the group is formatted (e.g. flex_bg or whatever may move allocated blocks around).

Cheers, Andreas--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 21bb2f6..efdc605 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -954,9 +954,7 @@  got:
 		ext4_lock_group(sb, group);
 		/* recheck and clear flag under lock if we still need to */
 		if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
-			free = ext4_free_blocks_after_init(sb, group, gdp);
 			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT);
-			ext4_free_blks_set(sb, gdp, free);
 			gdp->bg_checksum = ext4_group_desc_csum(sbi, group,
 								gdp);
 		}
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index b97a2d2..f92826b 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -2248,13 +2248,7 @@  int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 	 * initialize bb_free to be able to skip
 	 * empty groups without initialization
 	 */
-	if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
-		meta_group_info[i]->bb_free =
-			ext4_free_blocks_after_init(sb, group, desc);
-	} else {
-		meta_group_info[i]->bb_free =
-			ext4_free_blks_count(sb, desc);
-	}
+	meta_group_info[i]->bb_free = ext4_free_blks_count(sb, desc);
 
 	INIT_LIST_HEAD(&meta_group_info[i]->bb_prealloc_list);
 	init_rwsem(&meta_group_info[i]->alloc_sem);
@@ -2802,12 +2796,8 @@  ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 	}
 #endif
 	mb_set_bits(bitmap_bh->b_data, ac->ac_b_ex.fe_start,ac->ac_b_ex.fe_len);
-	if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
+	if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT))
 		gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT);
-		ext4_free_blks_set(sb, gdp,
-					ext4_free_blocks_after_init(sb,
-					ac->ac_b_ex.fe_group, gdp));
-	}
 	len = ext4_free_blks_count(sb, gdp) - ac->ac_b_ex.fe_len;
 	ext4_free_blks_set(sb, gdp, len);
 	gdp->bg_checksum = ext4_group_desc_csum(sbi, ac->ac_b_ex.fe_group, gdp);