On Tue, 15 May 2012 16:00:25 +0800 Yuanhan Liu <yuanhan.liu@xxxxxxxxxxxxxxx>
wrote:
> The code does nothing but repating assigning sb->sb_csum with original
> value.
This code sets the sb_csum to 0 while the csum is calculated.
Without this, calculating the csum, setting it, then calculating again would
produce a different result.
So the code is correct as is.
Thanks,
NeilBrown
>
> Signed-off-by: Yuanhan Liu <yuanhan.liu@xxxxxxxxxxxxxxx>
> ---
> drivers/md/md.c | 4 ----
> 1 files changed, 0 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index 0bd44e3..f2d799a 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -1439,15 +1439,12 @@ super_90_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
>
> static __le32 calc_sb_1_csum(struct mdp_superblock_1 * sb)
> {
> - __le32 disk_csum;
> u32 csum;
> unsigned long long newcsum;
> int size = 256 + le32_to_cpu(sb->max_dev)*2;
> __le32 *isuper = (__le32*)sb;
> int i;
>
> - disk_csum = sb->sb_csum;
> - sb->sb_csum = 0;
> newcsum = 0;
> for (i=0; size>=4; size -= 4 )
> newcsum += le32_to_cpu(*isuper++);
> @@ -1456,7 +1453,6 @@ static __le32 calc_sb_1_csum(struct mdp_superblock_1 * sb)
> newcsum += le16_to_cpu(*(__le16*) isuper);
>
> csum = (newcsum & 0xffffffff) + (newcsum >> 32);
> - sb->sb_csum = disk_csum;
> return cpu_to_le32(csum);
> }
>
Attachment:
signature.asc
Description: PGP signature
