On Fri, Aug 10, Jan Blunck wrote:
> This patch removes the special round_down() to next power of 2 implementation
> used only at one place in the snapshot target. It is replaced by an equivalent
> 1 << fls() which might use an architecture specific implementation.
Err, ok this is the second fix for this patch. I hope it is the last. It
doesn't make any sense to have hash_size be defined as sector_t. Same goes for
max_buckets.
Regards,
Jan
--
Jan Blunck <[email protected]>
Subject: dm-snap: Replace special round_down()
This patch removes the special round_down() to next power of 2 implementation
used only at one place in the snapshot target. It is replaced by an equivalent
1 << fls() which might use an architecture specific implementation.
Signed-off-by: Jan Blunck <[email protected]>
---
drivers/md/dm-snap.c | 15 +++------------
1 file changed, 3 insertions(+), 12 deletions(-)
--- a/drivers/md/dm-snap.c
+++ b/drivers/md/dm-snap.c
@@ -333,21 +333,12 @@ static int calc_max_buckets(void)
}
/*
- * Rounds a number down to a power of 2.
- */
-static uint32_t round_down(uint32_t n)
-{
- while (n & (n - 1))
- n &= (n - 1);
- return n;
-}
-
-/*
* Allocate room for a suitable hash table.
*/
static int init_hash_tables(struct dm_snapshot *s)
{
- sector_t hash_size, cow_dev_size, origin_dev_size, max_buckets;
+ unsigned int hash_size, max_buckets;
+ sector_t cow_dev_size, origin_dev_size;
/*
* Calculate based on the size of the original volume or
@@ -361,7 +352,7 @@ static int init_hash_tables(struct dm_sn
hash_size = min(hash_size, max_buckets);
/* Round it down to a power of 2 */
- hash_size = round_down(hash_size);
+ hash_size = 1 << (fls(hash_size) - 1);
if (init_exception_table(&s->complete, hash_size))
return -ENOMEM;
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]