On Tue, Feb 26, 2013 at 04:10:22PM +0800, Qu Wenruo wrote:
> Though most of the btrfs codes are using ALIGN macro for page alignment,
> there are still some codes using open-coded alignment like the
> following:
> ------
> u64 mask = ((u64)root->stripesize - 1);
> u64 ret = (val + mask) & ~mask;
> ------
> Or even hidden one:
> ------
> num_bytes = (end - start + blocksize) & ~(blocksize - 1);
It's unobvious but not a bug. The 'end' values are sometimes inclusive,
sometimes not. The range we want to align is
RANGE = end + start + 1
then the alignment transformation does
(RANGE + ALIGNMENT - 1) & ~(ALIGNMENT - 1)
where
RANGE + ALIGNMENT - 1 = end + start + 1 + ALIGNMENT - 1
= end + start + ALIGNMENT
I have covered all cases in my original patch with a coccinelle semantic
patch http://www.spinics.net/lists/linux-btrfs/msg12750.html and also
removed the useless stripe_align helper and updated it's callers (and it
still applies after merging the raid56 code that updated the function
with 2 unused arguments).
> Also there is a previous patch from David Sterba with similar changes,
> but the patch is for 3.2 kernel and seems not merged.
> http://www.spinics.net/lists/linux-btrfs/msg12747.html
Yeah, amount of unmerged patches was high back then and I got no
feedback whether such changes are desired not, moreover a few of them
touched all sources likely to clash with every secnod patch in flight
(eg. "Btrfs: kill key type helpers").
If you used my patch I'd appreciate to keep the signed-off.
> Cc: David Sterba <dave@xxxxxxxx>
> Signed-off-by: Qu Wenruo <quwenruo@xxxxxxxxxxxxxx>
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
