Re: Block integrity patches for 2.6.28

[Jens Axboe <jens.axboe@oracle.com>]

On Thu, Oct 02 2008, Martin K. Petersen wrote:
> >>>>> "Jens" == Jens Axboe <jens.axboe@oracle.com> writes:
> 
> Jens> As far as I can tell, most of that commit is still fine. You
> Jens> want bdev_get_integrity() in blkdev.h, the 3 other moves and the
> Jens> unused bdev_get_tag_size() do not look like they are being used
> Jens> by this patch set.
> 
> bdev_get_integrity() and bdev_get_tag_size() are being used by
> stacking drivers and filesystems to prepare I/O.  It's correct that
> none of the in-tree stuff currently uses bdev_get_tag_size().  That's
> coming with the btrfs support.  If you want to pull that out for now
> and have me put that back later in that's ok.  Just adds another
> two-stage merge dependency for a later cycle.

Well, I would not have added it in the first place, but it was there. I
already did the bdev_get_integrity() addon instead of the revert, so
lets please just keep it at that.

> bdev_integrity_enabled() and blk_integrity_tuple_size() are only being
> used from within bio-integrity.c and can move there.  I originally put
> them in blkdev.h because they are block device functions and not bio
> ditto.
> 
> Want me to submit a new patch shuffling bdev_get_integrity() back
> where it came from?

Do we need any on top of current for-2.6.28?

I'll apply your series with the modified patch #5, it'll probably need a
hand edit or two since I didn't revert the commit in question, but
should be trivial to resolve.


-- 
Jens Axboe