[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [dm-devel] [PATCH v5 12/12] block: Only clone bio vecs that are in use

On Wed, Aug 08, 2012 at 04:30:07PM -0700, Tejun Heo wrote:
> Hello,
> On Mon, Aug 06, 2012 at 03:08:41PM -0700, Kent Overstreet wrote:
> > @@ -459,10 +460,10 @@ void __bio_clone(struct bio *bio, struct bio *bio_src)
> >  	bio->bi_sector = bio_src->bi_sector;
> >  	bio->bi_bdev = bio_src->bi_bdev;
> >  	bio->bi_flags |= 1 << BIO_CLONED;
> > +	bio->bi_flags &= ~(1 << BIO_SEG_VALID);
> This isn't obvious at all.  Why no explanation anywhere?  Also it
> would be nice to update comments of the updated functions so that it's
> clear that only partial cloning happens.

Because it's not obvious to me, either - I had to grep around through a
fair amount of code to figure out the semantics of BIO_SEG_VALID and I
doubt I have it 100%. I'm also pretty sure it's not used consistently in
the existing code...

If it means what I think it means, it should be obvious - nr_segs isn't
valid because the number of pages in the iovec changed (though we didn't
bother to copy it over anyways. So it's not necessary if nr_segs = 0
means nr_segs isn't valid, but bleh).

Anyways. yeah. BIO_SEG_VALID should be documented, and if it was I think
this code would be fine.

I will update the comment for the partial cloning thing:

 * 	__bio_clone	-	clone a bio
 * 	@bio: destination bio
 * 	@bio_src: bio to clone
 *	Clone a &bio. Caller will own the returned bio, but not
 *	the actual data it points to. Reference count of returned
 * 	bio will be one.
 *	We don't clone the entire bvec, just the part from bi_idx to b_vcnt
 *	(i.e. what the bio currently points to, so the new bio is still
 *	equivalent to the old bio).
void __bio_clone(struct bio *bio, struct bio *bio_src)
	       bio_segments(bio_src) * sizeof(struct bio_vec));

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]