Re: [PATCH 1/2] fs: Abstract calling the kiocb completion function

[Matthew Wilcox <willy@infradead.org>]

On Wed, Jul 08, 2020 at 04:37:21PM -0600, Jens Axboe wrote:
> On 7/8/20 4:26 PM, Matthew Wilcox (Oracle) wrote:
> > diff --git a/crypto/af_alg.c b/crypto/af_alg.c
> > index b1cd3535c525..590dbbcd0e9f 100644
> > --- a/crypto/af_alg.c
> > +++ b/crypto/af_alg.c
> > @@ -1045,7 +1045,7 @@ void af_alg_async_cb(struct crypto_async_request *_req, int err)
> >  	af_alg_free_resources(areq);
> >  	sock_put(sk);
> >  
> > -	iocb->ki_complete(iocb, err ? err : (int)resultlen, 0);
> > +	complete_kiocb(iocb, err ? err : (int)resultlen, 0);
> 
> I'd prefer having it called kiocb_complete(), seems more in line with
> what you'd expect in terms of naming for an exported interface.

Happy to make that change.  It seemed like you preferred the opposite
way round with is_sync_kiocb() and init_sync_kiocb() already existing.

Should I switch register_kiocb_completion and unregister_kiocb_completion
to kiocb_completion_register or kiocb_register_completion?