Re: [PATCH net-next v2 08/10] crypto: af_alg: Support MSG_SPLICE_PAGES

From: David Howells <dhowells@redhat.com>
To: Paolo Abeni <pabeni@redhat.com>
Cc: dhowells@redhat.com, netdev@vger.kernel.org,
	Herbert Xu <herbert@gondor.apana.org.au>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>,
	Willem de Bruijn <willemdebruijn.kernel@gmail.com>,
	David Ahern <dsahern@kernel.org>,
	Matthew Wilcox <willy@infradead.org>,
	Jens Axboe <axboe@kernel.dk>,
	linux-crypto@vger.kernel.org, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH net-next v2 08/10] crypto: af_alg: Support MSG_SPLICE_PAGES
Date: Thu, 01 Jun 2023 12:35:12 +0100	[thread overview]
Message-ID: <822308.1685619312@warthog.procyon.org.uk> (raw)
In-Reply-To: <bd2750e52b47af1782233e254114eb8d627f1073.camel@redhat.com>

Paolo Abeni <pabeni@redhat.com> wrote:

> > +	if ((msg->msg_flags & MSG_SPLICE_PAGES) &&
> > +	    !iov_iter_is_bvec(&msg->msg_iter))
> > +		return -EINVAL;
> > +
> ...
> It looks like the above expect/supports only ITER_BVEC iterators, what
> about adding a WARN_ON_ONCE(<other iov type>)?

Meh.  I relaxed that requirement as I'm now using tools to extract stuff from
any iterator (extract_iter_to_sg() in this case) rather than walking the
bvec[] directly.  I forgot to remove the check from af_alg.  I can add an
extra patch to remove it.  Also, it probably doesn't matter for AF_ALG since
that's only likely to be called from userspace, either directly (which will
not set MSG_SPLICE_PAGES) or via splice (which will pass a BVEC).  Internal
kernel code will use crypto API directly.

> Also, I'm keeping this series a bit more in pw to allow Herbert or
> others to have a look.

Thanks.

David