Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230519074047.1739879-1-dhowells@redhat.com> <20230519074047.1739879-4-dhowells@redhat.com>
In-Reply-To: <20230519074047.1739879-4-dhowells@redhat.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Fri, 19 May 2023 09:31:17 -0700
Message-ID: <CAHk-=whX+mAESz01NJZssoLMsgEpFjx7LDLO1_uW1qaDY2Jidw@mail.gmail.com>
Subject: Re: [PATCH v20 03/32] splice: Make direct_read_splice() limit to eof
 where appropriate
To:     David Howells <dhowells@redhat.com>
Cc:     Jens Axboe <axboe@kernel.dk>, Al Viro <viro@zeniv.linux.org.uk>,
        Christoph Hellwig <hch@infradead.org>,
        Matthew Wilcox <willy@infradead.org>, Jan Kara <jack@suse.cz>,
        Jeff Layton <jlayton@kernel.org>,
        David Hildenbrand <david@redhat.com>,
        Jason Gunthorpe <jgg@nvidia.com>,
        Logan Gunthorpe <logang@deltatee.com>,
        Hillf Danton <hdanton@sina.com>,
        Christian Brauner <brauner@kernel.org>,
        linux-fsdevel@vger.kernel.org, linux-block@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        Christoph Hellwig <hch@lst.de>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, May 19, 2023 at 12:41=E2=80=AFAM David Howells <dhowells@redhat.com=
> wrote:
>
> +
> +       if (S_ISREG(file_inode(in)->i_mode) ||
> +           S_ISBLK(file_inode(in)->i_mode)) {

This really feels fundamentally wrong to me.

If block and regular files have this limit, they should have their own
splice_read() function that implements that limit.

Not make everybody else check it.

IOW, this should be a separate function ("block_splice_read()" or
whatever), not inside a generic function that other users use.

The zero size checking looks fine, although I wondered about that too.
Some special files do traditionally have special meanings for
zero-sized reads (as in "packet boundary"). But I suspect that isn't
an issue for splice, and perhaps more importantly, I think the same
rule should be in place: special files that want special rules
shouldn't be using this generic function directly then.

                 Linus