Re: [PATCH v14 08/17] splice: Do splice read from a file without using ITER_PIPE

From: Ming Lei
Date: Fri Feb 17 2023 - 04:19:11 EST


On Fri, Feb 17, 2023 at 4:22 PM Ming Lei <ming.lei@xxxxxxxxxx> wrote:
>
> On Tue, Feb 14, 2023 at 05:13:21PM +0000, David Howells wrote:
> > Make generic_file_splice_read() use filemap_splice_read() and
> > direct_splice_read() rather than using an ITER_PIPE and call_read_iter().
> >
> > With this, ITER_PIPE is no longer used.
> >
> > Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
> > cc: Jens Axboe <axboe@xxxxxxxxx>
> > cc: Christoph Hellwig <hch@xxxxxx>
> > cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
> > cc: David Hildenbrand <david@xxxxxxxxxx>
> > cc: John Hubbard <jhubbard@xxxxxxxxxx>
> > cc: linux-mm@xxxxxxxxx
> > cc: linux-block@xxxxxxxxxxxxxxx
> > cc: linux-fsdevel@xxxxxxxxxxxxxxx
> > ---
> >
> > Notes:
> > ver #14)
> > - Split out filemap_splice_read() into a separate patch.
> >
> > fs/splice.c | 30 +++++++-----------------------
> > 1 file changed, 7 insertions(+), 23 deletions(-)
> >
> > diff --git a/fs/splice.c b/fs/splice.c
> > index 4c6332854b63..a93478338cec 100644
> > --- a/fs/splice.c
> > +++ b/fs/splice.c
> > @@ -391,29 +391,13 @@ ssize_t generic_file_splice_read(struct file *in, loff_t *ppos,
> > struct pipe_inode_info *pipe, size_t len,
> > unsigned int flags)
> > {
> > - struct iov_iter to;
> > - struct kiocb kiocb;
> > - int ret;
> > -
> > - iov_iter_pipe(&to, ITER_DEST, pipe, len);
> > - init_sync_kiocb(&kiocb, in);
> > - kiocb.ki_pos = *ppos;
> > - ret = call_read_iter(in, &kiocb, &to);
> > - if (ret > 0) {
> > - *ppos = kiocb.ki_pos;
> > - file_accessed(in);
> > - } else if (ret < 0) {
> > - /* free what was emitted */
> > - pipe_discard_from(pipe, to.start_head);
> > - /*
> > - * callers of ->splice_read() expect -EAGAIN on
> > - * "can't put anything in there", rather than -EFAULT.
> > - */
> > - if (ret == -EFAULT)
> > - ret = -EAGAIN;
> > - }
> > -
> > - return ret;
> > + if (unlikely(*ppos >= file_inode(in)->i_sb->s_maxbytes))
> > + return 0;
> > + if (unlikely(!len))
> > + return 0;
> > + if (in->f_flags & O_DIRECT)
> > + return direct_splice_read(in, ppos, pipe, len, flags);
>
> Hello David,
>
> I have one question, for dio, pages need to map to userspace
> memory, but direct_splice_read() just allocates pages and not
> see when the user mapping is setup, can you give one hint?

oops, it is ->splice_read, not ->read_iter, and pipe buffer isn't
user-backed, sorry for the noise.

Thanks,
Ming