[PATCH v2 08/10] io_uring/rw: add support to send meta along with read/write

[Anuj gupta]

On Wed, Jun 26, 2024 at 10:47 PM Gabriel Krisman Bertazi
<krisman at suse.de> wrote:
>
> Anuj Gupta <anuj20.g at samsung.com> writes:
>
> >  static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
> >                     int ddir, bool do_import)
> >  {
> > @@ -269,11 +307,16 @@ static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe,
> >               rw->kiocb.ki_ioprio = get_current_ioprio();
> >       }
> >       rw->kiocb.dio_complete = NULL;
> > +     rw->kiocb.ki_flags = 0;
> >
> >       rw->addr = READ_ONCE(sqe->addr);
> >       rw->len = READ_ONCE(sqe->len);
> >       rw->flags = READ_ONCE(sqe->rw_flags);
> > -     return io_prep_rw_setup(req, ddir, do_import);
> > +     ret = io_prep_rw_setup(req, ddir, do_import);
> > +
> > +     if (unlikely(req->ctx->flags & IORING_SETUP_SQE128 && !ret))
> > +             ret = io_prep_rw_meta(req, sqe, rw, ddir);
> > +     return ret;
>
> Would it be useful to have a flag to differentiate a malformed SQE from
> a SQE with io_uring_meta, instead of assuming sqe->cmd has it? We don't
> check for addr3 at the moment and differently from uring_cmd, userspace
> will be mixing writes commands with and without metadata to different
> files, so it would be useful to catch that.
>
Yes, but I couldn't find a good place to keep that flag. sqe->rw_flags are RWF
flags and are meant for generic read/write. sqe->flags are generic io_uring
flags and are not opcode specific. Do you see a place where this flag
could fit in?

> --
> Gabriel Krisman Bertazi
>
--
Anuj Gupta