Re: [PATCH v4 08/25] io_uring: add io_uring_fixed_index_get() and io_uring_fixed_index_put()

[Joanne Koong <joannelkoong@gmail.com>]

On Thu, Jan 22, 2026 at 1:02 PM Caleb Sander Mateos
<csander@purestorage.com> wrote:
>
> On Fri, Jan 16, 2026 at 3:31 PM Joanne Koong <joannelkoong@gmail.com> wrote:
> >
> > Add two new helpers, io_uring_fixed_index_get() and
> > io_uring_fixed_index_put(). io_uring_fixed_index_get() constructs an
> > iter for a fixed buffer at a given index and acquires a refcount on
> > the underlying node. io_uring_fixed_index_put() decrements this
> > refcount. The caller is responsible for ensuring
> > io_uring_fixed_index_put() is properly called for releasing the refcount
> > after it is done using the iter it obtained through
> > io_uring_fixed_index_get().
> >
> > The struct io_rsrc_node pointer needs to be returned in
> > io_uring_fixed_index_get() because the buffer at the index may be
> > unregistered/replaced in the meantime between this and the
> > io_uring_fixed_index_put() call. io_uring_fixed_index_put() takes in the
> > struct io_rsrc_node pointer as an arg.
> >
> > This is a preparatory patch needed for fuse-over-io-uring support, as
> > the metadata for fuse requests will be stored at the last index, which
> > will be different from the buf index set on the sqe.
> >
> > Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
> > ---
> >  include/linux/io_uring/cmd.h | 20 ++++++++++++
> >  io_uring/rsrc.c              | 59 ++++++++++++++++++++++++++++++++++++
> >  2 files changed, 79 insertions(+)
> >
> > diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c
> > index 41c89f5c616d..fa41cae5e922 100644
> > --- a/io_uring/rsrc.c
> > +++ b/io_uring/rsrc.c
> > @@ -1152,6 +1152,65 @@ int io_import_reg_buf(struct io_kiocb *req, struct iov_iter *iter,
> >         return io_import_fixed(ddir, iter, node->buf, buf_addr, len);
> >  }
> >
> > +struct io_rsrc_node *io_uring_fixed_index_get(struct io_uring_cmd *cmd,
> > +                                             int buf_index, unsigned int off,
> > +                                             size_t len, int ddir,
> > +                                             struct iov_iter *iter,
> > +                                             unsigned int issue_flags)
> > +{
> > +       struct io_ring_ctx *ctx = cmd_to_io_kiocb(cmd)->ctx;
> > +       struct io_rsrc_node *node;
> > +       struct io_mapped_ubuf *imu;
> > +       u64 addr;
> > +       int err;
> > +
> > +       io_ring_submit_lock(ctx, issue_flags);
> > +
> > +       node = io_rsrc_node_lookup(&ctx->buf_table, buf_index);
> > +       if (!node) {
> > +               io_ring_submit_unlock(ctx, issue_flags);
> > +               return ERR_PTR(-EINVAL);
> > +       }
> > +
> > +       node->refs++;
> > +
> > +       io_ring_submit_unlock(ctx, issue_flags);
> > +
> > +       imu = node->buf;
> > +       if (!imu) {
>
> How is this possible?

You're right, this null check is unnecessary. I'll drop it.

Thank you for reviewing the patches, Caleb.

>
> Other than that,
> Reviewed-by: Caleb Sander Mateos <csander@purestorage.com>
>
> > +               err = -EFAULT;
> > +               goto error;
> > +       }