Re: [PATCH 04/26] net: add a new sockptr_t type

[Eric Dumazet]

On Wed, Jul 22, 2020 at 11:09 PM Christoph Hellwig <hch@xxxxxx> wrote:
>
> Add a uptr_t type that can hold a pointer to either a user or kernel
> memory region, and simply helpers to copy to and from it.
>
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> ---
>  include/linux/sockptr.h | 104 ++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 104 insertions(+)
>  create mode 100644 include/linux/sockptr.h
>
> diff --git a/include/linux/sockptr.h b/include/linux/sockptr.h
> new file mode 100644
> index 00000000000000..700856e13ea0c4
> --- /dev/null
> +++ b/include/linux/sockptr.h
> @@ -0,0 +1,104 @@
> +/* SPDX-License-Identifier: GPL-2.0-only */
> +/*
> + * Copyright (c) 2020 Christoph Hellwig.
> + *
> + * Support for "universal" pointers that can point to either kernel or userspace
> + * memory.
> + */
> +#ifndef _LINUX_SOCKPTR_H
> +#define _LINUX_SOCKPTR_H
> +
> +#include <linux/slab.h>
> +#include <linux/uaccess.h>
> +
> +typedef struct {
> +       union {
> +               void            *kernel;
> +               void __user     *user;
> +       };
> +       bool            is_kernel : 1;
> +} sockptr_t;
>

I am not sure why you chose sockptr_t   for something that really seems generic.

Or is it really meant to be exclusive to setsockopt() and/or getsockopt() ?

If the first user of this had been futex code, we would have used
futexptr_t, I guess.