Re: [PATCH] Compiler Attributes: Introduce __access_*() function attribute

[Miguel Ojeda]

On Sat, Sep 24, 2022 at 1:54 AM Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>
> are specified. While it is legal to provide only the pointer argument
> position and access type, design the kernel macros to require also the
> bounds (element count) argument position: if a function has no bounds
> argument, refactor the code to include one.

Should this bit be included in the comment of the attribute? (since it
is specific to the kernel)

> These can be used multiple times. For example:
>
> __access_wo(2, 3) __access_ro(4, 5)
> int copy_something(struct context *ctx, u32 *dst, size_t dst_count,
>                    u8 *src, int src_len);
>
> (And if "dst" will also be read, it could use __access_rw(2, 3) instead.)

Also maybe the example could be nice there too, since it uses the
syntax for the kernel and you took the time to write it :)

By the way, shouldn't `src` typically be `const u8 *`? Given it is an
example, I would qualify it.

> +#if __has_attribute(__access__)
> +#define __access_rw(ptr, count)        __attribute__((__access__(read_write, ptr, count)))
> +#define __access_ro(ptr, count)        __attribute__((__access__(read_only,  ptr, count)))
> +#define __access_wo(ptr, count)        __attribute__((__access__(write_only, ptr, count)))
> +#else
> +#define __access_rw(ptr, count)
> +#define __access_ro(ptr, count)
> +#define __access_wo(ptr, count)
> +#endif

If you do a v2 for the above, please take the chance to indent with a
single space after the `#` (like `# define`) for consistency.

Thanks!

Cheers,
Miguel