Re: [Qemu-block] [PATCH 3/4] osdep: Document differences in rounding macros

[Paolo Bonzini]

On 21/07/2016 21:34, Eric Blake wrote:
> Make it obvious which macros are safe in which situations.
> 
> Useful since QEMU_ALIGN_UP and ROUND_UP both purport to do
> the same thing, but differ on whether the alignment must be
> a power of 2.
> 
> Signed-off-by: Eric Blake <address@hidden>
> ---
>  include/qemu/osdep.h | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
> index fbb8759..9991fb0 100644
> --- a/include/qemu/osdep.h
> +++ b/include/qemu/osdep.h
> @@ -158,7 +158,8 @@ extern int daemon(int, int);
>  /* Round number down to multiple */
>  #define QEMU_ALIGN_DOWN(n, m) ((n) / (m) * (m))
> 
> -/* Round number up to multiple */
> +/* Round number up to multiple. Safe when m is not a power of 2 (see
> + * ROUND_UP for a faster version when a power of 2 is guaranteed) */
>  #define QEMU_ALIGN_UP(n, m) QEMU_ALIGN_DOWN((n) + (m) - 1, (m))
> 
>  /* Check if n is a multiple of m */
> @@ -175,6 +176,9 @@ extern int daemon(int, int);
>  /* Check if pointer p is n-bytes aligned */
>  #define QEMU_PTR_IS_ALIGNED(p, n) QEMU_IS_ALIGNED((uintptr_t)(p), (n))
> 
> +/* Round number up to multiple. Requires that d be a power of 2 (see
> + * QEMU_ALIGN_UP for a safer but slower version on arbitrary
> + * numbers) */
>  #ifndef ROUND_UP
>  #define ROUND_UP(n,d) (((n) + (d) - 1) & -(d))
>  #endif

Ouch, this is ugly, especially since DIV_ROUND_UP does not require
alignment!  Not your fault of course, and the patch is arguably an
improvement.

Paolo