Re: [PATCH v6 11/17] openrisc: use fallback for random_get_entropy() instead of zero

From: Stafford Horne
Date: Wed Apr 27 2022 - 17:42:40 EST


On Sat, Apr 23, 2022 at 11:26:17PM +0200, Jason A. Donenfeld wrote:
> In the event that random_get_entropy() can't access a cycle counter or
> similar, falling back to returning 0 is really not the best we can do.
> Instead, at least calling random_get_entropy_fallback() would be
> preferable, because that always needs to return _something_, even
> falling back to jiffies eventually. It's not as though
> random_get_entropy_fallback() is super high precision or guaranteed to
> be entropic, but basically anything that's not zero all the time is
> better than returning zero all the time.
>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Arnd Bergmann <arnd@xxxxxxxx>
> Cc: Jonas Bonn <jonas@xxxxxxxxxxxx>
> Cc: Stefan Kristiansson <stefan.kristiansson@xxxxxxxxxxxxx>
> Cc: Stafford Horne <shorne@xxxxxxxxx>
> Signed-off-by: Jason A. Donenfeld <Jason@xxxxxxxxx>
> ---
> arch/openrisc/include/asm/timex.h | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/arch/openrisc/include/asm/timex.h b/arch/openrisc/include/asm/timex.h
> index d52b4e536e3f..115e89b336cd 100644
> --- a/arch/openrisc/include/asm/timex.h
> +++ b/arch/openrisc/include/asm/timex.h
> @@ -23,6 +23,9 @@ static inline cycles_t get_cycles(void)
> {
> return mfspr(SPR_TTCR);
> }
> +#define get_cycles get_cycles
> +
> +#define random_get_entropy() (((unsigned long)get_cycles()) ?: random_get_entropy_fallback())
>
> /* This isn't really used any more */
> #define CLOCK_TICK_RATE 1000
> --
> 2.35.1

Hi Jason,

This looks OK, but I am wondering why we cannot add this to
"include/linux/timex.h" as the default implementation of random_get_entropy
if get_cycles is defined. I see there are 2 cases:

1. architectures that define get_cycles, and random_get_entropy is defined in
include/linux/timex.h.
(openrisc, sparc*, xtensa*, nios2, um*, arm)

2. architectures that define random_get_entropy explicitly
(mips, m68k, riscv, x86)

* Those marked with * just define get_cycles as 0.

I would think in "include/linux/timex.h" we could define.

#ifndef random_get_entropy
#ifdef get_cycles
#define random_get_entropy() (((unsigned long)get_cycles()) ?: random_get_entropy_fallback())
#else
#define random_get_entropy() random_get_entropy_fallback()
#endif
#endif

For architectures that define get_cycles as 0, they can be cleaned up.

-Stafford