mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: David Laight <david.laight.linux@gmail.com>
To: Kent Overstreet <kent.overstreet@linux.dev>
Cc: linux-kernel@vger.kernel.org, "Theodore Ts'o" <tytso@mit.edu>,
	"Jason A. Donenfeld" <Jason@zx2c4.com>
Subject: Re: [PATCH] random: get_random_u64_below()
Date: Sat, 15 Mar 2025 13:52:34 +0000	[thread overview]
Message-ID: <20250315135234.65423e07@pumpkin> (raw)
In-Reply-To: <20250313163810.60564-1-kent.overstreet@linux.dev>

On Thu, 13 Mar 2025 12:38:10 -0400
Kent Overstreet <kent.overstreet@linux.dev> wrote:

> bcachefs needs this, for sampling devices to read from based on squared
> device latencies.
> 
> this uses the same algorithm as get_random_u32_below: since the multiply
> uses the top and bottom halves separately, it works out fairly well.

Adding two separate copies of much the same code is silly.
Given what the code is doing, does it ever make any sense to inline it.

Inlining the original get_random_u32_below(ceil) that did
	(random_u32() * ((1ull << 32) / ceil) >> 32
(for constant ceil) made sense.
While good enough for most purposes it was replaced by the much more
expensive function that guarantees that all the output values are
equally likely - rather than just evenly distributed.

	David

> 
> Cc: "Theodore Ts'o" <tytso@mit.edu> (maintainer:RANDOM NUMBER DRIVER)
> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com> (maintainer:RANDOM NUMBER DRIVER)
> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
> ---
>  drivers/char/random.c  | 22 ++++++++++++++++++++++
>  include/linux/random.h | 22 ++++++++++++++++++++++
>  2 files changed, 44 insertions(+)
> 
> diff --git a/drivers/char/random.c b/drivers/char/random.c
> index 2581186fa61b..84808300044c 100644
> --- a/drivers/char/random.c
> +++ b/drivers/char/random.c
> @@ -588,6 +588,28 @@ u32 __get_random_u32_below(u32 ceil)
>  }
>  EXPORT_SYMBOL(__get_random_u32_below);
>  
> +u64 __get_random_u64_below(u64 ceil)
> +{
> +	if (unlikely(!ceil))
> +		return get_random_u64();
> +	if (ceil <= U32_MAX)
> +		return __get_random_u32_below(ceil);
> +
> +	u64 rand = get_random_u64();
> +	u64 mult = ceil * rand;
> +
> +	if (unlikely(mult < ceil)) {
> +		u64 bound = -ceil % ceil;
> +		while (unlikely(mult < bound)) {
> +			rand = get_random_u64();
> +			mult = ceil * rand;
> +		}
> +	}
> +
> +	return mul_u64_u64_shr(ceil, rand, 64);
> +}
> +EXPORT_SYMBOL(__get_random_u64_below);
> +
>  #ifdef CONFIG_SMP
>  /*
>   * This function is called when the CPU is coming up, with entry
> diff --git a/include/linux/random.h b/include/linux/random.h
> index 333cecfca93f..b025bf3d8f27 100644
> --- a/include/linux/random.h
> +++ b/include/linux/random.h
> @@ -6,6 +6,7 @@
>  #include <linux/bug.h>
>  #include <linux/kernel.h>
>  #include <linux/list.h>
> +#include <linux/math64.h>
>  
>  #include <uapi/linux/random.h>
>  
> @@ -90,6 +91,27 @@ static inline u32 get_random_u32_below(u32 ceil)
>  	}
>  }
>  
> +u64 __get_random_u64_below(u64 ceil);
> +
> +static inline u64 get_random_u64_below(u32 ceil)
> +{
> +	if (!__builtin_constant_p(ceil))
> +		return __get_random_u64_below(ceil);
> +
> +	BUILD_BUG_ON_MSG(!ceil, "get_random_u64_below() must take ceil > 0");
> +	if (ceil <= 1)
> +		return 0;
> +	if (ceil <= U32_MAX)
> +		return get_random_u32_below(ceil);
> +
> +	for (;;) {
> +		u64 rand = get_random_u64();
> +		u64 mult = ceil * rand;
> +		if (likely(mult >= -ceil % ceil))
> +			return mul_u64_u64_shr(ceil, rand, 64);
> +	}
> +}
> +
>  /*
>   * Returns a random integer in the interval (floor, U32_MAX], with uniform
>   * distribution, suitable for all uses. Fastest when floor is a constant, but


  reply	other threads:[~2025-03-15 13:52 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-03-13 16:38 Kent Overstreet
2025-03-15 13:52 ` David Laight [this message]
2025-03-15 13:58   ` David Laight
2025-03-15 18:20   ` Kent Overstreet
2025-03-15 20:55     ` David Laight
2025-03-15 21:32       ` Kent Overstreet
2025-03-16 13:35         ` David Laight

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250315135234.65423e07@pumpkin \
    --to=david.laight.linux@gmail.com \
    --cc=Jason@zx2c4.com \
    --cc=kent.overstreet@linux.dev \
    --cc=linux-kernel@vger.kernel.org \
    --cc=tytso@mit.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®