public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: David Laight <David.Laight@ACULAB.COM>
To: "Jason A. Donenfeld" <zx2c4@kernel.org>
Cc: "kbuild-all@lists.01.org" <kbuild-all@lists.01.org>,
	"zx2c4@kernel.org" <zx2c4@kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: RE: [crng-random:jd/get_random_u32_below 23/23] include/linux/random.h:64:69: sparse: sparse: cast truncates bits from constant value (1f4 becomes f4)
Date: Mon, 10 Oct 2022 17:18:40 +0000	[thread overview]
Message-ID: <a473cc4da4874235b2b98fd1a57e782e@AcuMS.aculab.com> (raw)
In-Reply-To: <202210100722.Wc2gmfS6-lkp@intel.com>

From: kernel test robot <lkp@intel.com>
> Sent: 10 October 2022 00:32
> To: Jason A. Donenfeld <zx2c4@kernel.org>
...

I'm missing the main mailing list email for this change.
I'm guessing the non-inlined code for non-constant ceil
is similar.

> vim +64 include/linux/random.h
> 
>     53
>     54	u32 __get_random_u32_below(u32 ceil);
>     55	/* Returns a random integer in the interval [0, ceil), with uniform distribution. */
>     56	static inline u32 get_random_u32_below(u32 ceil)
>     57	{
>     58		if (!__builtin_constant_p(ceil))
>     59			return __get_random_u32_below(ceil);
>     60
>     61		for (;;) {
>     62			if (ceil <= 1U << 8) {
>     63				u32 mult = ceil * get_random_u8();
>   > 64				if (is_power_of_2(ceil) || (u8)mult >= -(u8)ceil % ceil)
>     65					return mult >> 8;

If you are going to check is_power_of_2() then you might as well do:
		u32 val = get_random_u8();
		if (is_power_of_2(ceil))
			return ceil == 0x100 ? val : val & (ceil - 1);
Except that you don't want to loop on zero - so !(ceil & (ceil - 1))
is arguably better since it is absolutely explicit.
Or (for the constant case) a BUILD_BUG_ON(ceil == 0)?

Notwithstanding the completely untested code the bot complained about
doing a division here is unnecessary and expensive.
(Except this is the constant path...)
I think:
		val *= ceil;
		if ((val + ceil - 1) >> 8 == val >> 8)
			return val >> 8;
has the desired property.

It is also definitely worth a comment like:
	/* In the worst case this loops 50% of the time.
	 * While the loop is unbounded the average number
	 * of iterations (for the worst ceil) is 2. */

There are also a lot of places where having the values
evenly spread is enough - even if some values are returned
one more time than others.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)


  reply	other threads:[~2022-10-10 17:18 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-09 23:32 [crng-random:jd/get_random_u32_below 23/23] include/linux/random.h:64:69: sparse: sparse: cast truncates bits from constant value (1f4 becomes f4) kernel test robot
2022-10-10 17:18 ` David Laight [this message]
2022-10-10 17:53   ` Jason A. Donenfeld
2022-10-11 23:24     ` David Laight
2022-10-11 23:31       ` Jason A. Donenfeld

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=a473cc4da4874235b2b98fd1a57e782e@AcuMS.aculab.com \
    --to=david.laight@aculab.com \
    --cc=kbuild-all@lists.01.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=zx2c4@kernel.org \
    /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