Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:118351
Message-ID: <5d5939059e4dbad172b25101fd9517d46af9e86d.camel@sandfox.me>
To: Hans Henrik Bergan <divinity76@gmail.com>
Cc: PHP Internals List <internals@lists.php.net>, Go Kudo
 <zeriyoshi@gmail.com>,  Tim =?ISO-8859-1?Q?D=FCsterhus?= <tim@bastelstu.be>
Date: Thu, 04 Aug 2022 23:40:16 +0300
In-Reply-To: <CAJmy8YEGzTnr9=XxUTkWU9fcBOzs1x-dXNov1ks-EzfcQoUXtQ@mail.gmail.com>
References: <157e4f6b6e96376e61194b9a358c87b02880ec28.camel@sandfox.me>
	 <67f9981b-de29-2f83-cecf-08ea3c1199ae@bastelstu.be>
	 <CAJmy8YEGzTnr9=XxUTkWU9fcBOzs1x-dXNov1ks-EzfcQoUXtQ@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.44.3 (3.44.3-1.fc36) 
MIME-Version: 1.0
Subject: Re: [PHP-DEV] xoshiro** edge case (all zeros)
From: sandfox@sandfox.me (Anton Smirnov)

On Thu, 2022-08-04 at 22:32 +0200, Hans Henrik Bergan wrote:
> dangerous to be sure, but it's also a technically valid seed,
> are you sure we should disallow a valid seed?

Reference implementation defines it as invalid:
https://prng.di.unimi.it/xoshiro256starstar.c

Some implementations choose to seed the RNG with 0x0000...0001 but that
kinda inroduces bias

> On Thu, 4 Aug 2022 at 20:33, Tim D=C3=BCsterhus <tim@bastelstu.be> wrote:
> > Hi
> >=20
> > On 8/4/22 10:09, Anton Smirnov wrote:
> > > xoshiro** has a known edge case: all-zero seed
> >=20
> > Indeed, good catch. I had that in mind, but forgot about it.
> >=20
> > > <?php
> > >=20
> > > $engine =3D new \Random\Engine\Xoshiro256StarStar(str_repeat("\0",
> > > 32));
> > >=20
> > > while (true) {
> > > =C2=A0 =C2=A0 =C2=A0 echo hex2bin($engine->generate()), PHP_EOL; //
> > > 0000000000000000
> > > }
> > >=20
> > > It should be documented and/or handled
> > >=20
> > > It's only for a string seed, int seed is not affected
> > >=20
> >=20
> > I've created a PR here:
> >=20
> > https://github.com/php/php-src/pull/9250
> >=20
> > I've opted to throw a ValueError in that case, as that's the only
> > safe=20
> > option that does not introduce a bias.
> >=20
> > The 32xNUL seed basically should only happen for manually written=20
> > testing input and not happen otherwise. An actual random seed will=20
> > result in 32 NUL bytes with just a 2**-256 chance and when relying
> > on=20
> > the implicit CSPRNG seeding (`null` as seed parameter) my PR will
> > just=20
> > retry to catch even that edge case.
> >=20
> > Best regards
> > Tim D=C3=BCsterhus
> >=20