Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 17 Nov 2022 03:05:14 +0100
From:   "Jason A. Donenfeld" <Jason@zx2c4.com>
To:     Kees Cook <keescook@chromium.org>
Cc:     linux-kernel@vger.kernel.org, patches@lists.linux.dev,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jakub Kicinski <kuba@kernel.org>,
        Russell King <linux@armlinux.org.uk>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
        Heiko Carstens <hca@linux.ibm.com>,
        Herbert Xu <herbert@gondor.apana.org.au>,
        Christoph =?utf-8?Q?B=C3=B6hmwalder?= 
        <christoph.boehmwalder@linbit.com>,
        Jani Nikula <jani.nikula@linux.intel.com>,
        Jason Gunthorpe <jgg@nvidia.com>,
        Sakari Ailus <sakari.ailus@linux.intel.com>,
        "Martin K . Petersen" <martin.petersen@oracle.com>,
        Theodore Ts'o <tytso@mit.edu>,
        Andreas Dilger <adilger.kernel@dilger.ca>,
        Jaegeuk Kim <jaegeuk@kernel.org>,
        Richard Weinberger <richard@nod.at>,
        "Darrick J . Wong" <djwong@kernel.org>,
        SeongJae Park <sj@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Andrew Morton <akpm@linux-foundation.org>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Helge Deller <deller@gmx.de>, netdev@vger.kernel.org,
        linux-crypto@vger.kernel.org, linux-block@vger.kernel.org,
        linux-fsdevel@vger.kernel.org, linux-media@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, loongarch@lists.linux.dev,
        linux-mips@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
        linux-mmc@vger.kernel.org, linux-parisc@vger.kernel.org,
        ydroneaud@opteya.com
Subject: Re: [PATCH v2 3/3] treewide: use get_random_u32_between() when
 possible
Message-ID: <Y3WW2lOgoYLKQeve@zx2c4.com>
References: <20221114164558.1180362-1-Jason@zx2c4.com>
 <20221114164558.1180362-4-Jason@zx2c4.com>
 <202211161436.A45AD719A@keescook>
 <Y3V4g8eorwiU++Y3@zx2c4.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <Y3V4g8eorwiU++Y3@zx2c4.com>
Precedence: bulk

On Thu, Nov 17, 2022 at 12:55:47AM +0100, Jason A. Donenfeld wrote:
> 1) How/whether to make f(0, UR2_MAX) safe,
>    - without additional 64-bit arithmetic,
>    - minimizing the number of branches.
>    I have a few ideas I'll code golf for a bit.
> I think I can make progress with (1) alone by fiddling around with
> godbolt enough, like usual.

The code gen is definitely worse.

Original half-open interval:

    return floor + get_random_u32_below(ceil - floor);

Suggested fully closed interval:
	
    ceil = ceil - floor + 1;
    return likely(ceil) ? floor + get_random_u32_below(ceil) : get_random_u32();

Is the worse code gen actually worth it? Options:

 a) Decide worse codegen is worth it.
 b) Declare f(0, U32_MAX) undefined and just not handle it.
 c) Stick with original half-open interval that doesn't have this problem.

Jason