linux-kernel - Re: [PATCH net-next v3 01/17] asm: simd context helper API

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 13 Sep 2018 05:03:01 +0000
From:   Kevin Easton <kevin@...rana.org>
To:     "Jason A. Donenfeld" <Jason@...c4.com>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Netdev <netdev@...r.kernel.org>,
        David Miller <davem@...emloft.net>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Andrew Lutomirski <luto@...nel.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Samuel Neves <sneves@....uc.pt>, linux-arch@...r.kernel.org
Subject: Re: [PATCH net-next v3 01/17] asm: simd context helper API

On Wed, Sep 12, 2018 at 08:10:41PM +0200, Jason A. Donenfeld wrote:
> On Wed, Sep 12, 2018 at 8:14 AM Kevin Easton <kevin@...rana.org> wrote:
> > Given that it's always supposed to be used like that, mightn't it be
> > better if simd_relax() took a pointer to the context, so the call is
> > just
> >
> >     simd_relax(&simd_context);
> >
> > ?
> >
> > The inlining means that there won't actually be a pointer dereference in
> > the emitted code.
> >
> > If simd_put() also took a pointer then it could set the context back to
> > HAVE_NO_SIMD as well?
> 
> That's sort of a neat idea. I guess in this scheme, you'd envision:
> 
>    simd_context_t simd_context;
> 
>    simd_get(&simd_context);
>    simd_relax(&simd_context);
>    simd_put(&simd_context);
> 
> And this way, if simd_context ever becomes a heavier struct, it can be
> modified in place rather than returned by value from the function. On
> the other hand, it's a little bit more annoying to type and makes it
> harder to do declaration and initialization on the same line.

Yes.  It's also how most get/put APIs already work in the kernel, eg
kref_get/put (mostly because they tend to be 'getting/putting' an
already-initialized object, though).

    - Kevin