linux-kernel - Re: [PATCH REBASED RESEND] x86/cpu: Move early cpu initialization into a separate translation unit

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] [day] [month] [year] [list]

Message-ID: <20180622174359.GD1882@zn.tnic>
Date:   Fri, 22 Jun 2018 19:43:59 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
        Ingo Molnar <mingo@...hat.com>, x86@...nel.org,
        "H. Peter Anvin" <hpa@...or.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH REBASED RESEND] x86/cpu: Move early cpu initialization
 into a separate translation unit

On Fri, Jun 22, 2018 at 06:50:56PM +0200, Thomas Gleixner wrote:
> > The magic, which I cut out, will rewrite the "jmp 6f, nops" thing to
> > "jmp %l[y_{yes,no}]" at the alternative patching and we'll loose the
> > dynamic test, pinning the condition forever more.
> 
> Hrm. Memory seems have to tricked me. So yes, it should work then.

Yes, but do verify that it does still. Because depending on how early
you call it, boot_cpu_data.x86_capability might not be populated
properly yet and then the dynamic case is wrong too. So check the order
pls.

> Though I still prefer the two liners fixup of the cpu_init() section
> mismatch thingy for now over the whole code move. Especially since Borislav
> and I have plans to rework that insanity completely once the speculative
> distractions are subsiding.

Hell yeah.

-- 
Regards/Gruss,
    Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.