linux-kernel - Re: [Xen-devel] [PATCH v2] xen: get rid of paravirt op adjust_exception

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]

Message-ID: <5c14a9a9-7b13-50d0-2338-2bccac7b7ac3@citrix.com>
Date:   Wed, 2 Aug 2017 00:38:59 +0100
From:   Andrew Cooper <andrew.cooper3@...rix.com>
To:     Andy Lutomirski <luto@...capital.net>,
        Juergen Gross <jgross@...e.com>
Cc:     X86 ML <x86@...nel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Ingo Molnar <mingo@...hat.com>,
        "H. Peter Anvin" <hpa@...or.com>,
        "xen-devel@...ts.xenproject.org" <xen-devel@...ts.xenproject.org>,
        Boris Ostrovsky <boris.ostrovsky@...cle.com>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [Xen-devel] [PATCH v2] xen: get rid of paravirt op
 adjust_exception_frame

On 01/08/2017 20:45, Andy Lutomirski wrote:
> Also, IMO it would be nice to fully finish the job.  Remaining steps are:
>
> 1. Unsuck the SYSCALL entries on Xen PV.
> 2. Unsuck the SYENTER entry on Xen PV.
> 3. Make a xen_nmi that's actually correct (should be trivial)
>
> #1 is here:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/luto/linux.git/commit/?h=x86/entry_syscall&id=14fee348e3e86c994400d68085217d1232a637d6

If the

/* Zero-extending 32-bit regs, do not remove */
movl    %eax, %eax

comments are to be believed, then the entry logic needs reordering
slightly to:

GLOBAL(entry_*_compat_after_hwframe)
    movl    %eax, %eax    /* Zero-extending 32-bit regs, do not remove */
    pushq    %rax                /* pt_regs->orig_ax */

~Andrew

(It is unfortunate this can't be simplified to pushq %eax)