linux-kernel - Re: espfix and 5-level paging

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

Date:   Wed, 31 May 2017 06:56:51 -0700
From:   Andy Lutomirski <luto@...nel.org>
To:     Yuhong Bao <YuhongBao_386@...mail.com>
Cc:     "kirill.shutemov@...ux.intel.com" <kirill.shutemov@...ux.intel.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "luto@...nel.org" <luto@...nel.org>
Subject: Re: espfix and 5-level paging

On Sat, May 27, 2017 at 2:30 PM, Yuhong Bao <YuhongBao_386@...mail.com> wrote:
> I don't like espfix of course, and would probably manually fix it to N in the kernel config unless I really had to run buggy DPMI apps.
> Win8 decided to just disable NTVDM by default instead.
> ________________________________________
> From: Yuhong Bao
> Sent: Saturday, May 13, 2017 2:39:59 AM
> To: kirill.shutemov@...ux.intel.com; linux-kernel@...r.kernel.org
> Subject: espfix and 5-level paging
>
> Instead of porting espfix to 5-level paging, why not fix the IRET instruction?
> It probably should not require even a CR4 flag to enable.

I have no context here.  It would be great if Intel and/or AMD were to
fix IRET.  Meanwhile, we can work around it in the kernel, which we do
if MODIFY_LDT is on and configured to allow 16-bit segments.  If you
don't want 16-bit segments, you can turn them off.

(This has nothing to do with buggy DPMI apps.  It affects pretty much
all 16-bit code.)