linux-kernel - Re: [PATCH] x86/timers: Move simple_udelay_calibration past init_hypervisor

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: <ed94ee98-04ce-8c23-3fd8-84ff8c97ee1f@siemens.com>
Date:   Sun, 28 May 2017 18:41:40 +0200
From:   Jan Kiszka <jan.kiszka@...mens.com>
To:     Andy Shevchenko <andy.shevchenko@...il.com>,
        Mathias Nyman <mathias.nyman@...ux.intel.com>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>,
        "H. Peter Anvin" <hpa@...or.com>, x86 <x86@...nel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Lu Baolu <baolu.lu@...ux.intel.com>
Subject: Re: [PATCH] x86/timers: Move simple_udelay_calibration past
 init_hypervisor_platform

On 2017-05-27 01:04, Andy Shevchenko wrote:
> +Cc Mathias
> 
> On Wed, May 24, 2017 at 9:04 PM, Jan Kiszka <jan.kiszka@...mens.com> wrote:
>> This ensures that adjustments to x86_platform done by the hypervisor
>> setup is already respected by this simple calibration.
>>
>> The current user of this, introduced by 1b5aeebf3a92 ("x86/earlyprintk:
>> Add support for earlyprintk via USB3 debug port"), comes much later
>> into play.
>>
> 
> Can you elaborate a bit what is the issue you are experience?

I stumbled over this while rebasing the patches to boot Linux over the
Jailhouse hypervisor (as non-root cell). Those come with hypervisor
setup code that changes relevant callbacks because Jailhouse does not
virtualize most of the platform devices and simply denies access.

While these patches are still out-of-tree (will propose later this
year), the issue seen here is conceptual: e.g., calibration is pointless
if the hypervisor has a PV channel to report the required parameters -
like vmware.

Jan

-- 
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux