All of lore.kernel.org
 help / color / mirror / Atom feed
From: Zachary Amsden <zach@vmware.com>
To: Arjan van de Ven <arjan@infradead.org>
Cc: Chris Wright <chrisw@sous-sol.org>,
	linux-kernel@vger.kernel.org, virtualization@lists.osdl.org,
	xen-devel@lists.xensource.com,
	Jeremy Fitzhardinge <jeremy@goop.org>, Andi Kleen <ak@suse.de>,
	Andrew Morton <akpm@osdl.org>,
	Rusty Russell <rusty@rustcorp.com.au>,
	Ian Pratt <ian.pratt@xensource.com>,
	Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Subject: Re: [RFC PATCH 18/33] Subarch support for CPUID instruction
Date: Tue, 18 Jul 2006 04:33:52 -0700	[thread overview]
Message-ID: <44BCC720.7050601@vmware.com> (raw)
In-Reply-To: <1153217686.3038.37.camel@laptopd505.fenrus.org>

Arjan van de Ven wrote:
> On Tue, 2006-07-18 at 00:00 -0700, Chris Wright wrote:
>   
>> plain text document attachment (i386-cpuid)
>> Allow subarchitectures to modify the CPUID instruction.  This allows
>> the subarch to provide a limited set of CPUID feature flags during CPU
>> identification.  Add a subarch implementation for Xen that traps to the
>> hypervisor where unsupported feature flags can be hidden from guests.
>>     
>
> Hi,
>
> I'm wondering if this is entirely the wrong level of abstraction; to me
> it feels the subarch shouldn't override the actual cpuid, but the cpu
> feature flags that linux uses. That's a lot less messy: cpuid has many
> many pieces of information which are near impossible to filter in
> practice, however filtering the USAGE of it is trivial; linux basically
> flattens the cpuid namespace into a simple bitmap of "what the kernel
> can use". That is really what the subarch should filter/fixup, just like
> we do for cpu quirks etc etc.
>   

You really need a CPUID hook.  The instruction is non-virtualizable, and 
anything claiming to be a hypervisor really has to support masking and 
flattening the cpuid namespace for the instruction itself.  It is used 
in assembler code and very early in boot.  The alternative is injecting 
a bunch of Xen-specific code to filter feature bits into the i386 layer, 
which is both bad for Linux and bad for Xen - and was quite ugly in the 
last set of Xen patches.

Zach

WARNING: multiple messages have this Message-ID (diff)
From: Zachary Amsden <zach@vmware.com>
To: Arjan van de Ven <arjan@infradead.org>
Cc: Andrew Morton <akpm@osdl.org>,
	Jeremy Fitzhardinge <jeremy@goop.org>,
	xen-devel@lists.xensource.com,
	Ian Pratt <ian.pratt@xensource.com>,
	Rusty Russell <rusty@rustcorp.com.au>,
	linux-kernel@vger.kernel.org, Andi Kleen <ak@suse.de>,
	Chris Wright <chrisw@sous-sol.org>,
	virtualization@lists.osdl.org,
	Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Subject: Re: [RFC PATCH 18/33] Subarch support for CPUID instruction
Date: Tue, 18 Jul 2006 04:33:52 -0700	[thread overview]
Message-ID: <44BCC720.7050601@vmware.com> (raw)
In-Reply-To: <1153217686.3038.37.camel@laptopd505.fenrus.org>

Arjan van de Ven wrote:
> On Tue, 2006-07-18 at 00:00 -0700, Chris Wright wrote:
>   
>> plain text document attachment (i386-cpuid)
>> Allow subarchitectures to modify the CPUID instruction.  This allows
>> the subarch to provide a limited set of CPUID feature flags during CPU
>> identification.  Add a subarch implementation for Xen that traps to the
>> hypervisor where unsupported feature flags can be hidden from guests.
>>     
>
> Hi,
>
> I'm wondering if this is entirely the wrong level of abstraction; to me
> it feels the subarch shouldn't override the actual cpuid, but the cpu
> feature flags that linux uses. That's a lot less messy: cpuid has many
> many pieces of information which are near impossible to filter in
> practice, however filtering the USAGE of it is trivial; linux basically
> flattens the cpuid namespace into a simple bitmap of "what the kernel
> can use". That is really what the subarch should filter/fixup, just like
> we do for cpu quirks etc etc.
>   

You really need a CPUID hook.  The instruction is non-virtualizable, and 
anything claiming to be a hypervisor really has to support masking and 
flattening the cpuid namespace for the instruction itself.  It is used 
in assembler code and very early in boot.  The alternative is injecting 
a bunch of Xen-specific code to filter feature bits into the i386 layer, 
which is both bad for Linux and bad for Xen - and was quite ugly in the 
last set of Xen patches.

Zach

  parent reply	other threads:[~2006-07-18 11:33 UTC|newest]

Thread overview: 135+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-07-18  9:18 [RFC PATCH 00/33] Xen i386 paravirtualization support Chris Wright
2006-07-18  9:18 ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 01/33] Add apply_to_page_range() function Chris Wright
2006-07-18 10:38   ` Adrian Bunk
2006-07-18 19:29     ` Chris Wright
2006-07-18 19:29       ` Chris Wright
2006-07-20  6:17       ` Adrian Bunk
2006-07-18  7:00 ` [RFC PATCH 02/33] Add sync bitops Chris Wright
2006-07-18  9:56   ` Arjan van de Ven
2006-07-18 10:18     ` Keir Fraser
2006-07-18 10:18       ` Keir Fraser
2006-07-19 12:54     ` Andi Kleen
2006-07-18 10:34   ` Adrian Bunk
2006-07-18  7:00 ` [RFC PATCH 03/33] Add nosegneg capability to the vsyscall page notes Chris Wright
2006-07-18  7:00 ` [RFC PATCH 04/33] Add XEN config options and disable unsupported config options Chris Wright
2006-07-18  9:59   ` Arjan van de Ven
2006-07-18 10:21     ` Keir Fraser
2006-07-18 10:21       ` Keir Fraser
2006-07-18  7:00 ` [RFC PATCH 05/33] Makefile support to build Xen subarch Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18 10:00   ` Arjan van de Ven
2006-07-18 11:40     ` Andrew Morton
2006-07-18 20:41       ` Jeremy Fitzhardinge
2006-07-18 20:15     ` Jeremy Fitzhardinge
2006-07-18  7:00 ` [RFC PATCH 06/33] Add Xen interface header files Chris Wright
2006-07-18  7:00 ` [RFC PATCH 07/33] Hypervisor " Chris Wright
2006-07-18  7:00 ` [RFC PATCH 08/33] Add vmlinuz build target Chris Wright
2006-07-18  7:00 ` [RFC PATCH 09/33] Add start-of-day setup hooks to subarch Chris Wright
2006-07-18 10:03   ` Arjan van de Ven
2006-07-18 20:49     ` Jeremy Fitzhardinge
2006-07-20  6:07   ` Adrian Bunk
2006-07-20 12:10     ` Keir Fraser
2006-07-20 13:27       ` Jeremy Fitzhardinge
2006-07-18  7:00 ` [RFC PATCH 10/33] add support for Xen feature queries Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 11/33] Add Xen-specific memory management definitions Chris Wright
2006-07-18  7:00 ` [RFC PATCH 12/33] Change __FIXADDR_TOP to leave room for the hypervisor Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 13/33] Add a new head.S start-of-day file for booting on Xen Chris Wright
2006-07-18 10:06   ` Arjan van de Ven
2006-07-18 20:13     ` Jeremy Fitzhardinge
2006-07-18  7:00 ` [RFC PATCH 14/33] subarch support for controlling interrupt delivery Chris Wright
2006-07-18  7:00 ` [RFC PATCH 15/33] move segment checks to subarch Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18 10:09   ` Arjan van de Ven
2006-07-18 11:28     ` Zachary Amsden
2006-07-18 11:28       ` Zachary Amsden
2006-07-18 19:06   ` Rusty Russell
2006-07-18 19:06     ` Rusty Russell
2006-07-18 19:25     ` Chris Wright
2006-07-18 19:25       ` Chris Wright
2006-07-18 20:00       ` [Xen-devel] " Rusty Russell
2006-07-18  7:00 ` [RFC PATCH 16/33] Add support for Xen to entry.S Chris Wright
2006-07-18 10:11   ` Arjan van de Ven
2006-07-18 20:04     ` Jeremy Fitzhardinge
2006-07-18 19:17   ` Rusty Russell
2006-07-18 19:17     ` Rusty Russell
2006-07-18 20:43     ` Chris Wright
2006-07-18 20:43       ` Chris Wright
2006-07-18 23:03       ` Jeremy Fitzhardinge
2006-07-19  5:30         ` Chris Wright
2006-07-19  5:30           ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 17/33] Support loading an initrd when running on Xen Chris Wright
2006-07-18  7:00 ` [RFC PATCH 18/33] Subarch support for CPUID instruction Chris Wright
2006-07-18 10:14   ` Arjan van de Ven
2006-07-18 10:26     ` Keir Fraser
2006-07-18 10:26       ` Keir Fraser
2006-07-18 10:38       ` Arjan van de Ven
2006-07-18 11:33     ` Zachary Amsden [this message]
2006-07-18 11:33       ` Zachary Amsden
2006-07-18 20:46       ` David Miller
2006-07-18 21:00         ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 19/33] Support gdt/idt/ldt handling on Xen Chris Wright
2006-07-18  7:00 ` [RFC PATCH 20/33] subarch support for interrupt and exception gates Chris Wright
2006-07-18  7:00 ` [RFC PATCH 21/33] subarch support for control register accesses Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 22/33] subarch stack pointer update Chris Wright
2006-07-18  7:00 ` [RFC PATCH 23/33] subarch TLB support Chris Wright
2006-07-18 20:39   ` David Miller
2006-07-18 21:00     ` Chris Wright
2006-07-18 21:00       ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 24/33] Add support for Xen event channels Chris Wright
2006-07-18  7:00 ` [RFC PATCH 25/33] Implement timekeeping for Xen Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-25  2:49   ` john stultz
2006-07-25 20:05     ` Jeremy Fitzhardinge
2006-07-18  7:00 ` [RFC PATCH 26/33] subarch suport for idle loop (NO_IDLE_HZ for Xen) Chris Wright
2006-07-18  7:00 ` [RFC PATCH 27/33] Add the Xen virtual console driver Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18 10:24   ` Arjan van de Ven
2006-07-18 10:31     ` Keir Fraser
2006-07-18 10:31       ` Keir Fraser
2006-07-27 15:05   ` Hollis Blanchard
2006-07-18  7:00 ` [RFC PATCH 28/33] Add Xen grant table support Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-19 10:04   ` [Xen-devel] " Harry Butterworth
2006-07-19 10:04     ` Harry Butterworth
2006-07-25 18:30   ` Hollis Blanchard
2006-07-25 18:30     ` Hollis Blanchard
2006-07-25 18:45     ` Keir Fraser
2006-07-25 18:45       ` Keir Fraser
2006-07-25 19:06     ` Segher Boessenkool
2006-07-25 19:06       ` [XenPPC] " Segher Boessenkool
2006-07-18  7:00 ` [RFC PATCH 29/33] Add Xen driver utility functions Chris Wright
2006-07-18  7:00 ` [RFC PATCH 30/33] Add the Xenbus sysfs and virtual device hotplug driver Chris Wright
2006-07-18  7:00   ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 31/33] Add Xen subarch reboot support Chris Wright
2006-07-20  6:16   ` Adrian Bunk
2006-07-18  7:00 ` [RFC PATCH 32/33] Add the Xen virtual network device driver Chris Wright
2006-07-18 10:27   ` Arjan van de Ven
2006-07-18 10:35     ` Keir Fraser
2006-07-18 10:35       ` Keir Fraser
2006-07-18 10:42       ` Arjan van de Ven
2006-07-18 12:18         ` Dave Boutcher
2006-07-18 12:39     ` jamal
2006-07-18 13:08       ` Herbert Xu
2006-07-18 13:08         ` Herbert Xu
2006-07-18 13:25         ` John Haller
2006-07-18 15:22           ` Herbert Xu
2006-07-18 15:22             ` Herbert Xu
2006-07-18 15:44   ` Stephen Hemminger
2006-07-19  3:55     ` Herbert Xu
2006-07-19  3:55       ` Herbert Xu
2006-07-19  3:55       ` Herbert Xu
2006-07-18 20:42   ` David Miller
2006-07-18 21:09     ` Chris Wright
2006-07-18  7:00 ` [RFC PATCH 33/33] Add Xen virtual block " Chris Wright
2006-07-18 10:34   ` Arjan van de Ven
2006-07-18 20:57     ` Jeremy Fitzhardinge
2006-07-18 13:01   ` Dave Boutcher
2006-07-18 16:25     ` Jeff Garzik
2006-07-18 16:25       ` Jeff Garzik
2006-07-18 19:28     ` Chris Wright
2006-07-18 19:28       ` Chris Wright
2006-07-18 21:22     ` Jeremy Fitzhardinge

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=44BCC720.7050601@vmware.com \
    --to=zach@vmware.com \
    --cc=Christian.Limpach@cl.cam.ac.uk \
    --cc=ak@suse.de \
    --cc=akpm@osdl.org \
    --cc=arjan@infradead.org \
    --cc=chrisw@sous-sol.org \
    --cc=ian.pratt@xensource.com \
    --cc=jeremy@goop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rusty@rustcorp.com.au \
    --cc=virtualization@lists.osdl.org \
    --cc=xen-devel@lists.xensource.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.