Xen devel - Jul 2007 - [Q]Which part of Xen is in slated for 2.6.23 kernel release?

Hi Keir, Jeremy  and all,

Is Xen really a part of official 2.6.23 kernel now?
If yes i would like to ask
 - what is accepted by Linus in the mainline kernel?
 - Is it just the hypervisor code?
 - What about dom0?

The gitweb does shows Jeremy as comitter but it looks to me as its
just the hypervisor part.

What is the status BTW?

thanks
Regards
-- 
---
pradeep singh rautela

"Genius is 1% inspiration, and 99% perspiration" - not me :)

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 21/7/07 10:50, "pradeep singh rautela" <rautelap@gmail.com>
wrote:
> Is Xen really a part of official 2.6.23 kernel now?
Yes, assuming the patches are not dropped again from Linus''s tree
before
2.6.23 is released. But that is unlikely. They''ve had lots of testing
and
review.
> If yes i would like to ask
>  - what is accepted by Linus in the mainline kernel?
>  - Is it just the hypervisor code?
>  - What about dom0?
The scope of the patches is domU guest support. Vanilla linux 2.6.23 will
not support running as dom0 (Jeremy is working on this for a future Linux
release). Nor does it contain the hypervisor code itself -- you have to
obtain the Xen hypervisor separately from the kernel sources.

 -- Keir
 
> The gitweb does shows Jeremy as comitter but it looks to me as its
> just the hypervisor part.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Hi Keir,
On 7/21/07, Keir Fraser <keir@xensource.com>
wrote:
> On 21/7/07 10:50, "pradeep singh rautela"
<rautelap@gmail.com> wrote:
>
> > Is Xen really a part of official 2.6.23 kernel now?
>
> Yes, assuming the patches are not dropped again from Linus''s tree
before
> 2.6.23 is released. But that is unlikely. They''ve had lots of
testing and
> review.
great news!!!
>
> > If yes i would like to ask
> >  - what is accepted by Linus in the mainline kernel?
> >  - Is it just the hypervisor code?
> >  - What about dom0?
>
> The scope of the patches is domU guest support. Vanilla linux 2.6.23 will
> not support running as dom0 (Jeremy is working on this for a future Linux
> release). Nor does it contain the hypervisor code itself -- you have to
> obtain the Xen hypervisor separately from the kernel sources.
Great!!!
So this means 2.6.23 vanilla kernel can be compiled as a domU natively, right?

Thanks for the info Keir.

Regards

---
pradeep singh rautela

"Genius is 1% inspiration, and 99% perspiration" - not me :)

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

At 12:21 21/07/2007, pradeep singh rautela wrote:
>Hi Keir,
>On 7/21/07, Keir Fraser <keir@xensource.com> wrote:
>>On 21/7/07 10:50, "pradeep singh rautela"
<rautelap@gmail.com> wrote:
>>
>> > Is Xen really a part of official 2.6.23 kernel now?
>>
>>Yes, assuming the patches are not dropped again from Linus''s
tree before
>>2.6.23 is released. But that is unlikely. They''ve had lots of
testing and
>>review.
>
>great news!!!
>>
>> > If yes i would like to ask
>> >  - what is accepted by Linus in the mainline kernel?
>> >  - Is it just the hypervisor code?
>> >  - What about dom0?
>>
>>The scope of the patches is domU guest support. Vanilla linux 2.6.23
will
>>not support running as dom0 (Jeremy is working on this for a future
Linux
>>release). Nor does it contain the hypervisor code itself -- you have to
>>obtain the Xen hypervisor separately from the kernel sources.
>
>Great!!!
>So this means 2.6.23 vanilla kernel can be compiled as a domU natively,
right?

I think for now, only modules can be shared, kernel itself needs to 
be compiled for DomU or Native use, but ultimately, the idea is to 
have one kernel that does both DomU and Native in the same kernel.

Here''s what Jeremy wrote on the subject:
----Jeremy----
Hm, I should prepare a documentation patch.

If you build the kernel with CONFIG_PARAVIRT=y, CONFIG_XEN=y and enable
the Xen console, block and net devices (they will default =y), then you
should be able to use the resulting kernel with an existing Xen
installation (it will also boot native).  Unfortunately you can''t use
the same actual kernel image file in both cases, but you can use the
same modules.

To boot native: boot the bzImage file as usual.

To boot under Xen, use the vmlinux file as your kernel image (you can
strip and gzip it if you wish).

The console is hvc0, so you may need to pass console=hvc0 in your Xen
config.

Similarly, the blockfront device is /dev/xvd[abc...][123...]; you''ll
probably need to pass the appropriate root= kernel command line option.
It does not support stealing the hdX or sdX names, so that part of the
Xen config will be ignored.

Netfront appears as ethX as usual.

The kernel does not support balloon or suspend/migrate/resume yet, so
that won''t work.  You should be able to cleanly shutdown/reboot the
domain via xm.

---End Jeremy---

--
Mats

>Thanks for the info Keir.
>
>Regards
>
>---
>pradeep singh rautela
>
>"Genius is 1% inspiration, and 99% perspiration" - not me :)
>
>_______________________________________________
>Xen-devel mailing list
>Xen-devel@lists.xensource.com
>http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 21/7/07 12:21, "pradeep singh rautela" <rautelap@gmail.com>
wrote:
>> The scope of the patches is domU guest support. Vanilla linux 2.6.23
will
>> not support running as dom0 (Jeremy is working on this for a future
Linux
>> release). Nor does it contain the hypervisor code itself -- you have to
>> obtain the Xen hypervisor separately from the kernel sources.
> 
> Great!!!
> So this means 2.6.23 vanilla kernel can be compiled as a domU natively,
right?
Yes. In fact the paravirt_ops framework means that you can build a single
kernel that will work natively or on Xen (or on VMware or lguest or ...).

 -- Keir


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On Sat, Jul 21, 2007 at 01:42:06PM +0100, Keir Fraser
wrote:
> 
> 
> 
> On 21/7/07 12:21, "pradeep singh rautela"
<rautelap@gmail.com> wrote:
> 
> >> The scope of the patches is domU guest support. Vanilla linux
2.6.23 will
> >> not support running as dom0 (Jeremy is working on this for a
future Linux
> >> release). Nor does it contain the hypervisor code itself -- you
have to
> >> obtain the Xen hypervisor separately from the kernel sources.
> > 
> > Great!!!
> > So this means 2.6.23 vanilla kernel can be compiled as a domU
natively, right?
> 
> Yes. In fact the paravirt_ops framework means that you can build a single
> kernel that will work natively or on Xen (or on VMware or lguest or ...).
I thought that with this initial merge, we stll had to build separate 
images for bare metal vs Xen paravirt because of the way Linux uses 
the bzImage, while Xen is in ELF format ? Or has this bootup difference
already been resolved ?


Dan.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=| 

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

At 19:24 21/07/2007, Daniel P. Berrange wrote:
>On Sat, Jul 21, 2007 at 01:42:06PM +0100, Keir Fraser wrote:
> >
> >
> >
> > On 21/7/07 12:21, "pradeep singh rautela"
<rautelap@gmail.com> wrote:
> >
> > >> The scope of the patches is domU guest support. Vanilla linux
> 2.6.23 will
> > >> not support running as dom0 (Jeremy is working on this for a 
> future Linux
> > >> release). Nor does it contain the hypervisor code itself --
you have to
> > >> obtain the Xen hypervisor separately from the kernel sources.
> > >
> > > Great!!!
> > > So this means 2.6.23 vanilla kernel can be compiled as a domU 
> natively, right?
> >
> > Yes. In fact the paravirt_ops framework means that you can build a
single
> > kernel that will work natively or on Xen (or on VMware or lguest or
...).
>
>I thought that with this initial merge, we stll had to build separate
>images for bare metal vs Xen paravirt because of the way Linux uses
>the bzImage, while Xen is in ELF format ? Or has this bootup difference
>already been resolved ?

I think that is correct. Work is (from what Jeremy said) in progress 
to make a "compatible format" for both uses.

--
Mats


>Dan.
>--
>|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496
-=|
>|=-           Perl modules: http://search.cpan.org/~danberr/             
-=|
>|=-               Projects: http://freshmeat.net/~danielpb/              
-=|
>|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 
-=|
>
>_______________________________________________
>Xen-devel mailing list
>Xen-devel@lists.xensource.com
>http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 21/7/07 19:24, "Daniel P. Berrange" <berrange@redhat.com>
wrote:
>> Yes. In fact the paravirt_ops framework means that you can build a
single
>> kernel that will work natively or on Xen (or on VMware or lguest or
...).
> 
> I thought that with this initial merge, we stll had to build separate
> images for bare metal vs Xen paravirt because of the way Linux uses
> the bzImage, while Xen is in ELF format ? Or has this bootup difference
> already been resolved ?
Actually yes, it''s still in progress. So currently it is true that you
can
target native or Xen from a single *build*, but you then take different
image files (vmlinux vs bzImage).

 -- Keir


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

> >I thought that with this initial merge, we stll had to build separate
> >images for bare metal vs Xen paravirt because of the way Linux uses
> >the bzImage, while Xen is in ELF format ? Or has this bootup difference
> >already been resolved ?
>
> I think that is correct. Work is (from what Jeremy said) in progress
> to make a "compatible format" for both uses.
Yep...  But the same build produces both images (the vmlinux and the bzImage) 
anyhow, so you don''t actually have to build the core kernel twice or
anything
like that.  It''s just a bit annoying having to manage two different
types of
image files - the work underway is to make a bzImage that is native bootable 
and loadable by the Xen domain builder.

cheers,
Mark
> --
> Mats
>
> >Dan.
> >--
> >
> >|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392
2496
> >| -=| =-           Perl modules: http://search.cpan.org/~danberr/
> >|     -=| =-               Projects: http://freshmeat.net/~danielpb/
> >|         -=| =-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF
> >| F742 7D3B 9505  -=|
> >
> >_______________________________________________
> >Xen-devel mailing list
> >Xen-devel@lists.xensource.com
> >http://lists.xensource.com/xen-devel
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
a

-- 
Dave: Just a question. What use is a unicyle with no seat?  And no pedals!
Mark: To answer a question with a question: What use is a skateboard?
Dave: Skateboards have wheels.
Mark: My wheel has a wheel!

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Mats Petersson wrote:
> At 12:21 21/07/2007, pradeep singh rautela wrote:
>> Hi Keir,
>> On 7/21/07, Keir Fraser <keir@xensource.com> wrote:
>>> On 21/7/07 10:50, "pradeep singh rautela"
<rautelap@gmail.com> wrote:
>>>
>>> > Is Xen really a part of official 2.6.23 kernel now?
>>>
>>> Yes, assuming the patches are not dropped again from
Linus''s tree
>>> before
>>> 2.6.23 is released. But that is unlikely. They''ve had lots
of
>>> testing and
>>> review.
>>
>> great news!!!
>>>
>>> > If yes i would like to ask
>>> >  - what is accepted by Linus in the mainline kernel?
>>> >  - Is it just the hypervisor code?
>>> >  - What about dom0?
>>>
>>> The scope of the patches is domU guest support. Vanilla linux
2.6.23
>>> will
>>> not support running as dom0 (Jeremy is working on this for a future
>>> Linux
>>> release). Nor does it contain the hypervisor code itself -- you
have to
>>> obtain the Xen hypervisor separately from the kernel sources.
>>
>> Great!!!
>> So this means 2.6.23 vanilla kernel can be compiled as a domU
>> natively, right?
>
>
> I think for now, only modules can be shared, kernel itself needs to be
> compiled for DomU or Native use, but ultimately, the idea is to have
> one kernel that does both DomU and Native in the same kernel.
To clarify, you build the kernel once, and two build products are
vmlinux and bzImage.  If you have a bootloader which can handle it, you
can boot vmlinux natively, or under Xen.  Xen doesn''t currently support
booting bzImage, so its really only usable for native execution.  But
it''s just a matter of packaging; the kernel code and data is
bit-for-bit
the same in both kernel images.

    J

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Luiz Vitor Martinez Cardoso wrote:
> Hi,
>
> Congratulations for the work! But i have a question. When xen gets stable
at kernel.org how we can configure xen options? Today i can edit Config.mk
before build my own kernel, how it will be done?
>   
Hm, not quite sure what you''re asking.  Xen itself is the same as it
always has been.  The kernel.org kernel has been Xenified for use as a
domU guest.  There''s no build dependency between them.

    J

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Hi,

Congratulations for the work! But i have a question. When xen gets stable at
kernel.org how we can configure xen options? Today i can edit Config.mk before
build my own kernel, how it will be done?

Regards,
Luiz Vitor Martinez Cardoso aka Grabber



Em Tue, 24 Jul 2007 22:16:16 -0700
Jeremy Fitzhardinge <jeremy@goop.org> escreveu:
> Mats Petersson wrote:
> > At 12:21 21/07/2007, pradeep singh rautela wrote:
> >> Hi Keir,
> >> On 7/21/07, Keir Fraser <keir@xensource.com> wrote:
> >>> On 21/7/07 10:50, "pradeep singh rautela"
<rautelap@gmail.com> wrote:
> >>>
> >>> > Is Xen really a part of official 2.6.23 kernel now?
> >>>
> >>> Yes, assuming the patches are not dropped again from
Linus''s tree
> >>> before
> >>> 2.6.23 is released. But that is unlikely. They''ve had
lots of
> >>> testing and
> >>> review.
> >>
> >> great news!!!
> >>>
> >>> > If yes i would like to ask
> >>> >  - what is accepted by Linus in the mainline kernel?
> >>> >  - Is it just the hypervisor code?
> >>> >  - What about dom0?
> >>>
> >>> The scope of the patches is domU guest support. Vanilla linux
2.6.23
> >>> will
> >>> not support running as dom0 (Jeremy is working on this for a
future
> >>> Linux
> >>> release). Nor does it contain the hypervisor code itself --
you have to
> >>> obtain the Xen hypervisor separately from the kernel sources.
> >>
> >> Great!!!
> >> So this means 2.6.23 vanilla kernel can be compiled as a domU
> >> natively, right?
> >
> >
> > I think for now, only modules can be shared, kernel itself needs to be
> > compiled for DomU or Native use, but ultimately, the idea is to have
> > one kernel that does both DomU and Native in the same kernel.
> 
> To clarify, you build the kernel once, and two build products are
> vmlinux and bzImage.  If you have a bootloader which can handle it, you
> can boot vmlinux natively, or under Xen.  Xen doesn''t currently
support
> booting bzImage, so its really only usable for native execution.  But
> it''s just a matter of packaging; the kernel code and data is
bit-for-bit
> the same in both kernel images.
> 
>     J
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel