[PATCH] allow xenfb initialization for hvm guests

[PATCH] allow xenfb initialization for hvm guests

[Stefano Stabellini]

There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
Given that no Xen toolstacks configure a xenfb backend for x86 HVM
guests, they are not affected.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

diff --git a/drivers/video/xen-fbfront.c b/drivers/video/xen-fbfront.c
index cd005c2..02e1c01 100644
--- a/drivers/video/xen-fbfront.c
+++ b/drivers/video/xen-fbfront.c
@@ -692,7 +692,7 @@ static DEFINE_XENBUS_DRIVER(xenfb, ,
 
 static int __init xenfb_init(void)
 {
-	if (!xen_pv_domain())
+	if (!xen_domain())
 		return -ENODEV;
 
 	/* Nothing to do if running in dom0. */

Re: [Xen-devel] [PATCH] allow xenfb initialization for hvm guests

[Ian Campbell]

On Tue, 2013-12-17 at 17:53 +0000, Stefano Stabellini wrote:
> There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> guests, they are not affected.

and if a toolstack did I think it would be reasonable to expect the
kernel to at least try and drive it!

> Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Acked-by: Ian Campbell <ian.campbell@citrix.com>

Re: [Xen-devel] [PATCH] allow xenfb initialization for hvm guests

[Konrad Rzeszutek Wilk]

On Wed, Dec 18, 2013 at 11:51:22AM +0000, Ian Campbell wrote:
> On Tue, 2013-12-17 at 17:53 +0000, Stefano Stabellini wrote:
> > There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> > As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> > Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> > guests, they are not affected.
> 
> and if a toolstack did I think it would be reasonable to expect the
> kernel to at least try and drive it!
> 
> > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> 
> Acked-by: Ian Campbell <ian.campbell@citrix.com>

I think David Vrabel is going to Nack it unless there is a test-case.

I am saying that because in his previous statement for Wei's multi-page
patches was that if there are no users (or at least no test-cases) then
it should not be part of the kernel.

Re: [Xen-devel] [PATCH] allow xenfb initialization for hvm guests

[Stefano Stabellini]

On Wed, 18 Dec 2013, Konrad Rzeszutek Wilk wrote:
> On Wed, Dec 18, 2013 at 11:51:22AM +0000, Ian Campbell wrote:
> > On Tue, 2013-12-17 at 17:53 +0000, Stefano Stabellini wrote:
> > > There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> > > As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> > > Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> > > guests, they are not affected.
> > 
> > and if a toolstack did I think it would be reasonable to expect the
> > kernel to at least try and drive it!
> > 
> > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> > 
> > Acked-by: Ian Campbell <ian.campbell@citrix.com>
> 
> I think David Vrabel is going to Nack it unless there is a test-case.
> 
> I am saying that because in his previous statement for Wei's multi-page
> patches was that if there are no users (or at least no test-cases) then
> it should not be part of the kernel.

Well there is one: ARM. There isn't one on x86.

http://marc.info/?l=xen-devel&m=138730150806639&w=2

Re: [Xen-devel] [PATCH] allow xenfb initialization for hvm guests

[Konrad Rzeszutek Wilk]

On Wed, Dec 18, 2013 at 03:18:45PM +0000, Stefano Stabellini wrote:
> On Wed, 18 Dec 2013, Konrad Rzeszutek Wilk wrote:
> > On Wed, Dec 18, 2013 at 11:51:22AM +0000, Ian Campbell wrote:
> > > On Tue, 2013-12-17 at 17:53 +0000, Stefano Stabellini wrote:
> > > > There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> > > > As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> > > > Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> > > > guests, they are not affected.
> > > 
> > > and if a toolstack did I think it would be reasonable to expect the
> > > kernel to at least try and drive it!
> > > 
> > > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> > > 
> > > Acked-by: Ian Campbell <ian.campbell@citrix.com>
> > 
> > I think David Vrabel is going to Nack it unless there is a test-case.
> > 
> > I am saying that because in his previous statement for Wei's multi-page
> > patches was that if there are no users (or at least no test-cases) then
> > it should not be part of the kernel.
> 
> Well there is one: ARM. There isn't one on x86.
> 
> http://marc.info/?l=xen-devel&m=138730150806639&w=2

Ah I missed that. I think that this particular patch can wait till the
next merge window?

Would it make sense to have the XenFB also work on x86? Is it difficult
to make it work on it?

Re: [Xen-devel] [PATCH] allow xenfb initialization for hvm guests

[Andrew Cooper]

On 18/12/13 15:14, Konrad Rzeszutek Wilk wrote:
> On Wed, Dec 18, 2013 at 11:51:22AM +0000, Ian Campbell wrote:
>> On Tue, 2013-12-17 at 17:53 +0000, Stefano Stabellini wrote:
>>> There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
>>> As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
>>> Given that no Xen toolstacks configure a xenfb backend for x86 HVM
>>> guests, they are not affected.
>> and if a toolstack did I think it would be reasonable to expect the
>> kernel to at least try and drive it!
>>
>>> Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
>> Acked-by: Ian Campbell <ian.campbell@citrix.com>
> I think David Vrabel is going to Nack it unless there is a test-case.
>
> I am saying that because in his previous statement for Wei's multi-page
> patches was that if there are no users (or at least no test-cases) then
> it should not be part of the kernel.

Without trying to presume too much, I really don't think he would.

This patch is taking an existing thing and permitting it to work in more
cases (which is a good change IMO).

The multi-page rings was completely brand new functionality with no
consumers at all, so no ability to verify the implementation.  It is a
completely different context.

~Andrew

Re: [Xen-devel] [PATCH] allow xenfb initialization for hvm guests

[Stefano Stabellini]

On Wed, 18 Dec 2013, Konrad Rzeszutek Wilk wrote:
> On Wed, Dec 18, 2013 at 03:18:45PM +0000, Stefano Stabellini wrote:
> > On Wed, 18 Dec 2013, Konrad Rzeszutek Wilk wrote:
> > > On Wed, Dec 18, 2013 at 11:51:22AM +0000, Ian Campbell wrote:
> > > > On Tue, 2013-12-17 at 17:53 +0000, Stefano Stabellini wrote:
> > > > > There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> > > > > As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> > > > > Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> > > > > guests, they are not affected.
> > > > 
> > > > and if a toolstack did I think it would be reasonable to expect the
> > > > kernel to at least try and drive it!
> > > > 
> > > > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> > > > 
> > > > Acked-by: Ian Campbell <ian.campbell@citrix.com>
> > > 
> > > I think David Vrabel is going to Nack it unless there is a test-case.
> > > 
> > > I am saying that because in his previous statement for Wei's multi-page
> > > patches was that if there are no users (or at least no test-cases) then
> > > it should not be part of the kernel.
> > 
> > Well there is one: ARM. There isn't one on x86.
> > 
> > http://marc.info/?l=xen-devel&m=138730150806639&w=2
> 
> Ah I missed that. I think that this particular patch can wait till the
> next merge window?

Surely it can wait.


> Would it make sense to have the XenFB also work on x86? Is it difficult
> to make it work on it?

The issue is that you need stdvga anyway in order to boot and xenfb is
not much better than stdvga after all (except that it doesn't require
QEMU's pc machine emulator).

Re: [PATCH] allow xenfb initialization for hvm guests

[Konrad Rzeszutek Wilk]

On Tue, Dec 17, 2013 at 05:53:13PM +0000, Stefano Stabellini wrote:
> There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> guests, they are not affected.

Could you reference the upstream git commit that enables this.

And also CC the maintainers of drivers/video/*

And lastly lets CC also David and Boris on it.

> 
> Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> 
> diff --git a/drivers/video/xen-fbfront.c b/drivers/video/xen-fbfront.c
> index cd005c2..02e1c01 100644
> --- a/drivers/video/xen-fbfront.c
> +++ b/drivers/video/xen-fbfront.c
> @@ -692,7 +692,7 @@ static DEFINE_XENBUS_DRIVER(xenfb, ,
>  
>  static int __init xenfb_init(void)
>  {
> -	if (!xen_pv_domain())
> +	if (!xen_domain())
>  		return -ENODEV;
>  
>  	/* Nothing to do if running in dom0. */

Re: [PATCH] allow xenfb initialization for hvm guests

[David Vrabel]

On 31/12/13 14:50, Konrad Rzeszutek Wilk wrote:
> On Tue, Dec 17, 2013 at 05:53:13PM +0000, Stefano Stabellini wrote:
>> There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
>> As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
>> Given that no Xen toolstacks configure a xenfb backend for x86 HVM
>> guests, they are not affected.
> 
> Could you reference the upstream git commit that enables this.
> 
> And also CC the maintainers of drivers/video/*
> 
> And lastly lets CC also David and Boris on it.
> 
>>
>> Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Acked-by: David Vrabel <david.vrabel@citrix.com>

David

Re: [PATCH] allow xenfb initialization for hvm guests

[Stefano Stabellini]

On Tue, 31 Dec 2013, Konrad Rzeszutek Wilk wrote:
> On Tue, Dec 17, 2013 at 05:53:13PM +0000, Stefano Stabellini wrote:
> > There is no reasons why an HVM guest shouldn't be allowed to use xenfb.
> > As a matter of fact ARM guests, HVM from Linux POV, can use xenfb.
> > Given that no Xen toolstacks configure a xenfb backend for x86 HVM
> > guests, they are not affected.
> 
> Could you reference the upstream git commit that enables this.

The toolstack can already setup a xenfb frontend/backend pair for ARM
guests. However QEMU xenpv machine still needs few unapplied fixes to be
built on ARM:

http://marc.info/?l=qemu-devel&m=138739419700837&w=2


> And also CC the maintainers of drivers/video/*
>
> And lastly lets CC also David and Boris on it.

OK

> > 
> > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> > 
> > diff --git a/drivers/video/xen-fbfront.c b/drivers/video/xen-fbfront.c
> > index cd005c2..02e1c01 100644
> > --- a/drivers/video/xen-fbfront.c
> > +++ b/drivers/video/xen-fbfront.c
> > @@ -692,7 +692,7 @@ static DEFINE_XENBUS_DRIVER(xenfb, ,
> >  
> >  static int __init xenfb_init(void)
> >  {
> > -	if (!xen_pv_domain())
> > +	if (!xen_domain())
> >  		return -ENODEV;
> >  
> >  	/* Nothing to do if running in dom0. */
>