Xen devel - Jul 2013 - [PATCH v3] xen/arm: enable PV control for ARM

Enable power management from the toolstack for ARM guest.

Signed-off-by: Julien Grall <julien.grall@linaro.org>

---
    Changes in v3:
        - Fix condition to compile cpu_hotplug.o
    Changes in v2:
        - Don''t compile xen/cpu_hotplug.o with ARM64
---
 drivers/xen/Makefile |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
index 2bf461a..f185e8d 100644
--- a/drivers/xen/Makefile
+++ b/drivers/xen/Makefile
@@ -1,9 +1,8 @@
-ifneq ($(CONFIG_ARM),y)
-obj-y	+= manage.o
+ifeq ($(filter y, $(CONFIG_ARM) $(CONFIG_ARM64)),)
 obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
 endif
 obj-$(CONFIG_X86)			+= fallback.o
-obj-y	+= grant-table.o features.o events.o balloon.o time.o
+obj-y	+= grant-table.o features.o events.o balloon.o time.o manage.o
 obj-y	+= xenbus/
 
 nostackp := $(call cc-option, -fno-stack-protector)
-- 
1.7.10.4

Hi,

[Adding Catalin to Cc]

On Mon, Jul 22, 2013 at 11:57:34AM +0100, Julien Grall
wrote:
> Enable power management from the toolstack for ARM guest.
> 
> Signed-off-by: Julien Grall <julien.grall@linaro.org>
> 
> ---
>     Changes in v3:
>         - Fix condition to compile cpu_hotplug.o
>     Changes in v2:
>         - Don''t compile xen/cpu_hotplug.o with ARM64
My arm64 cpu hotplug series is going to be dependent on not compiling
xen/cpu_hotplug.o. Currently I''m not able to compile an arm64 kernel
with my patches and Xen support.

How do you intend to push this? I''d like to ensure my cpu hotplug
series
is able to go in for the next merge window.
> ---
>  drivers/xen/Makefile |    5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> index 2bf461a..f185e8d 100644
> --- a/drivers/xen/Makefile
> +++ b/drivers/xen/Makefile
> @@ -1,9 +1,8 @@
> -ifneq ($(CONFIG_ARM),y)
> -obj-y	+= manage.o
> +ifeq ($(filter y, $(CONFIG_ARM) $(CONFIG_ARM64)),)
>  obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
>  endif
Might it make sense to enable this based on a list the architectures
that actually need this, rather than disabling it based on the list of
architectures that don''t?

Thanks,
Mark.
>  obj-$(CONFIG_X86)			+= fallback.o
> -obj-y	+= grant-table.o features.o events.o balloon.o time.o
> +obj-y	+= grant-table.o features.o events.o balloon.o time.o manage.o
>  obj-y	+= xenbus/
>  
>  nostackp := $(call cc-option, -fno-stack-protector)
> -- 
> 1.7.10.4
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>

On Mon, 22 Jul 2013, Julien Grall wrote:
> Enable power management from the toolstack for ARM guest.
> 
> Signed-off-by: Julien Grall <julien.grall@linaro.org>
> 
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

>     Changes in v3:
>         - Fix condition to compile cpu_hotplug.o
>     Changes in v2:
>         - Don''t compile xen/cpu_hotplug.o with ARM64
> ---
>  drivers/xen/Makefile |    5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> index 2bf461a..f185e8d 100644
> --- a/drivers/xen/Makefile
> +++ b/drivers/xen/Makefile
> @@ -1,9 +1,8 @@
> -ifneq ($(CONFIG_ARM),y)
> -obj-y	+= manage.o
> +ifeq ($(filter y, $(CONFIG_ARM) $(CONFIG_ARM64)),)
>  obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
>  endif
>  obj-$(CONFIG_X86)			+= fallback.o
> -obj-y	+= grant-table.o features.o events.o balloon.o time.o
> +obj-y	+= grant-table.o features.o events.o balloon.o time.o manage.o
>  obj-y	+= xenbus/
>  
>  nostackp := $(call cc-option, -fno-stack-protector)
> -- 
> 1.7.10.4
>

On Mon, 22 Jul 2013, Mark Rutland wrote:
> Hi,
> 
> [Adding Catalin to Cc]
> 
> On Mon, Jul 22, 2013 at 11:57:34AM +0100, Julien Grall wrote:
> > Enable power management from the toolstack for ARM guest.
> > 
> > Signed-off-by: Julien Grall <julien.grall@linaro.org>
> > 
> > ---
> >     Changes in v3:
> >         - Fix condition to compile cpu_hotplug.o
> >     Changes in v2:
> >         - Don''t compile xen/cpu_hotplug.o with ARM64
> 
> My arm64 cpu hotplug series is going to be dependent on not compiling
> xen/cpu_hotplug.o. Currently I''m not able to compile an arm64
kernel
> with my patches and Xen support.
> 
> How do you intend to push this? I''d like to ensure my cpu hotplug
series
> is able to go in for the next merge window.
This patch does exactly that: it makes sure cpu_hotplug is not compiled
on ARM64.
Given that this patch is for drivers/xen/Makefile, normally it goes via
Konrad''s tree.

> > ---
> >  drivers/xen/Makefile |    5 ++---
> >  1 file changed, 2 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> > index 2bf461a..f185e8d 100644
> > --- a/drivers/xen/Makefile
> > +++ b/drivers/xen/Makefile
> > @@ -1,9 +1,8 @@
> > -ifneq ($(CONFIG_ARM),y)
> > -obj-y	+= manage.o
> > +ifeq ($(filter y, $(CONFIG_ARM) $(CONFIG_ARM64)),)
> >  obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
> >  endif
> 
> Might it make sense to enable this based on a list the architectures
> that actually need this, rather than disabling it based on the list of
> architectures that don''t?
This is more philosophical than practical, but I would expect this
limitation to go away soon, and be able to compile cpu_hotplug without
issues on ARM and ARM64 too.

On Mon, Jul 22, 2013 at 03:09:35PM +0100, Stefano Stabellini
wrote:
> On Mon, 22 Jul 2013, Mark Rutland wrote:
> > Hi,
> > 
> > [Adding Catalin to Cc]
> > 
> > On Mon, Jul 22, 2013 at 11:57:34AM +0100, Julien Grall wrote:
> > > Enable power management from the toolstack for ARM guest.
> > > 
> > > Signed-off-by: Julien Grall <julien.grall@linaro.org>
> > > 
> > > ---
> > >     Changes in v3:
> > >         - Fix condition to compile cpu_hotplug.o
> > >     Changes in v2:
> > >         - Don''t compile xen/cpu_hotplug.o with ARM64
> > 
> > My arm64 cpu hotplug series is going to be dependent on not compiling
> > xen/cpu_hotplug.o. Currently I''m not able to compile an arm64
kernel
> > with my patches and Xen support.
> > 
> > How do you intend to push this? I''d like to ensure my cpu
hotplug series
> > is able to go in for the next merge window.
> 
> This patch does exactly that: it makes sure cpu_hotplug is not compiled
> on ARM64.
Sorry, what I meant was that my series is going to be entirely dependent
on this patch or it''ll break bisection. That''ll make merging
things a
little painful as I''ll need a stable branch with this patch to base my
series upon.

Another option would be to split this patch into two parts, one enabling
power management and one fixing things up for HOTPLUG_CPU on arm64. That
way I could take the arm64 HOTPLUG_CPU fix in my series adding arm64
HOTPLUG_CPU support.

Would that be acceptable? Have I missed something?
> Given that this patch is for drivers/xen/Makefile, normally it goes via
> Konrad''s tree.
> 
> 
> > > ---
> > >  drivers/xen/Makefile |    5 ++---
> > >  1 file changed, 2 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> > > index 2bf461a..f185e8d 100644
> > > --- a/drivers/xen/Makefile
> > > +++ b/drivers/xen/Makefile
> > > @@ -1,9 +1,8 @@
> > > -ifneq ($(CONFIG_ARM),y)
> > > -obj-y	+= manage.o
> > > +ifeq ($(filter y, $(CONFIG_ARM) $(CONFIG_ARM64)),)
> > >  obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
> > >  endif
> > 
> > Might it make sense to enable this based on a list the architectures
> > that actually need this, rather than disabling it based on the list of
> > architectures that don''t?
> 
> This is more philosophical than practical, but I would expect this
> limitation to go away soon, and be able to compile cpu_hotplug without
> issues on ARM and ARM64 too.
> 
Ok, that''s fine then.

Thanks,
Mark.

On Mon, 22 Jul 2013, Mark Rutland wrote:
> On Mon, Jul 22, 2013 at 03:09:35PM +0100, Stefano Stabellini wrote:
> > On Mon, 22 Jul 2013, Mark Rutland wrote:
> > > Hi,
> > > 
> > > [Adding Catalin to Cc]
> > > 
> > > On Mon, Jul 22, 2013 at 11:57:34AM +0100, Julien Grall wrote:
> > > > Enable power management from the toolstack for ARM guest.
> > > > 
> > > > Signed-off-by: Julien Grall <julien.grall@linaro.org>
> > > > 
> > > > ---
> > > >     Changes in v3:
> > > >         - Fix condition to compile cpu_hotplug.o
> > > >     Changes in v2:
> > > >         - Don''t compile xen/cpu_hotplug.o with
ARM64
> > > 
> > > My arm64 cpu hotplug series is going to be dependent on not
compiling
> > > xen/cpu_hotplug.o. Currently I''m not able to compile an
arm64 kernel
> > > with my patches and Xen support.
> > > 
> > > How do you intend to push this? I''d like to ensure my
cpu hotplug series
> > > is able to go in for the next merge window.
> > 
> > This patch does exactly that: it makes sure cpu_hotplug is not
compiled
> > on ARM64.
> 
> Sorry, what I meant was that my series is going to be entirely dependent
> on this patch or it''ll break bisection. That''ll make
merging things a
> little painful as I''ll need a stable branch with this patch to
base my
> series upon.
> 
> Another option would be to split this patch into two parts, one enabling
> power management and one fixing things up for HOTPLUG_CPU on arm64. That
> way I could take the arm64 HOTPLUG_CPU fix in my series adding arm64
> HOTPLUG_CPU support.
> 
> Would that be acceptable? Have I missed something?
As I am very sensitive to painful merging situations, I think we could
split the patch in two and send upstream the fix (not compiling
cpu_hotplug.o on ARM64) in 3.11-rc2 or rc3.  Konrad, what do you think?

On Mon, Jul 22, 2013 at 03:46:34PM +0100, Stefano Stabellini
wrote:
> On Mon, 22 Jul 2013, Mark Rutland wrote:
> > On Mon, Jul 22, 2013 at 03:09:35PM +0100, Stefano Stabellini wrote:
> > > On Mon, 22 Jul 2013, Mark Rutland wrote:
> > > > Hi,
> > > > 
> > > > [Adding Catalin to Cc]
> > > > 
> > > > On Mon, Jul 22, 2013 at 11:57:34AM +0100, Julien Grall
wrote:
> > > > > Enable power management from the toolstack for ARM
guest.
> > > > > 
> > > > > Signed-off-by: Julien Grall
<julien.grall@linaro.org>
> > > > > 
> > > > > ---
> > > > >     Changes in v3:
> > > > >         - Fix condition to compile cpu_hotplug.o
> > > > >     Changes in v2:
> > > > >         - Don''t compile xen/cpu_hotplug.o with
ARM64
> > > > 
> > > > My arm64 cpu hotplug series is going to be dependent on not
compiling
> > > > xen/cpu_hotplug.o. Currently I''m not able to
compile an arm64 kernel
> > > > with my patches and Xen support.
> > > > 
> > > > How do you intend to push this? I''d like to ensure
my cpu hotplug series
> > > > is able to go in for the next merge window.
> > > 
> > > This patch does exactly that: it makes sure cpu_hotplug is not
compiled
> > > on ARM64.
> > 
> > Sorry, what I meant was that my series is going to be entirely
dependent
> > on this patch or it''ll break bisection. That''ll make
merging things a
> > little painful as I''ll need a stable branch with this patch
to base my
> > series upon.
> > 
> > Another option would be to split this patch into two parts, one
enabling
> > power management and one fixing things up for HOTPLUG_CPU on arm64.
That
> > way I could take the arm64 HOTPLUG_CPU fix in my series adding arm64
> > HOTPLUG_CPU support.
> > 
> > Would that be acceptable? Have I missed something?
> 
> As I am very sensitive to painful merging situations, I think we could
> split the patch in two and send upstream the fix (not compiling
> cpu_hotplug.o on ARM64) in 3.11-rc2 or rc3.  Konrad, what do you think?
Sure, prep a patch and send it my way.