Current git for me (b1d0e4f535e10775cffde922208b49629169aeaa) doesn't
let me build KVM. In fact, I don't seem to have *ANY* kvm-related
symbols in my .config at all.
I've verified that arch/x86/kvm/Kconfig was getting run by putting some
comments in it. It seems to me like just putting:
config HAVE_KVM
bool
isn't letting anything come into my .config at all. I'm not sure why we
do that, and then have:
select HAVE_KVM
in arch/x86/Kconfig.
This patch just defines HAVE_KVM in the x86 Kconfig and is done with it.
Seems to work for me.
Was there some reason that it was done this way? Was it ever tested?
-- Dave
Signed-off-by: Dave Hansen <[email protected]>
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 65a70b7..cbbf35d 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -115,7 +115,8 @@ config ARCH_HAS_CPU_RELAX
config HAVE_SETUP_PER_CPU_AREA
def_bool X86_64
-select HAVE_KVM
+config HAVE_KVM
+ def_bool y
config ARCH_HIBERNATION_POSSIBLE
def_bool y
diff --git a/arch/x86/kvm/Kconfig b/arch/x86/kvm/Kconfig
index 41962e7..f2177ec 100644
--- a/arch/x86/kvm/Kconfig
+++ b/arch/x86/kvm/Kconfig
@@ -1,8 +1,6 @@
#
# KVM configuration
#
-config HAVE_KVM
- bool
menuconfig VIRTUALIZATION
bool "Virtualization"
* Dave Hansen <[email protected]> wrote:
> config HAVE_SETUP_PER_CPU_AREA
> def_bool X86_64
>
> -select HAVE_KVM
> +config HAVE_KVM
> + def_bool y
i think it might be better to just move the select to after the "config
X86" option. (this is a legitimate use of select) Sam, do you concur?
(but, your current fix should go in ASAP, so no objections from me.)
Ingo
On Sat, Feb 09, 2008 at 10:24:01PM +0100, Ingo Molnar wrote:
>
> * Dave Hansen <[email protected]> wrote:
>
> > config HAVE_SETUP_PER_CPU_AREA
> > def_bool X86_64
> >
> > -select HAVE_KVM
> > +config HAVE_KVM
> > + def_bool y
>
> i think it might be better to just move the select to after the "config
> X86" option. (this is a legitimate use of select) Sam, do you concur?
Correct.
We should see it used like this:
config X86
select HAVE_KVM
Sam
Hi Dave.
>
> I've verified that arch/x86/kvm/Kconfig was getting run by putting some
> comments in it. It seems to me like just putting:
>
> config HAVE_KVM
> bool
>
> isn't letting anything come into my .config at all. I'm not sure why we
> do that, and then have:
>
> select HAVE_KVM
>
> in arch/x86/Kconfig.
See rationale in Documentation/kbuild/kconfig-language.txt - last chapter.
In this case the select was just placed wrongly.
Sam
On Feb 10, 2008 2:29 AM, Dave Hansen <[email protected]> wrote:
> Current git for me (b1d0e4f535e10775cffde922208b49629169aeaa) doesn't
> let me build KVM. In fact, I don't seem to have *ANY* kvm-related
> symbols in my .config at all.
>
> I've verified that arch/x86/kvm/Kconfig was getting run by putting some
> comments in it. It seems to me like just putting:
>
> config HAVE_KVM
> bool
>
> isn't letting anything come into my .config at all. I'm not sure why we
> do that, and then have:
>
> select HAVE_KVM
>
> in arch/x86/Kconfig.
>
> This patch just defines HAVE_KVM in the x86 Kconfig and is done with it.
> Seems to work for me.
>
> Was there some reason that it was done this way? Was it ever tested?
>
> -- Dave
>
> Signed-off-by: Dave Hansen <[email protected]>
>
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index 65a70b7..cbbf35d 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -115,7 +115,8 @@ config ARCH_HAS_CPU_RELAX
> config HAVE_SETUP_PER_CPU_AREA
> def_bool X86_64
>
> -select HAVE_KVM
> +config HAVE_KVM
> + def_bool y
>
> config ARCH_HIBERNATION_POSSIBLE
> def_bool y
> diff --git a/arch/x86/kvm/Kconfig b/arch/x86/kvm/Kconfig
> index 41962e7..f2177ec 100644
> --- a/arch/x86/kvm/Kconfig
> +++ b/arch/x86/kvm/Kconfig
> @@ -1,8 +1,6 @@
> #
> # KVM configuration
> #
> -config HAVE_KVM
> - bool
>
> menuconfig VIRTUALIZATION
> bool "Virtualization"
>
Hi, Dave,
I did not see your email, until just now. I've posted a patch at
http://lkml.org/lkml/2008/2/10/26. Avi has acked it
Balbir