[PATCH] x86/vdso: Fix build when CONFIG_KVM_GUEST is not set

[PATCH] x86/vdso: Fix build when CONFIG_KVM_GUEST is not set

[Borislav Petkov]

From: Borislav Petkov <bp@suse.de>

Doing randbuilds fails like this here at link time:

  arch/x86/built-in.o: In function `arch_setup_additional_pages':
  /home/boris/kernel/linux/arch/x86/entry/vdso/vma.c:174: undefined reference to `pvclock_pvti_cpu0_va'
  make: *** [vmlinux] Error 1

This is because pvclock_pvti_cpu0_va() is defined in
arch/x86/kernel/kvmclock.c as an accessor to the private hv_clock thing
but if CONFIG_KVM_GUEST is not set, we don't have that function to link
to.

So add CONFIG_KVM_GUEST to the ifdeffery.

Signed-off-by: Borislav Petkov <bp@suse.de>
---
 arch/x86/include/asm/pvclock.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/include/asm/pvclock.h b/arch/x86/include/asm/pvclock.h
index 66df22b2e0c9..1af6bee77fb0 100644
--- a/arch/x86/include/asm/pvclock.h
+++ b/arch/x86/include/asm/pvclock.h
@@ -4,7 +4,7 @@
 #include <linux/clocksource.h>
 #include <asm/pvclock-abi.h>
 
-#ifdef CONFIG_PARAVIRT_CLOCK
+#if defined(CONFIG_PARAVIRT_CLOCK) && defined(CONFIG_KVM_GUEST)
 extern struct pvclock_vsyscall_time_info *pvclock_pvti_cpu0_va(void);
 #else
 static inline struct pvclock_vsyscall_time_info *pvclock_pvti_cpu0_va(void)
-- 
2.3.5

Re: [PATCH] x86/vdso: Fix build when CONFIG_KVM_GUEST is not set

[Borislav Petkov]

On Tue, Dec 22, 2015 at 07:50:51PM +0100, Borislav Petkov wrote:
> From: Borislav Petkov <bp@suse.de>
> 
> Doing randbuilds fails like this here at link time:
> 
>   arch/x86/built-in.o: In function `arch_setup_additional_pages':
>   /home/boris/kernel/linux/arch/x86/entry/vdso/vma.c:174: undefined reference to `pvclock_pvti_cpu0_va'
>   make: *** [vmlinux] Error 1
> 
> This is because pvclock_pvti_cpu0_va() is defined in
> arch/x86/kernel/kvmclock.c as an accessor to the private hv_clock thing
> but if CONFIG_KVM_GUEST is not set, we don't have that function to link
> to.
> 
> So add CONFIG_KVM_GUEST to the ifdeffery.
> 
> Signed-off-by: Borislav Petkov <bp@suse.de>
> ---
>  arch/x86/include/asm/pvclock.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/include/asm/pvclock.h b/arch/x86/include/asm/pvclock.h
> index 66df22b2e0c9..1af6bee77fb0 100644
> --- a/arch/x86/include/asm/pvclock.h
> +++ b/arch/x86/include/asm/pvclock.h
> @@ -4,7 +4,7 @@
>  #include <linux/clocksource.h>
>  #include <asm/pvclock-abi.h>
>  
> -#ifdef CONFIG_PARAVIRT_CLOCK
> +#if defined(CONFIG_PARAVIRT_CLOCK) && defined(CONFIG_KVM_GUEST)
>  extern struct pvclock_vsyscall_time_info *pvclock_pvti_cpu0_va(void);
>  #else
>  static inline struct pvclock_vsyscall_time_info *pvclock_pvti_cpu0_va(void)
> -- 

Ok, ignore this one for now - we need to sort out the tip:x86/vdso mess
first and in particular:

  677a73a9aa54 ("x86/kvm: On KVM re-enable (e.g. after suspend), update clocks")

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.