Re: [PATCH 35/38] unicore32: clean-up uapi Kbuild file

[James Hogan]

On 14/07/14 16:09, Sam Ravnborg wrote:
> Remove redundant assignments.
> 
> Signed-off-by: Sam Ravnborg 
> Cc: Guan Xuetao 
> ---
>  arch/unicore32/include/uapi/asm/Kbuild | 7 +--
>  1 file changed, 1 insertion(+), 6 deletions(-)
> 
> diff --git a/arch/unicore32/include/uapi/asm/Kbuild 
> b/arch/unicore32/include/uapi/asm/Kbuild
> index 0514d7a..aff532f 100644
> --- a/arch/unicore32/include/uapi/asm/Kbuild
> +++ b/arch/unicore32/include/uapi/asm/Kbuild
> @@ -1,10 +1,5 @@
>  # UAPI Header export list
>  include include/uapi/asm-generic/Kbuild.asm
>  
> -header-y += byteorder.h
> -header-y += kvm_para.h
> -header-y += ptrace.h
> -header-y += sigcontext.h
> -header-y += unistd.h
> -
> +# Wapper from asm-generic
>  generic-y += kvm_para.h
> 

You dropped the generic-y += kvm_para.h for arm64 & c6x, but not for
unicore, avr32 & ia64 patch, which seems a bit inconsistent. Was there
some reason behind that?

AFACT the non-uapi kvm_para.h is needed (e.g. by using the generic
non-uapi one) by kernel/watchdog.c for
kvm_check_and_clear_guest_paused(), so all arches should have at least
generic-y += kvm_para.h (or their own one) for non-uapi, even if they
have a uapi one too.

For UAPI, they're presumably only needed for architectures which support
KVM (s390, ia64, arm, powerpc, arm64, mips, x86 all provide an asm/kvm.h).

Is that right? Clearly there's some cleanup to do for kvm_para.h (which
is what made me look at the header issues previously).

Cheers
James
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 35/38] unicore32: clean-up uapi Kbuild file

[Sam Ravnborg]

Hi James.

> > +# Wapper from asm-generic
> >  generic-y += kvm_para.h
> > 
> 
> You dropped the generic-y += kvm_para.h for arm64 & c6x, but not for
> unicore, avr32 & ia64 patch, which seems a bit inconsistent. Was there
> some reason behind that?
I was fooling around with kvm_para.h and in both cases dropping
them was wrong.

kvm_para.h shows the current nightmare when you want to add a header
file that for most architectures has a generic implementation, but
in a few cases have a dedicated implementation.
This is the same point that Arnd raises.

I will fix it up so kvm_para.h is again specified for arm64 + c6x.
And then will address the whole mess later if I manage to come up
with something that is acceptable (simple enough).

Thanks for the review!

Sam
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html