Re: [PATCH 0/14] KVM: Kernel-based Virtual Machine (v4)

[Avi Kivity <avi@qumranet.com>]

Arjan van de Ven wrote:
>> The problem with that is that the test comes too late: after we've 
>> configured.  Andrew wants to keep allmodconfig working, and for that we 
>> need to deselect CONFIG_KVM before compilation starts.
>>     
>
> not really. You can also select to just not compile kvm at all *from the
> Makefile*
>   

Yes.  And then a real user (not an allmodconfig user) selects CONFIG_KVM 
and wonders where it went.

For users, we want it to fail fatally.  For allmodconfig, we want it to 
succeed but don't care about the output.


>> gcc.*protector.sh only affects the Makefile, not the configuration, AFAICT.
>>     
>
> but it is the Makefile that goes into the kvm directory and compiles
> stuff!
>
> yes it's ugly and not so elegant, but it's effective and you can warn
> bigtime via nasty messages if you want ;)
>   

We already have an ugly solution:

    #define VMLAUNCH ".byte 0x0f, foo, bar"

etc.  We were looking for an elegant one.

-- 
error compiling committee.c: too many arguments to function