On 05/11/19 11:04, Paolo Bonzini wrote:
> On 04/11/19 23:59, Andrea Arcangeli wrote:
>> kvm_x86_set_hv_timer and kvm_x86_cancel_hv_timer needs to be defined
>> to succeed the 32bit kernel build, but they can't be called.
>>
>> Signed-off-by: Andrea Arcangeli <aarcange@xxxxxxxxxx>
>> ---
>> arch/x86/kvm/vmx/vmx.c | 11 +++++++++++
>> 1 file changed, 11 insertions(+)
>>
>> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
>> index bd17ad61f7e3..1a58ae38c8f2 100644
>> --- a/arch/x86/kvm/vmx/vmx.c
>> +++ b/arch/x86/kvm/vmx/vmx.c
>> @@ -7195,6 +7195,17 @@ void kvm_x86_cancel_hv_timer(struct kvm_vcpu *vcpu)
>> {
>> to_vmx(vcpu)->hv_deadline_tsc = -1;
>> }
>> +#else
>> +int kvm_x86_set_hv_timer(struct kvm_vcpu *vcpu, u64 guest_deadline_tsc,
>> + bool *expired)
>> +{
>> + BUG();
>> +}
>> +
>> +void kvm_x86_cancel_hv_timer(struct kvm_vcpu *vcpu)
>> +{
>> + BUG();
>> +}
>> #endif
>>
>> void kvm_x86_sched_in(struct kvm_vcpu *vcpu, int cpu)
>>
>
> I'll check for how long this has been broken. It may be the proof that
> we can actually drop 32-bit KVM support.
Ah no, I was confused because this series is not bisectable (in addition
to doing two things at a same time, namely the monolithic kvm.ko and the
retpoline eliminations).
I have picked up the patches that are independent of the monolithic
kvm.ko work or can be considered bugfixes.
For the rest, please do this before posting again:
- ensure that everything is bisectable
- look into how to remove the modpost warnings. A simple (though
somewhat ugly) way is to keep a kvm.ko module that includes common
virt/kvm/ code as well as, for x86 only, page_track.o. A few functions,
such as kvm_mmu_gfn_disallow_lpage and kvm_mmu_gfn_allow_lpage, would
have to be moved into mmu.h, but that's not a big deal.
- provide at least some examples of replacing the NULL kvm_x86_ops
checks with error codes in the function (or just early "return"s). I
can help with the others, but remember that for the patch to be merged,
kvm_x86_ops must be removed completely.
Thanks,
Paolo
