FWIW, this is a kernel fallout of the valgrind related QEMU discussion.
On 11/21/2017 04:08 PM, Christian Borntraeger wrote:
> Old kernels did not check for zero in the irq_state.flags field and old
> QEMUs did not zero the flag field when calling KVM_S390_*_IRQ_STATE.
> Let's add a comment and dummy code to prevent future usage of flags
> and pad.
>
> Signed-off-by: Christian Borntraeger <borntraeger@xxxxxxxxxx>
> ---
> arch/s390/kvm/kvm-s390.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
>
> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> index 6a5e02f..1baa393 100644
> --- a/arch/s390/kvm/kvm-s390.c
> +++ b/arch/s390/kvm/kvm-s390.c
> @@ -3834,6 +3834,16 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
> r = -EINVAL;
> break;
> }
> + if (irq_state.flags) {
> + /*
> + * This is a placeholder to make sure that nobody uses
> + * flags and pad. Old kernels did not check for zero
> + * and old QEMUs did not zero the flag field.
> + * That means that we cannot use the flags field for
> + * any possible extension.
> + */
> + irq_state.flags = 0;
> + }
> r = kvm_s390_set_irq_state(vcpu,
> (void __user *) irq_state.buf,
> irq_state.len);
> @@ -3849,6 +3859,10 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
> r = -EINVAL;
> break;
> }
> + if (irq_state.flags) {
> + /* see above */
> + irq_state.flags = 0;
> + }
> r = kvm_s390_get_irq_state(vcpu,
> (__u8 __user *) irq_state.buf,
> irq_state.len);
>
