On 12/01/21 10:29, Kunkun Jiang wrote:
The KVM_CLEAR_DIRTY_LOG ioctl lacks the check whether the capability KVM_CAP_MANUAL_DIRTY_LOG_PROTECT is enabled or not. This may cause some problems if userspace calls the KVM_CLEAR_DIRTY_LOG ioctl, but dose't enable this capability. So we'd better to add it.Fixes: 2a31b9db15353 ("kvm: introduce manual dirty log reprotect") Signed-off-by: Kunkun Jiang <jiangkun...@huawei.com> --- virt/kvm/kvm_main.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index fa9e3614d30e..8f5633d8a0e8 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -1602,6 +1602,9 @@ static int kvm_clear_dirty_log_protect(struct kvm *kvm, unsigned long *dirty_bitmap_buffer; bool flush;+ if (!kvm->manual_dirty_log_protect)+ return -EPERM; + /* Dirty ring tracking is exclusive to dirty log tracking */ if (kvm->dirty_ring_size) return -ENXIO;
This is not a problem, KVM_CLEAR_DIRTY_LOG can always be used even if KVM_CAP_MANUAL_DIRTY_LOG_PROTECT was not enabled. The meaning is the same: clear the bits without returning them.
Paolo
