On Tue, Sep 12, 2017 at 12:08:27PM +0200, Auger Eric wrote: > Hi Wanghaibin, > > On 12/09/2017 10:50, wanghaibin wrote: > > On 2017/9/6 21:05, wanghaibin wrote: > > > >> We slightly refactor vgic_its_destroy, separate vgic_its_free_list() > >> function for later patch invoke. > >> > >> The patch also take a functional change. If the its->device_list.next > >> is NULL, we still should free the its. > > > > > > Hi, Eric > > > > Does this its->device_list.next is NULL can happened ? > > I don't get why we have this check. > > The kvm device is removed by kvm_destroy_devices which loops on all > devices added to kvm->devices. kvm_ioctl_create_device only adds the > device to kvm_devices once the lists have been initialized (in > vgic_create_its). So it looks safe to me without the check. > > André, do we miss something? Eric, I think I agree with you, and even if we need to check for initialization, we should have a separate flag instead of piggy-backing on the list_head. Thanks, -Christoffer _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
