On 27/11/19 22:57, Leonardo Bras wrote:
> But on the above case, kvm_put_kvm{,_no_destroy}() would be called
> with refcount == 1, and if reorder patch is applied, it would not cause
> any use-after-free error, even on kvm_put_kvm() case.
I think this is what you're missing: kvm_put_kvm_no_destroy() does not
protect against bugs in the code that uses it. It protect against bugs
_elsewhere_.
Therefore, kvm_put_kvm_no_destroy() is always a better choice when
applicable, because it turns bugs in _other parts of the code_ from
use-after-free to WARN+leak.
Paolo
Attachment:
signature.asc
Description: OpenPGP digital signature
