>>> On 31.03.15 at 17:57, wrote:
> Currently, _xmalloc() supports zero-sized allocations by returning a sentinel
> poisoned pointer.
>
> I posit that there are no legitimate situation for any code in the
> hypervisor
> to make a zero sized allocation.
I'm afraid there are, and we ran into them
At 16:57 +0100 on 31 Mar (1427821072), Andrew Cooper wrote:
> Currently, _xmalloc() supports zero-sized allocations by returning a sentinel
> poisoned pointer.
>
> I posit that there are no legitimate situation for any code in the hypervisor
> to make a zero sized allocation.
>
> Furthermore, the
Currently, _xmalloc() supports zero-sized allocations by returning a sentinel
poisoned pointer.
I posit that there are no legitimate situation for any code in the hypervisor
to make a zero sized allocation.
Furthermore, the sentinel value will pass a NULL pointer check, and
introduces an unnecess
