On Thu, 27 Sep 2018 13:58:45 +0200
David Hildenbrand wrote:
> On 27/09/2018 13:53, Igor Mammedov wrote:
> > On Thu, 27 Sep 2018 10:58:47 +0200
> > David Hildenbrand wrote:
> >
> >> On 27/09/2018 10:47, Igor Mammedov wrote:
> >>> On Thu, 27 Sep 2018 10:13:07 +0200
> >>> David Hildenbrand wr
On 27/09/2018 13:53, Igor Mammedov wrote:
> On Thu, 27 Sep 2018 10:58:47 +0200
> David Hildenbrand wrote:
>
>> On 27/09/2018 10:47, Igor Mammedov wrote:
>>> On Thu, 27 Sep 2018 10:13:07 +0200
>>> David Hildenbrand wrote:
>>>
On 27/09/2018 10:02, Igor Mammedov wrote:
> On Wed, 26 Se
On Thu, 27 Sep 2018 10:58:47 +0200
David Hildenbrand wrote:
> On 27/09/2018 10:47, Igor Mammedov wrote:
> > On Thu, 27 Sep 2018 10:13:07 +0200
> > David Hildenbrand wrote:
> >
> >> On 27/09/2018 10:02, Igor Mammedov wrote:
> >>> On Wed, 26 Sep 2018 11:41:59 +0200
> >>> David Hildenbrand wr
On 27/09/2018 10:47, Igor Mammedov wrote:
> On Thu, 27 Sep 2018 10:13:07 +0200
> David Hildenbrand wrote:
>
>> On 27/09/2018 10:02, Igor Mammedov wrote:
>>> On Wed, 26 Sep 2018 11:41:59 +0200
>>> David Hildenbrand wrote:
>>>
Make address_space_end point at the real end, instead of end +
On Thu, 27 Sep 2018 10:13:07 +0200
David Hildenbrand wrote:
> On 27/09/2018 10:02, Igor Mammedov wrote:
> > On Wed, 26 Sep 2018 11:41:59 +0200
> > David Hildenbrand wrote:
> >
> >> Make address_space_end point at the real end, instead of end + 1, so we
> >> don't
> >> have to handle special
On 27/09/2018 10:02, Igor Mammedov wrote:
> On Wed, 26 Sep 2018 11:41:59 +0200
> David Hildenbrand wrote:
>
>> Make address_space_end point at the real end, instead of end + 1, so we don't
>> have to handle special cases like it being 0. This will allow us to
>> place a memory device at the very
On Wed, 26 Sep 2018 11:41:59 +0200
David Hildenbrand wrote:
> Make address_space_end point at the real end, instead of end + 1, so we don't
> have to handle special cases like it being 0. This will allow us to
> place a memory device at the very end of the guest physical 64bit address
> space (if
Although unlikely in practice, we could have integer overflows on some
calculations based on addresses and sizes, leading to error checks not
triggering.
Let's properly handle this whenever we do an addition. Make
address_space_end point at the real end, instead of end + 1, so we don't
have to han