Re: [Qemu-devel] Fwd: QEMU: AMD IOMMU implementation bugs

[David Kiarie]

On Mon, Feb 18, 2019 at 9:59 PM David Kiarie <address@hidden> wrote:

>
>
>> >
>> > 3. And this might be something specific to my AMD IOMMU (1022:1437,
>> Family
>> > 16h), but the PS4 OS is checks the entire word stored in these registers
>> > and computes the expression as:
>> > config[AMDVI_CAPAB_BAR_LOW] | cap[AMDVI_CAPAB_BAR_HIGH] << 32
>> > So just storing half words should be wrong.
>> >
>>
>
> also,  as much as try, i cannot exactly figure out what you're talking
> about here and i'm under the impression i could not be the only person.
>
> would you mind elaborating? :-)
>

i can't express how grateful i am about your report.

having being the author of the culprit code, it's a fair bit of
irresponsibility on my side for me to leave issue unresolved, though it
does look easy to fix. it was probably not as easy to "dig" out the bugs.

on the other hand, Mr Alexandro, it does look like Peter(who most probably
still actively works on the Qemu IOMMUs, both Intel and AMD)will be
positively considering any IOMMU issues from you in the future.

4 days has been a wait long enough and i had better be out.