Re: ovmf / PCI passthrough impaired due to very limiting PCI64 aperture

[Laszlo Ersek]

On 06/17/20 18:01, Guilherme Piccoli wrote:
> On Wed, Jun 17, 2020 at 12:57 PM Laszlo Ersek <lersek@redhat.com> wrote:
>> [...]
>> I don't necessarily see compat issues -- large-BAR PCI device assignment
>> might simply stop working under those circumstances, because you could
>> no longer use X-PciMmio64Mb, and the new way wouldn't be supported by
>> (old) QEMU.
>>
>> This would indeed be a regression relative to "X-PciMmio64Mb", but
>> that's exactly why there's an "X" in "X-PciMmio64Mb".
>>
> 
> Are you planning to remove that option, with this improvement?

Yes.

> I think
> we could keep it for a while, as a way to override the automatic
> mechanism we might implement. This is even for "safe" purposes, in
> case there's some corner case with the auto-sized aperture that we
> ignore upfront.

I disagree. The knob is called "experimental" specifically so we don't
have to introduce even worse complexity for compatibility's sake than
what we have now. Graceful deprecation is for options that used to be
supported; "X-PciMmio64Mb" has been explicitly experimental from the start.

I'm not saying that I'll kill "X-PciMmio64Mb" for sure, just that I very
likely will.

Thanks
Laszlo