Re: [Qemu-devel] [PATCH] hw/arm/virt: mark the PCIe host controller as DMA coherent in the DT

[Peter Maydell]

On 1 July 2016 at 12:40, Andrew Jones <address@hidden> wrote:
> I might have mentioned in the commit message that the ACPI generation
> already does this, as _CCA is set to 1, added with commit bc64b96c
> (assuming I'm right, and a value of 1 there is the ACPI equivalent of
>  this patch)
>
> bc64b96c's commit message is also lacking, in the fact it doesn't
> state why the value of 1 is chosen, only that the attribute is
> compulsory, which I presume could have been added with the value 0
> to satisfy that.
>
> Anyway, I just wanted to point out that I *think* we're fine for
> ACPI. Perhaps I'm the only one who didn't know that already though...

Thanks for the update: I've edited the commit message to add:
# This brings the DT description into line with the ACPI description,
# which already marks the PCI bridge as cache coherent (see commit
# bc64b96c984abf).

thanks
-- PMM