Re: xlnx-versal-virt smmu

[Edgar E. Iglesias]

On Fri, Nov 08, 2024 at 04:26:20PM +0000, Ho, Nelson via wrote:
> Hi Edgar,
> 
> I am working on bringing up the Wind River Helix hypervisor on the 
> xlnx-versal-virt machine, which expects to find MMU-500 SMMU where it lives 
> on the Versal SoC. I understand the -virt machine is not intended to fully 
> emulate Versal SoC.
> 
> I found an implementation of the MMU-500 on the Xilinx QEMU fork, and am 
> curious why this was not upstreamed, and if you know what issues I might 
> encounter if I try to integrate this MMU-500 implementation into the 
> versal-virt machine?
>

Hi Nelson,

There's no particular reason other than lack of time.

The tricky parts will likely be to hook it up without the device-tree
bindings that Xilinx uses to create machines. Perhaps also how
we use Memory Regions to describe the DMA ports towards DMA masters
and towards SMMU TBUs in a non-PCI specific manner.

Cheers,
Edgar