Re: [RFC PATCH 1/3] content: Add VIRTIO_F_SWIOTLB to negotiate use of SWIOTLB bounce buffers

[David Woodhouse]

On Thu, 2025-04-03 at 00:24 -0700, Christoph Hellwig wrote:
> On Wed, Apr 02, 2025 at 12:04:45PM +0100, David Woodhouse wrote:
> > Define a VIRTIO_F_SWIOTLB feature which allows the device and driver to
> > agree on the use of the SWIOTLB, if present. This enables the device to
> > refuse to operate further if the driver does not support the SWIOTLB
> > requirement expressed in the device-tree.
> 
> This makes no sense whatsoever.  Swiotlb is a Linux guest implementation
> detail, virtio is a transport protocol.  Mixing them in the same spec
> doesn't even compute.  Please describe the actual on the wire semantics
> you want, and don't use the term swiotlb.

Linux has 'SWIOTLB' support, but I didn't see it as a term specific to
that particular implementation; it seemed like a concept would would be
understood elsewhere. But sure, I'm happy to rename it to just 'bounce
buffers' or something like that. I'll see what I can come up with that
is succinct enough to use in VIRTIO_F_xxx and VIRTIO_PCI_CAP_xxx names.

Attachment: smime.p7s
Description: S/MIME cryptographic signature