Re: [PATCH v2 2/4] vhost-user: introduce virtio API

On 17/07/2024 07:21, David Gibson wrote:

On Fri, Jul 12, 2024 at 05:32:42PM +0200, Laurent Vivier wrote:

...

Add virtio.c and virtio.h that define the functions needed to manage virtqueues.

Signed-off-by: Laurent Vivier --- Makefile | 4 +- util.h | 11 + virtio.c | 611 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ virtio.h | 190 +++++++++++++++++ 4 files changed, 814 insertions(+), 2 deletions(-) create mode 100644 virtio.c create mode 100644 virtio.h

...

...

+ +/** + * struct vu_dev_region - guest shared memory region + * @gpa: Guest physical address of the region + * @size: Memory size in bytes + * @qva: QEMU virtual address

Is this actually the qemu virtual address? Or is it our virtual address?

It is actually QEMU virtual address, it's a virtual address provided by the front-end (in our case QEMU). https://qemu-project.gitlab.io/qemu/interop/vhost-user.html#memory-region-de... It's used in qva_to_va() to convert vring addresses that are in QEMU userspace (vhost-user) to our process userspace mapped address, while vu_gpa_to_va() is used to convert guest physical address (virtio) of buffers address stored vring descritors that are in guest physical address space to our process address space. vhost-user addresses are in QEMU virtual adress space, virtio addresses are in guest physical address space. Thansk, Laurent