qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v5 1/4] vhost-user: add new vhost user messages


From: Michael S. Tsirkin
Subject: Re: [Qemu-devel] [PATCH v5 1/4] vhost-user: add new vhost user messages to support virtio config space
Date: Wed, 22 Nov 2017 16:49:15 +0200

On Fri, Nov 17, 2017 at 04:44:37AM +0800, Changpeng Liu wrote:
> Add VHOST_USER_GET_CONFIG/VHOST_USER_SET_CONFIG messages which can be
> used for live migration of vhost user devices, also vhost user devices
> can benefit from the messages to get/set virtio config space from/to the
> I/O target. For the purpose to support virtio config space change,
> VHOST_USER_SET_CONFIG_FD message is added as the event notifier
> in case virtio config space change in the I/O target.
> 
> Signed-off-by: Changpeng Liu <address@hidden>
> ---
>  docs/interop/vhost-user.txt       | 39 ++++++++++++++++
>  hw/virtio/vhost-user.c            | 98 
> +++++++++++++++++++++++++++++++++++++++
>  hw/virtio/vhost.c                 | 63 +++++++++++++++++++++++++
>  include/hw/virtio/vhost-backend.h |  8 ++++
>  include/hw/virtio/vhost.h         | 16 +++++++
>  5 files changed, 224 insertions(+)
> 
> diff --git a/docs/interop/vhost-user.txt b/docs/interop/vhost-user.txt
> index 954771d..1b98388 100644
> --- a/docs/interop/vhost-user.txt
> +++ b/docs/interop/vhost-user.txt
> @@ -116,6 +116,16 @@ Depending on the request type, payload can be:
>      - 3: IOTLB invalidate
>      - 4: IOTLB access fail
>  
> + * Virtio device config space
> +   ---------------------------
> +   | offset | size | payload |
> +   ---------------------------
> +
> +   Offset: a 32-bit offset of virtio device's configuration space
> +   Size: a 32-bit size of configuration space that master wanted to change

I guess only legal values here are 1-256? But also see below. Also, we
already know the structure size.

> +   Payload: a 256-bytes array holding the contents of the virtio
> +       device's configuration space
> +

Why not *size* bytes? These are not performance critical but still,
why waste cycles.

>  In QEMU the vhost-user message is implemented with the following struct:
>  
>  typedef struct VhostUserMsg {

Virtio spec says:
        For device configuration access, the driver MUST use 8-bit wide 
accesses for 8-bit wide fields, 16-bit wide
        and aligned accesses for 16-bit wide fields and 32-bit wide and aligned 
accesses for 32-bit and 64-bit wide
        fields. For 64-bit fields, the driver MAY access each of the high and 
low 32-bit parts of the field independently.

So if these commands mirror guest accesses, they are always 1,2,4 or 8 bytes,
and aligned.


-- 
MST



reply via email to

[Prev in Thread] Current Thread [Next in Thread]