qemu-devel
[Top][All Lists]
Advanced

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

Re: [RFC PATCH v1 0/2] vhost-vdpa: cache Virtio states


From: Shao-Chien Chiang
Subject: Re: [RFC PATCH v1 0/2] vhost-vdpa: cache Virtio states
Date: Mon, 24 Apr 2023 22:05:00 +0800

Hi,
The "devices" means there are multiple vhost_dev.
After further research, I found the problem only happens in status.
Vhot-vdpa layer set status ACKNOWLEDGE,  DRIVER, and FEATURES_OK
only once by checking vhost_vdpa_first_dev().
However, it set status DRIVER_OK only once by checking if the device
is the last,
whose code is dev->vq_index + dev->nvqs != dev->vq_index_end in
vhost_vdpa_dev_start().
I think these are the causes of the problem.
Thank you!

On Mon, Apr 24, 2023 at 3:04 PM Eugenio Perez Martin
<eperezma@redhat.com> wrote:
>
> On Sat, Apr 22, 2023 at 8:39 AM Shao-Chien Chiang <ray90514@gmail.com> wrote:
> >
> > Hi,
> >
> > I found a problem about cache.
> >
> > If there are several devices operating the same backend device, the
> > cache might be inconsistent.
> >
>
> Hi Shao-Chien,
>
> What do you mean by "several devices operating the same backend
> device". The guest only sees "one virtio device per vhost-vdpa backend
> device".
>
> Thanks!
>
> > I think we could handle this by checking if a device is the first
> > device, but I'm not sure it will be feasible.
> >
> > Is there any better approach to this problem?
> >
> > Thank you!
> >
>



reply via email to

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