virtio-comment

 View Only
  • 1.  Re: [VIRTIO GPU PATCH 1/1] virtio-gpu: Add new feature flag VIRTIO_GPU_F_FREEZING

    Posted 07-06-2023 11:54
    Hi,

    > +\item[VIRTIO_GPU_CMD_STATUS_FREEZING]
    > +Notify freezing status through controlq.
    > +Request data is \field{struct virtio_gpu_status_freezing}.
    > +Response type is VIRTIO_GPU_RESP_OK_NODATA.
    > +
    > +Guest notifies QEMU if virtio-gpu is in freezing status or not in \field{freezing}.
    > +zero means it is not in freezing status, none-zero is the opposite.
    > +When guest is in freezing status, QEMU will not destroy resources.

    This should be more clear which resources are meant here. Just
    resources created by VIRTIO_GPU_CMD_RESOURCE_CREATE_* commands?
    Or all objects (including cursors, scanouts, 3D contexts, ...)?
    Or some other set?

    I've seen on the (code) patch discussion the topic of different
    suspend modes (s3/s4) came up.
    What was the conclusion there?
    Should a note about this be added to the spec too?

    take care,
    Gerd




  • 2.  Re: [virtio-dev] Re: [VIRTIO GPU PATCH 1/1] virtio-gpu: Add new feature flag VIRTIO_GPU_F_FREEZING

    Posted 07-06-2023 12:14
    Hi Gerd Hoffmann,

    Thank you for your review and suggestions.

    On 2023/7/6 19:54, Gerd Hoffmann wrote:
    > Hi,
    >
    >> +\item[VIRTIO_GPU_CMD_STATUS_FREEZING]
    >> +Notify freezing status through controlq.
    >> +Request data is \field{struct virtio_gpu_status_freezing}.
    >> +Response type is VIRTIO_GPU_RESP_OK_NODATA.
    >> +
    >> +Guest notifies QEMU if virtio-gpu is in freezing status or not in \field{freezing}.
    >> +zero means it is not in freezing status, none-zero is the opposite.
    >> +When guest is in freezing status, QEMU will not destroy resources.
    >
    > This should be more clear which resources are meant here. Just
    > resources created by VIRTIO_GPU_CMD_RESOURCE_CREATE_* commands?
    > Or all objects (including cursors, scanouts, 3D contexts, ...)?
    > Or some other set?

    Yes, you are right, it should be clearer about resource here. I will elaborate on the types of resources in the next version.

    >
    > I've seen on the (code) patch discussion the topic of different
    > suspend modes (s3/s4) came up.
    > What was the conclusion there?
    > Should a note about this be added to the spec too?

    My modification was originally to support xen's s3 function, but I will try s4 locally, if s4 does not work, I will add instructions here.

    >
    > take care,
    > Gerd
    >
    >
    > ---------------------------------------------------------------------
    > To unsubscribe, e-mail: virtio-dev-unsubscribe@lists.oasis-open.org
    > For additional commands, e-mail: virtio-dev-help@lists.oasis-open.org
    >

    --
    Best regards,
    Jiqian Chen.



  • 3.  Re: [virtio-dev] Re: [VIRTIO GPU PATCH 1/1] virtio-gpu: Add new feature flag VIRTIO_GPU_F_FREEZING

    Posted 07-06-2023 12:14
    Hi Gerd Hoffmann,

    Thank you for your review and suggestions.

    On 2023/7/6 19:54, Gerd Hoffmann wrote:
    > Hi,
    >
    >> +\item[VIRTIO_GPU_CMD_STATUS_FREEZING]
    >> +Notify freezing status through controlq.
    >> +Request data is \field{struct virtio_gpu_status_freezing}.
    >> +Response type is VIRTIO_GPU_RESP_OK_NODATA.
    >> +
    >> +Guest notifies QEMU if virtio-gpu is in freezing status or not in \field{freezing}.
    >> +zero means it is not in freezing status, none-zero is the opposite.
    >> +When guest is in freezing status, QEMU will not destroy resources.
    >
    > This should be more clear which resources are meant here. Just
    > resources created by VIRTIO_GPU_CMD_RESOURCE_CREATE_* commands?
    > Or all objects (including cursors, scanouts, 3D contexts, ...)?
    > Or some other set?

    Yes, you are right, it should be clearer about resource here. I will elaborate on the types of resources in the next version.

    >
    > I've seen on the (code) patch discussion the topic of different
    > suspend modes (s3/s4) came up.
    > What was the conclusion there?
    > Should a note about this be added to the spec too?

    My modification was originally to support xen's s3 function, but I will try s4 locally, if s4 does not work, I will add instructions here.

    >
    > take care,
    > Gerd
    >
    >
    > ---------------------------------------------------------------------
    > To unsubscribe, e-mail: virtio-dev-unsubscribe@lists.oasis-open.org
    > For additional commands, e-mail: virtio-dev-help@lists.oasis-open.org
    >

    --
    Best regards,
    Jiqian Chen.