[virtio-dev] Re: [virtio-comment] Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-28 Thread Heng Qi
在 2023/2/28 下午3:49, Michael S. Tsirkin 写道: On Tue, Feb 28, 2023 at 10:41:18AM +0800, Heng Qi wrote: The problem is the global qualifier. And it's not even global - there are two sets for rx and for tx and does not apply to cvq at all. How about "RX/TX coalescing parameters"? As Parav

Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Michael S. Tsirkin
On Tue, Feb 28, 2023 at 10:41:18AM +0800, Heng Qi wrote: > > The problem is the global qualifier. And it's not even global - > > there are two sets for rx and for tx and does not apply to cvq at all. > > How about "RX/TX coalescing parameters"? > > As Parav suggested, we probably don't need to

[virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Michael S. Tsirkin
On Tue, Feb 28, 2023 at 01:40:29AM +, Parav Pandit wrote: > > > From: Heng Qi > > Sent: Sunday, February 26, 2023 4:37 AM > > > +# Read/Write attributes for coalescing parameters \begin{itemize} \item > Hash # above is not needed. Please remove. > > > +For commands

Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Heng Qi
在 2023/2/28 上午4:19, Michael S. Tsirkin 写道: On Mon, Feb 27, 2023 at 08:45:43PM +0200, Alvaro Karsz wrote: -If the VIRTIO_NET_F_NOTF_COAL feature is negotiated, the driver can -send control commands for dynamically changing the coalescing parameters. +If the VIRTIO_NET_F_NOTF_COAL feature is

[virtio-dev] RE: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Parav Pandit
> From: Heng Qi > Sent: Sunday, February 26, 2023 4:37 AM > +# Read/Write attributes for coalescing parameters \begin{itemize} \item Hash # above is not needed. Please remove. > +For commands VIRTIO_NET_CTRL_NOTF_COAL_TX_SET and > VIRTIO_NET_CTRL_NOTF_COAL_RX_SET, \field{max_usecs} > +

Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Alvaro Karsz
> > I see your point, but a reader will see the "global notifications > > coalescing parameter" concept, and won't know what it is until next > > paragraph. > > Maybe we can have a new list with the notification coalescing concepts > > (like the notification coalescing parameters)? Just throwing

Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Michael S. Tsirkin
On Mon, Feb 27, 2023 at 08:45:43PM +0200, Alvaro Karsz wrote: > > >> -If the VIRTIO_NET_F_NOTF_COAL feature is negotiated, the driver can > > >> -send control commands for dynamically changing the coalescing > > >> parameters. > > >> +If the VIRTIO_NET_F_NOTF_COAL feature is negotiated, the

Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Alvaro Karsz
> >> -If the VIRTIO_NET_F_NOTF_COAL feature is negotiated, the driver can > >> -send control commands for dynamically changing the coalescing parameters. > >> +If the VIRTIO_NET_F_NOTF_COAL feature is negotiated, the driver can send > >> the VIRTIO_NET_CTRL_NOTF_COAL_TX_SET > >> +and

[virtio-dev] RE: [virtio-comment] Re: [virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-27 Thread Parav Pandit
> From: virtio-comm...@lists.oasis-open.org open.org> On Behalf Of Heng Qi > >>>   \subsubsection{Legacy Interface: Framing > >>> Requirements}\label{sec:Device > >>>   Types / Network Device / Legacy Interface: Framing Requirements} > >>> -- > >> Let's wait for more comments before the next

[virtio-dev] Re: [PATCH v9] virtio-net: support the virtqueue coalescing moderation

2023-02-26 Thread Alvaro Karsz
Hi, > Currently, coalescing parameters are grouped for all transmit and receive > virtqueues. This patch supports setting or getting the parameters for a > specified virtqueue, and a typical application of this function is netdim[1]. > > When the traffic between virtqueues is unbalanced, for