On Fri, Feb 17, 2023 at 12:36:32PM +0200, Alvaro Karsz wrote:
> > sounds good. maybe "counting packets and microseconds"? to make clear
> > timer resets too.
> >
> > > I don't really mind, If you think that the device implementation is
> > > not essential here and we should add your suggestion,
> sounds good. maybe "counting packets and microseconds"? to make clear
> timer resets too.
>
> > I don't really mind, If you think that the device implementation is
> > not essential here and we should add your suggestion, I'm ok with
> > that.
>
> I don't mind, your proposal seems good too.
Ok,
On Fri, Feb 17, 2023 at 12:21:17PM +0200, Alvaro Karsz wrote:
> > Looks good, thanks! Yet something to improve, see below:
> > > ---
> > > v2:
> > > - Add the last 2 points to the patch.
> > > - Rephrase the "commands are best-effort" note.
> > > - Replace "notification" with
> Looks good, thanks! Yet something to improve, see below:
> > ---
> > v2:
> > - Add the last 2 points to the patch.
> > - Rephrase the "commands are best-effort" note.
> > - Replace "notification" with "used buffer notification" to be
> > more consistent.
> > v3:
> >
On Thu, Feb 16, 2023 at 04:56:11PM +0200, Alvaro Karsz wrote:
> This patch makes several improvements to the notification coalescing
> feature, including:
>
> - Consolidating virtio_net_ctrl_coal_tx and virtio_net_ctrl_coal_rx
> into a single struct, virtio_net_ctrl_coal, as they are identical.
> From: Alvaro Karsz
> Sent: Thursday, February 16, 2023 9:56 AM
[..]
> +When the device has \field{max_usecs} = 0 or \field{max_packets} = 0, the
> notification conditions are met after every packet received/sent.
> +
> +\subparagraph{RX Example}\label{sec:Device Types / Network Device /
>