> From: Jakub Kicinski <kuba@xxxxxxxxxx>
> Sent: Friday, April 12, 2024 9:21 PM
> To: Dan Jurgens <danielj@xxxxxxxxxx>
> Cc: netdev@xxxxxxxxxxxxxxx; mst@xxxxxxxxxx; jasowang@xxxxxxxxxx;
> xuanzhuo@xxxxxxxxxxxxxxxxx; virtualization@xxxxxxxxxxxxxxx;
> davem@xxxxxxxxxxxxx; edumazet@xxxxxxxxxx; pabeni@xxxxxxxxxx; Jiri
> Pirko <jiri@xxxxxxxxxx>
> Subject: Re: [PATCH net-next v3 5/6] virtio_net: Add a lock for per queue RX
> coalesce
>
> On Fri, 12 Apr 2024 14:53:08 -0500 Daniel Jurgens wrote:
> > Once the RTNL locking around the control buffer is removed there can
> > be contention on the per queue RX interrupt coalescing data. Use a
> > spin lock per queue.
>
> Does not compile on Clang.
Which version? It compiles for me with:
$ clang -v
clang version 15.0.7 (Fedora 15.0.7-2.fc37)
>
> > + scoped_guard(spinlock, &vi->rq[i].intr_coal_lock) {
> > + err = virtnet_send_rx_ctrl_coal_vq_cmd(vi, i,
> > + vi-
> >intr_coal_rx.max_usecs,
> > + vi-
> >intr_coal_rx.max_packets);
> > + if (err)
> > + return err;
> > + }
>
> Do you really think this needs a scoped guard and 4th indentation level,
> instead of just:
>
> ..lock(..);
> err = virtnet_send_rx_ctrl_coal_vq_cmd(vi, i,
> vi-
> >intr_coal_rx.max_usecs,
> vi-
> >intr_coal_rx.max_packets);
> ..unlock(..);
I'll change it in the next version.
> if (err)
> return err;
>
> > + scoped_guard(spinlock, &vi->rq[i].intr_coal_lock) {
> > + vi->rq[i].intr_coal.max_usecs = ec-
> >rx_coalesce_usecs;
> > + vi->rq[i].intr_coal.max_packets = ec-
> >rx_max_coalesced_frames;
> > + }
>
> :-|
> --
> pw-bot: cr
