netdev - Re: [RFC net-next 2/2] gve: Map NAPI instances to queues

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Zvb4h7gZvrCErJsi@LQ3V64L9R2>
Date: Fri, 27 Sep 2024 11:25:11 -0700
From: Joe Damato <jdamato@...tly.com>
To: Praveen Kaligineedi <pkaligineedi@...gle.com>
Cc: netdev@...r.kernel.org, Jeroen de Borst <jeroendb@...gle.com>,
	Shailend Chand <shailend@...gle.com>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Willem de Bruijn <willemb@...gle.com>,
	Harshitha Ramamurthy <hramamurthy@...gle.com>,
	Ziwei Xiao <ziweixiao@...gle.com>,
	open list <linux-kernel@...r.kernel.org>
Subject: Re: [RFC net-next 2/2] gve: Map NAPI instances to queues

On Fri, Sep 27, 2024 at 11:15:39AM -0700, Praveen Kaligineedi wrote:
> > diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c
> > index 661566db68c8..da811e90bdfa 100644
> > --- a/drivers/net/ethernet/google/gve/gve_main.c
> > +++ b/drivers/net/ethernet/google/gve/gve_main.c
> > @@ -1875,6 +1875,9 @@ static void gve_turndown(struct gve_priv *priv)
> >
> >                 if (!gve_tx_was_added_to_block(priv, idx))
> >                         continue;
> > +
> > +               netif_queue_set_napi(priv->dev, idx, NETDEV_QUEUE_TYPE_TX,
> > +                                    NULL);
> >                 napi_disable(&block->napi);
> >         }
> When XDP program is installed, the for loop iterates over both
> configured TX queues (idx <  priv->tx_cfg.num_queues) as well as
> dedicated XDP TX queues ( idx >= priv->tx_cfg.num_queues).
> Should add if (idx <  priv->tx_cfg.num_queues) check here.
> 
> > @@ -1909,6 +1915,9 @@ static void gve_turnup(struct gve_priv *priv)
> >                         continue;
> >
> >                 napi_enable(&block->napi);
> > +               netif_queue_set_napi(priv->dev, idx, NETDEV_QUEUE_TYPE_TX,
> > +                                    &block->napi);
> > +
> >                 if (gve_is_gqi(priv)) {
> >                         iowrite32be(0, gve_irq_doorbell(priv, block));
> >                 } else {
> 
> Same as above. When XDP program is installed, the for loop iterates
> over both configured TX queues (idx <  priv->tx_cfg.num_queues) as
> well as dedicated XDP TX queues ( idx >= priv->tx_cfg.num_queues)

Ah, OK. Thanks for the review. I'll make that change and retest it
on my GCP instance.

Since net-next is (I think) likely to reopen soon, I'll include that
change in the submission I send next week and mention it in the
changelog.