linux-kernel - Re: [net-next v3 2/2] tg3: Link queues to NAPIs

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] [day] [month] [year] [list]

Message-ID: <CACKFLi=TNxacUwG4BaFwRmfLq42fFxtrAmES5Gm4bt4yt+Y15Q@mail.gmail.com>
Date: Mon, 7 Oct 2024 15:39:06 -0700
From: Michael Chan <michael.chan@...adcom.com>
To: Joe Damato <jdamato@...tly.com>, Michael Chan <michael.chan@...adcom.com>, 
	netdev@...r.kernel.org, Pavan Chebbi <pavan.chebbi@...adcom.com>, 
	Michael Chan <mchan@...adcom.com>, "David S. Miller" <davem@...emloft.net>, 
	Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>, 
	open list <linux-kernel@...r.kernel.org>
Subject: Re: [net-next v3 2/2] tg3: Link queues to NAPIs

On Mon, Oct 7, 2024 at 7:23 AM Joe Damato <jdamato@...tly.com> wrote:
>
> On Mon, Oct 07, 2024 at 12:30:09AM -0700, Michael Chan wrote:
> > On Sat, Oct 5, 2024 at 7:57 AM Joe Damato <jdamato@...tly.com> wrote:
> > > +               if (tnapi->tx_buffers) {
> > > +                       netif_queue_set_napi(tp->dev, txq_idx,
> > > +                                            NETDEV_QUEUE_TYPE_TX,
> > > +                                            &tnapi->napi);
> > > +                       txq_idx++;
> > > +               } else if (tnapi->rx_rcb) {
> >
> > Shouldn't this be "if" instead of "else if" ?  A napi can be for both
> > a TX ring and an RX ring in some cases.
> > Thanks.
>
> BTW: tg3 set_channels doesn't seem to support combined queues;
> combined_count is not even examined in set_channels. But maybe
> the link queue can be a combined queue, I don't know.

tg3 is a little odd.  When there are more than 1 MSIX, TX starts from
vector 0 but RX starts at vector 1.  If there are more than 1 TX ring,
the 2nd TX ring is combined with the 1st RX ring and so on.

>
> Regardless, I'll still make the change you requested as there is
> similar code in tg3_request_irq.
>
> But what I really would like to get feedback on is the rxq and txq
> indexing with the running counters, please. That was called out
> explicitly in the cover letter.
>

It looks reasonable to me.  Thanks.

Download attachment "smime.p7s" of type "application/pkcs7-signature" (4209 bytes)