Received-SPF: pass (google.com: domain of linux-kernel+bounces-65598-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240202165315.2506384-1-leitao@debian.org> <CANn89iLWWDjp71R7zttfTcEvZEdmcA1qo47oXkAX5DuciYvOtQ@mail.gmail.com>
 <20240213100457.6648a8e0@kernel.org> <ZczSEBFtq6E6APUJ@gmail.com>
 <CANn89iJH5jpvBCw8csGux9U10HwM+ewnL1A7udBi6uwAX6VBYA@mail.gmail.com> <ZczvGJ90L7689F6J@gmail.com>
In-Reply-To: <ZczvGJ90L7689F6J@gmail.com>
From: Eric Dumazet <edumazet@google.com>
Date: Wed, 14 Feb 2024 17:58:37 +0100
Message-ID: <CANn89i+zF3k4OyhJsK3sg5zNsFzKAQ5G_ANYEaxOfc41B7S18w@mail.gmail.com>
Subject: Re: [PATCH net-next v3] net: dqs: add NIC stall detector based on BQL
To: Breno Leitao <leitao@debian.org>
Cc: Jakub Kicinski <kuba@kernel.org>, davem@davemloft.net, pabeni@redhat.com, 
	Steven Rostedt <rostedt@goodmis.org>, Masami Hiramatsu <mhiramat@kernel.org>, 
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>, Andrew Morton <akpm@linux-foundation.org>, 
	weiwan@google.com, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, 
	horms@kernel.org, Jonathan Corbet <corbet@lwn.net>, Randy Dunlap <rdunlap@infradead.org>, 
	Bjorn Helgaas <bhelgaas@google.com>, Johannes Berg <johannes.berg@intel.com>, 
	=?UTF-8?Q?Thomas_Wei=C3=9Fschuh?= <linux@weissschuh.net>, 
	"open list:TRACING" <linux-trace-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Feb 14, 2024 at 5:49=E2=80=AFPM Breno Leitao <leitao@debian.org> wr=
ote:
>
> On Wed, Feb 14, 2024 at 04:41:36PM +0100, Eric Dumazet wrote:
> > On Wed, Feb 14, 2024 at 3:45=E2=80=AFPM Breno Leitao <leitao@debian.org=
> wrote:
> > >
> > > On Tue, Feb 13, 2024 at 10:04:57AM -0800, Jakub Kicinski wrote:
> > > > On Tue, 13 Feb 2024 14:57:49 +0100 Eric Dumazet wrote:
> > > > > Please note that adding other sysfs entries is expensive for work=
loads
> > > > > creating/deleting netdev and netns often.
> > > > >
> > > > > I _think_ we should find a way for not creating
> > > > > /sys/class/net/<interface>/queues/tx-{Q}/byte_queue_limits  direc=
tory
> > > > > and files
> > > > > for non BQL enabled devices (like loopback !)
> > > >
> > > > We should try, see if anyone screams. We could use IFF_NO_QUEUE, an=
d
> > > > NETIF_F_LLTX as a proxy for "device doesn't have a real queue so BQ=
L
> > > > would be pointless"? Obviously better to annotate the drivers which
> > > > do have BQL support, but there's >50 of them on a quick count..
> > >
> > > Let me make sure I understand the suggestion above. We want to disabl=
e
> > > BQL completely for devices that has dev->features & NETIF_F_LLTX or
> > > dev->priv_flags & IFF_NO_QUEUE, right?
> > >
> > > Maybe we can add a ->enabled field in struct dql, and set it accordin=
g
> > > to the features above. Then we can created the sysfs and process the =
dql
> > > operations based on that field. This should avoid some unnecessary ca=
lls
> > > also, if we are not display sysfs.
> > >
> > > Here is a very simple PoC to represent what I had in mind. Am I in th=
e
> > > right direction?
> >
> > No, this was really about sysfs entries (aka dql_group)
> >
> > Partial patch would be:
>
> That is simpler than what I imagined. Thanks!
>

>
> for netdev_uses_bql(), would it be similar to what I proposed in the
> previous message? Let me copy it here.
>
>         static bool netdev_uses_bql(struct net_device *dev)
>         {
>                if (dev->features & NETIF_F_LLTX ||
>                    dev->priv_flags & IFF_NO_QUEUE)
>                        return false;
>
>                return true;
>         }

I think this should be fine, yes.