Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
MIME-Version: 1.0
Date:   Wed, 09 Feb 2022 02:20:07 +0000
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
From:   yajun.deng@linux.dev
Message-ID: <753bb02bfa8c2cf5c08c63c31f193f90@linux.dev>
Subject: Re: [PATCH net-next] net: dev: introduce netdev_drop_inc()
To:     "Jakub Kicinski" <kuba@kernel.org>
Cc:     davem@davemloft.net, rostedt@goodmis.org, mingo@redhat.com,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org
In-Reply-To: <20220208162729.41b62ae7@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
References: <20220208162729.41b62ae7@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
 <20220208064318.1075849-1-yajun.deng@linux.dev>
Precedence: bulk

February 9, 2022 8:27 AM, "Jakub Kicinski" <kuba@kernel.org> wrote:=0A=0A=
> On Tue, 8 Feb 2022 14:43:18 +0800 Yajun Deng wrote:=0A> =0A>> We will u=
se 'sudo perf record -g -a -e skb:kfree_skb' command to trace=0A>> the dr=
opped packets when dropped increase in the output of ifconfig.=0A>> But t=
here are two cases, one is only called kfree_skb(), another is=0A>> incre=
asing the dropped and called kfree_skb(). The latter is what=0A>> we need=
. So we need to separate these two cases.=0A>> =0A>> From the other side,=
 the dropped packet came from the core network and=0A>> the driver, we al=
so need to separate these two cases.=0A>> =0A>> Add netdev_drop_inc() and=
 add a tracepoint for the core network dropped=0A>> packets. use 'sudo pe=
rf record -g -a -e net:netdev_drop' and 'sudo perf=0A>> script' will reco=
red the dropped packets by the core network.=0A>> =0A>> Signed-off-by: Ya=
jun Deng <yajun.deng@linux.dev>=0A> =0A> Have you seen the work that's be=
ing done around kfree_skb_reason()?=0A=0AYes, I saw it. The focus of kfre=
e_skb_reason() is trace kfree_skb() and the reason, =0Abut the focus of t=
his patch only traces this case of the dropped packet.=0A=0AI don't want =
to trace all kfree_skb(), but I just want to trace the dropped packet.=0A=
=0AThis command 'sudo perf record -g -a -e skb:kfree_skb' would trace all=
 kfree_skb(),=0Akfree_skb() would drowned out the case of dropped packets=
 when the samples were too large.