Received-SPF: pass (google.com: domain of linux-kernel+bounces-68253-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Received-SPF: Pass (protection.outlook.com: domain of nvidia.com designates
 216.228.117.161 as permitted sender) receiver=protection.outlook.com;
 client-ip=216.228.117.161; helo=mail.nvidia.com; pr=C
References: <20240215160458.1727237-1-ast@fiberby.net>
 <20240215160458.1727237-2-ast@fiberby.net>
 <CAM0EoMndBjwC8Otx6th_dM_aV_r80NeLEke9C8PwzGt1q3vAMA@mail.gmail.com>
 <a4798b5d-1a8a-41ab-842f-52e8c7ac00ed@fiberby.net>
User-agent: mu4e 1.10.5; emacs 29.1.90
From: Vlad Buslov <vladbu@nvidia.com>
To: =?utf-8?Q?Asbj=C3=B8rn?= Sloth =?utf-8?Q?T=C3=B8nnesen?= <ast@fiberby.net>
CC: Jamal Hadi Salim <jhs@mojatatu.com>, Cong Wang <xiyou.wangcong@gmail.com>,
	Jiri Pirko <jiri@resnulli.us>, Daniel Borkmann <daniel@iogearbox.net>,
	<netdev@vger.kernel.org>, <linux-kernel@vger.kernel.org>, <llu@fiberby.dk>,
	Marcelo Ricardo Leitner <mleitner@redhat.com>
Subject: Re: [PATCH net-next 1/3] net: sched: cls_api: add skip_sw counter
Date: Fri, 16 Feb 2024 10:35:40 +0200
In-Reply-To: <a4798b5d-1a8a-41ab-842f-52e8c7ac00ed@fiberby.net>
Message-ID: <87a5o0dcpd.fsf@nvidia.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 16 Feb 2024 08:40:30.4618
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: abfd1028-6e7a-4ed0-3e17-08dc2ecaeee0
X-MS-Exchange-CrossTenant-Id: 43083d15-7273-40c1-b7db-39efd9ccc17a
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=43083d15-7273-40c1-b7db-39efd9ccc17a;Ip=[216.228.117.161];Helo=[mail.nvidia.com]
X-MS-Exchange-CrossTenant-AuthSource:
	CY4PEPF0000FCC0.namprd03.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: SN7PR12MB6789

On Thu 15 Feb 2024 at 23:34, Asbj=C3=B8rn Sloth T=C3=B8nnesen <ast@fiberby.=
net> wrote:
> Hi Jamal,
>
> Thank you for the review.
>
> On 2/15/24 17:39, Jamal Hadi Salim wrote:
>> +Cc Vlad and Marcelo..
>> On Thu, Feb 15, 2024 at 11:06=E2=80=AFAM Asbj=C3=B8rn Sloth T=C3=B8nnese=
n <ast@fiberby.net>
>> wrote:
>>>
>>> Maintain a count of skip_sw filters.
>>>
>>> This counter is protected by the cb_lock, and is updated
>>> at the same time as offloadcnt.
>>>
>>> Signed-off-by: Asbj=C3=B8rn Sloth T=C3=B8nnesen <ast@fiberby.net>
>>> ---
>>>   include/net/sch_generic.h | 1 +
>>>   net/sched/cls_api.c       | 4 ++++
>>>   2 files changed, 5 insertions(+)
>>>
>>> diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
>>> index 934fdb977551..46a63d1818a0 100644
>>> --- a/include/net/sch_generic.h
>>> +++ b/include/net/sch_generic.h
>>> @@ -476,6 +476,7 @@ struct tcf_block {
>>>          struct flow_block flow_block;
>>>          struct list_head owner_list;
>>>          bool keep_dst;
>>> +       atomic_t skipswcnt; /* Number of skip_sw filters */
>>>          atomic_t offloadcnt; /* Number of oddloaded filters */
>> For your use case is skipswcnt ever going to be any different than offlo=
adcnt?
>
> No, we only use skip_sw filters, since we only use TC as a control path to
> install skip_sw rules into hardware.
>
> AFAICT offloadcnt is the sum of skip_sw filters, and filters with no flag=
s which
> have implicitly been offloaded.
>
> The reason that I didn't just use offloadcnt, is that I'm not sure if it =
is
> acceptable to treat implicitly offloaded rules without skip_sw, as if the=
y were
> explicitly skip_sw. It sounds reasonable, given that the filters without =
skip_* flags
> shouldn't really care.

It is not acceptable since there are valid use-cases where packets need
to match sw filters that are supposedly also in-hw. For example, filters
with tunnel_key set action during neighbor update event.

>
> I tried to only trigger the TC bypass, in the cases that I was absolutely=
 sure would
> be safe as a first step.
>
>
>> cheers,
>> jamal
>>=20
>>>          unsigned int nooffloaddevcnt; /* Number of devs unable to do o=
ffload */
>>>          unsigned int lockeddevcnt; /* Number of devs that require rtnl=
 lock. */
>>> diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
>>> index ca5676b2668e..397c3d29659c 100644
>>> --- a/net/sched/cls_api.c
>>> +++ b/net/sched/cls_api.c
>>> @@ -3483,6 +3483,8 @@ static void tcf_block_offload_inc(struct tcf_bloc=
k *block, u32 *flags)
>>>          if (*flags & TCA_CLS_FLAGS_IN_HW)
>>>                  return;
>>>          *flags |=3D TCA_CLS_FLAGS_IN_HW;
>>> +       if (tc_skip_sw(*flags))
>>> +               atomic_inc(&block->skipswcnt);
>>>          atomic_inc(&block->offloadcnt);
>>>   }
>>>
>>> @@ -3491,6 +3493,8 @@ static void tcf_block_offload_dec(struct tcf_bloc=
k *block, u32 *flags)
>>>          if (!(*flags & TCA_CLS_FLAGS_IN_HW))
>>>                  return;
>>>          *flags &=3D ~TCA_CLS_FLAGS_IN_HW;
>>> +       if (tc_skip_sw(*flags))
>>> +               atomic_dec(&block->skipswcnt);
>>>          atomic_dec(&block->offloadcnt);
>>>   }
>>>
>>> --
>>> 2.43.0
>>>