[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y8FpfgtgG0yxApjC@nanopsycho>
Date: Fri, 13 Jan 2023 15:23:58 +0100
From: Jiri Pirko <jiri@...nulli.us>
To: Rahul Rameshbabu <rrameshbabu@...dia.com>
Cc: netdev@...r.kernel.org, Tariq Toukan <tariqt@...dia.com>,
Gal Pressman <gal@...dia.com>,
Saeed Mahameed <saeedm@...dia.com>,
Jamal Hadi Salim <jhs@...atatu.com>,
Cong Wang <xiyou.wangcong@...il.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
David Miller <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Maxim Mikityanskiy <maxtram95@...il.com>
Subject: Re: [PATCH net v3] sch_htb: Avoid grafting on
htb_destroy_class_offload when destroying htb
Fri, Jan 13, 2023 at 01:55:29AM CET, rrameshbabu@...dia.com wrote:
>Peek at old qdisc and graft only when deleting a leaf class in the htb,
>rather than when deleting the htb itself. Do not peek at the qdisc of the
>netdev queue when destroying the htb. The caller may already have grafted a
>new qdisc that is not part of the htb structure being destroyed.
>
>This fix resolves two use cases.
>
> 1. Using tc to destroy the htb.
> - Netdev was being prematurely activated before the htb was fully
> destroyed.
> 2. Using tc to replace the htb with another qdisc (which also leads to
> the htb being destroyed).
> - Premature netdev activation like previous case. Newly grafted qdisc
> was also getting accidentally overwritten when destroying the htb.
>
>Fixes: d03b195b5aa0 ("sch_htb: Hierarchical QoS hardware offload")
>Signed-off-by: Rahul Rameshbabu <rrameshbabu@...dia.com>
>Reviewed-by: Saeed Mahameed <saeedm@...dia.com>
>Cc: Eric Dumazet <edumazet@...gle.com>
>Cc: Maxim Mikityanskiy <maxtram95@...il.com>
Reviewed-by: Jiri Pirko <jiri@...dia.com>
Powered by blists - more mailing lists