Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 1 Aug 2023 17:53:38 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     "Lin Ma" <linma@zju.edu.cn>
Cc:     "Leon Romanovsky" <leon@kernel.org>, davem@davemloft.net,
        edumazet@google.com, pabeni@redhat.com, fw@strlen.de,
        yang.lee@linux.alibaba.com, jgg@ziepe.ca, markzhang@nvidia.com,
        phaddad@nvidia.com, yuancan@huawei.com, ohartoov@nvidia.com,
        chenzhongjin@huawei.com, aharonl@nvidia.com,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-rdma@vger.kernel.org
Subject: Re: [PATCH net v1 1/2] netlink: let len field used to parse
 type-not-care nested attrs
Message-ID: <20230801175338.74bc39c2@kernel.org>
In-Reply-To: <385b9766.f63d7.189b3a33c6b.Coremail.linma@zju.edu.cn>
References: <20230731121247.3972783-1-linma@zju.edu.cn>
        <20230731120326.6bdd5bf9@kernel.org>
        <20230801081117.GA53714@unreal>
        <20230801105726.1af6a7e1@kernel.org>
        <385b9766.f63d7.189b3a33c6b.Coremail.linma@zju.edu.cn>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Wed, 2 Aug 2023 08:26:06 +0800 (GMT+08:00) Lin Ma wrote:
> This is true. Actually, those check missing codes are mostly old codes and
> modern netlink consumers will choose the general netlink interface which
> can automatically get attributes description from YAML and never need to
> do things like *manual parsing* anymore.
> 
> However, according to my practice in auditing the code, I found there are
> some general netlink interface users confront other issues like choosing
> GENL_DONT_VALIDATE_STRICT without thinking or forgetting add a new
> nla_policy when introducing new attributes.
> 
> To this end, I'm currently writing a simple documentation about Netlink
> interface best practices for the kernel developer (the newly coming docs
> are mostly about the user API part). 

Keep in mind that even most of the genetlink stuff is pretty old
at this stage. ethtool is probably the first reasonably modern family.
But do send docs, we'll review and go from there :)