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;
Message-ID: <87fa06c9-d8a9-fda4-d069-6812605aa10b@linux.dev>
Date:   Tue, 25 Jul 2023 10:56:23 -0700
MIME-Version: 1.0
Subject: Re: [PATCH 2/4] io_uring/cmd: Introduce SOCKET_URING_OP_GETSOCKOPT
Content-Language: en-US
To:     Stanislav Fomichev <sdf@google.com>,
        Breno Leitao <leitao@debian.org>
Cc:     asml.silence@gmail.com, axboe@kernel.dk, io-uring@vger.kernel.org,
        netdev@vger.kernel.org, davem@davemloft.net, kuba@kernel.org,
        edumazet@google.com, pabeni@redhat.com,
        linux-kernel@vger.kernel.org, leit@meta.com, bpf@vger.kernel.org,
        ast@kernel.org
References: <20230724142237.358769-1-leitao@debian.org>
 <20230724142237.358769-3-leitao@debian.org> <ZL61cIrQuo92Xzbu@google.com>
 <ZL+VfRiJQqrrLe/9@gmail.com> <ZMAAMKTaKSIKi1RW@google.com>
From:   Martin KaFai Lau <martin.lau@linux.dev>
In-Reply-To: <ZMAAMKTaKSIKi1RW@google.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 7/25/23 10:02 AM, Stanislav Fomichev wrote:
> On 07/25, Breno Leitao wrote:
>> On Mon, Jul 24, 2023 at 10:31:28AM -0700, Stanislav Fomichev wrote:
>>> On 07/24, Breno Leitao wrote:
>>>> Add support for getsockopt command (SOCKET_URING_OP_GETSOCKOPT), where
>>>> level is SOL_SOCKET. This is leveraging the sockptr_t infrastructure,
>>>> where a sockptr_t is either userspace or kernel space, and handled as
>>>> such.
>>>>
>>>> Function io_uring_cmd_getsockopt() is inspired by __sys_getsockopt().
>>>
>>> We probably need to also have bpf bits in the new
>>> io_uring_cmd_getsockopt?

I also think this inconsistency behavior should be avoided.

>>
>> It might be interesting to have the BPF hook for this function as
>> well, but I would like to do it in a following patch, so, I can
>> experiment with it better, if that is OK.
> 
> We are not using io_uring, so fine with me. However, having a way to bypass
> get/setsockopt bpf might be problematic for some other heavy io_uring
> users.
> 
> Lemme CC a bunch of Meta folks explicitly. I'm not sure what that state
> of bpf support in io_uring.

We have use cases on the "cgroup/{g,s}etsockopt". It will be a surprise when the 
user moves from the syscall {g,s}etsockopt to SOCKET_URING_OP_*SOCKOPT and 
figured that the bpf handling is skipped.