From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 0B5A947279B for ; Fri, 15 May 2026 10:51:01 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1778842263; cv=none; b=BJMvC6KrwB1Nuwit8qngyhRINbhxpg/KD+7BSKlP+EJ4zZdDCoUUiNRmTqdbuFRk6yndCtcjGEp3WWKXE1yO6fCqyhgRJbu9vKI+cjmKDiWzm7ad8jVP0X7eC/0lszMApNwlXjBcPE7o/fIkr7+IqyAldOz9LUYbXWv+t+IcrQ0= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1778842263; c=relaxed/simple; bh=/ECf90eOrhXSOkdV53PEEz07VINGrfgsW10pHNon/cE=; h=Message-ID:Date:MIME-Version:Subject:To:Cc:References:From: In-Reply-To:Content-Type; b=Kf/a5+58Ev1A4Vc9CXnnkWcqT/hbGoRs6Nq5HxVnBNWvXOY571YSbTIudElnj3oSph14fJ25xSgJyShCA9CvI1j8BFlDe4NXTT1OlzbWo3vGXtfx7tCY0mlGQYVP78Lgn3DxFz3bwhbjDh1v87gzQUOOGTqcfaoMAgsXk0us80I= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=DmRaDcq1; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="DmRaDcq1" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1778842261; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=V2GcfIXT4hgjoWK/Zjyn99ZRplmcZPLBAZBxpyE5X50=; b=DmRaDcq1vLV92WTfmRXi3DVbAxid4iP8zXf7J9ix939UPN4uBQqyQrI0rGOInE507oSmxr NyuhmuyminkXz9CWEbiZZLJd2+PQQwt/XR8Sjbe8Iklqul52eEVOb4ponzh98CarmGEhtG /IPa1ZvvhzWpgI2AzO8heoaiqCBtjiQ= Received: from mail-wm1-f71.google.com (mail-wm1-f71.google.com [209.85.128.71]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-349-wnN53nSbO6uaGLgjnfsZOQ-1; Fri, 15 May 2026 06:50:59 -0400 X-MC-Unique: wnN53nSbO6uaGLgjnfsZOQ-1 X-Mimecast-MFC-AGG-ID: wnN53nSbO6uaGLgjnfsZOQ_1778842258 Received: by mail-wm1-f71.google.com with SMTP id 5b1f17b1804b1-48fd3dbd16aso22828785e9.0 for ; Fri, 15 May 2026 03:50:59 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1778842258; x=1779447058; h=content-transfer-encoding:in-reply-to:content-language:from :references:cc:to:subject:user-agent:mime-version:date:message-id :x-gm-gg:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=V2GcfIXT4hgjoWK/Zjyn99ZRplmcZPLBAZBxpyE5X50=; b=YqvGCj1voJKsnXJDIDbp9o6P17VxJI5ExNmyPeRkvSAtiKn7224c7DWYLuA/ooUFld 5uI8dFxCKi7sCR3GGqCzJqoMalZ1fdl1ffavirbGUji89zycqf+1HhTEttx0mT+FUfPU UwtzoqgOI7S3Za7CNNkEvsF/vGIgVNrmK48qH7xt1qQGLmLVt9AWt4DWt024VoyRV9hK SL81nnWIhh96IcOZkgcMg8ycGj+N+CqSoMlA8qWcuqJCj0yjU1nQczYWfUDc7AtR6JlU FvPriyr8nf9JNtutNjA9lludO4TqffchDpgJShQwouWpGTDybrdEbloFtX46wPdk3tIY eggw== X-Forwarded-Encrypted: i=1; AFNElJ/n6eZc9W8AAWEoH1hSfdHERUDFS0O785pycubc1AsoU66kWfBzgrHuMt3itpi9F5EoH6wVvA==@lists.linux.dev X-Gm-Message-State: AOJu0YxOFCck52oHA9ycAO377dMj+HJMVT4FR7WIIsrqVMNglWntrXsN IHfHMaoCNI9j76PB3svhiRI3Gnd01F+X0msmZsu3Pq70/rPGhGLMEp42e0XbUpMaJ8GDvx2dVT6 0WeYmKym2168U+/Y0HI7/qWidzoNosHldlvldAJtq1Q8F16IIdjaeb5FeTOcfQI5j X-Gm-Gg: Acq92OFW/ZNOwruGJWL6e9PlaT4FCtzRUEgLWpjf0l371YfvPq/nNl+MCNAqrHvpYdn i7QditS47F0mrENIBI02R6Dh+o/r5rRmmPDJnfVX9Y3tYr45wl/gXprB+rpWeBvw7nrVyKqO80D OVmcYHERnNJGkmj30wWduPPKEfb81hRPq7m65XE80l4RJBZsF5r1aqnphMG/hNsM7Okfm3xshWf 0YgCagm4cASobLuDCgzvUJm6mUodxvVRaJ3EZoR6XfivYQjqB+8nZsW07GbDpu2NQms924e4eKr wuoGGskqhqJnWHST785Ctx5GGKMMQvlvY0Ix7DOs33CYfqmqiEjSmjZnhBgI7d1b5x9MLGd0AAm 40uSR+kcyHcoAGRjoF+//NwwEVUXSz+HbtqsS8KeoBjaB+oIeLQ5ywRKouA== X-Received: by 2002:a05:600c:6992:b0:48e:6f39:f7be with SMTP id 5b1f17b1804b1-48fe60ea533mr44239595e9.10.1778842258215; Fri, 15 May 2026 03:50:58 -0700 (PDT) X-Received: by 2002:a05:600c:6992:b0:48e:6f39:f7be with SMTP id 5b1f17b1804b1-48fe60ea533mr44239305e9.10.1778842257815; Fri, 15 May 2026 03:50:57 -0700 (PDT) Received: from [192.168.88.32] ([212.105.155.231]) by smtp.gmail.com with ESMTPSA id 5b1f17b1804b1-48fe537ccf5sm55416005e9.14.2026.05.15.03.50.56 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Fri, 15 May 2026 03:50:57 -0700 (PDT) Message-ID: <9f183916-6db7-48fb-bf39-dfe87a31b7d4@redhat.com> Date: Fri, 15 May 2026 12:50:56 +0200 Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH mptcp-net v4 1/2] mptcp: use sockopt_lock(release)_sock in sockopt To: gang.yan@linux.dev, MPTCP Linux Cc: Gang Yan References: <20260509-sockopt_lock-v4-0-33f3a1c4d7a0@kylinos.cn> <20260509-sockopt_lock-v4-1-33f3a1c4d7a0@kylinos.cn> <2678c535-2b51-4a04-803e-14a2e3ac348b@redhat.com> <62bce132417a3c98a6ad6aea821dba9de1002916@linux.dev> From: Paolo Abeni In-Reply-To: <62bce132417a3c98a6ad6aea821dba9de1002916@linux.dev> X-Mimecast-Spam-Score: 0 X-Mimecast-MFC-PROC-ID: v-nzOOJtTRoKg3-L-4R1V7NuBS9NzW-CR-0jJ-PfRsE_1778842258 X-Mimecast-Originator: redhat.com Content-Language: en-US Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On 5/15/26 11:39 AM, gang.yan@linux.dev wrote: > May 15, 2026 at 5:06 PM, "Paolo Abeni" wrote: >> On 5/15/26 3:27 AM, gang.yan@linux.dev wrote: >>> The subflow locking concern is valid for future BPF support. >>> >>> That said, the observation about the subflow locking gap is a valid design >>> concern. As sashiko pointed out, when the MPTCP setsockopt handler is >>> eventually called from BPF context, the MPTCP meta socket (msk) is already >>> locked externally, but the subflow sockets (ssk) are not. The internal >>> delegation via sk_setsockopt(ssk, ...) or tcp_setsockopt(ssk, ...) calls >>> sockopt_lock_sock(ssk) which skips the lock due to the per-thread >>> has_current_bpf_ctx() flag, leaving ssk unprotected. >>> >>> However, solving this is not straightforward. In BPF context we don't have a >>> good way to call lock_sock() on the subflow — as sashiko noted in first point, >>> lock_sock() unconditionally calls might_sleep(), and will cause 'sleeping in >>> atomic context' problems in BPF context. So properly locking the subflow in >>> BPF context remains an open question that needs to be addressed when MPTCP >>> BPF setsockopt support is actually introduced. >>> > Hi Paolo, > > Thanks for your reply. > >> I'm sorry for the late feedback. I think such concern is actually quite >> relevant. Note that replacing >> >> bool slow = lock_sock_fast(ssk); >> >> with: >> >> lock_sock(ssk); >> >> does not change much WRT the sleeping issue. > > I'm not questioning your point, but just to confirm — when you mentioned the > "sleeping issue", are you referring to the BPF context? Yes. > In fact, this patch addresses an issue that occurs in non-BPF scenarios like > sashiko said in [1]: > > https://sashiko.dev/#/patchset/20260420093343.16443-1-gang.yan@linux.dev I think that bit/chunk should in a separate patch targeting net. >> I *think* the only viable path in the short term is making the sockopt >> requiring subflow-level lock fail when called by BPF. >> >> In a possible long term alternative could be delegating the >> subflow-level changes to a workqueue, but that really looks invasive and >> I would really like to avoid such option. > > I agree that the short-term approach sounds reasonable for now. We could > either implement the sockopt failure for BPF calls that require subflow-level > locking, or simply add a comment to mark this limitation clearly. I suggest avoid the comment, because application/BPF programs will trip on this. Returning an error will make it clear that is not working and will prevent racy accesses. > As far as I can see, MPTCP currently doesn’t have any real-world use cases > for setting sockopt via BPF anyway. The workqueue approach also seems too complex > and invasive for the current stage, so I’d prefer to avoid it for now. > > WDYT？ I'm very fine with keeping this as simple as possible. /P