Re: [PATCH] net: Fix suspicious RCU usage in bpf_sk_reuseport_detach()

From: David Howells <dhowells@redhat.com>
To: Hawkins Jiawei <yin31149@gmail.com>
Cc: dhowells@redhat.com, "David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Andrii Nakryiko <andrii@kernel.org>,
	Martin KaFai Lau <kafai@fb.com>, Song Liu <songliubraving@fb.com>,
	Yonghong Song <yhs@fb.com>,
	John Fastabend <john.fastabend@gmail.com>,
	KP Singh <kpsingh@kernel.org>,
	linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
	linux-kernel-mentees@lists.linuxfoundation.org,
	bpf@vger.kernel.org
Subject: Re: [PATCH] net: Fix suspicious RCU usage in bpf_sk_reuseport_detach()
Date: Tue, 16 Aug 2022 22:16:46 +0100	[thread overview]
Message-ID: <804153.1660684606@warthog.procyon.org.uk> (raw)
In-Reply-To: <20220816103452.479281-1-yin31149@gmail.com>

Hawkins Jiawei <yin31149@gmail.com> wrote:

> +__rcu_dereference_sk_user_data_with_flags_check(const struct sock *sk,
> +						uintptr_t flags, bool condition)

That doesn't work.  RCU_LOCKDEP_WARN() relies on anything passing on a
condition down to it to be a macro so that it can vanish the 'condition'
argument without causing an undefined symbol for 'lockdep_is_held' if lockdep
is disabled:

x86_64-linux-gnu-ld: kernel/bpf/reuseport_array.o: in function `bpf_sk_reuseport_detach':
/data/fs/linux-fs/build3/../kernel/bpf/reuseport_array.c:28: undefined reference to `lockdep_is_held'

So either __rcu_dereference_sk_user_data_with_flags_check() has to be a macro,
or we need to go with something like the first version of my patch where I
don't pass the condition through.  Do you have a preference?

David