Re: [PATCH bpf-next v2 3/3] bpf: Move nf_conn extern declarations to filter.h

From: Daniel Xu <dxu@dxuuu.xyz>
To: Martin KaFai Lau <martin.lau@linux.dev>
Cc: pablo@netfilter.org, fw@strlen.de, toke@kernel.org,
	netfilter-devel@vger.kernel.org, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org, bpf@vger.kernel.org,
	ast@kernel.org, daniel@iogearbox.net, andrii@kernel.org,
	memxor@gmail.com
Subject: Re: [PATCH bpf-next v2 3/3] bpf: Move nf_conn extern declarations to filter.h
Date: Tue, 20 Sep 2022 08:20:33 -0600	[thread overview]
Message-ID: <20220920142033.3k3yuupwyb5cxqxb@kashmir.localdomain> (raw)
In-Reply-To: <dc251395-78af-2ea3-9049-3b44cb831783@linux.dev>

On Mon, Sep 19, 2022 at 10:20:47PM -0700, Martin KaFai Lau wrote:
> On 9/19/22 12:44 PM, Daniel Xu wrote:
> > We're seeing the following new warnings on netdev/build_32bit and
> > netdev/build_allmodconfig_warn CI jobs:
> > 
> >      ../net/core/filter.c:8608:1: warning: symbol
> >      'nf_conn_btf_access_lock' was not declared. Should it be static?
> >      ../net/core/filter.c:8611:5: warning: symbol 'nfct_bsa' was not
> >      declared. Should it be static?
> > 
> > Fix by ensuring extern declaration is present while compiling filter.o.
> > 
> > Fixes: 864b656f82cc ("bpf: Add support for writing to nf_conn:mark")
> > Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
> > ---
> >   include/linux/filter.h                   | 6 ++++++
> >   include/net/netfilter/nf_conntrack_bpf.h | 7 +------
> >   2 files changed, 7 insertions(+), 6 deletions(-)
> > 
> > diff --git a/include/linux/filter.h b/include/linux/filter.h
> > index 75335432fcbc..98e28126c24b 100644
> > --- a/include/linux/filter.h
> > +++ b/include/linux/filter.h
> > @@ -567,6 +567,12 @@ struct sk_filter {
> >   DECLARE_STATIC_KEY_FALSE(bpf_stats_enabled_key);
> > +extern struct mutex nf_conn_btf_access_lock;
> > +extern int (*nfct_btf_struct_access)(struct bpf_verifier_log *log, const struct btf *btf,
> > +				     const struct btf_type *t, int off, int size,
> > +				     enum bpf_access_type atype, u32 *next_btf_id,
> > +				     enum bpf_type_flag *flag);
> > +
> >   typedef unsigned int (*bpf_dispatcher_fn)(const void *ctx,
> >   					  const struct bpf_insn *insnsi,
> >   					  unsigned int (*bpf_func)(const void *,
> > diff --git a/include/net/netfilter/nf_conntrack_bpf.h b/include/net/netfilter/nf_conntrack_bpf.h
> > index d1087e4da440..24d1ccc1f8df 100644
> > --- a/include/net/netfilter/nf_conntrack_bpf.h
> > +++ b/include/net/netfilter/nf_conntrack_bpf.h
> > @@ -5,6 +5,7 @@
> >   #include <linux/bpf.h>
> >   #include <linux/btf.h>
> > +#include <linux/filter.h>
> 
> The filter.h is only needed by nf_conntrack_bpf.c?  How about moving this
> include to nf_conntrack_bpf.c.  nf_conntrack_bpf.h is included by other
> conntrack core codes.  I would prefer not to spill over unnecessary bpf
> headers to them.  The same goes for the above bpf.h and btf.h which are only
> needed in nf_conntrack_bpf.c also?

Ah yeah. Thanks for catching. Will send out a v3.

Now I'm wondering if https://include-what-you-use.org/ would work with
kernel source. Might give it a try later.

> 
> >   #include <linux/kconfig.h>
> >   #include <linux/mutex.h>
> 
> Also, is mutex.h still needed?

Nope. But forgot to send that out in v3. I'll roll it into v4 if we need
another respin or otherwise I'll send out a separate patch after.

> 
> > @@ -14,12 +15,6 @@
> >   extern int register_nf_conntrack_bpf(void);
> >   extern void cleanup_nf_conntrack_bpf(void);
> > -extern struct mutex nf_conn_btf_access_lock;
> > -extern int (*nfct_btf_struct_access)(struct bpf_verifier_log *log, const struct btf *btf,
> > -				     const struct btf_type *t, int off, int size,
> > -				     enum bpf_access_type atype, u32 *next_btf_id,
> > -				     enum bpf_type_flag *flag);
> > -
> >   #else
> >   static inline int register_nf_conntrack_bpf(void)
> 