Feedback-ID: i083147f8:Fastmail
Date:   Wed, 8 Mar 2023 16:30:55 -0800
From:   Boris Burkov <boris@bur.io>
To:     sdf@google.com
Cc:     Rong Tao <rtoax@foxmail.com>, andrii@kernel.org, rongtao@cestc.cn,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Martin KaFai Lau <martin.lau@linux.dev>,
        Song Liu <song@kernel.org>, Yonghong Song <yhs@fb.com>,
        John Fastabend <john.fastabend@gmail.com>,
        KP Singh <kpsingh@kernel.org>, Hao Luo <haoluo@google.com>,
        Jiri Olsa <jolsa@kernel.org>,
        "open list:BPF [LIBRARY] (libbpf)" <bpf@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH bpf-next] libbpf: poison strlcpy()
Message-ID: <20230309003055.GA6586@zen>
References: <tencent_5695A257C4D16B4413036BA1DAACDECB0B07@qq.com>
 <Y7cbM1D2YvB9tdqg@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y7cbM1D2YvB9tdqg@google.com>
Precedence: bulk

On Thu, Jan 05, 2023 at 10:47:15AM -0800, sdf@google.com wrote:
> On 01/05, Rong Tao wrote:
> > From: Rong Tao <rongtao@cestc.cn>
> 
> > Since commit 9fc205b413b3("libbpf: Add sane strncpy alternative and use
> > it internally") introduce libbpf_strlcpy(), thus add strlcpy() to a poison
> > list to prevent accidental use of it.
> 
> > Signed-off-by: Rong Tao <rongtao@cestc.cn>
> 
> Acked-by: Stanislav Fomichev <sdf@google.com>
> 
> > ---
> >   tools/lib/bpf/libbpf_internal.h | 4 ++--
> >   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> > diff --git a/tools/lib/bpf/libbpf_internal.h
> > b/tools/lib/bpf/libbpf_internal.h
> > index 377642ff51fc..2d26ded383ca 100644
> > --- a/tools/lib/bpf/libbpf_internal.h
> > +++ b/tools/lib/bpf/libbpf_internal.h
> > @@ -20,8 +20,8 @@
> >   /* make sure libbpf doesn't use kernel-only integer typedefs */
> >   #pragma GCC poison u8 u16 u32 u64 s8 s16 s32 s64
> 
> > -/* prevent accidental re-addition of reallocarray() */
> > -#pragma GCC poison reallocarray
> > +/* prevent accidental re-addition of reallocarray()/strlcpy() */
> > +#pragma GCC poison reallocarray strlcpy

On my musl system, I believe this broke compilation, as string.h defines
strlcpy, and is included after this poisoning when compiling strset.c

FWIW, I could work around it by adding
#include <string.h>
above
#include <libbpf_internal.h>
in strset.c, since the poison doesn't apply to symbols that existed
before it ran, but this feels like a kludge, and not in the spirit of
the original poisoning patch..

I'm curious what the proper workaround should be for a libc that defines
strlcpy.

Thanks,
Boris

> 
> >   #include "libbpf.h"
> >   #include "btf.h"
> > --
> > 2.39.0
>