linux-kernel - Re: [PATCH bpf-next] libbpf: Fix macro redefined

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <b09a15b7-a57e-419f-8e78-28b5ef27ca86@linux.dev>
Date: Wed, 16 Jul 2025 08:24:04 -0700
From: Yonghong Song <yonghong.song@...ux.dev>
To: Feng Yang <yangfeng59949@....com>, ast@...nel.org, daniel@...earbox.net,
 andrii@...nel.org, martin.lau@...ux.dev, eddyz87@...il.com, song@...nel.org,
 john.fastabend@...il.com, kpsingh@...nel.org, sdf@...ichev.me,
 haoluo@...gle.com, jolsa@...nel.org, memxor@...il.com
Cc: bpf@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH bpf-next] libbpf: Fix macro redefined



On 7/16/25 1:06 AM, Feng Yang wrote:
> From: Feng Yang <yangfeng@...inos.cn>
>
> When compiling a program that include <linux/bpf.h> and <bpf/bpf_helpers.h>, (For example: make samples/bpf)
> the following warning will be generated:
> In file included from tcp_dumpstats_kern.c:7:
> samples/bpf/libbpf/include/bpf/bpf_helpers.h:321:9: warning: 'bpf_stream_printk' macro redefined [-Wmacro-redefined]
>    321 | #define bpf_stream_printk(stream_id, fmt, args...)                              \
>        |         ^
> include/linux/bpf.h:3626:9: note: previous definition is here
>   3626 | #define bpf_stream_printk(ss, ...) bpf_stream_stage_printk(&ss, __VA_ARGS__)
>        |         ^
>
> Therefore, similar to bpf_vprintk,
> two underscores are added to distinguish it from bpf_stream_printk in bpf.h.
>
> Fixes: 21a3afc76a31 ("libbpf: Add bpf_stream_printk() macro")
> Signed-off-by: Feng Yang <yangfeng@...inos.cn>
> ---
>   tools/lib/bpf/bpf_helpers.h                | 2 +-
>   tools/testing/selftests/bpf/progs/stream.c | 4 ++--
>   2 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/tools/lib/bpf/bpf_helpers.h b/tools/lib/bpf/bpf_helpers.h
> index 80c028540656..56391a7bee48 100644
> --- a/tools/lib/bpf/bpf_helpers.h
> +++ b/tools/lib/bpf/bpf_helpers.h
> @@ -318,7 +318,7 @@ enum libbpf_tristate {
>   extern int bpf_stream_vprintk(int stream_id, const char *fmt__str, const void *args,
>   			      __u32 len__sz, void *aux__prog) __weak __ksym;
>   
> -#define bpf_stream_printk(stream_id, fmt, args...)				\
> +#define __bpf_stream_printk(stream_id, fmt, args...)				\

I think we should not change here. If absolutely necessary, we should change
kernel side (which is not exposed to uapi). E.g., just remove this line
   #define bpf_stream_printk(ss, ...) bpf_stream_stage_printk(&ss, __VA_ARGS__)
and directly use bpf_stream_stage_printk(&ss, ...)

The main reason is due to below in sample/bpf/Makefile:

$(obj)/%.o: $(src)/%.c
         @echo "  CLANG-bpf " $@
         $(Q)$(CLANG) $(NOSTDINC_FLAGS) $(LINUXINCLUDE) $(BPF_EXTRA_CFLAGS) \
                 -I$(obj) -I$(srctree)/tools/testing/selftests/bpf/ \
                 -I$(LIBBPF_INCLUDE) $(CLANG_SYS_INCLUDES) \
                 -D__KERNEL__ -D__BPF_TRACING__ -Wno-unused-value -Wno-pointer-sign \
                 -D__TARGET_ARCH_$(SRCARCH) -Wno-compare-distinct-pointer-types \
                 -Wno-gnu-variable-sized-type-not-at-end \
                 -Wno-address-of-packed-member -Wno-tautological-compare \
                 -Wno-unknown-warning-option $(CLANG_ARCH_ARGS) \
                 -fno-asynchronous-unwind-tables \
                 -I$(srctree)/samples/bpf/ -include asm_goto_workaround.h \
                 -O2 -emit-llvm -Xclang -disable-llvm-passes -c $< -o - | \
                 $(OPT) -O2 -mtriple=bpf-pc-linux | $(LLVM_DIS) | \
                 $(LLC) -march=bpf $(LLC_FLAGS) -filetype=obj -o $@

Here, some kernel data structure is needed for some particular architecture so
the initial from source to IR is compiled with native arch and after IR optimization
is done, it is switched to bpf.

Since we have vmlinux.h now. Maybe we can remove such a hack at all.
Also, sample/bpf is not really tested. Maybe trying to convert some
useful things to selftests and discard others and eventually remove sample/bpf?

>   ({										\
>   	static const char ___fmt[] = fmt;					\
>   	unsigned long long ___param[___bpf_narg(args)];				\
> diff --git a/tools/testing/selftests/bpf/progs/stream.c b/tools/testing/selftests/bpf/progs/stream.c
> index 35790897dc87..1d0663d56c0a 100644
> --- a/tools/testing/selftests/bpf/progs/stream.c
> +++ b/tools/testing/selftests/bpf/progs/stream.c
> @@ -29,7 +29,7 @@ int stream_exhaust(void *ctx)
>   	/* Use global variable for loop convergence. */
>   	size = 0;
>   	bpf_repeat(BPF_MAX_LOOPS) {
> -		if (bpf_stream_printk(BPF_STDOUT, _STR) == -ENOSPC && size == 99954)
> +		if (__bpf_stream_printk(BPF_STDOUT, _STR) == -ENOSPC && size == 99954)
>   			return 0;
>   		size += sizeof(_STR) - 1;
>   	}
> @@ -72,7 +72,7 @@ SEC("syscall")
>   __success __retval(0)
>   int stream_syscall(void *ctx)
>   {
> -	bpf_stream_printk(BPF_STDOUT, "foo");
> +	__bpf_stream_printk(BPF_STDOUT, "foo");
>   	return 0;
>   }
>