Re: [PATCH v3 bpf-next 4/7] libbpf: use static const fmt string in __bpf_printk

[Dave Marchevsky <davemarchevsky@fb.com>]

On 8/28/21 1:20 AM, Dave Marchevsky wrote:   
> The __bpf_printk convenience macro was using a 'char' fmt string holder
> as it predates support for globals in libbpf. Move to more efficient
> 'static const char', but provide a fallback to the old way via
> BPF_NO_GLOBAL_DATA so users on old kernels can still use the macro.
> 
> Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
> ---
>  tools/lib/bpf/bpf_helpers.h | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/lib/bpf/bpf_helpers.h b/tools/lib/bpf/bpf_helpers.h
> index 5f087306cdfe..a1d5ec6f285c 100644
> --- a/tools/lib/bpf/bpf_helpers.h
> +++ b/tools/lib/bpf/bpf_helpers.h
> @@ -216,10 +216,16 @@ enum libbpf_tristate {
>  		     ___param, sizeof(___param));		\
>  })
>  
> +#ifdef BPF_NO_GLOBAL_DATA
> +#define BPF_PRINTK_FMT_TYPE char
> +#else
> +#define BPF_PRINTK_FMT_TYPE static const char

The reference_tracking prog test is failing as a result of this.
Specifically, it fails to load bpf_sk_lookup_test0 prog, which 
has a bpf_printk:

  47: (b4) w3 = 0
  48: (18) r1 = 0x0
  50: (b4) w2 = 7
  51: (85) call bpf_trace_printk#6
  R1 type=inv expected=fp, pkt, pkt_meta, map_key, map_value, mem, rdonly_buf, rdwr_buf

Setting BPF_NO_GLOBAL_DATA in the test results in a pass

> +#endif
> +
>  /* Helper macro to print out debug messages */
>  #define __bpf_printk(fmt, ...)				\
>  ({							\
> -	char ____fmt[] = fmt;				\
> +	BPF_PRINTK_FMT_TYPE ____fmt[] = fmt;		\
>  	bpf_trace_printk(____fmt, sizeof(____fmt),	\
>  			 ##__VA_ARGS__);		\
>  })
>