Re: [dpdk-dev] [PATCH] test-pmd: Fix "__BYTE_ORDER__" not defined error

["Qiu, Michael" <michael.qiu@intel.com>]

Hi all,

What about this patch?

It may be some network or mail client issue of me, so not sure this
patch posted to mail list successful.

If failed, I will re-post it later.

Thanks,
Michael

On 2014/12/3 19:28, Michael Qiu wrote:
> app/test-pmd/csumonly.c:84:5: error: "__BYTE_ORDER__" is not defined
> app/test-pmd/csumonly.c:84:23: error: "__ORDER_LITTLE_ENDIAN__" is not defined
>
> This because old gcc version, like 4.4.7, does not has these buildin macros.
>
> $ gcc -v
> Using built-in specs.
> Target: x86_64-redhat-linux
> ...
> gcc version 4.4.7 20120313 (Red Hat 4.4.7-4) (GCC)
>
> $ echo | gcc -dM -E -| grep "LITTLE"
> (none)
>
> In this situation, we should back to use the macros defined in
> <endian.h>
>
> Signed-off-by: Michael Qiu <michael.qiu@intel.com>
> ---
>  app/test-pmd/csumonly.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/app/test-pmd/csumonly.c b/app/test-pmd/csumonly.c
> index 6f43761..3fa81a2 100644
> --- a/app/test-pmd/csumonly.c
> +++ b/app/test-pmd/csumonly.c
> @@ -81,7 +81,13 @@
>  
>  /* we cannot use htons() from arpa/inet.h due to name conflicts, and we
>   * cannot use rte_cpu_to_be_16() on a constant in a switch/case */
> -#if  __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
> +#ifdef __BYTE_ORDER__
> +#define LITTLE_ENDIAN_CHECK (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)
> +#else
> +#define LITTLE_ENDIAN_CHECK (__BYTE_ORDER == __LITTLE_ENDIAN)
> +#endif
> +
> +#if LITTLE_ENDIAN_CHECK
>  #define _htons(x) ((uint16_t)((((x) & 0x00ffU) << 8) | (((x) & 0xff00U) >> 8)))
>  #else
>  #define _htons(x) (x)