[dpdk-dev] DPDK cpuflag check failure

[dpdk-dev] DPDK cpuflag check failure

[Zhang, Jerry]

Hi all, 

	Do you know why DPDK(shared library) constructor function rte_cpu_check_supported() needs to check some unnecessary CPU flags which is not used by current DPDK such as "AES"?
     
    On Hehalem platform with GCC4.8, using shard DPDK library, it will report AES is not supported. But actually AES instructions is not used by current DPDK.

    Thanks!

#ifndef __INTEL_COMPILER
void __attribute__ ((__constructor__))
#else
void
#endif
rte_cpu_check_supported(void)
{
	/* This is generated at compile-time by the build system */
	static const enum rte_cpu_flag_t compile_time_flags[] = {
			RTE_COMPILE_TIME_CPUFLAGS
	};
	unsigned i;
	int ret;

	for (i = 0; i < sizeof(compile_time_flags)/sizeof(compile_time_flags[0]); i++) {
		ret = rte_cpu_get_flag_enabled(compile_time_flags[i]);

		if (ret < 0) {
			fprintf(stderr,
				"ERROR: CPU feature flag lookup failed with error %d\n",
				ret);
			exit(1);
		}
		if (!ret) {
			fprintf(stderr,
			        "ERROR: This system does not support \"%s\".\n"
			        "Please check that RTE_MACHINE is set correctly.\n",
			        cpu_feature_table[compile_time_flags[i]].name);
			exit(1);
		}
	}
}

Re: [dpdk-dev] DPDK cpuflag check failure

[Neil Horman]

On Sun, Sep 28, 2014 at 07:39:57AM +0000, Zhang, Jerry wrote:
> Hi all, 
> 
> 	Do you know why DPDK(shared library) constructor function rte_cpu_check_supported() needs to check some unnecessary CPU flags which is not used by current DPDK such as "AES"?
>      
>     On Hehalem platform with GCC4.8, using shard DPDK library, it will report AES is not supported. But actually AES instructions is not used by current DPDK.
> 

The function you reference is meant to assert that all the cpu features that
were enabled during the build are supported on the processor.  I expect you
enabled the crypto test, and as such the AES instructions were enabled in the
build.  disable the examples build and you'll probably be ok.
Neil


>     Thanks!
> 
> #ifndef __INTEL_COMPILER
> void __attribute__ ((__constructor__))
> #else
> void
> #endif
> rte_cpu_check_supported(void)
> {
> 	/* This is generated at compile-time by the build system */
> 	static const enum rte_cpu_flag_t compile_time_flags[] = {
> 			RTE_COMPILE_TIME_CPUFLAGS
> 	};
> 	unsigned i;
> 	int ret;
> 
> 	for (i = 0; i < sizeof(compile_time_flags)/sizeof(compile_time_flags[0]); i++) {
> 		ret = rte_cpu_get_flag_enabled(compile_time_flags[i]);
> 
> 		if (ret < 0) {
> 			fprintf(stderr,
> 				"ERROR: CPU feature flag lookup failed with error %d\n",
> 				ret);
> 			exit(1);
> 		}
> 		if (!ret) {
> 			fprintf(stderr,
> 			        "ERROR: This system does not support \"%s\".\n"
> 			        "Please check that RTE_MACHINE is set correctly.\n",
> 			        cpu_feature_table[compile_time_flags[i]].name);
> 			exit(1);
> 		}
> 	}
> }