How to cross compile a c program to aarch64 with cpuid.h?

Question

I'm trying to cross-compile a simple C program to aarch64 (arm64) from a 64bit Ubuntu Linux. Can someone please help me why i'm getting this error.

It says 'cpuid.h' is not found. I've tried compiling it on the 64bit linux, it works fine. But when using aarch64-linux-gnu-gcc it is giving errors.

I'm getting the following error.

aarch64-linux-gnu-gcc -O1 -fno-stack-protector  -march=armv8-a  test.c -o test
test.c:4:10: fatal error: cpuid.h: No such file or directory
    4 | #include <cpuid.h>
      |          ^~~~~~~~~
compilation terminated.

The contents of test.c:

#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <cpuid.h>

// Requires that the user input the CPUID,
// plus the bytes "3" and "Q";

void succeed(char* string) {
    printf("Yes, %s is correct!\n", string);
    exit(0);
}

void fail(char* string) {
    printf("No, %s is not correct.\n", string);
    exit(1);
}

void shift_int_to_char(int i, char* buff) {
    buff[0] = (i) & 0xFF;
    buff[1] = (i >> 8) & 0xFF;
    buff[2] = (i >> 16) & 0xFF;
    buff[3] = (i >> 24) & 0xFF;
}

int main(int argc, char** argv) {

    if (argc != 2) {
        printf("Need exactly one argument.\n");
        return -1;
    }

    unsigned int eax, ebx, ecx, edx;
    char* buff = malloc(sizeof(char) * 15);
    __get_cpuid(0, &eax, &ebx, &ecx, &edx);
    shift_int_to_char(ebx, buff);
    shift_int_to_char(edx, buff + 4);
    shift_int_to_char(ecx, buff + 8);
    buff[12] = '3';
    buff[13] = 'Q';
    buff[14] = '\0';
    
    int correct = (strcmp(buff, argv[1]) == 0);
    free(buff);

    if (correct) {
        succeed(argv[1]);
    } else {
        fail(argv[1]);
    }
}

Answer 1

As explained in the comments, you need to port this program to the Aarch64 architecture, you cannot just compile the code as is. The features implemented in your SoC are exposed through the various AArch64 feature system registers or instruction set attribute registers, for example ID_AA64PFR1_EL1.

By convention, registers with names terminating with _EL1 cannot be accessed from a user-mode program running at EL0. Some support in the Operating System (running at EL1) is therefore required for reading those registers from a user-mode program - I used a Linux 5.15.0 kernel for the purpose of this answer.

On a recent Linux kernel, this can be achieved by using the HWCAP_CPUID API available in hwcaps - more details in this article , ARM64 CPU Feature Registers.

We can compile/execute the example code provided in Appendix I:

/*
 * Sample program to demonstrate the MRS emulation ABI.
 *
 * Copyright (C) 2015-2016, ARM Ltd
 *
 * Author: Suzuki K Poulose <[email protected]>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 */

#include <asm/hwcap.h>
#include <stdio.h>
#include <sys/auxv.h>

#define get_cpu_ftr(id) ({                                    \
              unsigned long __val;                            \
              asm("mrs %0, "#id : "=r" (__val));              \
              printf("%-20s: 0x%016lx\n", #id, __val);        \
      })

int main(void)
{

      if (!(getauxval(AT_HWCAP) & HWCAP_CPUID)) {
              fputs("CPUID registers unavailable\n", stderr);
              return 1;
      }

      get_cpu_ftr(ID_AA64ISAR0_EL1);
      get_cpu_ftr(ID_AA64ISAR1_EL1);
      get_cpu_ftr(ID_AA64MMFR0_EL1);
      get_cpu_ftr(ID_AA64MMFR1_EL1);
      get_cpu_ftr(ID_AA64PFR0_EL1);
      get_cpu_ftr(ID_AA64PFR1_EL1);
      get_cpu_ftr(ID_AA64DFR0_EL1);
      get_cpu_ftr(ID_AA64DFR1_EL1);

      get_cpu_ftr(MIDR_EL1);
      get_cpu_ftr(MPIDR_EL1);
      get_cpu_ftr(REVIDR_EL1);

#if 0
      /* Unexposed register access causes SIGILL */
      get_cpu_ftr(ID_MMFR0_EL1);
#endif

      return 0;
}
 

Execution:

gcc -o cpu-feature-registers  cpu-feature-registers.c
./cpu-feature-registers
ID_AA64ISAR0_EL1    : 0x0000000000010000
ID_AA64ISAR1_EL1    : 0x0000000000000000
ID_AA64MMFR0_EL1    : 0x00000111ff000000
ID_AA64MMFR1_EL1    : 0x0000000000000000
ID_AA64PFR0_EL1     : 0x0000000000000011
ID_AA64PFR1_EL1     : 0x0000000000000000
ID_AA64DFR0_EL1     : 0x0000000000000006
ID_AA64DFR1_EL1     : 0x0000000000000000
MIDR_EL1            : 0x00000000410fd034
MPIDR_EL1           : 0x0000000080000000
REVIDR_EL1          : 0x0000000000000000

You now just need to identify the feature/instructions set register(s) that will provide you with the same kind of information provided by your x86_64 code.