不可能发生了! 这是什么意思?

我遇到了一个有趣的运行时错误。 我认为这是某种内存泄漏。 我写了以下程序:

C代码:

 #include  #include  #include  #include  #include  #define PRECISION 4096 #define DECIMAL_POINT 1 #define NULL_TERMINATOR 1 void gatherSquares(const uint32_t limit, uint32_t ** const arr, uint32_t * const count); uint32_t inList(const uint32_t n, const uint32_t * const arr, const uint32_t count); void print_help(const char * const str); int main(int argc, char* argv[]) { uint32_t limit = 100; /* default */ uint32_t digits = 100; /* default */ uint32_t i,j,sum,bufLen,charCount,sqCount=10; uint32_t *squares; char *buffer; mpf_t irrational; if(argc > 1 && !strcmp(argv[1],"--help")) return print_help(argv[0]), EXIT_SUCCESS; if(argc > 1) sscanf(argv[1],"%u",&limit ); if(argc > 2) sscanf(argv[2],"%u",&digits); mpf_set_default_prec(PRECISION); mpf_init(irrational); charCount = digits + DECIMAL_POINT; bufLen = charCount + NULL_TERMINATOR; gatherSquares(limit,&squares,&sqCount); buffer = malloc(bufLen * sizeof *buffer); for(i=sum=0; i<=limit; ++i) { if(inList(i,squares,sqCount)) continue; mpf_sqrt_ui(irrational,i); gmp_snprintf(buffer,bufLen,"%.*Ff",charCount,irrational); for(j=0; j<digits+DECIMAL_POINT; ++j) if(buffer[j]!='.') sum += buffer[j]-'0'; } printf("%u\n",sum); return EXIT_SUCCESS; } void gatherSquares(const uint32_t limit, uint32_t ** const arr, uint32_t * const count) { uint32_t i,sq,size; size=10; *arr = malloc(size * sizeof **arr); for(i=0; (sq=i*i) size) *arr = realloc(*arr, (size*=2) * sizeof **arr); (*arr)[i] = sq; } *arr = realloc(*arr, i * sizeof **arr); *count = i; } uint32_t inList(const uint32_t n, const uint32_t * const arr, const uint32_t count) { uint32_t hi,mid,low; hi = count; /* exclusive upper bound */ low = 0; /* inclusive lower bound */ while(low n) hi = mid; else low = mid + 1; } return 0; } void print_help(const char * const str) { printf(" Usage: %s  \n",str); printf(" Calculates the digital sum of the first  digits \n"); printf(" of each irrational square root <= \n"); printf(" * limit : a decimal number\n"); printf(" : default = 100\n"); printf(" * digits : a decimal number\n"); printf(" : default = 100\n"); } 

我使用默认设置./a.out运行程序,一切都正常运行。 然后我决定尝试使用更高的搜索空间./a.out 1000来运行程序。 这产生了以下输出:

 user@comp:~/Current/Working/Directory ./a.out 1000 a.out: malloc.c:2369: sysmalloc: Assertion `(old_top == (((mbinptr) (((char *) &((av)->bins[((1) - 1) * 2])) - __builtin_offsetof (struct malloc_chunk,fd)))) && old_size == 0) || ((unsigned long) (old_size) >= (unsigned long)((((__builtin_offsetof (struct malloc_chunk, fd_nextsize))+ ((2 * (sizeof(size_t))) - 1)) & ~((2 * (sizeof(size_t))) - 1))) && ((old_top)->size & 0x1) && ((unsigned long)old_end & pagemask) == 0)' failed. Aborted (core dumped) 

我假设必须有一些内存泄漏,所以我启动了valgrind以帮助找到问题。 我对输出感到茫然:

 user@comp:~/Current/Working/Directory valgrind ./a.out 1000 ==4032== Memcheck, a memory error detector ==4032== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al. ==4032== Using Valgrind-3.8.1 and LibVEX; rerun with -h for copyright info ==4032== Command: ./ans-0080 1000 ==4032== ==4032== Invalid write of size 4 ==4032== at 0x8048920: gatherSquares (prog.c:59) ==4032== by 0x80486C4: main (prog.c:36) ==4032== Address 0x427b288 is 0 bytes after a block of size 40 alloc'd ==4032== at 0x402C418: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so) ==4032== by 0x804890F: gatherSquares (prog.c:55) ==4032== by 0x80486C4: main (prog.c:36) ==4032== --4032-- VALGRIND INTERNAL ERROR: Valgrind received a signal 11 (SIGSEGV) - exiting --4032-- si_code=1; Faulting address: 0x66656220; sp: 0x625b8a80 valgrind: the 'impossible' happened: Killed by fatal signal ==4032== at 0x380DA208: ??? (in /usr/lib/valgrind/memcheck-x86-linux) ==4032== by 0x380DAAB1: ??? (in /usr/lib/valgrind/memcheck-x86-linux) ==4032== by 0x38054133: ??? (in /usr/lib/valgrind/memcheck-x86-linux) ==4032== by 0x38054206: ??? (in /usr/lib/valgrind/memcheck-x86-linux) ==4032== by 0x38052115: ??? (in /usr/lib/valgrind/memcheck-x86-linux) ==4032== by 0x20657360: ??? sched status: running_tid=1 Thread 1: status = VgTs_Runnable ==4032== at 0x402C63E: realloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so) ==4032== by 0x8048979: gatherSquares (prog.c:61) ==4032== by 0x80486C4: main (prog.c:36) 

注意行valgrind: the 'impossible' happened: ! 发生了什么不可能的事 这是什么意思?


PS:
如果您知道该程序的编写内容,请不要发布对搜索引擎的引用以进行索引!

我怀疑这不是内存泄漏,而是你在分配缓冲区开始之前或结束之后编写,因此覆盖malloc用来维护数据结构的数据,因此malloc失败。

我无法编译您的代码进行测试,但在快速目视检查中,这看起来很可疑(在gatherSquares()中):

 if(*count > size) *arr = realloc(*arr, (size*=2) * sizeof **arr); 

您需要根据i重新分配,而不是*计数点。

这行在我的gdb中有问题

 *arr = realloc(*arr, i * sizeof **arr); 

gdb报道

*检测到glibc /run/shm/a.out:realloc():下一个大小无效:0x0000000000603230 * *