memcheck怎麼執行

valgrind --log-file=valgrind.log --tool=memcheck --leak-check=full --show-reachable=no --workaround-gcc296-bugs=yes ./mcsample arg1 arg2

–log-file 表示輸出報告檔案，可以是相對路徑或完全路徑
–tool=memcheck 做記憶體檢測就是memcheck，要知道valgrind是一個工具集
–leak-check=full 完整檢測
–show-reachable=no 是否顯示reachable詳見記憶體洩露部分，通常是no，也可以改成yes
–workaround-gcc296-bugs=yes 如果你的gcc存在對應的bug，則要設為yes，否則有誤報
最後是被檢測程式及其引數。

memcheck報告怎麼看

先來一段意外的寫錯

int main(int argc, char *argv[])
{
    char* bigBuff = (char*)malloc[1024];
    free(bigBuff);
}

==3498== Invalid free() / delete / delete[] / realloc()
==3498==    at 0x402B06C: free (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
==3498==    by 0x8048444: main (main.cpp:19)
==3498==  Address 0x40c0500 is in the Text segment of /lib/i386-linux-gnu/libc-2.15.so
 

程式碼錯誤的將malloc()寫成了malloc[]，相當於取得了malloc函式指標後面的地址，輸出報告告訴我們這個地址位於.text段。

可以看出報告的基本格式是：

{問題描述}   
at {地址、函式名、模組或程式碼行} 
by {地址、函式名、程式碼行}
by ...{逐層依次顯示呼叫堆疊}
Address 0x???????? {描述地址的相對關係}

而報告的輸出文件整體格式則可以總結為：

都有哪些常見異常報告

記憶體洩漏

int main(int argc, char *argv[])
{
    char* bigBuff = (char*)malloc
 
(1024);
}

1,024 bytes in 1 blocks are definitely lost in loss record 1 of 1
 at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x8048414: main (main.cpp:17)

definitely lost：記憶體沒有被釋放，且沒有任何指標指向這裡。肯定洩漏了。報告給出的堆疊是記憶體被分配時的呼叫堆疊，它可以基本明確記憶體是由什麼業務邏輯建立的。
still reachable：是說記憶體沒有被釋放，儘管如此仍有指標指向，記憶體仍在使用中，這可以不算洩露。(程式退出時仍在工作的非同步系統呼叫?)
possibly lost：是說可能有洩漏，一般是有二級指標（指標的指標）等複雜情況不易於追蹤時出現。
suppressed：統計了使用valgrind的某些引數取消了特定庫的某些錯誤，會被歸結到這裡

異常釋放

int main(int argc, char *argv[])
{
    char* bigBuff = (char*)malloc(1024);
    char* offsetBuff = bigBuff + 888;
    free(offsetBuff);
}

 Invalid free() / delete / delete[] / realloc()
  at 0x402B06C: free (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
  by 0x8048461: main (main.cpp:24)
 Address 0x41f23a0 is 888 bytes inside a block of size 1,024 alloc'd
  at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
  by 0x8048444: main (main.cpp:17)

free() / delete / delete[] / realloc() 四種中的任一種，這裡是free的非法釋放。在描述地址的相對關係時，使用了一個句子，句子的格式是：Address 0x???????? is {x} bytes {inside/before/after} a block of size {y} {alloc’d/free’d}

它表示了釋放的地址與一個y長度塊的相對位置關係。如果地址位於塊前，則用before，位於塊內則用inside，塊後則是after。而最後的alloc’d代表這個y長度的塊處於有效狀態，其分配時的棧如下；而free’d代表y長度塊已刪除，其刪除時的棧如下。

所以上面的報告可以解釋為：地址0x41f23a0位於一個長度1024的有效塊內+888處，其分配時的呼叫堆疊如下。

非法讀寫

int main(int argc, char *argv[])
{
    char* bigBuff = (char*)malloc(1024);
    uint64_t* bigNum = (uint64_t*)(bigBuff+1020);
    *bigNum = 0x12345678AABBCCDD;
    printf("bigNum is %llu\n",*bigNum);
    free(bigBuff);
}

Invalid write of size 4
 at 0x8048490: main (main.cpp:19)
Address 0x41f2428 is 0 bytes after a block of size 1,024 alloc'd
 at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x8048474: main (main.cpp:17)

Invalid read of size 4
 at 0x804849B: main (main.cpp:20)
Address 0x41f2428 is 0 bytes after a block of size 1,024 alloc'd
 at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x8048474: main (main.cpp:17)

對一個記憶體區的使用超過了分配的大小時，可以觸發Invalid write/read，同時被告知長度。本例中uint64_t有8位元組長，訪問超出了4位元組。如果將bigBuff+1020改成bigBuff-20，那麼報告中會準確的告訴你Address xxx is 20 bytes before a block of …

另外一個有趣的現象是，我發現對uint64_t的非法訪問會產生2次4位元組長度非法訪問的報告，這說明了什麼？

不匹配的釋放

int main(int argc, char *argv[])
{
    int unused;
    char* bigBuff = (char*)malloc(1024);
    delete[] bigBuff;
    printf("unused=%d",unused);
}

Mismatched free() / delete / delete []
 at 0x402A8DC: operator delete[](void*) (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x80484FB: main (main.cpp:19)
Address 0x4323028 is 0 bytes inside a block of size 1,024 alloc'd
 at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x80484E4: main (main.cpp:18)

Use of uninitialised value of size 4
 at 0x416E0DB: _itoa_word (_itoa.c:195)
 by 0x417221A: vfprintf (vfprintf.c:1629)
 by 0x4178B2E: printf (printf.c:35)
 by 0x41454D2: (below main) (libc-start.c:226)

不管malloc分配後用delete還是delete[]，又或者是new[]之後粗心用delete釋放，都會得到Mismatched free() / delete / delete []報告，且報告主體內容基本一致。

使用未初始的值

上例中int unused並未賦值即被使用，得到了Use of uninitialised value of size 4的報告，這樣的問題通常不致命，但是也需要排除。

可以觀察到一個有趣情況，堆疊最後一層首次出現了 (below main)，它表示程式碼位於main函式以外被執行，也並非來自於執行緒，我還不能明確解釋這種現象，但是我做了下面這個測試：…

靜態構造和釋放

class GlobalClass
{
public:
    GlobalClass()
    {
        char* buf = (char*)malloc(10);
        *(int*)(buf+8) = 100;
        free(buf);
    }
    ~GlobalClass()
    {
        char* buf = (char*)malloc(10);
        *(int*)(buf+8) = 100;
        free(buf);
    }
    void fake(){}
} g_globalClass;

int main(int argc, char *argv[])
{
    g_globalClass.fake();
}

Invalid write of size 4
 at 0x804857B: GlobalClass::GlobalClass() (main.cpp:21)
 by 0x804850F: __static_initialization_and_destruction_0(int, int) (main.cpp:31)
 by 0x8048551: _GLOBAL__sub_I_g_globalClass (main.cpp:55)
 by 0x8048631: __libc_csu_init (in /home/jinzeyu/codelocal/build-mcsample-Desktop_Qt_5_3_GCC_32bit-Debug/mcsample)
 by 0x4060469: (below main) (libc-start.c:185)
Address 0x41f2030 is 8 bytes inside a block of size 10 alloc'd
 at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x8048571: GlobalClass::GlobalClass() (main.cpp:20)
 by 0x804850F: __static_initialization_and_destruction_0(int, int) (main.cpp:31)
 by 0x8048551: _GLOBAL__sub_I_g_globalClass (main.cpp:55)
 by 0x8048631: __libc_csu_init (in /home/jinzeyu/codelocal/build-mcsample-Desktop_Qt_5_3_GCC_32bit-Debug/mcsample)
 by 0x4060469: (below main) (libc-start.c:185)

Invalid write of size 4
 at 0x80485B9: GlobalClass::~GlobalClass() (main.cpp:27)
 by 0x4079B80: __run_exit_handlers (exit.c:78)
 by 0x4079C0C: exit (exit.c:100)
 by 0x40604DA: (below main) (libc-start.c:258)
Address 0x41f2070 is 8 bytes inside a block of size 10 alloc'd
 at 0x402BE68: malloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
 by 0x80485AF: GlobalClass::~GlobalClass() (main.cpp:26)
 by 0x4079B80: __run_exit_handlers (exit.c:78)
 by 0x4079C0C: exit (exit.c:100)
 by 0x40604DA: (below main) (libc-start.c:258)

靜態類的構造和釋放都在main之外，所以都出現了(below main)的字樣，堆疊的函式名也很好的證實了這兩個過程。這裡我聯想到了另一個問題，就是靜態構造的順序不一定按預期，強烈建議靜態物件之間不要有依賴關係。

崩潰

如果在memcheck執行你的程式過程中遇到崩潰，它依然能夠提供一些有用的資訊

--16198-- VALGRIND INTERNAL ERROR: Valgrind received a signal 11 (SIGSEGV) - exiting
--16198-- si_code=1;  Faulting address: 0x74207972;  sp: 0x6564ca5c
valgrind: the 'impossible' happened:
   Killed by fatal signal
==16198==    at 0x380C0AD4: ??? (in /usr/lib/valgrind/memcheck-x86-linux)
==16198==    by 0x380C12C5: ??? (in /usr/lib/valgrind/memcheck-x86-linux)
==16198==    by 0x38040A63: ??? (in /usr/lib/valgrind/memcheck-x86-linux)
==16198==    by 0x38040B36: ??? (in /usr/lib/valgrind/memcheck-x86-linux)
==16198==    by 0x3803EA4B: ??? (in /usr/lib/valgrind/memcheck-x86-linux)
==16198==    by 0x20202E78: ???

sched status:
  running_tid=3

然後報告中依次羅列崩潰時各執行緒所處的堆疊和執行緒的執行狀態

Thread 1: status = VgTs_WaitSys
...

Thread 2: status = VgTs_WaitSys
...

Thread 3: status = VgTs_Runnable
==16198==    at 0x402C9B4: operator new(unsigned int) (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
==16198==    by 0x437D7D3: std::string::_Rep::_S_create(unsigned int, unsigned int, std::allocator<char> const&) (in /usr/lib/i386-linux-gnu/libstdc++.so.6.0.16)
==16198==    by 0x437FBB5: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::basic_string(char const*, std::allocator<char> const&) (in /usr/lib/i386-linux-gnu/libstdc++.so.6.0.16)
==16198==    by 0x82A76A3: DataChecker::handle_data_check_resp_msg(void*) (data_checker.c:55)
==16198==    by 0x8144411: main_thread(void*) (main_thread.c:198)
==16198==    by 0x82839CF: thread_manager_start_routine(void*) (thread_manager.c:72)
==16198==    by 0x42D3D4B: start_thread (pthread_create.c:308)
==16198==    by 0x450BFDD: clone (clone.S:130)

Thread 4: status = VgTs_WaitSys
...

那麼，執行中的執行緒自然是嫌疑最大的，我們可以提取它的堆疊資訊做進一步分析。