內存泄漏檢測工具Valgrind

1概述

1.1 介紹

Valgrind是一套Linux下,開放源代碼(GPL V2)的仿真調試工具的集合。Valgrind由內核(core)以及基於內核的其餘調試工具組成。內核相似於一個框架(framework),它模擬了一個CPU環境,並提供服務給其餘工具;而其餘工具則相似於插件 (plug-in),利用內核提供的服務完成各類特定的內存調試任務。Valgrind的體系結構以下圖所示:算法

 

圖1數組

1.2 工具

Valgrind的最新版是3.11.0,它通常包含下列工具: 緩存

1.Memcheck 多線程

    最經常使用的工具,用來檢測程序中出現的內存問題,全部對內存的讀寫都會被檢測到,一切對malloc()/free()/new/delete的調用都會被捕獲。因此,它能檢測如下問題: 框架

    對未初始化內存的使用; socket

    讀/寫釋放後的內存塊; ide

    讀/寫超出malloc分配的內存塊; 函數

    讀/寫不適當的棧中內存塊; 工具

    內存泄漏,指向一塊內存的指針永遠丟失; 優化

    不正確的malloc/free或new/delete匹配; 

    memcpy()相關函數中的dst和src指針重疊。 

2.Callgrind 

    和gprof相似的分析工具,但它對程序的運行觀察更是入微,能給咱們提供更多的信息。和gprof不一樣,它不須要在編譯源代碼時附加特殊選項,但加上調試選項是推薦的。Callgrind收集程序運行時的一些數據,創建函數調用關係圖,還能夠有選擇地進行cache模擬。在運行結束時,它會把分析數據寫入一個文件。callgrind_annotate能夠把這個文件的內容轉化成可讀的形式。 

3.Cachegrind 

    Cache分析器,它模擬CPU中的一級緩存I1,Dl和二級緩存,可以精確地指出程序中cache的丟失和命中。若是須要,它還可以爲咱們提供cache丟失次數,內存引用次數,以及每行代碼,每一個函數,每一個模塊,整個程序產生的指令數。這對優化程序有很大的幫助。 

4.Helgrind 

    它主要用來檢查多線程程序中出現的競爭問題。Helgrind尋找內存中被多個線程訪問,而又沒有一向加鎖的區域,這些區域每每是線程之間失去同步的地方,並且會致使難以發掘的錯誤。Helgrind實現了名爲「Eraser」的競爭檢測算法,並作了進一步改進,減小了報告錯誤的次數。不過,Helgrind仍然處於實驗階段。 

5.Massif 

    堆棧分析器,它能測量程序在堆棧中使用了多少內存,告訴咱們堆塊,堆管理塊和棧的大小。Massif能幫助咱們減小內存的使用,在帶有虛擬內存的現代系統中,它還可以加速咱們程序的運行,減小程序停留在交換區中的概率。 

此外,lackey和nulgrind也會提供。Lackey是小型工具,不多用到;Nulgrind只是爲開發者展現如何建立一個工具。

1.3 原理

Memcheck 可以檢測出內存問題,關鍵在於其創建了兩個全局表。Valid-Value 表

對於進程的整個地址空間中的每個字節(byte),都有與之對應的 8 個 bits;對於CPU的每一個寄存器,也有一個與之對應的bit向量。這些bits負責記錄該字節或者寄存器值是否具備有效的、已初始化的值。

Valid-Address 表

對於進程整個地址空間中的每個字節(byte),還有與之對應的1個bit,負責記錄該地址是否可以被讀寫。

檢測原理:

當要讀寫內存中某個字節時,首先檢查這個字節對應的 A bit。若是該A bit顯示該位置是無效位置,memcheck則報告讀寫錯誤。

內核(core)相似於一個虛擬的 CPU 環境,這樣當內存中的某個字節被加載到真實的 CPU 中時,該字節對應的 V bit 也被加載到虛擬的 CPU 環境中。一旦寄存器中的值,被用來產生內存地址,或者該值可以影響程序輸出,則 memcheck 會檢查對應的V bits,若是該值還沒有初始化,則會報告使用未初始化內存錯誤。

2 安裝使用

2.1安裝

從官網http://www.valgrind.org下載最新版本(當前3.11)

#tar xvf valgrind-3.11.1.tar.bz2
#cd valgrind-3.11.1
#./configure --prefix=/usr/local/valgrind--指定安裝目錄
#make
#make install

2.2 命令介紹

用法:valgrind[options] prog-and-args [options]: 經常使用選項,適用於全部Valgrind工具

  1. -tool=<name> 最經常使用的選項。運行 valgrind中名爲toolname的工具。默認memcheck。
  2. h –help 顯示幫助信息。
  3. -version 顯示valgrind內核的版本,每一個工具都有各自的版本。
  4. q –quiet 安靜地運行,只打印錯誤信息。
  5. v –verbose 更詳細的信息, 增長錯誤數統計。
  6. -trace-children=no|yes 跟蹤子線程? [no]
  7. -track-fds=no|yes 跟蹤打開的文件描述?[no]
  8. -time-stamp=no|yes 增長時間戳到LOG信息? [no]
  9. -log-fd=<number> 輸出LOG到描述符文件 [2=stderr]
  10. -log-file=<file> 將輸出的信息寫入到filename.PID的文件裏,PID是運行程序的進行ID
  11. -log-file-exactly=<file> 輸出LOG信息到 file
  12. -log-file-qualifier=<VAR> 取得環境變量的值來作爲輸出信息的文件名。 [none]
  13. -log-socket=ipaddr:port 輸出LOG到socket ,ipaddr:port

LOG信息輸出:

  1. -xml=yes 將信息以xml格式輸出,只有memcheck可用
  2. -num-callers=<number> show <number> callers in stack traces [12]
  3. -error-limit=no|yes 若是太多錯誤,則中止顯示新錯誤? [yes]
  4. -error-exitcode=<number> 若是發現錯誤則返回錯誤代碼 [0=disable]
  5. -db-attach=no|yes 當出現錯誤,valgrind會自動啓動調試器gdb。[no]
  6. -db-command=<command> 啓動調試器的命令行選項[gdb -nw %f %p]

適用於Memcheck工具的相關選項:

  1. -leak-check=no|summary|full 要求對leak給出詳細信息? [summary]
  2. -leak-resolution=low|med|high how much bt merging in leak check [low]
  3. -show-reachable=no|yes show reachable blocks in leak check? [no]

3 應用實踐

    下面經過介紹幾個範例來講明如何使用Memcheck (其餘工具暫不涉及,感興趣能夠交流),示例僅供參考,更多用途可在實際應用中不斷探索。

3.1數組越界/內存未釋放

#include<stdlib.h>

void k(void)

{

int *x = malloc(8 * sizeof(int));

x[9] = 0;               //數組下標越界

}                         //內存未釋放

 

int main(void)

{

    k();

return 0;

}

1)編譯程序test.c

gcc -Wall test.c -g -o test#Wall提示全部告警,-g gdb,-o輸出

2)使用Valgrind檢查程序BUG

valgrind --tool=memcheck --leak-check=full ./test

#--leak-check=full 全部泄露檢查

3) 運行結果以下:

==2989== Memcheck, a memory error detector

==2989== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward

et al.

==2989== Using Valgrind-3.8.1 and LibVEX; rerun with -h for

copyright info

==2989== Command: ./test

==2989==

==2989==  Invalid write of size 4

==2989==    at 0x4004E2: k (test.c:5)

==2989==    by 0x4004F2: main (test.c:10)

==2989==  Address 0x4c27064 is 4 bytes after a block of size 32 alloc'd

==2989==    at 0x4A06A2E: malloc (vg_replace_malloc.c:270)

==2989==    by 0x4004D5: k (test.c:4)

==2989==    by 0x4004F2: main (test.c:10)

==2989==

==2989==

==2989== HEAP SUMMARY:

==2989==     in use at exit: 32 bytes in 1 blocks

==2989==   total heap usage: 1 allocs, 0 frees, 32 bytes allocated

==2989==

==2989== 32 bytes in 1 blocks are definitely lost in loss record 1

of 1

==2989==    at 0x4A06A2E: malloc (vg_replace_malloc.c:270)

==2989==    by 0x4004D5: k (test.c:4)

==2989==    by 0x4004F2: main (test.c:10)

==2989==

==2989== LEAK SUMMARY:

==2989==    definitely lost: 32 bytes in 1 blocks

==2989==    indirectly lost: 0 bytes in 0 blocks

==2989==      possibly lost: 0 bytes in 0 blocks

==2989==    still reachable: 0 bytes in 0 blocks

==2989==suppressed: 0 bytes in 0 blocks

==2989==

==2989== For counts of detected and suppressed errors, rerun with: -v

==2989== ERROR SUMMARY: 2 errors from 2 contexts

(suppressed: 6 from 6)

3.2內存釋放後讀寫

#include <stdio.h>

#include <stdlib.h>

 

int main(void)

{

char *p = malloc(1);    //分配

*p = 'a';

 

char c = *p;

 

printf("\n [%c]\n",c);

 

free(p);         //釋放

    c = *p;         //取值

return 0;

}

1)編譯程序t2.c

gcc -Wall t2.c -g -o t2 

2)使用Valgrind檢查程序BUG

valgrind --tool=memcheck --leak-check=full ./t2

3) 運行結果以下:

      ==3058== Memcheck, a memory error detector

==3058== Copyright (C) 2002-2012, and GNU GPL'd, by Julian

Seward et al.

==3058== Using Valgrind-3.8.1 and LibVEX; rerun with -h

for copyright info

==3058== Command: ./t2

==3058==

 

      [a]

==3058== Invalid read of size 1

==3058==    at 0x4005A3: main (t2.c:14)

==3058==  Address 0x4c27040 is 0 bytes inside a block of size

1 free'd

==3058==    at 0x4A06430: free (vg_replace_malloc.c:446)

==3058==    by 0x40059E: main (t2.c:13)

==3058==

==3058==

==3058== HEAP SUMMARY:

==3058==     in use at exit: 0 bytes in 0 blocks

==3058==   total heap usage: 1 allocs, 1 frees, 1 bytes allocated

==3058==

==3058== All heap blocks were freed -- no leaks are possible

==3058==

==3058== For counts of detected and suppressed errors, rerun with:

 -v

==3058== ERROR SUMMARY: 1 errors from 1 contexts

(suppressed: 6 from 6)

從上輸出內容能夠看到,Valgrind檢測到無效的讀取操做而後輸出「Invalid read of size 1」。

3.3無效讀寫

#include <stdio.h>

#include <stdlib.h>

 

int main(void)

{

   char *p = malloc(1);    //分配1字節

   *p = 'a';

   char c = *(p+1);         //地址加1

   printf("\n [%c]\n",c); 

   free(p);

   return 0;

}

1)編譯程序t3.c

gcc -Wall t3.c -g -o t3 

2)使用Valgrind檢查程序BUG

valgrind --tool=memcheck --leak-check=full ./t3

3) 運行結果以下:

==3128== Memcheck, a memory error detector

==3128== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al.

==3128== Using Valgrind-3.8.1 and LibVEX; rerun with -h for copyright info

==3128== Command: ./t3

==3128==

==3128==  Invalid read of size 1        #無效讀取

==3128==at 0x400579: main (t3.c:9)

==3128==Address 0x4c27041 is 0 bytes after a block of size 1 alloc'd

==3128==at 0x4A06A2E: malloc (vg_replace_malloc.c:270)

==3128==by 0x400565: main (t3.c:6)

==3128==

 []

==3128==

==3128== HEAP SUMMARY:

==3128==in use at exit: 0 bytes in 0 blocks

==3128==total heap usage: 1 allocs, 1 frees, 1 bytes allocated

==3128==

==3128== All heap blocks were freed -- no leaks are possible

==3128==

==3128== For counts of detected and suppressed errors, rerun with: -v

==3128== ERROR SUMMARY: 1 errors from 1 contexts

(suppressed: 6 from 6)

3.4內存泄露

#include <stdio.h>

#include <stdlib.h>

 

int main(void)

{

int *p = malloc(1);

*p = 'x';

char c = *p;

printf("%c\n",c);        //申請後未釋放

    return 0;

}

1)編譯程序t4.c

gcc -Wall t4.c -g -o t4 

2)使用Valgrind檢查程序BUG

valgrind --tool=memcheck --leak-check=full ./t4

3) 運行結果以下:

==3221== Memcheck, a memory error detector

==3221== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al.

==3221== Using Valgrind-3.8.1 and LibVEX; rerun with -h for copyright info

==3221== Command: ./t4

==3221==

==3221== Invalid write of size 4

==3221==at 0x40051E: main (t4.c:7)

==3221==Address 0x4c27040 is 0 bytes inside a block of size 1 alloc'd

==3221==at 0x4A06A2E: malloc (vg_replace_malloc.c:270)

==3221==by 0x400515: main (t4.c:6)

==3221==

==3221== Invalid read of size 4

==3221==at 0x400528: main (t4.c:8)

==3221==Address 0x4c27040 is 0 bytes inside a block of size 1 alloc'd

==3221==at 0x4A06A2E: malloc (vg_replace_malloc.c:270)

==3221==by 0x400515: main (t4.c:6)

==3221==

x

==3221==

==3221== HEAP SUMMARY:

==3221==in use at exit: 1 bytes in 1 blocks

==3221==total heap usage: 1 allocs, 0 frees, 1 bytes allocated

==3221==

==3221== 1 bytes in 1 blocks are definitely lost in loss record 1 of 1

==3221==at 0x4A06A2E: malloc (vg_replace_malloc.c:270)

==3221==by 0x400515: main (t4.c:6)

==3221==

==3221== LEAK SUMMARY:

==3221==definitely lost: 1 bytes in 1 blocks

==3221==indirectly lost: 0 bytes in 0 blocks

==3221==      possibly lost: 0 bytes in 0 blocks

==3221==still reachable: 0 bytes in 0 blocks

==3221==         suppressed: 0 bytes in 0 blocks

==3221==

==3221== For counts of detected and suppressed errors, rerun with: -v

==3221== ERROR SUMMARY: 3 errors from 3 contexts

(suppressed: 6 from 6)

從檢查結果看,能夠發現內存泄露。

3.5內存屢次釋放

#include <stdio.h>

#include <stdlib.h>

int main(void) 

    char *p;

    p=(char *)malloc(100);    

    if(p)

        printf("Memory Allocated at: %s/n",p); 

    else

        printf("Not Enough Memory!/n"); 

    free(p);                           //重複釋放

    free(p);

    free(p);

    return 0;

1)編譯程序t5.c

gcc -Wall t5.c -g -o t5 

2)使用Valgrind檢查程序BUG

valgrind --tool=memcheck --leak-check=full ./t5

3) 運行結果以下:

==3294== Memcheck, a memory error detector

==3294== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward

et al.

==3294== Using Valgrind-3.8.1 and LibVEX; rerun with -h for

copyright info

==3294== Command: ./t5

==3294==

==3294== Conditional jump or move depends on uninitialised value(s)

==3294==    at 0x3CD4C47E2C: vfprintf (in /lib64/libc-2.12.so)

==3294==    by 0x3CD4C4F189: printf (in /lib64/libc-2.12.so)

==3294==    by 0x400589: main (t5.c:9)

==3294==

==3294== Invalid free() / delete / delete[] / realloc()

==3294==    at 0x4A06430: free (vg_replace_malloc.c:446)

==3294==    by 0x4005B5: main (t5.c:13)

==3294==  Address 0x4c27040 is 0 bytes inside a block of size

100 free'd

==3294==    at 0x4A06430: free (vg_replace_malloc.c:446)

==3294==    by 0x4005A9: main (t5.c:12)

==3294==

==3294== Invalid free() / delete / delete[] / realloc()

==3294==    at 0x4A06430: free (vg_replace_malloc.c:446)

==3294==    by 0x4005C1: main (t5.c:14)

==3294==  Address 0x4c27040 is 0 bytes inside a block of size

100 free'd

==3294==    at 0x4A06430: free (vg_replace_malloc.c:446)

==3294==    by 0x4005A9: main (t5.c:12)

==3294==

Memory Allocated at: /n==3294==

==3294== HEAP SUMMARY:

==3294==     in use at exit: 0 bytes in 0 blocks

==3294==   total heap usage: 1 allocs, 3 frees, 100 bytes allocated

從上面的輸出能夠看到(標註), 該功能檢測到咱們對同一個指針調用了3次釋放內存操做。

3.6內存動態管理

常見的內存分配方式分三種:靜態存儲,棧上分配,堆上分配。全局變量屬於靜態存儲,它們是在編譯時就被分配了存儲空間,函數內的局部變量屬於棧上分配,而最靈活的內存使用方式當屬堆上分配,也叫作內存動態分配了。經常使用的內存動態分配函數包括:malloc, alloc, realloc, new等,動態釋放函數包括free, delete。

一旦成功申請了動態內存,咱們就須要本身對其進行內存管理,而這又是最容易犯錯誤的。下面的一段程序,就包括了內存動態管理中常見的錯誤。

#include <stdio.h>

#include <stdlib.h>

int main(int argc,char *argv[])

{

int i;

char* p = (char*)malloc(10);

char* pt=p;

for(i = 0;i < 10;i++)

    {

p[i] = 'z';

    }

free(p);

pt[1] = 'x';

free(pt);

return 0;

}

1)編譯程序t6.c

gcc -Wall t6.c -g -o t6 

2)使用Valgrind檢查程序BUG

valgrind --tool=memcheck --leak-check=full ./t6

3) 運行結果以下:

==3380== Memcheck, a memory error detector

==3380== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al.

==3380== Using Valgrind-3.8.1 and LibVEX; rerun with -h for copyright info

==3380== Command: ./t6

==3380==

==3380==  Invalid write of size 1

==3380==at 0x40055C: main (t6.c:14)

==3380==Address 0x4c27041 is 1 bytes inside a block of size 10 free'd

==3380==at 0x4A06430: free (vg_replace_malloc.c:446)

==3380==by 0x400553: main (t6.c:13)

==3380==

==3380==  Invalid free() / delete / delete[] / realloc()

==3380==at 0x4A06430: free (vg_replace_malloc.c:446)

==3380==by 0x40056A: main (t6.c:15)

==3380==Address 0x4c27040 is 0 bytes inside a block of size 10 free'd

==3380==at 0x4A06430: free (vg_replace_malloc.c:446)

==3380==by 0x400553: main (t6.c:13)

==3380==

==3380==

==3380== HEAP SUMMARY:

==3380==in use at exit: 0 bytes in 0 blocks

==3380==total heap usage: 1 allocs, 2 frees, 10 bytes allocated

   申請內存在使用完成後就要釋放。若是沒有釋放,或少釋放了就是內存泄露;多釋放也會產生問題。上述程序中,指針p和pt指向的是同一塊內存,卻被前後釋放兩次。系統會在堆上維護一個動態內存鏈表,若是被釋放,就意味着該塊內存能夠繼續被分配給其餘部分,若是內存被釋放後再訪問,就可能覆蓋其餘部分的信息,這是一種嚴重的錯誤,上述程序第14行中就在釋放後仍然寫這塊內存。

    輸出結果顯示,第13行分配和釋放函數不一致;第14行發生非法寫操做,也就是往釋放後的內存地址寫值;第15行釋放內存函數無效。

相關文章
相關標籤/搜索