为什么我可以在/ proc / pid / maps输出中看到几个相同的段? [英] Why I can see the several same segments in the /proc/pid/maps output?
问题描述
测试是在32位Linux上
Test is on the 32 bit Linux
代码如下:
int foo(int a, int b)
{
int c = a + b;
return c;
}
int main()
{
int e = 0;
int d = foo(1, 2);
printf("%d\n", d);
scanf("%d", &e);
return 0;
}
当我使用cat / proc / pid / maps查看内存布局时,似乎我可以看到我的代码和库中有三个
text 段。
and when I use cat /proc/pid/maps to see the memory layout, it seems that I can see three text segment for my code and the library.
ubuntu% cat /proc/2191/maps
08048000-08049000 r-xp 00000000 08:01 1467306 /home/shuai/work/asm/test1
08049000-0804a000 r--p 00000000 08:01 1467306 /home/shuai/work/asm/test1
0804a000-0804b000 rw-p 00001000 08:01 1467306 /home/shuai/work/asm/test1
09137000-09158000 rw-p 00000000 00:00 0 [heap]
b75c6000-b75c7000 rw-p 00000000 00:00 0
b75c7000-b776b000 r-xp 00000000 08:01 3149924 /lib/i386-linux-gnu/libc-2.15.so
b776b000-b776d000 r--p 001a4000 08:01 3149924 /lib/i386-linux-gnu/libc-2.15.so
b776d000-b776e000 rw-p 001a6000 08:01 3149924 /lib/i386-linux-gnu/libc-2.15.so
b776e000-b7771000 rw-p 00000000 00:00 0
b7780000-b7784000 rw-p 00000000 00:00 0
b7784000-b7785000 r-xp 00000000 00:00 0 [vdso]
b7785000-b77a5000 r-xp 00000000 08:01 3149914 /lib/i386-linux-gnu/ld-2.15.so
b77a5000-b77a6000 r--p 0001f000 08:01 3149914 /lib/i386-linux-gnu/ld-2.15.so
b77a6000-b77a7000 rw-p 00020000 08:01 3149914 /lib/i386-linux-gnu/ld-2.15.so
bfd47000-bfd68000 rw-p 00000000 00:00 0 [stack]
有人可以给我一些关于这个问题的指南吗?非常感谢!
Could any one give me some guide about this issue? Thank you a lot!
推荐答案
请注意第3列(起始偏移)和2(权限)中的值。实际上你有相同的部分映射两次,在你的二进制文件的第1行和第2行,但是,在第3行,它是不同的。允许多次分别映射同一个文件;不同的系统可以跳过将其合并到一个VM映射条目中,因此它可以反映映射历史记录而不是当前状态jist。
Please mind the values in columns 3 (starting offset) and 2 (permissions). Really you have the same part mapped twice, in lines 1 and 2 for your binary file, but, in line 3, it's different. It's permitted to map the same file separately multiple times; different systems could skip merging this into one VM map entry, so it could reflect mapping history but not the current state jist.
如果您看到库映射,您可以轻松找到任何库分别映射的法则:
If you see at library mappings you could easily find the law that any library is mapped separately:
- 具有读取和执行权限:不应更改的主要代码。
- 具有读取权限:不允许代码的常量数据区。
- 具有读写权限:它结合了非常量数据区和重定位表共享对象。
具有相同的起始4K二进制文件区域映射两次可以使用RTLD逻辑来解释,该逻辑与任意库逻辑到期不同引导需求。我并不认为它如此重要,因此它可能很容易因平台细节而异。
Having the same starting 4K binary file area mapped twice could be explained with RTLD logic which differs from an arbitrary library logic due to bootstrapping needs. I don't treat it so important, more so it could easily differ on platform specifics.
这篇关于为什么我可以在/ proc / pid / maps输出中看到几个相同的段?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!