Linux kernel, Android, Golang & Recsys

之前使用Graphviz生成了一张mongoose的函数调用图，这里将讲是怎么生成的。

通过Graphvis生成mongoose的函数调用关系
基本的理论可以参考这里(http://www.ibm.com/developerworks/cn/linux/l-graphvis/), 思路如下:
1. 通过GCC的 -finstrument-functions 和 -g 选项，获得函数调用关系的地址.
2. 使用 Addr2line 将函数地址解析为函数名.
3. 精简函数跟踪数据(可以参考pvtrace, http://download.boulder.ibm.com/ibmdl/pub/software/dw/library/l-graphvis/pvtrace.zip)
4. 使用Graphvis生成图片.

具体的步骤:
1. 在我们编译mongoose的时候，加上 instrument.c 文件。
instrument.c 文件的作用是在我们执行文件的时候，会自动记录函数入口和出口的函数地址，并保存在当前目录下面的trace.txt文件里面。 注意: 在用gcc编译的时候一定要加 -finstrument-functions 和 -g 参数。

在mongoose的Makefile文件中增加instrument.c文件。
原文：

$(CC) $(LINFLAGS) mongoose.c main.c -s -o $(PROG)

改为:

$(CC) $(LINFLAGS) instrument.c mongoose.c  main.c -finstrument-functions -g -s -o $(PROG)

2. 编译mongoose.
make linux

3. 运行编译好的mongoose程序，并在当前目录生成一个trace.txt 文件。

3. 下载 pvtrace.zip 并编译。
回生成一个pvtrace程序，这个程序主要是从trace.txt和mongoose中提取函数地址对应的函数名，并生成graphivz的语法树。

4. 使用你编译好的文件pvtrace, 运行mongoose, 来获得graph.dot文件。

5. 使用Graph程序，生成图片, 前提是你必须安装graphivz
Debian/Ubuntu

sudo apt-get install graphviz.

安装完后运行:

dot -Tjpg graph.dot -o graph.jpg

6. Over.
效果图参考: http://wifihack.net/blog/2009/04/mongoose-start-function-call-use-graph/

instrument.c:

/********************************************************************
* File: instrument.c
*
* Instrumentation source — link this with your application, and
*  then execute to build trace data file (trace.txt).
*
* Author: M. Tim Jones <mtj@mtjones.com>
*
*/

#include <stdio.h>
#include <stdlib.h>

/* Function prototypes with attributes */
void main_constructor( void )
__attribute__ ((no_instrument_function, constructor));

void main_destructor( void )
__attribute__ ((no_instrument_function, destructor));

void __cyg_profile_func_enter( void *, void * )
__attribute__ ((no_instrument_function));

void __cyg_profile_func_exit( void *, void * )
__attribute__ ((no_instrument_function));

static FILE *fp;

void main_constructor( void )
{
fp = fopen( “trace.txt”, “w” );
if (fp == NULL) exit(-1);
}

void main_deconstructor( void )
{
fclose( fp );
}

void __cyg_profile_func_enter( void *this, void *callsite )
{
fprintf(fp, “E%p\n”, (int *)this);
}

void __cyg_profile_func_exit( void *this, void *callsite )
{
fprintf(fp, “X%p\n”, (int *)this);
}

Ref:
1. 用 Graphviz 可视化函数调用 http://www.ibm.com/developerworks/cn/linux/l-graphvis/

–EOF–

mongoose start function call use graph

I constructed a mongoose start function call use graphviz.

Links:

1.   Visualize function calls with Graphviz http://www.ibm.com/developerworks/library/l-graphvis/

这几天在解析ARP，ICMP 的时候, 要对数据进行移位，由于之前对unsigned 和 singned没有很好的认识，有些莫名奇怪的问题。

比如下面的代码：

#include <stdio.h>
#include <stdint.h>
#include <string.h>

#define ARRAY_SIZE(arr)        (sizeof(arr) / sizeof((arr)[0]))

int main(void)
{
char mac[6] = {0×80, 0x1E, 0×37, 0x1A, 0x3D, 0x7F};

int i = 0;
for(i = 0; i<ARRAY_SIZE(mac); i++) {
printf(“0x%X “, mac[i]);
}
printf(“\n”);

write1(mac, 6);
return 0;
}

void write1(unsigned char *buf, uint16_t buf_size)
{
int i = 0;
uint16_t cur_word = 0;

for (i = 0; i < buf_size; i += 2) {
cur_word = (uint16_t)((buf[i] <<8 ) | buf[i+1]);
printf(“0x%04X buf[i] 0x%02X  buf[i+1] 0x%02X\n”, cur_word, buf[i], buf[i+1]);
}
}

border@ubuntu:~$ uname -a
Linux ubuntu 2.6.24-21-generic #1 SMP Tue Oct 21 23:43:45 UTC 2008 i686 GNU/Linux
border@ubuntu:~$ ./hello
0xFFFFFF80 0x1E 0×37 0x1A 0x3D 0x7F
0x801E buf[i] 0×80  buf[i+1] 0x1E
0x371A buf[i] 0×37  buf[i+1] 0x1A
0x3D7F buf[i] 0x3D  buf[i+1] 0x7F

怎么打印结果不同, 在主函数里面第一个打印出来的是 0xFFFFFF80, 但是在write1函数里面就是正常的0×80了，但是你有可能说
是0x%x 和 0x%02x 的问题，于是程序改为:

char mac[6] = {0x80, 0x1E, 0x37, 0x1A, 0x3D, 0x7F};
int i = 0;
for(i = 0; iprintf("0x%02X ", mac[i]);
}
printf("\n");

就算是改为 %02X，也是输出0xFFFFFF80，和%x的结果一样。

但是如果把char mac[6] 改为 unsigned char mac[6] 主函数就没有问题。

在write1函数里面，如果你把参数unsigned char *buf 改为char *buf, 当大于0x7F的时候，也会溢出，是因为C 里面如果没有定义
类型符合的话默认为singned.

singned的取之范围是-128-127(0×80-0x7F), 所以当大于0x7F的时候会溢出。

unsigned 的取值范围是0-255(0-0xFF).

如果你要进行高低段位移的话，以后要定义为unsigned.

疑问:  为什么%02X输出的也是8个字节,而不是两个字节?

Bian Jiang

http://www.wifihack.net

–EOF–

1. linux/unix
OpenSSH 支持SOCKS4和SOCKS5, 我们可以通过参数 -D 在本地创建一个代理端口. 例如:

ssh -D  12345 myuser@remote_ssh_server

我们已经在本地创建的一个SOCKS的端口12345, 现在你可以通过修改你的IE, Firefox 中的连接方式来用代理上网。比如:

IP 改为 127.0.0.1
端口: 12345
方式改为: SOCKS5

这样我们就可以继续访问 youtube 了。
2. 如果是windows 可以使用 putty的后台命令行程序plink(http://www.chiark.greenend.org.uk/~sgtatham/putty/download.html)

plink -N username@remote.ssh.server -D 127.0.0.1:7070

其中 -N 表示不需要shell
username@remote.ssh.server 换成你ssh帐户名和主机域名或者地址
或者替换成 -load sessionname 也可以，用dreamhost上的帐号试了一下，真的很快
如果你用的是 Firefox 可以用SwitchProxy(https://addons.mozilla.org/en-US/firefox/addon/125), 非常方便的切换。
ssh -D 参数的描述:

-D [bind_address:]port
Specifies a local “dynamic” application-level port forwarding.  This works by allocating a
socket to listen to port on the local side, optionally bound to the specified bind_address.
Whenever a connection is made to this port, the connection is forwarded over the secure chan-
nel, and the application protocol is then used to determine where to connect to from the
remote machine.  Currently the SOCKS4 and SOCKS5 protocols are supported, and ssh will act as
a SOCKS server.  Only root can forward privileged ports.  Dynamic port forwardings can also
be specified in the configuration file.

IPv6 addresses can be specified with an alternative syntax: [bind_address/]port or by enclos-
ing the address in square brackets.  Only the superuser can forward privileged ports.  By
default, the local port is bound in accordance with the GatewayPorts setting.  However, an
explicit bind_address may be used to bind the connection to a specific address.  The
bind_address of “localhost” indicates that the listening port be bound for local use only,
while an empty address or ‘*’ indicates that the port should be available from all inter-
faces.

3.摘自chedong.com

ssh -qTfnN -D 7070 remotehost.

All the added options are for a ssh session that’s used for tunneling.

-q :- be very quite, we are acting only as a tunnel.
-T :- Do not allocate a pseudo tty, we are only acting a tunnel.
-f :- move the ssh process to background, as we don’t want to interact with this ssh session directly.
-N :- Do not execute remote command.
-n :- redirect standard input to /dev/null.

Links:
1. Use ssh create http proxy (http://www.linuxjournal.com/content/use-ssh-create-http-proxy)
2. http://www.chedong.com/blog/archives/001246.html
3. Putty (http://www.chiark.greenend.org.uk/~sgtatham/putty/download.html)
Bian Jiang

http://www.wifihack.net

–EOF–

本文参考: http://notes.jimlindley.com/2008/3/25/git-svn-that-works-for-me
作者: Jim Lindley

The canonical git-svn workflow that I’ve seen goes like this:
标准的git-svn工作流程参考这里:

[code lang="C"]
git svn clone
git checkout -b
...hack...hack...

git commit -a
git checkout master
git merge #NOTE: no need for --squash anymore
git svn rebase
git svn dcommit -e # -e will let you enter a commit message for SVN

[/code]

I’ve had more luck with the following workflow, when integrating changes via SVN from other team members:
我从团队的其他成员学到了更好的工作流程:

[code lang="C"]
# initial setup
git svn clone

# 99% of daily workflow
git checkout -b
...hack...hack...
git commit -a

# switch back to master, then rebase against
# any revisions in the svn repo
git checkout master
git svn rebase

# now that master is current with svn,
# sync working branch to local master
git checkout # These two are the added steps
git rebase master # which help prevent conflicts

# final upstream commit after rebasing
git checkout master
git svn rebase # one last check for new svn check ins
git merge
git svn dcommit -e

[/code]

The extra rebase step seems to do a better job of integrating your patches into the tree. Merge should do the same thing, if I’m reading the man pages right, but splitting the steps is more idiot proof (me-proof) this way.

It also keeps the master local branch from getting messy dealing with conflicts. Instead conflict is kept in the side working branch.

上面虽然很罗嗦但是做的好处在于，我们把所有的工作和从svn代码的合并都放在本地的分支上进行。这样就能保证本地主干与svn服务器进行同步，并且在本地主干没有任何修改。

更多有关git, git-svn,  git-reset 的文章:
1. git-svn workflow
2. An introduction to git-svn for Subversion/SVK users and deserters
3. Git reset in depth

2009.4.3
Bian Jiang
–EOF–