日历
| |||||||||
| 日 | 一 | 二 | 三 | 四 | 五 | 六 | |||
| 1 | 2 | 3 | |||||||
| 4 | 5 | 6 | 7 | 8 | 9 | 10 | |||
| 11 | 12 | 13 | 14 | 15 | 16 | 17 | |||
| 18 | 19 | 20 | 21 | 22 | 23 | 24 | |||
| 25 | 26 | 27 | 28 | 29 | 30 | 31 | |||
搜索标题
统计信息
- 访问量: 1257
- 日志数: 7
- 建立时间: 2005-10-16
- 更新时间: 2007-10-30
RSS订阅
我的最新日志
-
用对象关系去思考 LEFT JOIN 的用法
2007-10-30
-
用C来实现base64编码与解码
2007-7-31
base64编码是邮件传输中比较常用的一种简单编码，它把8bit的编码，转换成6bit有效的编码，头两个bit置零（和邮件传输网关有关，高位为1会被过滤），因为只有6bit有效，所以有64个码，另外，base64希望编码能在ascii范围内，以便打印出来，所以会设置一个码表，编码映射到码表来。
也就是说，base64把每3个8bit,转换成4个8bit,其中每个8bit里面的高2bit是恒0。
这个是码表：CODE:
static const char *codes =我写了一个编码和解码的程序，根据rfc来写的，也许因为我理解的漏洞，或者有逻辑错误，程序不一定是正确的（但是我测试多多个图片和email文本的编码，是正确的）， 请用者自己检查正确性。
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
编码程序：CODE:
// 输入串，输入串长，输出串，输出串长。解码过程需要弄一个编码反查表：
void base64_encode(char *in, const int in_len, char *out, int out_len)
{
int base64_len = 4 * ((in_len+2)/3); // 要保证输出串的长度。
assert(out_len >= base64_len);
char *p = out;
int times = in_len / 3;
for(int i=0; i<times; ++i) {
*p++ = codes[(in[0] >> 2) & 0x3f];
*p++ = codes[((in[0] & 0x3) << 4) + (in[1] >> 4)];
*p++ = codes[((in[1] & 0xf) << 2) + (in[2] >> 6)];
*p++ = codes[in[2] & 0x3f];
in += 3;
}
// pad .. 如果不够3个8bit来，后面要补充'='，rfc里面称为pad
if(times * 3 + 1 == in_len) {
*p++ = codes[(in[0] >> 2) & 0x3f];
*p++ = codes[((in[0] & 0x3) << 4)];
*p++ = '=';
*p++ = '=';
}
if(times * 3 + 2 == in_len) {
*p++ = codes[(in[0] >> 2) & 0x3f];
*p++ = codes[((in[0] & 0x3) << 4) + (in[1] >> 4)];
*p++ = codes[((in[1] & 0xf) << 2)];
*p++ = '=';
}
*p = 0;
}CODE:
char ords[128];这样可以根据明文，把编码位置查出来，然后把多余的00都去掉，最后的一个反pad过程要注意一下下
#define PAD -1
void init_ords()
{
ords['A'] = 0; ords['B'] = 1; ords['C'] = 2; ords['D'] = 3; ords['E'] = 4;
ords['F'] = 5; ords['G'] = 6; ords['H'] = 7; ords['I'] = 8; ords['J'] = 9;
ords['K'] = 10; ords['L'] = 11; ords['M'] = 12; ords['N'] = 13; ords['O'] = 14;
ords['P'] = 15; ords['Q'] = 16; ords['R'] = 17; ords['S'] = 18; ords['T'] = 19;
ords['U'] = 20; ords['V'] = 21; ords['W'] = 22; ords['X'] = 23; ords['Y'] = 24;
ords['Z'] = 25; ords['a'] = 26; ords['b'] = 27; ords['c'] = 28; ords['d'] = 29;
ords['e'] = 30; ords['f'] = 31; ords['g'] = 32; ords['h'] = 33; ords['i'] = 34;
ords['j'] = 35; ords['k'] = 36; ords['l'] = 37; ords['m'] = 38; ords['n'] = 39;
ords['o'] = 40; ords['p'] = 41; ords['q'] = 42; ords['r'] = 43; ords['s'] = 44;
ords['t'] = 45; ords['u'] = 46; ords['v'] = 47; ords['w'] = 48; ords['x'] = 49;
ords['y'] = 50; ords['z'] = 51; ords['0'] = 52; ords['1'] = 53; ords['2'] = 54;
ords['3'] = 55; ords['4'] = 56; ords['5'] = 57; ords['6'] = 58; ords['7'] = 59;
ords['8'] = 60; ords['9'] = 61; ords['+'] = 62; ords['/'] = 63;
ords['='] = PAD;
}CODE:
void base64_decode(char *in, const int in_len, char *out, int *out_len)我写的文件，仅仅考虑了自己测试的需要，没有考虑通用性，把主要实现发上来供讨论。
{
int decode_len = in_len * 3 / 4;
assert(*out_len > decode_len);
char tmp[in_len]; // for ords[]
char *p = in;
for(int i=0; i<in_len; ++i,++p) {
tmp[i] = ords[*p];
}
char *q = out;
p = tmp;
*out_len = 0;
for(int i=0; i<in_len-4; i+=4) {
*q++ = (p[0] << 2) + (p[1] >> 4);
*q++ = (p[1] << 4) + (p[2] >> 2);
*q++ = (p[2] << 6) + p[3];
p += 4;
*out_len += 3;
}
// deal with pad
if(p[3] != PAD) { // no pad
*q++ = (p[0] << 2) + (p[1] >> 4);
*q++ = (p[1] << 4) + (p[2] >> 2);
*q++ = (p[2] << 6) + p[3];
*out_len += 3;
} else if(p[2] != PAD) { // one pad
*q++ = (p[0] << 2) + (p[1] >> 4);
*q++ = (p[1] << 4) + (p[2] >> 2);
*out_len += 2;
} else if(p[1] != PAD) { // two pads
*q++ = (p[0] << 2) + (p[1] >> 4);
*q++ = (p[1] << 4);
*out_len += 2;
}
*q++ = 0;
}
base64.rar
(2007-07-31 14:43:54, Size: 8.71 kB, Downloads: 0) -
给BlackBerry充电的模块
2007-7-12
新买了台bb,在linux上默认充不了电，因为bb需要500mA的电流。
于是找了个驱动。是2.6.21内核新增的，不过我在2.6.19上测试通过(实际上这个代码的想法来自barry项目--可以上sf去找，这个项目在很早的内核就有了)。
这些天都考试，今天又想找工作了，没有时间给代码上注释。明天或者后天我上来的时候，补上注释
CODE:
#include <linux/kernel.h>Makefile可以这样写：
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/usb.h>
#define RIM_VENDOR 0x0fca
#define BLACKBERRY 0x0001
static int debug;
#ifdef dbg
#undef dbg
#endif
#define dbg(dev, format, arg...) \
if (debug) \
dev_printk(KERN_DEBUG , dev , format , ## arg)
static struct usb_device_id id_table [] = {
{ USB_DEVICE(RIM_VENDOR, BLACKBERRY) },
{ }, /* Terminating entry */
};
MODULE_DEVICE_TABLE(usb, id_table);
static int magic_charge(struct usb_device *udev)
{
char *dummy_buffer = kzalloc(2, GFP_KERNEL);
int retval;
if (!dummy_buffer)
return -ENOMEM;
/* send two magic commands and then set the configuration. The device
* will then reset itself with the new power usage and should start
* charging. */
/* Note, with testing, it only seems that the first message is really
* needed (at least for the 8700c), but to be safe, we emulate what
* other operating systems seem to be sending to their device. We
* really need to get some specs for this device to be sure about what
* is going on here.
*/
dbg(&udev->dev, "Sending first magic command\n");
retval = usb_control_msg(udev, usb_rcvctrlpipe(udev, 0),
0xa5, 0xc0, 0, 1, dummy_buffer, 2, 100);
if (retval != 2) {
dev_err(&udev->dev, "First magic command failed: %d.\n",
retval);
return retval;
}
dbg(&udev->dev, "Sending second magic command\n");
retval = usb_control_msg(udev, usb_rcvctrlpipe(udev, 0),
0xa2, 0x40, 0, 1, dummy_buffer, 0, 100);
if (retval != 0) {
dev_err(&udev->dev, "Second magic command failed: %d.\n",
retval);
return retval;
}
dbg(&udev->dev, "Calling set_configuration\n");
retval = usb_driver_set_configuration(udev, 1);
if (retval)
dev_err(&udev->dev, "Set Configuration failed :%d.\n", retval);
return retval;
}
static int berry_probe(struct usb_interface *intf,
const struct usb_device_id *id)
{
struct usb_device *udev = interface_to_usbdev(intf);
dbg(&udev->dev, "ower is set to %dmA\n",
udev->actconfig->desc.bMaxPower * 2);
/* check the power usage so we don't try to enable something that is
* already enabled */
if ((udev->actconfig->desc.bMaxPower * 2) == 500) {
dbg(&udev->dev, "device is already charging, power is "
"set to %dmA\n", udev->actconfig->desc.bMaxPower * 2);
return -ENODEV;
}
/* turn the power on */
magic_charge(udev);
/* we don't really want to bind to the device, userspace programs can
* handle the syncing just fine, so get outta here. */
return -ENODEV;
}
static void berry_disconnect(struct usb_interface *intf)
{
}
static struct usb_driver berry_driver = {
.name = "berry_charge",
.probe = berry_probe,
.disconnect = berry_disconnect,
.id_table = id_table,
};
static int __init berry_init(void)
{
return usb_register(&berry_driver);
}
static void __exit berry_exit(void)
{
usb_deregister(&berry_driver);
}
module_init(berry_init);
module_exit(berry_exit);
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Greg Kroah-Hartman <gregkh@suse.de>");
module_param(debug, bool, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(debug, "Debug enabled or not");CODE:
#Makefile驱动里面命令2对7290貌似是有点错误的。我有空也调试一下。
obj-m := berry_charge.o
KERNELDIR := /lib/modules/$(shell uname -r)/build
PWD := $(shell pwd)
default:
make -C $(KERNELDIR) SUBDIRS=$(PWD) modules -
用dante+pam_mysql搭建socks5代理
2007-6-22
dante是一个很好的socks代理，同时有服务器和客户端的功能，现在我只说服务器部分。
pam是linux的通用认证模块，可以ls一下/etc/pam.d/，看看内容就明白了。
pam_mysql: http://pam-mysql.sourceforge.net/Documentation/install.php?seemore=y ， 用于把pam认证的信息，映射到mysql里面，方便web存取。
我主要参考了参考文章[1]，这里给出我的配置。
1.需要安装的软件 dante(后面我附上一个archlinux的PKGBUILD)，mysql，pam_mysql。
2. mysql的数据表:
表 logs:
+---------+--------------+------+-----+-------------------+-------+
| Field | Type | Null | Key | Default | Extra |
+---------+--------------+------+-----+-------------------+-------+
| msg | varchar(255) | YES | | NULL | |
| user | varchar(50) | YES | | NULL | |
| pid | int(10) | YES | | NULL | |
| host | char(32) | YES | | NULL | |
| rhost | char(32) | YES | | NULL | |
| logtime | timestamp | NO | | CURRENT_TIMESTAMP | |
+---------+--------------+------+-----+-------------------+-------+
这个表，我发现应该是没有作用的，不过建立也没关系。
表 user:
+----------+--------------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+----------+--------------+------+-----+---------+-------+
| username | varchar(50) | YES | | NULL | |
| password | varchar(30) | YES | | NULL | |
| comment | varchar(255) | YES | | NULL | |
+----------+--------------+------+-----+---------+-------+
用于存放用户数据，密码的加密方式，情看后面的配置。
3. pam_mysql 的配置:
建立 /etc/pam.d/sockd 文件，内容：auth required /lib/security/pam_mysql.so user=sockd passwd=8123196 host=localhost db=sockd table=user usercolumn=username passwdcolumn=password crypt=0 logtable=logs logmsgcolumn=msg logusercolumn=user logpidcolumn=pid loghostcolumn=host logtimecolumn=logtime
account required /lib/security/pam_mysql.so user=sockd passwd=8123196 host=localhost db=sockd table=user usercolumn=username passwdcolumn=password crypt=2 logtable=logs logmsgcolumn=msg logusercolumn=user logpidcolumn=pid loghostcolumn=host logrhostcolumn=rhost logtimecolumn=logtime
上面的内容，指定了认证的项目，pam_mysql的路径，mysql存取用户，以及最表和字段到认证字段的映射。
4. /etc/sockd.conf 的内容：
internal: eth0 port = 1080 # 内网入口
external: eth0 # 出口网卡
method: username none pam # 认证方式
user.notprivileged: nobody
logoutput: /var/log/danted.log
connecttimeout: 30
# 通过的客户端
client pass {
from: 192.168.0.0/16 to: 0.0.0.0/0
log: connect disconnect # 记录的内容
}
# 阻止其他客户端
client block {
from: 0.0.0.0/0 to: 0.0.0.0/0
log: connect error
}
# 允许的数据流
pass {
from: 0.0.0.0/0 to: 0.0.0.0/0 port gt 0
command: bind
log: connect disconnect
method: pam # 这里指定pam认证，然后/etc/pam.d/sockd指定pam认证的方式
}
pass {
from: 0.0.0.0/0 to: 0.0.0.0/0
protocol: tcp udp
log: connect disconnect
method:pam
}
我对内网进行用户认证，如果需要ip认证，直接指定单独的client pass就行了。
5. mysql添加新用户
grant select,update,insert on sockd.* to sockd@localhost identified by "8123196"
完毕了，请提意见
参考：
[1] http://www.360doc.com/showWeb/0/0/251512.aspx
[2] man sockd
[3] man sockd.conf -
Gedit中简单实用的External Tools - 配置详解
2007-6-22
-
Linux下监视进程运行状况(一)
2006-8-05
-
Fluke's Blog
2006-8-02