kk Blog —— 通用基础

<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="viewport" content="width=device-width,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no"/>
<meta name="MobileOptimized" content="320">
<meta name="format-detection" content="telephone=no">
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">

<!DOCTYPE html>
<html>
<head>
	<title>proxy</title>
</head>
<body>

<?php
	$region = "hk";
	if (isset($_GET['region']))
		$region = $_GET['region'];
	$count = "0";
	if (isset($_GET['count']))
		$count = $_GET['count'];
?>
<form action="index.php" method="get">
<input name="region" value="<?php echo $region; ?>" style="width:100px">
<input name="count" value="0" style="width:100px">
<input class="submit" type="submit" value="创建">
</form>
<?php
	$cmd = "python CVM.py $region $count";

	echo date("Y-m-d H:i:s") . " " . $cmd;
	echo "</br><br>";

	exec($cmd);

	if (filesize("show.log") > 0) {
		$myfile = fopen("show.log", "r");
		echo fread($myfile, filesize("show.log"));
		fclose($myfile);
	}
?>
</body>
</html>

web配置cache_peer, 方案二中的guangzhou才需要

cachemgr.cgi 没找到接口，可惜

填写 IP

<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="viewport" content="width=device-width,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no"/>
<meta name="MobileOptimized" content="320">
<meta name="format-detection" content="telephone=no">
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">

<!DOCTYPE html>
<html>
<head>
	<title>squid</title>
</head>
<body>

<?php
	$up = "";
	if (isset($_GET['up']))
		$up = $_GET['up'];
?>
<form action="squid.php" method="get">
<input name="up" value="<?php echo $up; ?>" style="width:200px">
<input class="submit" type="submit" value="更新">
</form>
<?php

	echo date("Y-m-d H:i:s");
	echo "</br><br>";

	if (strlen($up) > 3 && strlen($up) < 128) {
		$myfile = fopen("squid.log", "w");
		fwrite($myfile, $up . " up.com\n");
		fclose($myfile);
	}
	$myfile = fopen("squid.log", "r");
	echo fread($myfile, filesize("squid.log"));
	fclose($myfile);
?>
</body>
</html>

刷新squid

# */1 * * * * cd /var/www/html/; /var/www/html/squid.sh
# */1 * * * * sleep 30; cd /var/www/html/; /var/www/html/squid.sh
# chown -R www-data:www-data ../html/
cat /etc/hosts | grep up.com > hosts_tmp
diff hosts_tmp squid.log > /dev/null
if [ $? -eq 0 ]; then exit; fi

cat /etc/hosts | grep -v up.com > hosts
cat squid.log >> hosts 
cp hosts /etc/

/usr/sbin/squid -k reconfigure -f /root/squid/squid.conf

skb 数据异或

2020-09-10 04:21:00

skb 数据异或，可以改进成单独的option

diff --git a/run.sh b/run.sh
index 138e664..0eb5286 100755
--- a/run.sh
+++ b/run.sh
@@ -19,3 +19,4 @@ echo 10000 > /proc/sys/net/tcp/digest/digest_err_retry
 echo 0 > /proc/sys/net/ipv4/tcp_timestamps # for mptcp+digest
 
 echo 0 > /proc/sys/net/tcp/mss_adjust
+echo 0 > /proc/sys/net/tcp/digest/digest_xor
diff --git a/src/4.15.18/tcp_ipv4.c b/src/4.15.18/tcp_ipv4.c
index b1b5c9d..b94ba6e 100644
--- a/src/4.15.18/tcp_ipv4.c
+++ b/src/4.15.18/tcp_ipv4.c
@@ -1848,6 +1848,12 @@ int tcp_v4_rcv(struct sk_buff *skb)
  th = (const struct tcphdr *)skb->data;
  iph = ip_hdr(skb);
 
+ if (digest_xor && digest_skb_check(NULL, NULL, skb, 0) != 0) {
+     int idx = 0;
+     skb->ip_summed = CHECKSUM_UNNECESSARY;
+     skb_change_bits(skb, th->doff * 4, skb->len - th->doff * 4, &idx);
+ }
+
 lookup:
  sk = __inet_lookup_skb(&tcp_hashinfo, skb, __tcp_hdrlen(th), th->source,
                 th->dest, sdif, &refcounted);
diff --git a/src/4.15.18/tcp_ipv6.c b/src/4.15.18/tcp_ipv6.c
index 3e27967..e975356 100644
--- a/src/4.15.18/tcp_ipv6.c
+++ b/src/4.15.18/tcp_ipv6.c
@@ -1619,6 +1619,12 @@ int tcp_v6_rcv(struct sk_buff *skb)
  th = (const struct tcphdr *)skb->data;
  hdr = ipv6_hdr(skb);
 
+ if (digest_xor && digest_skb_check(NULL, NULL, skb, 0) != 0) {
+     int idx = 0;
+     skb->ip_summed = CHECKSUM_UNNECESSARY;
+     skb_change_bits(skb, th->doff * 4, skb->len - th->doff * 4, &idx);
+ }
+
 lookup:
  sk = __inet6_lookup_skb(&tcp_hashinfo, skb, __tcp_hdrlen(th),
              th->source, th->dest, inet6_iif(skb), sdif,
diff --git a/src/4.15.18/tcp_output.c b/src/4.15.18/tcp_output.c
index dab8703..f766ff4 100644
--- a/src/4.15.18/tcp_output.c
+++ b/src/4.15.18/tcp_output.c
@@ -1118,6 +1118,9 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
      oskb = skb;
 
      tcp_skb_tsorted_save(oskb) {
+         if (digest_xor)
+             skb = skb_copy(oskb, gfp_mask);
+         else
          if (unlikely(skb_cloned(oskb)))
              skb = pskb_copy(oskb, gfp_mask);
          else
@@ -1264,6 +1267,13 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
      sk_nocaps_add(sk, NETIF_F_GSO_MASK);
 
      merge = !(tcb->tcp_flags & TCPHDR_SYN);
+
+     if (digest_xor) {
+         int idx = 0;
+         skb_change_bits(skb, tcp_header_size, skb->len - tcp_header_size, &idx);
+         skb->ip_summed = CHECKSUM_NONE;
+         skb->csum = skb_checksum(skb, tcp_header_size, skb->len - tcp_header_size, 0);
+     }
  }
 
  tcp_options_write((__be32 *)(th + 1), tp, &opts, skb, merge);
@@ -3122,6 +3132,9 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs)
      struct sk_buff *nskb;
 
      tcp_skb_tsorted_save(skb) {
+         if (digest_xor)
+             nskb = skb_copy(skb, GFP_ATOMIC);
+         else
          nskb = __pskb_copy(skb, MAX_TCP_HEADER, GFP_ATOMIC);
          err = nskb ? tcp_transmit_skb(sk, nskb, 0, GFP_ATOMIC) :
                   -ENOBUFS;
diff --git a/src/digest_core.c b/src/digest_core.c
index 54af869..50840b1 100644
--- a/src/digest_core.c
+++ b/src/digest_core.c
@@ -419,9 +419,106 @@ out:
  return 0;
 }
 
+
+int digest_xor = 0;
+unsigned char MY_XOR[65536];
+
+int __skb_change_bits(unsigned char *from, int len, int *idx)
+{
+ int i;
+ for (i = 0; i < len; i ++) {
+     *(from + i) ^= MY_XOR[*idx];
+     *idx = *idx + 1;
+ }
+ return 0;
+}
+
+int skb_change_bits(struct sk_buff *skb, int offset, int len, int *idx)
+{
+ int start = skb_headlen(skb);
+ struct sk_buff *frag_iter;
+ int i, copy;
+
+ if (offset > (int)skb->len - len)
+     goto fault;
+
+ /* Copy header. */
+ if ((copy = start - offset) > 0) {
+     if (copy > len)
+         copy = len;
+     // skb_copy_from_linear_data_offset(skb, offset, to, copy);
+     __skb_change_bits(skb->data + offset, copy, idx);
+     if ((len -= copy) == 0)
+         return 0;
+     offset += copy;
+ }
+
+ for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+     int end;
+     skb_frag_t *f = &skb_shinfo(skb)->frags[i];
+
+     WARN_ON(start > offset + len);
+
+     end = start + skb_frag_size(f);
+     if ((copy = end - offset) > 0) {
+         u32 p_off, p_len, copied;
+         struct page *p;
+         u8 *vaddr;
+
+         if (copy > len)
+             copy = len;
+
+#if LINUX_VERSION_CODE > KERNEL_VERSION(4, 0, 0)
+         skb_frag_foreach_page(f,
+                       f->page_offset + offset - start,
+                       copy, p, p_off, p_len, copied) {
+             vaddr = kmap_atomic(p);
+             //memcpy(to + copied, vaddr + p_off, p_len);
+             __skb_change_bits(vaddr + p_off, p_len, idx);
+             kunmap_atomic(vaddr);
+         }
+#else
+         vaddr = kmap_atomic(skb_frag_page(f));
+         __skb_change_bits(vaddr + f->page_offset + offset - start, copy, idx);
+         kunmap_atomic(vaddr);
+#endif
+
+         if ((len -= copy) == 0)
+             return 0;
+         offset += copy;
+     }
+     start = end;
+ }
+
+ skb_walk_frags(skb, frag_iter) {
+     int end;
+
+     WARN_ON(start > offset + len);
+
+     end = start + frag_iter->len;
+     if ((copy = end - offset) > 0) {
+         if (copy > len)
+             copy = len;
+         if (skb_change_bits(frag_iter, offset - start, copy, idx))
+             goto fault;
+         if ((len -= copy) == 0)
+             return 0;
+         offset += copy;
+     }
+     start = end;
+ }
+
+ if (!len)
+     return 0;
+
+fault:
+ return -EFAULT;
+}
+
+
 int digest_init(void)
 {
  int ret;
+ int i;
+ unsigned long int next = 1;
+ for (i = 0; i < 65536; i ++) {
+     next = next * 2 + 1; // TODO
+     MY_XOR[i] = next % 251;
+ }
 
  ret = digest_trace_init();
  if (ret) {
diff --git a/src/digest_core.h b/src/digest_core.h
index 007628c..3ba8c80 100644
--- a/src/digest_core.h
+++ b/src/digest_core.h
@@ -79,6 +79,8 @@ extern atomic_t digest_trace_head_num;
 extern atomic_t digest_trace_free_num;
 extern int digest_trace_log_total;
 
+extern int digest_xor;
+
 extern atomic64_t digest_current, digest_total1, digest_total2, digest_drop_total;
 extern atomic64_t digest_drop_connect;
 
diff --git a/src/digest_sysctl.c b/src/digest_sysctl.c
index 27adfa3..3bfe0ae 100644
--- a/src/digest_sysctl.c
+++ b/src/digest_sysctl.c
@@ -478,6 +478,13 @@ static struct ctl_table digest_sysctl_table[] = {
      .mode = 0644,
      .proc_handler = digest_trace_show_handler,
  },
+ {
+     .procname = "digest_xor",
+     .data = &digest_xor,
+     .maxlen = sizeof(int),
+     .mode = 0644,
+     .proc_handler = proc_dointvec
+ },
  {}
 };

MPTCP sched

2020-08-15 01:02:00

$ git grep mptcp_register_scheduler net/mptcp/
net/mptcp/mptcp_blest.c:        if (mptcp_register_scheduler(&mptcp_sched_blest))
net/mptcp/mptcp_redundant.c:    if (mptcp_register_scheduler(&mptcp_sched_red))
net/mptcp/mptcp_rr.c:   if (mptcp_register_scheduler(&mptcp_sched_rr))
net/mptcp/mptcp_sched.c:int mptcp_register_scheduler(struct mptcp_sched_ops *sched)

blest

redundant

rr

MPTCP cong

2020-08-15 01:00:00

不设置时，默认的就是cubic

$ git grep tcp_congestion_ops net/mptcp/
net/mptcp/mctcp_desync.c:static struct tcp_congestion_ops mctcp_desync = {
net/mptcp/mptcp_balia.c:static struct tcp_congestion_ops mptcp_balia = {
net/mptcp/mptcp_coupled.c:static struct tcp_congestion_ops mptcp_ccc = {
net/mptcp/mptcp_olia.c:static struct tcp_congestion_ops mptcp_olia = {
net/mptcp/mptcp_wvegas.c:static struct tcp_congestion_ops mptcp_wvegas __read_mostly = {

desync

balia

ccc

olia

wvegas

将松散源路由（Loose Source and Record Routing, LSRR）选项添加到MPTCP，以确保MPTCP子流使用所有可用网关，即采用LSRR分发子流的数据包。Binder路径管理算法允许用户终端设备从网关直接获得更高的网络吞吐量，而不需要进行任何修改，可用图a和b谁明其工作原理。理论上，只有在网络中设置了LSRR选项，才能使其B>0，从而使得Binder路径管理算法发挥良好效果。

松散源路由：允许相邻两个IP地址之间跳过多个网络。

通俗解释：发送端指明了流量或者数据包必须经过的IP地址清单，但如果需要，也可以经过一些其他的地址。换句话说，不用考虑数据包经过的确切地址，只要它经过这些地址就可以。

← Older Blog Archives Newer →

kk Blog —— 通用基础

date [-d @int|str] [+%s|"+%F %T"]
netstat -ltunp

vpn

方案一

方案二

通过web，实现跨平台随时随地购买使用

购买CVM