kk Blog —— 通用基础

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

centos系统各种包下载

0 centos 系統原包

1
https://buildlogs.centos.org/c7.1611.u/kernel/20170620122143/3.10.0-514.21.2.el7.x86_64/

1、系统包

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
http://mirror.centos.org/centos/6.5/os/x86_64/Packages/
国内地址
http://isoredirect.centos.org/centos/6.5/isos/x86_64/  
ex:
	http://mirror.symnds.com/distributions/CentOS-vault/5.5/isos/x86_64/  
	http://mirrors.stuhome.net/centos/6.5/isos/x86_64/  
	http://mirrors.neusoft.edu.cn/centos/6.5/isos/x86_64/
	http://mirrors.163.com/centos/6.5/isos/x86_64/
	http://mirrors.hust.edu.cn/centos/6.5/isos/x86_64/
	http://centos.ustc.edu.cn/centos/6.5/isos/x86_64/
	http://mirror.bit.edu.cn/centos/6.5/isos/x86_64/
	http://mirrors.tuna.tsinghua.edu.cn/centos/6.5/isos/x86_64/
	http://mirrors.grandcloud.cn/centos/6.5/isos/x86_64/
	http://mirror.neu.edu.cn/centos/6.5/isos/x86_64/
	http://mirrors.btte.net/centos/6.5/isos/x86_64/
	http://mirrors.hustunique.com/centos/6.5/isos/x86_64/
	http://mirrors.aliyun.com/centos/6.5/isos/x86_64/

2、debuginfo包:

1
http://debuginfo.centos.org/6/x86_64/

3、src.prm包

1
2
3
4
ftp://ftp.redhat.com/pub/redhat/linux/enterprise
ftp://ftp.redhat.com/pub/redhat/linux/enterprise/5Client/en/os/SRPMS/kexec-tools-1.102pre-154.el5.src.rpm
ftp://ftp.redhat.com/pub/redhat/linux/enterprise/5Client/en/os/SRPMS/kexec-tools-1.102pre-164.el5.src.rpm
http://vault.centos.org/5.11/os/SRPMS/kexec-tools-1.102pre-165.el5.src.rpm

4、各种包

1
pkgs/org

tsc时钟初始化

tsc时钟源初始化
1
2
3
4
5
6
7
8
9
//    调用路径:time_init->tsc_init
//    函数任务:
//        1.矫正tsc,获取tsc频率,设置cpu频率等于tsc频率
//        2.初始化基于tsc的延迟函数
//        3.检查tsc的特性
//            3.1 tsc之间是否同步
//                3.1.1 如果tsc之间不同步,标记tsc不稳定,设置rating=0
//            3.2 tsc是否稳定
//        4.注册tsc时钟源设备
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
void __init tsc_init(void)
{
	u64 lpj;
	int cpu;

	//矫正tsc,获取tsc频率
	tsc_khz = x86_platform.calibrate_tsc();
	//cpu频率等于tsc频率
	cpu_khz = tsc_khz;
	//计算辅助cycle到ns转换的辅助参数scale
	for_each_possible_cpu(cpu)
	    set_cyc2ns_scale(cpu_khz, cpu);
	//初始化基于tsc的延迟函数,ndely,udelay,mdelay
	use_tsc_delay();
	//检查cpu之间tsc是否同步
	if (unsynchronized_tsc())
	    mark_tsc_unstable("TSCs unsynchronized");
	//检查tsc是否可靠
	check_system_tsc_reliable();
	//注册tsc时钟源设备
	init_tsc_clocksource();
}
延迟函数ndelay,udelay,mdelay

通过tsc实现短延迟

1
2
3
4
5
void use_tsc_delay(void)
{
	//通过tsc进行短延迟
	delay_fn = delay_tsc;
}
tsc延迟函数

通过rep_nop实现轮询时的短延迟,查询tsc时禁止内核抢占,确保不受不同cpu间影响。

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
32
static void delay_tsc(unsigned long loops)
{
	unsigned long bclock, now;
	int cpu;
	//短延迟,禁止内核抢占
	preempt_disable();
	//delay_tsc当前运行的cpu
	cpu = smp_processor_id();
	rdtsc_barrier();
	rdtscl(bclock);
	for (;;) {
	    rdtsc_barrier();
	    rdtscl(now);
	    if ((now - bclock) >= loops)
	        break;
	    //允许rt策略进程运行
	    preempt_enable();
	    //空操作
	    rep_nop();
	    preempt_disable();

	    //delay_tsc在运行过程中,可能会迁移到不同的cpu
	    //tsc
	    if (unlikely(cpu != smp_processor_id())) {
	        loops -= (now - bclock);
	        cpu = smp_processor_id();
	        rdtsc_barrier();
	        rdtscl(bclock);
	    }
	}
	preempt_enable();
}
检查tsc是否同步
1
2
3
4
5
6
//    调用路径:tsc_init->unsynchronized_tsc
//    检查办法:
//        1.如果apic在多块板卡,则tsc不同步
//        2.如果cpuid显示具有稳定的tsc,则tsc同步
//        3.intel cpu的tsc都是同步的
//        4.默认其他品牌的多核的tsc不同步
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
__cpuinit int unsynchronized_tsc(void)
{
	//如果apic分布在多块板卡上,tsc可能不同步
	if (apic_is_clustered_box())
	    return 1;
	//cpu具有稳定的tsc
	if (boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
	    return 0;
	//intel cpu的tsc都是同步的
	if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL) {
	    //非intel cpu,如果cpu个数>1,则认为不同步
	    if (num_possible_cpus() > 1)
	        tsc_unstable = 1;
	}
	return tsc_unstable;
}
标记tsc不稳定
1
2
3
4
//    调用路径:tsc_init->mark_tsc_unstable
//    函数任务:
//        1.如果tsc时钟已经注册,异步设置tsc的rating=0,标识其不稳定
//        2.如果tsc时钟还未注册,同步设置tsc的rating=0,标识其不稳定
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
void mark_tsc_unstable(char *reason)
{
	if (!tsc_unstable) {
	    tsc_unstable = 1;
	    sched_clock_stable = 0;
	    //tsc已经注册,
	    if (clocksource_tsc.mult)
	    {
	        clocksource_mark_unstable(&clocksource_tsc);
	    }
	    //如果tsc时钟源未注册,修改rating为最低,从而不会被当做最佳的时钟源
	    else {
	        clocksource_tsc.flags |= CLOCK_SOURCE_UNSTABLE;
	        clocksource_tsc.rating = 0;
	    }
	}
}
注册tsc时钟源
1
2
3
4
5
6
//    函数任务:
//        1.计算tsc的mult
//        2.检查tsc是否稳定
//            2.1 如果tsc不稳定,降低其rating,清除时钟源连续标志
//        3.向系统注册tsc clocksource
//    调用路径:tsc_init->init_tsc_clocksource
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
static void __init init_tsc_clocksource(void)
{
	// 计算tsc的mult
	clocksource_tsc.mult = clocksource_khz2mult(tsc_khz,
	        clocksource_tsc.shift);
	// 如果tsc的可靠性已经验证,则清除 必须验证 标记
	if (tsc_clocksource_reliable)
	    clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY;
	
	// 检查tsc是否稳定
	// 在tsc_init前通过全局变量标记tsc是否稳定,可靠
	if (check_tsc_unstable()) {
	    // 如果tsc不稳定,则降低rating最低,清除连续标记
	    clocksource_tsc.rating = 0;
	    clocksource_tsc.flags &= ~CLOCK_SOURCE_IS_CONTINUOUS;
	}
	// 向系统注册tsc clocksource
	clocksource_register(&clocksource_tsc);
}

TSC时间错误

arch/x86/kernel/tsc.c:
开机初始化会调用tsc_init() -> set_cyc2ns_scale() 设置per_cpu变量cyc2ns、cyc2ns_offset。以供后面shced_clock()->native_sched_clock()->__cycles_2_ns()调用。

在cpufreq_tsc()中如果
//cpu具有稳定的tsc
if (boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
return 0;
所以一般不会注册time_cpufreq_notifier函数,也就不会再调用set_cyc2ns_scale。

  • 现象:top、ps出来的TIME和CPU的值非常异常。
1
2
3
4
5
6
7
8
9
10
11
// 查看TSC寄存器的值
#include <stdio.h>

int main()
{
	    unsigned long low, high, val;
	    asm volatile("rdtsc": "=a" (low), "=d" (high));
	    val = ((low) | ((unsigned long)(high) << 32));
	    printf("%lu\n", val);
	    return 0;
}

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=733043

Xeon E5 has a bug, it doesn’t reset TSC on warm reboot, just keep it instead. see “BT81. X X X No Fix TSC is Not Affected by Warm Reset” http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/xeon-e5-family-spec-update.pdf

And also kernel 2.6.32 has a bug.
Xeon bug + kernel bug = hung after warm reboot (or kexec) after 208.5 days
since booting. So, administrators should shutdown it once at all, then
boot it again because “shutdown -r” causes hang up.

Red Hat has released a fix for this as kernel 2.6.32-220, 2.6.32-279
and 2.6.32-358 series (RHEL6.x) https://access.redhat.com/site/solutions/433883 (for detail subscriber only :-(

Attached patch is based on upstream patch.
see http://kernel.opensuse.org/cgit/kernel/patch/?id=9993bc635d01a6ee7f6b833b4ee65ce7c06350b1


Red Hat Enterprise Linux 6.1 (kernel-2.6.32-131.26.1.el6 and newer)
Red Hat Enterprise Linux 6.2 (kernel-2.6.32-220.4.2.el6 and newer)
Red Hat Enterprise Linux 6.3 (kernel-2.6.32-279 series)
Red Hat Enterprise Linux 6.4 (kernel-2.6.32-358 series)
Any Intel® Xeon® E5, Intel® Xeon® E5 v2, or Intel® Xeon® E7 v2 series processor


From 9993bc635d01a6ee7f6b833b4ee65ce7c06350b1 Mon Sep 17 00:00:00 2001
From: Salman Qazi <sqazi@google.com>
Date: Sat, 10 Mar 2012 00:41:01 +0000
Subject: sched/x86: Fix overflow in cyc2ns_offset

When a machine boots up, the TSC generally gets reset. However, when kexec is used to boot into a kernel, the TSC value would be carried over from the previous kernel. The computation of cycns_offset in set_cyc2ns_scale is prone to an overflow, if the machine has been up more than 208 days prior to the kexec. The overflow happens when we multiply *scale, even though there is enough room to store the final answer.

We fix this issue by decomposing tsc_now into the quotient and remainder of division by CYC2NS_SCALE_FACTOR and then performing the multiplication separately on the two components.

Refactor code to share the calculation with the previous fix in __cycles_2_ns().

Signed-off-by: Salman Qazi <sqazi@google.com>
Acked-by: John Stultz <john.stultz@linaro.org>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Turner <pjt@google.com>
Cc: john stultz <johnstul@us.ibm.com>
Link: http://lkml.kernel.org/r/20120310004027.19291.88460.stgit@dungbeetle.mtv.corp.google.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>


patch: http://kernel.opensuse.org/cgit/kernel/patch/?id=9993bc635d01a6ee7f6b833b4ee65ce7c06350b1

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
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
diff --git a/arch/x86/include/asm/timer.h b/arch/x86/include/asm/timer.h
index 431793e..34baa0e 100644
--- a/arch/x86/include/asm/timer.h
+++ b/arch/x86/include/asm/timer.h
@@ -57,14 +57,10 @@ DECLARE_PER_CPU(unsigned long long, cyc2ns_offset);
 
 static inline unsigned long long __cycles_2_ns(unsigned long long cyc)
 {
- unsigned long long quot;
- unsigned long long rem;
  int cpu = smp_processor_id();
  unsigned long long ns = per_cpu(cyc2ns_offset, cpu);
- quot = (cyc >> CYC2NS_SCALE_FACTOR);
- rem = cyc & ((1ULL << CYC2NS_SCALE_FACTOR) - 1);
- ns += quot * per_cpu(cyc2ns, cpu) +
-     ((rem * per_cpu(cyc2ns, cpu)) >> CYC2NS_SCALE_FACTOR);
+ ns += mult_frac(cyc, per_cpu(cyc2ns, cpu),
+         (1UL << CYC2NS_SCALE_FACTOR));
  return ns;
 }
 
diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c
index a62c201..183c592 100644
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -620,7 +620,8 @@ static void set_cyc2ns_scale(unsigned long cpu_khz, int cpu)
 
  if (cpu_khz) {
      *scale = (NSEC_PER_MSEC << CYC2NS_SCALE_FACTOR)/cpu_khz;
-     *offset = ns_now - (tsc_now * *scale >> CYC2NS_SCALE_FACTOR);
+     *offset = ns_now - mult_frac(tsc_now, *scale,
+                      (1UL << CYC2NS_SCALE_FACTOR));
  }
 
  sched_clock_idle_wakeup_event(0);
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index e834342..d801acb 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -85,6 +85,19 @@
 }                            \
 )
 
+/*
+ * Multiplies an integer by a fraction, while avoiding unnecessary
+ * overflow or loss of precision.
+ */
+#define mult_frac(x, numer, denom)(          \
+{                            \
+ typeof(x) quot = (x) / (denom);         \
+ typeof(x) rem  = (x) % (denom);         \
+ (quot * (numer)) + ((rem * (numer)) / (denom)); \
+}                            \
+)
+
+
 #define _RET_IP_     (unsigned long)__builtin_return_address(0)
 #define _THIS_IP_  ({ __label__ __here; __here: (unsigned long)&&__here; }) 

mark problem

C语言中宽字符和多字节字符 MB_CUR_MAX

stdlib.h
MB_CUR_MAX 当前locale中多字节字符的最大字节数目
如果把字符串当成多字节字符处理会慢很多
sort命令的一些版本中会根据locale中的LANG来设置比较函数,如果没设置好会导致sort跑的很慢。

内核栈溢出

因为内核栈中栈底保存thread_info,所以如果内核栈溢出会破坏thread_info,这样当进程发生睡眠、中断、抢占等调度时就会出错。
容易报try_to_wakeup+XXX错误,等等

定时器 timer 注意

如果mod_timer设置成马上执行,然后下面又是setup_timer。
有可能的情况是:进入了中断执行过程,可是setup_timer又改变timer结构,导致继续执行时用到timer结构出错。

查看内核栈

用 echo t > /proc/sysrq-trigger 把内核栈整个打出来。
可以看/proc/{pid}/wchan,里面是该进程阻塞位置的内核函数名,在所有办法都没戏的时候可以看它。

mark

修复ext4日志(jbd2)bug
rhel6再次发现jbd2的bug
rhel6 的软RAID问题
stable pages
追踪CPU跑满

VirtualBox摄像头功能

1、下载Oracle VM VirtualBox Extension Pack。下载网址如下:

https://www.virtualbox.org/wiki/Downloads

2、安装Oracle VM VirtualBox Extension Pack

安装外部扩展包很简单

第一:打开Oracle VirtualBox,在菜单栏中找到“管理”–>“全局设定”
第二:打开“全局设定”,参考下图:

第三:“Extension Packages”中添加下载的Oracle VM VirtualBox Extension Pack,根据提示安装即可,

然后虚拟机就能用摄像头了