delphij's Chaos

今天偶然发现以前做的一套CMS里面的显示发布日期的功能被干掉了。

这已经不是第一、第二或者第三次在我毫不知情的情况下作的commit了。而且我注意到commit log里面还赫然写着"Approved by: delphij"。

Yes, finally postfix is supporting DSN. To quote Witese, this is the first postfix snapshot which provides “full support for RFC 3461..3464 Delivery Status Notifications”!

I have done several experiments/asserts to ensure that the panic is related to ULE+PREEMPTION case only.

ULE: no panic
4BSD: no panic
4BSD+PREEMPTION: no panic

My latest progress is that I found that sched_switch in ULE for the threaded case, can decide whether subsequent call would panic. Investigating.

My company has offered me some test equipment (namely, a HP 360 with ILO, which is *great* for debugging at home!), and I has started to trip over the ULE code and find for bugs.

You can obtain latest list from sys/sys/proc.h:

 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

 * Below is a key of locks used to protect each member of struct proc.  The
 * lock is indicated by a reference to a specific character in parens in the
 * associated comment.
 *      * - not yet protected
 *      a - only touched by curproc or parent during fork/wait
 *      b - created at fork, never changes
 *              (exception aiods switch vmspaces, but they are also
 *              marked 'P_SYSTEM' so hopefully it will be left alone)
 *      c - locked by proc mtx
 *      d - locked by allproc_lock lock
 *      e - locked by proctree_lock lock
 *      f - session mtx
 *      g - process group mtx       
 *      h - callout_lock mtx
 *      i - by curproc or the master session mtx
 *      j - locked by sched_lock mtx
 *      k - only accessed by curthread
 *      l - the attaching proc or attaching proc parent
 *      m - Giant
 *      n - not locked, lazy
 *      o - ktrace lock
 *      p - select lock (sellock)
 *      q - td_contested lock
 *      r - p_peers lock
 *      x - created at fork, only changes during single threading in exec
 *      z - zombie threads/ksegroup lock

Today Giorgos Keramidas (keramida@) has noted that the following skill that can minimize lock period:

1. save the commit message to a file (say ~/commitlog)
2. cvs ci -F ~/commitlog file[…]

From: http://y.sina.com.cn 2005年06月17日 11:24 榕树下 文/流浪猫

这是我的一个同事给我讲的他上高中时一个真实的故事。

那一年他上高二，因为他家住在外地，所以他从上高中就在学校住校。他们高中大部分都是长春本市的学生，住校的同学相对较少，他们寝室也就只住了四个人。从生日排行，他排老四。他们寝的老大跟他是同乡，他们俩的关系比较好。

那是快要放暑假的时候，他们考完试，等着发表成绩。这时候已经没有什么课了，学习也比较轻松。那天是星期五，晚上吃完饭没有什么事，几个人就在寝室里侃大山。说着说着，就说到一些奇异的事情上面来了。

老三就说：「哎，你们听没听过一个游戏，说是比‘碟仙’更吓人。」

老二问：「啥游戏啊？」

「拍鬼。你们听过吗？」

我们同事就说：「没听过。咋玩呀？」

有一天，我们发现了一个bug。

然是有人决定把它workaround掉，大家都很开心。

突然有一天，那个bug被修了。

然后workaround变成了bug……所以说坚决抵制workaround。。。

It seems that I have hit a bug, where kseq_nice_rem is being called too much. Have read the code and added some KASSERT/MPASS’es.

pav@ has put together this which looks quite interesting. Have a read of it :-)

(Got that from arved@’s blog)