系统技术非业余研究

Erlang R16B03发布了，通常03版本是bug fix版本，进入生产版本，官方的说明如下：

OTP R16B03 is a service release with mostly a number of small corrections and user contributions. But there are some new functions worth mentioning as well, here are some of them:

A new memory allocation feature called “super carrier” has been introduced. It can for example be used for pre-allocation of all memory that the runtime system should be able to use. It is enabled by passing the +MMscs (size in MB) command line argument. For more information see the documentation of the +MMsco, +MMscrfsd, +MMscrpm, +MMscs, +MMusac, and, +Mlpm command line arguments in the erts_alloc(3) documentation.
The LDAP client (eldap application) now supports the start_tls operation. This upgrades an existing tcp connection to encryption using TLS, see eldap:start_tls/2 and /3.
The FTP client (inets application) now supports FTP over TLS (ftps).

其中最大的改进就是super carrier， 参见我之前写的博文, 这个特性在于专机专用的系统，内存的使用效率就高很多，同时这个版本对多个核心之间内存倒腾的效率和利用率也高很多，值得大家去用。内存的利用率和碎片率通常不会引起大家的关注，在生产的服务器中，这点非常值得关注。

R16B03发布了后，官方马不停蹄的就进入R17的开发。其中最大的期待就是语言方面的改进，包括eep37和map数据结构。 eep37特性已经进入master, commit在这里， 该特性的具体描述在这里.

简单的说：就是你过去在shell下写如下的fun过去是不可能的。

fun Fact(N) when N > 0 ->
            N * Fact(N - 1);
        Fact(0) ->
            1
end.

但是我们又经常需要这样的匿名函数，比如spawn函数，很不方便。 eep37就是解决这样的事情的。

我们来演示下，首先安装R17:

$ kerl build git git://github.com/erlang/otp.git master r17 && kerl install r17  r17
$ r17/bin/erl   
Erlang/OTP 17.0-rc0 [erts-6.0] [source-7d4e5e2] [64-bit] [smp:16:16] [async-threads:10] [hipe] [kernel-poll:false] [type-assertions] [debug-compiled] [lock-checking] [systemtap]

Eshell V6.0  (abort with ^G)
1> fun Fact(N) when N > 0 ->
1>             N * Fact(N - 1);
1>         Fact(0) ->
1>             1
1>     end.
#Fun<erl_eval.29.42696066>
2> F=e(-1).
#Fun<erl_eval.29.42696066>
3> F(10).
3628800

eep37这个特性涉及到的改变还是很大的:语言规格，编译器，调试器，dialyzer, tracer， shell, emacs插件等等，全系列的改变，所以要放在R17里面来，虽然看提交，代码早在1年前准备好了。我们再来体验下：

$ cat eep37.erl 
-module(eep37).

-compile(export_all).

-spec self() -> fun(() -> fun()).
self() ->
    fun Self() -> Self end.

-spec fact() -> fun((non_neg_integer()) -> non_neg_integer()).
fact() ->
    fun Fact(N) when N > 0 ->
            N * Fact(N - 1);
        Fact(0) ->
            1
    end.

$ r17/bin/erl 
Erlang/OTP 17.0-rc0 [erts-6.0] [source-7d4e5e2] [64-bit] [smp:16:16] [async-threads:10] [hipe] [kernel-poll:false] [type-assertions] [debug-compiled] [lock-checking] [systemtap]

Eshell V6.0  (abort with ^G)
1> eep37:fact().    
#Fun<eep37.1.16748913>
2> F=e(-1).
#Fun<eep37.1.16748913>
3> F(10).
3628800
4> 

小结： 语言也在不停的让用户爽。

祝玩得开心！

Erlang内存分配的框架一句话总结，从erts_alloc文档摘抄如下：

erts_alloc is an Erlang Run-Time System internal memory allocator library. erts_alloc provides the Erlang Run-Time System with a number of memory allocators.

可见Erlang的内存分配体系是非常复杂的，有很深的层次，erts内部开发人员面对的是erts_alloc来提供服务，比如分配port相关的数据结构代码如下：

pdhp = erts_alloc(ERTS_ALC_T_PORT_DATA_HEAP,
sizeof(ErtsPortDataHeap) + hsize*(sizeof(Eterm)-1));

使用起来非常简单。但是Erlang系统是个靠消息传递的语言，每个消息传递都需要分配内存，在自动Gc的时候需要释放内存，在典型的服务器上比如proxy, 每天单binary数据类型的分配和释放达到1亿次之多，所以内存分配器的效率就显的特别的重要。 所以erlang采用了一套非常庞杂的内存分配系统来满足这种需求，见下图：

粗粗的讲，内存分配器从sys_alloc和mseg_alloc批发内存，然后再零售给终端用户。其中sys_alloc就是libc的malloc, mseg_alloc就是mmap, 通过这二个接口从操作系统大批量申请内存，我们把上图的相关部分放大下看：

我们今天要讲的就是红框的那部分，erlang系统偏向于从mmap申请内存，因为过程比libc或者tcmalloc比较可控。所以如果Erlang的应用内存使用非常密集和需求变化很大的时候，就需要经常从操作系统那里批发和归还内存。而批发通常是通过mmap来的，这就是为什么我们strace beam的时候，进程会发现有很多mmap系统调用。

我们知道mmap系统调用是要进入内核再出来的。内核在内核空间维护了一颗树（比如红黑树）来管理虚拟内存。当系统调用次数非常多的时候，开销就出来了。既然mmap是用树在内核空间，那为什么我们不能在erlang内存分配器里面自己来维护呢？这样算法是一样的，但是减少了进出内核的开销。基于这个思路，最近rickard-sverker同学为Erlang R16B03添加了supercarrier， 具体参见这里。

这个super carrier的原理就是通过一次向内核申请大量的内存自己管理，进一步减少mmap的调用次数，虽然mseg_alloc已经做了简单的段cache有点效果了.

我们来看下supercarrier的使用文档：

+MMscmgc
Set super carrier max guaranteed no of carriers. This parameter defaults to 65536. This parameter determines an amount of pre-allocated structures that is needed in order to keep track of different areas in the super carrier. When the system runs out of such structures it may crash due to an out of memory condition.
+MMsco true|false
Set super carrier only flag. This flag defaults to true. When a super carrier is used and this flag is true, the system will crash when a carrier request cannot be satisfied by the super carrier. When the flag is false the system will try to create requested carrier by other means.

NOTE: Setting this flag to false may not be supported on all systems. This flag will in that case be ignored.

NOTE: The super carrier cannot be enabled nor disabled on halfword heap systems. This flag will be ignored on halfword heap systems.
+MMscrpm true|false
Set super carrier reserve physical memory flag. This flag defaults to true. When this flag is true, physical memory will be reserved for the whole super carrier at once when it is created. The reservation will after that be left unchanged. When this flag is set to false only virtual address space will be reserved for the super carrier upon creation. The system will attempt to reserve physical memory upon carrier creations in the super carrier, and attempt to unreserve physical memory upon carrier destructions in the super carrier.

NOTE: What reservation of physical memory actually means highly depends on the operating system, and how it is configured. For example, different memory overcommit settings on Linux drastically change the behaviour. Also note, setting this flag to false may not be supported on all systems. This flag will in that case be ignored.

NOTE: The super carrier cannot be enabled nor disabled on halfword heap systems. This flag will be ignored on halfword heap systems.
+MMscs
Set super carrier size (in MB). The super carrier size defaults to zero; i.e, the super carrier is by default disabled. The super carrier is a large continuous area in the virtual address space. The system will always try to create new carriers in the super carrier.

NOTE: The super carrier cannot be enabled nor disabled on halfword heap systems. This flag will be ignored on halfword heap systems.

关键参数有二个：MMscs控制一次向内核申请的内存的总量，MMscrpm控制申请的内存要不要马上兑现（马上分配物理内存）。

我们来演示下supercarrier的使用，我们一次性给到erts 16G内存，用到的beam版本是2013/11/02号github上的erlang/otp master分支：
Read more…