webclient在gc的时候可能触发崩溃

[xzzwandi]

服务器使用的是skynet框架，最近有时出现崩溃问题，不好重现
收集几个coredump，gdb里看到的都是在webclient_destory的时候触发崩溃
以下是gdb的栈：
(gdb) bt
#0 0x00007fdfdbc1988a in ?? ()
#1 0x00007fdfe942098f in ?? () from /lib/x86_64-linux-gnu/libcrypto.so.1.0.0
#2 0x00007fdfe941ed56 in ?? () from /lib/x86_64-linux-gnu/libcrypto.so.1.0.0
#3 0x00007fdfe942e440 in sk_pop_free () from /lib/x86_64-linux-gnu/libcrypto.so.1.0.0
#4 0x00007fdfe941f0dc in ENGINE_cleanup () from /lib/x86_64-linux-gnu/libcrypto.so.1.0.0
#5 0x00007fdfe01e352e in ?? () from /usr/lib/x86_64-linux-gnu/libcurl.so.4
#6 0x00007fdfe01e3ec9 in ?? () from /usr/lib/x86_64-linux-gnu/libcurl.so.4
#7 0x00007fdfe01bf695 in curl_global_cleanup () from /usr/lib/x86_64-linux-gnu/libcurl.so.4
#8 0x00007fdfe03fe47c in webclient_destory (l=0x7fdfe1b5ed08) at lualib-src/lua-webclient.c:87
#9 0x0000000000414739 in luaD_precall (L=L@entry=0x7fdfe1b5ed08, func=0x7fdfe1a79930, nresults=0) at ldo.c:434
#10 0x0000000000414a03 in luaD_call (L=L@entry=0x7fdfe1b5ed08, func=, nResults=) at ldo.c:498
#11 0x0000000000414a61 in luaD_callnoyield (L=0x7fdfe1b5ed08, func=, nResults=) at ldo.c:509
#12 0x0000000000413e6f in luaD_rawrunprotected (L=L@entry=0x7fdfe1b5ed08, f=0x416010 , ud=0x0) at ldo.c:142
#13 0x0000000000414d4d in luaD_pcall (L=0x7fdfe1b5ed08, func=, u=, old_top=48, ef=) at ldo.c:729
#14 0x0000000000415f1c in GCTM (L=L@entry=0x7fdfe1b5ed08, propagateerrors=propagateerrors@entry=0) at lgc.c:833
#15 0x00000000004175aa in callallpendingfinalizers (L=0x7fdfe1b5ed08) at lgc.c:872
#16 luaC_freeallobjects (L=0x7fdfe1b5ed08) at lgc.c:981
#17 0x000000000041bbce in close_state (L=0x7fdfe1b5ed08) at lstate.c:245
#18 0x000000000041c0c0 in lua_close (L=) at lstate.c:344
#19 0x00007fdfe7dfc7ac in snlua_release (l=0x7fdfe8b79470) at service-src/service_snlua.c:192
#20 0x000000000040abfa in delete_context (ctx=0x7fdfe0991400) at skynet-src/skynet_server.c:213
#21 skynet_context_release (ctx=ctx@entry=0x7fdfe0991400) at skynet-src/skynet_server.c:223
#22 0x000000000040b1fd in skynet_context_message_dispatch (sm=sm@entry=0x7fdfe8a168a0, q=0x7fdfe1b9c500, q@entry=0x0, weight=weight@entry=0) at skynet-src/skynet_server.c:348
#23 0x000000000040b87d in thread_worker (p=) at skynet-src/skynet_start.c:162
#24 0x00007fdfe9e3e184 in start_thread (arg=0x7fdfe2df0700) at pthread_create.c:312
#25 0x00007fdfe907d03d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111

请问这可能是什么问题？从哪个方向查？谢谢！

[dpull]

了解一下

1. 进程中存在几个webclient实例？
2. curl版本是多少？OpenSSL版本是多少？

[xzzwandi]

1、我们的webclient有很多， 但只有一个有实例有通信，且这个不会摧毁， 其他都是只用了urlencode功能
2、curl 7.35.0 (x86_64-pc-linux-gnu) libcurl/7.35.0
OpenSSL/1.0.1f zlib/1.2.8 libidn/1.28 librtmp/2.3

[dpull]

/* true globals -- for curl_global_init() and curl_global_cleanup() */
static unsigned int initialized;

中这个变量是没有加锁的，可否尝试将 webclient_destory 和 webclient_create 增加一个Mutex测试一下，看可否解决问题

[xzzwandi]

我现在尝试只使用一个webclient实例，这个问题不是很好重现，我再运行观察一段时间看看是否解决吧。谢谢。

[dpull]

https://curl.haxx.se/libcurl/c/curl_global_init.html

This function is not thread safe. You must not call it when any other thread in the program (i.e. a thread sharing the same memory) is running. This doesn't just mean no other thread that is using libcurl. Because curl_global_init calls functions of other libraries that are similarly thread unsafe, it could conflict with any other thread that uses these other libraries.

我本周末补充上mutex 保证线程安全

[dpull]

我提交了 你看看

[xzzwandi]

好的，多谢了！