From: Eric Wong <normalperson@...>
Date: 2018-07-29T21:39:28+00:00
Subject: [ruby-core:88170] Re: [ruby-alerts:8205] failure alert on trunk-vm-asserts@silicon-docker (NG (r64107))

ko1c-failure@atdot.net wrote:
> Alert on trunk-vm-asserts@silicon-docker
> rsult : NG (r64107)
> detail: https://gist.github.com/e8a541a114cb252685dcbd7fa0fba11e

Not much detail, there...

> test_all             [BUG] pthread_mutex_destroy: Device or resource busy (EBUSY)
> ruby 2.6.0dev (2018-07-30 trunk 64106) [x86_64-linux]
> 
> -- Control frame information -----------------------------------------------
> 
> [BUG] Segmentation fault at 0xffffffffffffffa0
> ruby 2.6.0dev (2018-07-30 trunk 64106) [x86_64-linux]

I can't find a backtrace, but I suspect it's gvl_destroy.
gvl_destroy seems pointless, anyways; and maybe we should
skip it...

It seems timer-thread removal speeds up shutdown and finds
old race conditions and bugs.  I found several old problems
the same way:

r64102 (trunk): thread_pthread.c: clear altstacks in thread cache at GVL destruction
[Bug #14939] lost trap interrupts during ec_switch
[Bug #14945] unregister_ubf_list ordering
...

Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>