Project

General

Profile

Bug #6009

Rapid signal delivery via kill(2) causes SystemStackError

Added by drbrain (Eric Hodel) over 8 years ago. Updated over 7 years ago.

Status:
Closed
Priority:
Normal
Target version:
-
ruby -v:
ruby 2.0.0dev (2012-02-11 trunk 34547) [x86_64-darwin11.2.0]
Backport:
[ruby-core:42524]

Description

=begin
Running the following program with a trivial signal handler can crash with a SystemStackError if signals are delivered rapidly:

$ ruby -ve 'p Process.pid; trap "USR1" do 10 ** 100; end; sleep'
ruby 2.0.0dev (2012-02-11 trunk 34547) [x86_64-darwin11.2.0]
4504
-e: SystemStackError

In a separate terminal:

ruby -e 'loop do Process.kill "USR1", 4504 end'

As each signal is delivered, ruby interrupts the current signal handler to perform the newly arrived one and quickly runs out of stack.

With an empty system handler the SystemStackError takes slightly longer to occur.

This also occurs with 1.9.3-p0:

$ ~/.multiruby/install/1.9.3-p0/bin/ruby -ve 'p Process.pid; trap "USR1" do 10 ** 100; end; sleep'
ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-darwin11.2.0]
4529
-e: SystemStackError

and with ruby-1.9.2-p290:

$ ~/.multiruby/install/1.9.2-p290/bin/ruby -ve 'p Process.pid; trap "USR1" do 10 ** 100; end; sleep'
ruby 1.9.2p290 (2011-07-09 revision 32553) [x86_64-darwin11.2.0]
4534

ruby-1.8.7-p330 exits with a zero exit code:

$ ~/.multiruby/install/1.8.7-p330/bin/ruby -ve 'p Process.pid; trap "USR1" do 10 ** 100; end; sleep'
ruby 1.8.7 (2010-12-23 patchlevel 330) [i686-darwin11.2.0]
4564
$ echo $?
0

=end


Related issues

Has duplicate Ruby master - Bug #4909: trapハンドラは再入されてはいけないのではないか?Closedkosaki (Motohiro KOSAKI)06/20/2011Actions

Also available in: Atom PDF