Last active
September 28, 2017 17:50
-
-
Save mbj/31163a8e712573877268 to your computer and use it in GitHub Desktop.
segfault.rb
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| class Foo | |
| include Enumerable | |
| def each(&block) | |
| [].__send__(:each, &block) | |
| rescue Exception | |
| end | |
| def more | |
| to_a # any method from enumerable | |
| [self] | |
| end | |
| end | |
| def call(resource) | |
| resource.more.each(&method(__method__)) | |
| end | |
| Thread.new do | |
| Kernel.catch(Exception.new) do | |
| call(Foo.new) | |
| end | |
| end.join | |
| # It is expected to: | |
| # a) Raise a system stack error (expected) | |
| # b) Segfault (Bug in MRI I'm hunting) | |
| # c) Hang (Serve bug in MRI I'm hunting) please also check via strace / gdb if a segfault happened and the signal handler is trapped in a futex | |
| # Context: https://github.com/mbj/mutant/issues/265 |
~ ❯❯❯ lsb_release -a
No LSB modules are available.
Distributor ID: Debian
Description: Debian GNU/Linux testing (jessie)
Release: testing
Codename: jessie
~ ❯❯❯ uname -a ⏎
Linux matt-desktop 3.2.0-4-amd64 #1 SMP Debian 3.2.63-2 x86_64 GNU/Linux
~ ❯❯❯ cat <<EOF | ruby ⏎
pipe heredoc> class Foo
pipe heredoc> include Enumerable
pipe heredoc>
pipe heredoc> def each(&block)
pipe heredoc> [].__send__(:each, &block)
pipe heredoc> rescue Exception
pipe heredoc> end
pipe heredoc>
pipe heredoc> def more
pipe heredoc> to_a # any method from enumerable
pipe heredoc> [self]
pipe heredoc> end
pipe heredoc> end
pipe heredoc>
pipe heredoc> def call(resource)
pipe heredoc> resource.more.each(&method(__method__))
pipe heredoc> end
pipe heredoc>
pipe heredoc> Thread.new do
pipe heredoc> Kernel.catch(Exception.new) do
pipe heredoc> call(Foo.new)
pipe heredoc> end
pipe heredoc> end.join
pipe heredoc> EOF
[1] 2066 done cat <<<'' |
2067 segmentation fault ruby
~ ❯❯❯ ruby --version ⏎
ruby 2.0.0p353 (2013-11-22 revision 43784) [x86_64-linux]
~ ❯❯❯ cat <<EOF | ruby
class Foo
include Enumerable
def each(&block)
[].__send__(:each, &block)
rescue Exception
end
def more
to_a # any method from enumerable
[self]
end
end
def call(resource)
resource.more.each(&method(__method__))
end
Thread.new do
Kernel.catch(Exception.new) do
call(Foo.new)
end
end.join
EOF
[1] 2265 done cat <<<'' |
2266 segmentation fault ruby
~ ❯❯❯ ruby --version ⏎
ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-linux]
ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-darwin13.0]
Mac OS x 10.10
SystemStackError: stack level too deep
from /Users/tj/.rvm/rubies/ruby-2.1.3/lib/ruby/2.1.0/irb/workspace.rb:86
pete@balloon:/tmp$ uname -a
Linux balloon 3.13.0-37-generic #64-Ubuntu SMP Mon Sep 22 21:28:38 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
pete@balloon:/tmp$ for i in 1.9.3 2.1.2 2.1.3 ; do rvm $i do ruby -v segfault.rb ; done
ruby 1.9.3p547 (2014-05-14 revision 45962) [x86_64-linux]
segfault.rb:19: stack level too deep (SystemStackError)
ruby 2.1.2p95 (2014-05-08 revision 45877) [x86_64-linux]
Segmentation fault
ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-linux]
segfault.rb:10: [BUG] vm_call_cfunc - cfp consistency error
ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-linux]
Segmentation fault
Additionally, 2.0.0-p576 just hangs.
Fedora Linux x86_64, compiled with ruby-install:
- ruby-2.0.0-p353: stack level too deep (SystemStackError)
- ruby-2.1.2(-p95): stack level too deep (SystemStackError)
- ruby-2.1.3(-p242): hangs
This seems to imply the Segfault is due to the version of gcc or custom CFLAGS? Here is my gcc version:
gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-7)
OS X 10.10 (14A389)
-=>> ruby --version
ruby 2.1.2p95 (2014-05-08 revision 45877) [x86_64-darwin13.0]
Just hangs. Can provide a spindump.
% system_profiler SPSoftwareDataType -detailLevel mini
# ...
System Version: OS X 10.9.5 (13F34)
Kernel Version: Darwin 13.4.0
% gcc -v
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 6.0 (clang-600.0.54) (based on LLVM 3.5svn)
Target: x86_64-apple-darwin13.4.0
Thread model: posix
% ruby-install --version
ruby-install: 0.4.3
% ruby -v
2.0.0p594 (2014-10-27) [x86_64-darwin13.4.0]
% ruby segfault.rb
segfault.rb:23: stack level too deep (SystemStackError)
% ruby -v
ruby 2.1.0p0 (2013-12-25 revision 44422) [x86_64-darwin13.0]
% ruby segfault.rb
(hangs)
% ruby -v
ruby 2.1.1p76 (2014-02-24 revision 45161) [x86_64-darwin13.0]
% ruby segfault.rb
(hangs)
% ruby -v
ruby 2.1.2p95 (2014-05-08 revision 45877) [x86_64-darwin13.0]
% ruby segfault.rb
(hangs)
% ruby -v
ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-darwin13.0]
% ruby segfault.rb
segfault.rb:23: stack level too deep (SystemStackError)
% ruby -v
ruby 2.1.4p265 (2014-10-27 revision 48166) [x86_64-darwin13.0]
% ruby segfault.rb
segfault.rb:23: stack level too deep (SystemStackError)
% ruby -v
ruby 2.2.0preview1 (2014-09-17 trunk 47616) [x86_64-darwin13]
% ruby segfault.rb
(hangs)
Ubuntu 14.04
Kernel Linux lutze 3.13.0-39-generic #66-Ubuntu SMP Tue Oct 28 13:30:27 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
gcc (Ubuntu 4.8.2-19ubuntu1) 4.8.2
1.9.3 to 2.1.2 : normal behaviour
ruby-1.9.3-p484
ruby 1.9.3p484 (2013-11-22) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:19: stack level too deep (SystemStackError)
ruby-1.9.3-p545
ruby 1.9.3p545 (2014-02-24) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:19: stack level too deep (SystemStackError)
ruby-1.9.3-p547
ruby 1.9.3p547 (2014-05-14) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:19: stack level too deep (SystemStackError)
ruby-2.0.0-p451
ruby 2.0.0p451 (2014-02-24) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:23: stack level too deep (SystemStackError)
ruby-2.1.1
ruby 2.1.1p76 (2014-02-24 revision 45161) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:23: stack level too deep (SystemStackError)
ruby-2.1.1-tco
ruby 2.1.1p76 (2014-02-24 revision 45161) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:23: stack level too deep (SystemStackError)
ruby-2.1.2
ruby 2.1.2p95 (2014-05-08 revision 45877) [x86_64-linux]
/home/arne/projects/ruby-tmp/00209.rb:23: stack level too deep (SystemStackError)
Getting a segfault on 2.1.3, 2.1.4, and trunk
ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-linux]
ruby 2.1.4p265 (2014-10-27 revision 48166) [x86_64-linux]
ruby 2.2.0dev (2014-10-29 trunk 48188) [x86_64-linux]
Reported to ruby-lang: https://bugs.ruby-lang.org/issues/10460
I knew that never cleaning up would come in handy:
OS X 10.9.5, all rubies built with ruby-install
[ skade ~ ] chruby-exec ruby-1.9.3-p484 segfault.rb
chruby-exec: COMMAND required
[ skade ~ ] chruby-exec ruby-1.9.3-p484 ruby segfault.rb
chruby-exec: COMMAND required
[ skade ~ ] chruby-exec ruby-1.9.3-p484 -- ruby segfault.rb
segfault.rb:19: stack level too deep (SystemStackError)
[ skade ~ ] chruby-exec ruby-2.0.0-p247 -- ruby segfault.rb
segfault.rb:23: stack level too deep (SystemStackError)
[ skade ~ ] chruby-exec ruby-2.0.0-p353 -- ruby segfault.rb
segfault.rb:23: stack level too deep (SystemStackError)
[ skade ~ ] chruby-exec ruby-2.1.0 -- ruby segfault.rb
<hangs>
[ skade ~ ] chruby-exec ruby-2.1.1 -- ruby segfault.rb
<hangs>
gcc is clang by the way:
[ skade ~ ] gcc -v
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer//usr --with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 6.0 (clang-600.0.51) (based on LLVM 3.5svn)
Target: x86_64-apple-darwin13.4.0
Thread model: posix
dtruss output for ruby-2.1.0:
@ALL When it hangs, can you post an strace of the hanging process?
@mbj: I used dtruss (OS X) and added a link to the gist with the output.
@skade, thx.
Some more info
On trunk from a week ago, compiled with -O0, -O1, -O2, it would work correctly (SystemStackError). This is for this commit
commit 98ea62756142b6c9cc030e99b6d936d986ce463c
Author: nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
Date: Wed Oct 29 12:13:26 2014 +0000
parse.y: warn circular argument reference
* parse.y (gettable_gen): warn circular argument reference, for
transition from 2.1 and earlier. [ruby-core:65990] [Bug #10314]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48188 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
With -O3 it would segfault in rb_const_get (variable.c:2658) on this line
if (tbl && st_lookup(tbl, (st_data_t)id, &val)) {On current trunk
commit 2f00182634eaa0389b46322301795de475f46cec
Author: nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
Date: Tue Nov 4 07:25:41 2014 +0000
suppress warnings
* vm.c (rb_vm_make_proc): cast to suppress warning by VC6.
* ext/-test-/win32/console/attribute.c (console_set_attribute):
ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48266 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
It works correctly with -O0 and -O2, gives a segfault with -O1, and hangs with -O3.
Straces of the two problematic runs:
I just tried with freshly released 2.1.5 and it still segfaults. :(
This problem was probably solved with: ruby/ruby@8fe95fe
2.2.3 and 2.2.2 give me the same results.
ruby segfault.rb
segfault.rb:10:in `to_a': stack level too deep (SystemStackError)
from segfault.rb:10:in `more'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
... 1189 levels...
from segfault.rb:16:in `call'
from segfault.rb:21:in `block (2 levels) in <main>'
from segfault.rb:20:in `catch'
from segfault.rb:20:in `block in <main>'
ruby --version
ruby 2.2.3p173 (2015-08-18 revision 51636) [x86_64-darwin14]
of note, 2.3dev just hangs for me.
[mathieu@hz550 loyalty]$ rvm use 2.2.8
Using /home/mathieu/.rvm/gems/ruby-2.2.8
[mathieu@hz550 loyalty]$ ruby segfault.rb
segfault.rb:10:in `to_a': stack level too deep (SystemStackError)
from segfault.rb:10:in `more'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
... 1103 levels...
from segfault.rb:16:in `call'
from segfault.rb:21:in `block (2 levels) in <main>'
from segfault.rb:20:in `catch'
from segfault.rb:20:in `block in <main>'
[mathieu@hz550 loyalty]$ ruby -v
ruby 2.2.8p477 (2017-09-14 revision 59906) [x86_64-linux]
[mathieu@hz550 loyalty]$ rvm use 2.3.5
Using /home/mathieu/.rvm/gems/ruby-2.3.5
[mathieu@hz550 loyalty]$ ruby segfault.rb
segfault.rb:10:in `to_a': stack level too deep (SystemStackError)
from segfault.rb:10:in `more'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
from segfault.rb:16:in `each'
from segfault.rb:16:in `call'
... 809 levels...
from segfault.rb:16:in `call'
from segfault.rb:21:in `block (2 levels) in <main>'
from segfault.rb:20:in `catch'
from segfault.rb:20:in `block in <main>'
[mathieu@hz550 loyalty]$ rvm use 2.4.2
Using /home/mathieu/.rvm/gems/ruby-2.4.2
[mathieu@hz550 loyalty]$ ruby segfault.rb
segfault.rb:10: [BUG] vm_call_cfunc - cfp consistency error
ruby 2.4.2p198 (2017-09-14 revision 59899) [x86_64-linux]
-- Control frame information -----------------------------------------------
Segmentation fault (core dumped)
[mathieu@hz550 loyalty]$
ArchLinux 64bit
Linux hz550 4.12.12-1-ARCH #1 SMP PREEMPT Sun Sep 10 09:41:14 CEST 2017 x86_64 GNU/Linux
on debian 14.04 I am getting segfault.rb:10: [BUG] vm_call_cfunc - cfp consistency error on 2.2.8
mathieu@ubuntu-1404 (i-123):~$ ruby segfault.rb
Segmentation fault (core dumped)
mathieu@ubuntu-1404 (i-123):~$
with brighbox compiled packages
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Ubuntu 12.04 using Ruby 2.1.3 from the Brightbox PPA: