Bugs: Browse | Submit New | Admin
Ubuntu 9.04 on x86_64 ruby 1.8.7 (2008-08-11 patchlevel 72) [x86_64-linux] ruby-serialport 0.7 via rubygems Ruby-serial port caused a segmentation fault when #read_timeout= was called. Example: irb(main):001:0> require 'rubygems' => true irb(main):002:0> require 'serialport' => true irb(main):003:0> sp = SerialPort.new("/dev/ttyUSB0", 9600) => #<SerialPort:0x7f9641de1910> irb(main):004:0> sp.read_timeout = 100 (irb):5: [BUG] Segmentation fault ruby 1.8.7 (2008-08-11 patchlevel 72) [x86_64-linux] When I attached gdb, following call trace was shown. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7f43b5b596f0 (LWP 484)] 0x00007f43b56b8f40 in rb_io_taint_check () from /usr/lib/libruby1.8.so.1.8 (gdb) bt #0 0x00007f43b56b8f40 in rb_io_taint_check () from /usr/lib/libruby1.8.so.1.8 #1 0x00007f43b2da9a76 in get_fd_helper (obj=18446744072429324056) at impl/posix_serialport.c:43 #2 0x00007f43b2da9c62 in sp_set_read_timeout (self=-1280227560, val=<value optimized out>) at impl/posix_serialport.c:490 #3 0x00007f43b569ac40 in ?? () from /usr/lib/libruby1.8.so.1.8 #4 0x00007f43b569adfa in ?? () from /usr/lib/libruby1.8.so.1.8
Add A Comment:
Date: 2010-02-08 16:38 Sender: Ryan Payne I concur with Tim regarding the status of this project. It appears that most of the Rubyforge projects are being phased out and GemCutter is the new gem repo. I worked with hparra to integrate the fixes that I made into his fork of serial port. Also, you no longer have to add gemcutter.org to your gem sources as the gems.rubyforge.org address is automagically forwarded to gemcutter.
Date: 2010-02-08 16:13 Sender: Tim Becker In case anyone else is having this problem: This project is/seems to be/may be more or less abandoned. There is a fork(?) of serialport at github (http://github.com/hparra/ruby-serialport/) which may be installed via gemcutter (gem sources -a http://gemcutter.org; gem install serialport). This contains the necessary fixes. I can't tell whether the Rubyforge project is abandoned or not, anyway, the github/gemcutter gem works fine with no need for patching.
Date: 2010-02-08 15:59 Sender: Tim Becker Ryan basically described it. As stated below, this also affects OSX, which is the only plattform I tested on. Patch: ================================================================= == --- ext/serialport.c (revision 79) +++ ext/serialport.c (working copy) @@ -%ld,%ld +%ld,%ld @@ * Set the state (0 or 1) of the DTR line */ static VALUE sp_set_dtr(self, val) + VALUE self, val; { return sp_set_dtr_impl(self, val); } @@ -%ld,%ld +%ld,%ld @@ * supported. */ static VALUE sp_set_flow_control(self, val) + VALUE self, val; { return sp_set_flow_control_impl(self, val); } @@ -%ld,%ld +%ld,%ld @@ * Note: Read timeouts don't mix well with multi-threading. */ static VALUE sp_set_read_timeout(self, val) + VALUE self, val; { return sp_set_read_timeout_impl(self, val); } @@ -%ld,%ld +%ld,%ld @@ * Set the state (0 or 1) of the RTS line */ static VALUE sp_set_rts(self, val) + VALUE self, val; { return sp_set_rts_impl(self, val); } @@ -%ld,%ld +%ld,%ld @@ * Note: Under Posix, write timeouts are not implemented. */ static VALUE sp_set_write_timeout(self, val) + VALUE self, val; { return sp_set_write_timeout_impl(self, val); }
Date: 2010-01-02 02:41 Sender: Ryan Payne OK. Fixed it. The problem is in serialport.c. The arguments to the sp_set_dt, sp_set_flow_control, sp_set_read_timeout, sp_set_rts and sp_set_write_timeout methods were not being defined as VALUEs. I'll be submitting the patch here in a few minutes.
Date: 2010-01-02 02:08 Sender: Ryan Payne The following all cause the same seg fault: flow_control= read_timeout= rts= dtr= I am trying to debug this and I'll keep you all posted if I get it figured out.
Date: 2010-01-01 06:52 Sender: Ryan Payne Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: KERN_INVALID_ADDRESS at address: 0x0000000001512641 0x000000010003ea8e in rb_io_taint_check () (gdb) bt #0 0x000000010003ea8e in rb_io_taint_check () #1 0x00000001001ad0a0 in get_fd_helper (obj=<value temporarily unavailable, due to optimizations>) at posix_serialport.c:47 #2 0x00000001001ad553 in sp_set_read_timeout_impl [inlined] () at /Library/Ruby/Gems/1.8/gems/ruby-serialport-0.7.1/ext/impl/ posix_serialport.c:510 #3 0x00000001001ad553 in sp_set_read_timeout (self=22095424, val=<value temporarily unavailable, due to optimizations>) at posix_serialport.c:157 #4 0x00000001000254d5 in rb_proc_call () #5 0x0000000100026297 in rb_proc_call () #6 0x0000000100020f15 in rb_provided () #7 0x00000001000311a9 in rb_eval_string () #8 0x00000001000311e3 in ruby_exec () #9 0x0000000100031205 in ruby_run () #10 0x0000000100000ef0 in main () Note: I am running the latest code out of SVN, which includes the #ifdefs for Ruby 1.9
Date: 2010-01-01 06:24 Sender: Ryan Payne Pretty much any of the setters (rts=, dts=, read_timeout=) seem to have this same issue.
Date: 2010-01-01 05:46 Sender: Ryan Payne I am also getting the same crash on OS X 10.6.2, also x86_64.
Date: 2009-11-11 07:39 Sender: Nick Wimpney irb >> require 'serialport' => true >> sp = SerialPort.new("/dev/ttyUSB0", 9600) => #<SerialPort:0x7f442bb24a78> >> sp.read_timeout = 100 (irb):3: [BUG] Segmentation fault ruby 1.8.7 (2009-06-12 patchlevel 174) [x86_64-linux]
Date: 2009-11-11 07:34 Sender: Nick Wimpney I get the same crash. I was working with this same module today on my work computer and it works fine there. Possibly only a problem on x86_64?
