www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - rdmd main.d leads to Segmentation fault

reply Timoses <timosesu gmail.com> writes:
Hey,

simple hello world crashes with segfault:

import std.stdio;
void main()
{
     writeln("hi");
}

$ rdmd main.d
Segmentation fault

Same problem with a vibe.d project.

Just set up this VirtualBox


$ dmd --version
DMD32 D Compiler v2.078.1
$ rdmd --version
rdmd build 20180121
...
$ uname -a

4.9.65-3+deb9u1 (2017-12-23) i686 GNU/Linux


$ dmd main.d
$ gdb ./main
(gdb) r
Starting program: /home/timoses/test3/main
[Thread debugging using libthread_db enabled]
Using host libthread_db library 
"/lib/i386-linux-gnu/libthread_db.so.1".

Program received signal SIGSEGV, Segmentation fault.
0x00432e04 in _d_dso_registry ()
(gdb) bt




argc=1, argv=0xbffff784,
     init=0x45c040 <__libc_csu_init>, fini=0x45c0a0 
<__libc_csu_fini>,
     rtld_fini=0xb7feb080 <_dl_fini>, stack_end=0xbffff77c) at 
../csu/libc-start.c:247




Any ideas?
Jan 26 2018
next sibling parent reply Fra Mecca <me francescomecca.eu> writes:
On Friday, 26 January 2018 at 22:40:29 UTC, Timoses wrote:

 Hey,

 simple hello world crashes with segfault:

 [...]


Where did you get the D toolchain?
Does the same segmentation fault happen with dmd or gdc or ldc? 
(dmd should be the more probable)

Can you compile a dub project?
Jan 26 2018
parent reply Timoses <timosesu gmail.com> writes:
On Saturday, 27 January 2018 at 01:23:44 UTC, Fra Mecca wrote:

 On Friday, 26 January 2018 at 22:40:29 UTC, Timoses wrote:

 Hey,

 simple hello world crashes with segfault:

 [...]


 Where did you get the D toolchain?


Got it from here:
http://d-apt.sourceforge.net/
with

$ apt-get install dmd-compiler


 Does the same segmentation fault happen with dmd or gdc or ldc? 
 (dmd should be the more probable)


dmd: generated executable also leads to segfault
gdc: works...
ldc: didn't try

$ gdc --version
gdc (Debian 6.3.0-18) 6.3.0 20170516


 Can you compile a dub project?


No, leads to executable causing a segfault as well


So, it does compile, but compiled executable with dmd seems to 
cause segfault.
Jan 28 2018
parent Kagamin <spam here.lot> writes:
On Sunday, 28 January 2018 at 21:41:39 UTC, Timoses wrote:

 Got it from here:
 http://d-apt.sourceforge.net/
 with

 $ apt-get install dmd-compiler


Sometimes such crashes can be caused by files left from previous 
installation. Or maybe confusion between gdc and dmd files.
Jan 30 2018
prev sibling next sibling parent reply Kagamin <spam here.lot> writes:
On Friday, 26 January 2018 at 22:40:29 UTC, Timoses wrote:

 Program received signal SIGSEGV, Segmentation fault.
 0x00432e04 in _d_dso_registry ()
 (gdb) bt




 argc=1, argv=0xbffff784,
     init=0x45c040 <__libc_csu_init>, fini=0x45c0a0 
 <__libc_csu_fini>,
     rtld_fini=0xb7feb080 <_dl_fini>, stack_end=0xbffff77c) at 
 ../csu/libc-start.c:247




 Any ideas?


What `disas` command shows at the point of failure?
Jan 27 2018
parent reply Timoses <timosesu gmail.com> writes:
On Saturday, 27 January 2018 at 21:04:07 UTC, Kagamin wrote:

 On Friday, 26 January 2018 at 22:40:29 UTC, Timoses wrote:

 Program received signal SIGSEGV, Segmentation fault.
 0x00432e04 in _d_dso_registry ()
 (gdb) bt




 argc=1, argv=0xbffff784,
     init=0x45c040 <__libc_csu_init>, fini=0x45c0a0 
 <__libc_csu_fini>,
     rtld_fini=0xb7feb080 <_dl_fini>, stack_end=0xbffff77c) at 
 ../csu/libc-start.c:247




 Any ideas?


 What `disas` command shows at the point of failure?


How would I do that?

$ (gdb) info line main
No line number information available for address 0x431c24 <main>
$ (gdb) info line 0x431c24
Function "0x431c24" not defined.

?
Jan 28 2018
parent reply Kagamin <spam here.lot> writes:
On Sunday, 28 January 2018 at 22:02:11 UTC, Timoses wrote:

 How would I do that?


https://forum.dlang.org/post/mailman.39.1510078013.9493.digitalmars
d-ldc puremagic.com like this
Jan 30 2018
parent reply Timoses <timosesu gmail.com> writes:
On Tuesday, 30 January 2018 at 14:08:35 UTC, Kagamin wrote:

 On Sunday, 28 January 2018 at 22:02:11 UTC, Timoses wrote:

 How would I do that?


 https://forum.dlang.org/post/mailman.39.1510078013.9493.digitalmars
d-ldc puremagic.com like this


Thanks!

I did

$ gdb main
$ (gdb) set logging on
$ (gdb) r
$ (gdb) disas

Output:
https://pastebin.com/raw/SSx0P1Av

Helps?
Jan 30 2018
parent reply Kagamin <spam here.lot> writes:
On Tuesday, 30 January 2018 at 16:56:28 UTC, Timoses wrote:

 Output:
 https://pastebin.com/raw/SSx0P1Av

 Helps?


Looks like TLS is not initialized.
Jan 31 2018
parent reply Timoses <timosesu gmail.com> writes:
On Wednesday, 31 January 2018 at 08:40:26 UTC, Kagamin wrote:

 On Tuesday, 30 January 2018 at 16:56:28 UTC, Timoses wrote:

 Output:
 https://pastebin.com/raw/SSx0P1Av

 Helps?


 Looks like TLS is not initialized.


And I would need to do what about it?

Sorry, I'm not familiar with assembly code stuff in detail.
Jan 31 2018
parent reply Kagamin <spam here.lot> writes:
On Wednesday, 31 January 2018 at 16:59:15 UTC, Timoses wrote:

 And I would need to do what about it?

 Sorry, I'm not familiar with assembly code stuff in detail.


You can try to see if it works on another distro or version.
Feb 01 2018
parent reply Timoses <timosesu gmail.com> writes:
On Thursday, 1 February 2018 at 09:01:34 UTC, Kagamin wrote:

 On Wednesday, 31 January 2018 at 16:59:15 UTC, Timoses wrote:

 And I would need to do what about it?

 Sorry, I'm not familiar with assembly code stuff in detail.


 You can try to see if it works on another distro or version.


It does work on debian jessie. Stretch failed..
Feb 04 2018
parent reply Martin Tschierschke <mt smartdolphin.de> writes:
On Sunday, 4 February 2018 at 11:50:05 UTC, Timoses wrote:

 On Thursday, 1 February 2018 at 09:01:34 UTC, Kagamin wrote:

 On Wednesday, 31 January 2018 at 16:59:15 UTC, Timoses wrote:

 And I would need to do what about it?

 Sorry, I'm not familiar with assembly code stuff in detail.


 You can try to see if it works on another distro or version.


 It does work on debian jessie. Stretch failed..


It looks like I have the same issue, I got:
gdb hiworld core
Core was generated by `./hiworld'.
Program terminated with signal SIGSEGV, Segmentation fault.


ldc2 is working.

My system: Ubuntu 17.10 32 Bit

i686 i686 GNU/Linux

(No problems on our office system Ubuntu 16.04 64Bit, too)

I tried older DMD (from 2017 and 2016!), because I already used 
DMD on this system, but I got the same same effect.

I am unfamiliar with debugging (gdb etc.) so any hint would be 
appreciated!
Feb 12 2018
parent reply Kagamin <spam here.lot> writes:
On Tuesday, 13 February 2018 at 06:53:46 UTC, Martin Tschierschke 
wrote:

 I am unfamiliar with debugging (gdb etc.) so any hint would be 
 appreciated!


https://issues.dlang.org/show_bug.cgi?id=18350 - maybe adjust bug 
severity.
Feb 14 2018
parent Martin Tschierschke <mt smartdolphin.de> writes:
On Wednesday, 14 February 2018 at 10:28:51 UTC, Kagamin wrote:

 On Tuesday, 13 February 2018 at 06:53:46 UTC, Martin 
 Tschierschke wrote:

 I am unfamiliar with debugging (gdb etc.) so any hint would be 
 appreciated!


 https://issues.dlang.org/show_bug.cgi?id=18350 - maybe adjust 
 bug severity.


I added a comment and increased prio, OK to do it like this?
I may post my gdb results in the MET evening later, should I?
Feb 14 2018
prev sibling parent reply Stefan Koch <uplink.coder googlemail.com> writes:
On Friday, 26 January 2018 at 22:40:29 UTC, Timoses wrote:

 Hey,

 simple hello world crashes with segfault:

 [...]


I can not reproduce this.
Jan 29 2018
parent Timoses <timosesu gmail.com> writes:
On Monday, 29 January 2018 at 15:03:48 UTC, Stefan Koch wrote:

 On Friday, 26 January 2018 at 22:40:29 UTC, Timoses wrote:

 Hey,

 simple hello world crashes with segfault:

 [...]


 I can not reproduce this.


Well, the cause is not yet uncovered, I suppose...

Any more ideas?
Jan 29 2018