SEGV/GNAT-3.13p/Florist/RedHat-7.1

SEGV/GNAT-3.13p/Florist/RedHat-7.1

[Matt Majka]

Hi!

I was wondering if anyone else has run into this and has a solution/workaround.
I'm using gnat-3.13p on a fresh RedHat-7.1 installation.  I've built
florist-3.13p
and installed it as well.  The GNAT compiler seems to be working fine with some
test programs I brought over from Solaris.  However, as soon as I try to use the
POSIX bindings, the program causes a segmentation fault prior to executing any
of my code (I don't even think it makes it to the elaboration stage). 
I've included
my simple test program and the output of GDB when I ran it with the resulting
core file.  I'm trying to port some serial IO code I wrote under a proprietary
real-time Unix to Linux and the low-level code makes extensive use of the
IEEE binding interface.  I tried generating the binder file in C to see if
I could
get a better feel for what is going on, but I had the same result.

Thanks!

Matt

-----------------

with Ada.Text_IO;
with POSIX;
with POSIX_Process_Identification;

procedure POSIX_Test is
begin

   Ada.Text_IO.Put_Line (

      POSIX.To_String ( POSIX_Process_Identification.Get_Login_Name )

      );

end POSIX_Test;

---------------------------

barnyard 13% ./posix_test 
Segmentation fault (core dumped)
barnyard 14% gdb posix_test core
GNU gdb 4.17.gnat.3.13p -1
Copyright 1998 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "i386-pc-linux"...
Core was generated by `./posix_test'.
Program terminated with signal 11, Segmentation fault.
find_solib: Can't read pathname for load map: Input/output error

#0  pthread_sigcpyset2set (dst=0xbffff588, src=0x308) at signal.c:141
signal.c:141: No such file or directory.
(gdb) backtrace
#0  pthread_sigcpyset2set (dst=0xbffff588, src=0x308) at signal.c:141
#1  0x807202f in sigprocmask (how=0, set=0xbffff5b4, oset=0x0) at signal.c:1139
#2  0x40066529 in ?? () from /lib/i686/libpthread.so.0
#3  0x4006ad25 in ?? () from /lib/i686/libpthread.so.0
#4  0x40062a1a in ?? () from /lib/i686/libpthread.so.0
#5  0x4000e0f7 in ?? ()
Current language:  auto; currently c
(gdb) quit

Re: SEGV/GNAT-3.13p/Florist/RedHat-7.1

[Florian Weimer]

matthewmajka@mac.com (Matt Majka) writes:

> I was wondering if anyone else has run into this and has a
> solution/workaround.  I'm using gnat-3.13p on a fresh RedHat-7.1
> installation.  I've built florist-3.13p and installed it as well.
> The GNAT compiler seems to be working fine with some test programs I
> brought over from Solaris.  However, as soon as I try to use the
> POSIX bindings, the program causes a segmentation fault prior to
> executing any of my code (I don't even think it makes it to the
> elaboration stage).

Have you tried to run some code which uses the tasking runtime and
does not use FLORIST?

Re: SEGV/GNAT-3.13p/Florist/RedHat-7.1

[Matt Majka]

In article <87r8x3w3wz.fsf@deneb.enyo.de>, Florian Weimer
<fw@deneb.enyo.de> wrote:

> matthewmajka@mac.com (Matt Majka) writes:
> 
> > I was wondering if anyone else has run into this and has a
> > solution/workaround.  I'm using gnat-3.13p on a fresh RedHat-7.1
> > installation.  I've built florist-3.13p and installed it as well.
> > The GNAT compiler seems to be working fine with some test programs I
> > brought over from Solaris.  However, as soon as I try to use the
> > POSIX bindings, the program causes a segmentation fault prior to
> > executing any of my code (I don't even think it makes it to the
> > elaboration stage).
> 
> Have you tried to run some code which uses the tasking runtime and
> does not use FLORIST?

Yes.  I have a simple plotting application that uses OpenGL and
tasks.  I haven't tested it extensively, but it does not crash in this
manner.  I'll try it a little more on Monday.

I also got a suggestion to try the RPMs from gnuada.org.  I'll
give that a shot also.

Re: SEGV/GNAT-3.13p/Florist/RedHat-7.1

[Tony Gair]

Are you any storage size pragmas in the task specs?
i.e. you may need to calclate the space needed to reserve space for you
variables within any tasks...
all the best
Tony
"Matt Majka" <matthewmajka@mac.com> wrote in message
news:matthewmajka-0106011720490001@10.0.1.2...
> Hi!
>
> I was wondering if anyone else has run into this and has a
solution/workaround.
> I'm using gnat-3.13p on a fresh RedHat-7.1 installation.  I've built
> florist-3.13p
> and installed it as well.  The GNAT compiler seems to be working fine with
some
> test programs I brought over from Solaris.  However, as soon as I try to
use the
> POSIX bindings, the program causes a segmentation fault prior to executing
any
> of my code (I don't even think it makes it to the elaboration stage).
> I've included
> my simple test program and the output of GDB when I ran it with the
resulting
> core file.  I'm trying to port some serial IO code I wrote under a
proprietary
> real-time Unix to Linux and the low-level code makes extensive use of the
> IEEE binding interface.  I tried generating the binder file in C to see if
> I could
> get a better feel for what is going on, but I had the same result.
>
> Thanks!
>
> Matt
>
> -----------------
>
> with Ada.Text_IO;
> with POSIX;
> with POSIX_Process_Identification;
>
> procedure POSIX_Test is
> begin
>
>    Ada.Text_IO.Put_Line (
>
>       POSIX.To_String ( POSIX_Process_Identification.Get_Login_Name )
>
>       );
>
> end POSIX_Test;
>
> ---------------------------
>
> barnyard 13% ./posix_test
> Segmentation fault (core dumped)
> barnyard 14% gdb posix_test core
> GNU gdb 4.17.gnat.3.13p -1
> Copyright 1998 Free Software Foundation, Inc.
> GDB is free software, covered by the GNU General Public License, and you
are
> welcome to change it and/or distribute copies of it under certain
conditions.
> Type "show copying" to see the conditions.
> There is absolutely no warranty for GDB.  Type "show warranty" for
details.
> This GDB was configured as "i386-pc-linux"...
> Core was generated by `./posix_test'.
> Program terminated with signal 11, Segmentation fault.
> find_solib: Can't read pathname for load map: Input/output error
>
> #0  pthread_sigcpyset2set (dst=0xbffff588, src=0x308) at signal.c:141
> signal.c:141: No such file or directory.
> (gdb) backtrace
> #0  pthread_sigcpyset2set (dst=0xbffff588, src=0x308) at signal.c:141
> #1  0x807202f in sigprocmask (how=0, set=0xbffff5b4, oset=0x0) at
signal.c:1139
> #2  0x40066529 in ?? () from /lib/i686/libpthread.so.0
> #3  0x4006ad25 in ?? () from /lib/i686/libpthread.so.0
> #4  0x40062a1a in ?? () from /lib/i686/libpthread.so.0
> #5  0x4000e0f7 in ?? ()
> Current language:  auto; currently c
> (gdb) quit

Re: SEGV/GNAT-3.13p/Florist/RedHat-7.1

[Matt Majka]

In article <1u6T6.5434$IY1.918786@news1.cableinet.net>, "Tony Gair"
<tonygair@nospammy.blueyonder.co.uk> wrote:

> Are you any storage size pragmas in the task specs?
> i.e. you may need to calclate the space needed to reserve space for you
> variables within any tasks...
> all the best
> Tony

No.  I installed the RPMs for Florist and things worked better, although
Get_Login
doesn't work because my getlogin() is broken.