GNU bug report logs - #11645
Unable to cross compile guile 2.0.5 for MIPS/PPC on x86_64

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 11645 in the body.
You can then email your comments to 11645 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: "Marinescu, Bogdan A" <bogdan.a.marinescu <at> intel.com>
To: bug-guile <at> gnu.org
Subject: Unable to cross compile guile 2.0.5 for MIPS/PPC on x86_64
Date: Thu, 7 Jun 2012 14:00:19 +0300

Compiling guile 2.0.5 for MIPS and PowerPC (both in big endian mode)
fails with a segmentation fault. guile 2.0.5 is built first for x86_64
and this version is used to cross-compile guile for MIPS. The
compilation stops with this error:

GUILE_AUTO_COMPILE=0      \
  ../meta/uninstalled-env     \
  guild compile --target="mips-poky-linux-gnu" -Wunbound-variable
-Warity-mismatch -Wformat \
    -L "/ssd/build/tmp/work/mips32-poky-linux/guile-2.0.5-r0/guile-2.0.5/module"
-L "/ssd/build/tmp/work/mips32-poky-linux/guile-2.0.5-r0/guile-2.0.5/module"
  \
    -L "/ssd/build/tmp/work/mips32-poky-linux/guile-2.0.5-r0/guile-2.0.5/guile-readline"
     \
    -o "ice-9/occam-channel.go" "ice-9/occam-channel.scm"
make[2]: *** [ice-9/occam-channel.go] Segmentation fault (core dumped)

This is the stack trace:

Core was generated by
`/ssd/build/tmp/sysroots/x86_64-linux/usr/bin/x86_64-linux-guile.real
-e (@@ (gu'.
Program terminated with signal 11, Segmentation fault.
#0  0x00007f76a81450fd in vm_regular_engine (vm=0x26b76c0,
program=0x2d0a8e0, argv=0x0, nargs=0) at vm-i-loader.c:80
80    NEXT;
(gdb) bt
#0  0x00007f76a81450fd in vm_regular_engine (vm=0x26b76c0,
program=0x2d0a8e0, argv=0x0, nargs=0) at vm-i-loader.c:80
#1  0x00007f76a8160e09 in scm_c_vm_run (vm=0x26b76c0,
program=0x2b8dc60, argv=0x0, nargs=0) at vm.c:567
#2  0x00007f76a8161ae5 in scm_load_compiled_with_vm (file=0x2b8dca0) at vm.c:878
#3  0x00007f76a8096a86 in scm_primitive_load_path (args=0x2aec5d0) at load.c:913
#4  0x00007f76a8136555 in vm_regular_engine (vm=0x26b76c0,
program=0x7f76a83e36a0, argv=0x7fff60d8d860, nargs=1) at
vm-i-system.c:892
#5  0x00007f76a8160e09 in scm_c_vm_run (vm=0x26b76c0,
program=0x26c69c0, argv=0x7fff60d8d858, nargs=1) at vm.c:567
#6  0x00007f76a800010c in scm_primitive_eval (exp=0x28907c0) at eval.c:685
#7  0x00007f76a80001d2 in scm_eval (exp=0x28907c0,
module_or_state=0x2871d80) at eval.c:719
#8  0x00007f76a80f6fd4 in scm_shell (argc=19, argv=0x7fff60d8eb68) at
script.c:441
#9  0x0000000000400bc8 in inner_main (closure=0x0, argc=19,
argv=0x7fff60d8eb68) at guile.c:62
#10 0x00007f76a808ad8b in invoke_main_func (body_data=0x7fff60d8ea30)
at init.c:336
#11 0x00007f76a7ff25f2 in c_body (d=0x7fff60d8e910) at continuations.c:512
#12 0x00007f76a812a2e4 in apply_catch_closure (clo=0x2870b40,
args=0x304) at throw.c:146
#13 0x00007f76a8136d10 in vm_regular_engine (vm=0x26b76c0,
program=0x2870a00, argv=0x7fff60d8e7d0, nargs=1) at vm-i-system.c:960
#14 0x00007f76a8160e09 in scm_c_vm_run (vm=0x26b76c0,
program=0x279ea50, argv=0x7fff60d8e7b0, nargs=4) at vm.c:567
#15 0x00007f76a7fff7d3 in scm_call_4 (proc=0x279ea50, arg1=0x404,
arg2=0x2870b40, arg3=0x2870b20, arg4=0x2870b00) at eval.c:507
#16 0x00007f76a8129f67 in scm_catch_with_pre_unwind_handler
(key=0x404, thunk=0x2870b40, handler=0x2870b20,
pre_unwind_handler=0x2870b00) at throw.c:86
#17 0x00007f76a812a3cc in scm_c_catch (tag=0x404, body=0x7f76a7ff25ca
<c_body>, body_data=0x7fff60d8e910, handler=0x7f76a7ff2601
<c_handler>, handler_data=0x7fff60d8e910,
pre_unwind_handler=0x7f76a7ff2655 <pre_unwind_handler>,
    pre_unwind_handler_data=0x26b75f0) at throw.c:213
#18 0x00007f76a7ff2466 in scm_i_with_continuation_barrier
(body=0x7f76a7ff25ca <c_body>, body_data=0x7fff60d8e910,
handler=0x7f76a7ff2601 <c_handler>, handler_data=0x7fff60d8e910,
pre_unwind_handler=0x7f76a7ff2655 <pre_unwind_handler>,
    pre_unwind_handler_data=0x26b75f0) at continuations.c:450
#19 0x00007f76a7ff26ec in scm_c_with_continuation_barrier
(func=0x7f76a808ad38 <invoke_main_func>, data=0x7fff60d8ea30) at
continuations.c:546
#20 0x00007f76a8126b82 in with_guile_and_parent (base=0x7fff60d8e980,
data=0x7fff60d8e9c0) at threads.c:902
#21 0x00007f76a7d28d75 in GC_call_with_stack_base () from
/ssd/build/tmp/sysroots/x86_64-linux/usr/bin/../../usr/lib/libgc.so.1
#22 0x00007f76a8126c62 in scm_i_with_guile_and_parent
(func=0x7f76a808ad38 <invoke_main_func>, data=0x7fff60d8ea30,
parent=0x0) at threads.c:945
#23 0x00007f76a8126c8e in scm_with_guile (func=0x7f76a808ad38
<invoke_main_func>, data=0x7fff60d8ea30) at threads.c:951
#24 0x00007f76a808ad19 in scm_boot_guile (argc=19,
argv=0x7fff60d8eb68, main_func=0x400ba4 <inner_main>, closure=0x0) at
init.c:319
#25 0x0000000000400c20 in main (argc=19, argv=0x7fff60d8eb68) at guile.c:79
(gdb)

While debugging this issue, I modified the "NEXT" macro in vm-engine.h
to output information about the current IP and the next instruction:

#define NEXT          \
{           \
  NEXT_HOOK ();         \
  CHECK_STACK_LEAK ();                          \
  if(getenv("GUILELOGS")) fprintf(stderr,"->NEXT:%d IP=%p\n", (*ip) &
SCM_VM_INSTRUCTION_MASK, ip); \
  NEXT_JUMP ();         \
}

Below are the last lines of the output that I get with this
modification (the complete output is too large to attach):

->NEXT:11 IP=0x260df6d
->NEXT:1 IP=0x260df6e
->NEXT:52 IP=0x7f76a85e2d0a
->NEXT:21 IP=0x7f76a85e2d0b
->NEXT:62 IP=0x7f76a85e2d0d
->NEXT:93 IP=0x7f76a84e3548
->NEXT:52 IP=0x7f76a84e354a
->NEXT:27 IP=0x7f76a84e354b
->NEXT:76 IP=0x7f76a84e354d
->NEXT:62 IP=0x7f76a84e354f
->NEXT:44 IP=0x7f76a8187c50
->NEXT:19 IP=0x7f76a8187c53
->NEXT:55 IP=0x7f76a8187c55
->NEXT:2 IP=0x7f76a84e3554
->NEXT:34 IP=0x7f76a84e3555
->NEXT:52 IP=0x7f76a84e355c
->NEXT:76 IP=0x7f76a84e355d
->NEXT:53 IP=0x7f76a84e355f
->NEXT:93 IP=0x28acaa0
->NEXT:6 IP=0x28acaa2
->NEXT:6 IP=0x28acaa3
->NEXT:18 IP=0x28acaa4
->NEXT:29 IP=0x28acaa7
->NEXT:52 IP=0x28acaa9
->NEXT:21 IP=0x28acaaa
->NEXT:0 IP=0x28acaac
->NEXT:0 IP=0x28acaad
->NEXT:0 IP=0x28acaae
->NEXT:104 IP=0x28acaaf
Segmentation fault (core dumped)

Message #10 received at 11645-done <at> debbugs.gnu.org (full text, mbox):

From: ludo <at> gnu.org (Ludovic Courtès)
To: "Marinescu\, Bogdan A" <bogdan.a.marinescu <at> intel.com>
Cc: 11645-done <at> debbugs.gnu.org
Subject: Re: bug#11645: Unable to cross compile guile 2.0.5 for MIPS/PPC on
	x86_64
Date: Wed, 20 Jun 2012 15:34:09 +0200

Hi,

"Marinescu, Bogdan A" <bogdan.a.marinescu <at> intel.com> skribis:

> Compiling guile 2.0.5 for MIPS and PowerPC (both in big endian mode)
> fails with a segmentation fault. guile 2.0.5 is built first for x86_64
> and this version is used to cross-compile guile for MIPS. The
> compilation stops with this error:
>
> GUILE_AUTO_COMPILE=0      \
>   ../meta/uninstalled-env     \
>   guild compile --target="mips-poky-linux-gnu" -Wunbound-variable
> -Warity-mismatch -Wformat \
>     -L "/ssd/build/tmp/work/mips32-poky-linux/guile-2.0.5-r0/guile-2.0.5/module"
> -L "/ssd/build/tmp/work/mips32-poky-linux/guile-2.0.5-r0/guile-2.0.5/module"
>   \
>     -L "/ssd/build/tmp/work/mips32-poky-linux/guile-2.0.5-r0/guile-2.0.5/guile-readline"
>      \
>     -o "ice-9/occam-channel.go" "ice-9/occam-channel.scm"
> make[2]: *** [ice-9/occam-channel.go] Segmentation fault (core dumped)

This is fixed by commit f3b312a19d70293d7a3407fc4ef479183edd7cca, which
will be in 2.0.6 (the fix was simple, but the investigation trickier ;-)).

There were only little-endian cross-compilation targets under continuous
integration, so I’ve added this one:
<http://hydra.nixos.org/jobset/gnu/guile-2-0>.

Thanks for the report!

Ludo’.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.