topic Re: aCC 3.65 core dump in Operating System - Linux

aCC 3.65 core dump

StefanoIntroini — Fri, 07 Apr 2006 03:35:52 GMT

Hi,
whilst using aCC 3.65 (trial version), i've got this error:

Signal 11
( 0) 0x00297cc8 sighandler__FiT1 + 0x138 [/opt/aCC/lbin/ctcom.pa20]
( 1) 0xc020ba00 _sigreturn [/usr/lib/libc.2]
( 2) 0x002abd04 isNRVDtor__FP4Callb + 0xb8 [/opt/aCC/lbin/ctcom.pa20]
( 3) 0x002abdd0 removeAllNRVDtors__FP8Function + 0xc0 [/opt/aCC/lbin/ctcom.pa20]
( 4) 0x00168d80 expand__8FunctionFv + 0x4c0 [/opt/aCC/lbin/ctcom.pa20]
( 5) 0x00212360 ProcessExternalDeclaration__8CompilerFP11Declarationb + 0x124 [/opt/aCC/lbin/ctcom.pa20]
( 6) 0x0020b958 ProcessExternalDeclarationList__8CompilerFP11Declaration + 0x2a4 [/opt/aCC/lbin/ctcom.pa20]
( 7) 0x001e6ca8 doReduction1__FiR12ScannerValueP12ScannerValue + 0x6e0 [/opt/aCC/lbin/ctcom.pa20]
( 8) 0x001e5868 yyparse__Fv + 0xc54 [/opt/aCC/lbin/ctcom.pa20]
( 9) 0x00274878 DoCompile__8CompilerFv + 0x224 [/opt/aCC/lbin/ctcom.pa20]
(10) 0x00286ec4 DoCompile__8CompilerFP6Buffer + 0x34 [/opt/aCC/lbin/ctcom.pa20]
(11) 0x00283c6c DoCompileFile__8CompilerFPc + 0x110 [/opt/aCC/lbin/ctcom.pa20]
(12) 0x000cc844 main + 0x404 [/opt/aCC/lbin/ctcom.pa20]
(13) 0xc0143460 _start + 0xc0 [/usr/lib/libc.2]
(14) 0x000b7780 $START$ + 0x178 [/opt/aCC/lbin/ctcom.pa20]
Error (system problem) 689: # Compiler received signal 11
*** Error exit code 3

Any ideas on the reason ??, i tried to install the PHSS_34411 (in order to upgrade the compiler to 3.67) but it fails (maybe it needs the unlimited 3.65 aCC version), the error is:

The specification for this base fileset is "ACXX.ACXX,fr=C.03.65,v=HP". Since there is no fileset on the target or selected from the source which matches this specification, fileset "PHSS_34411.ACXX,r=1.0" will be excluded.

Thanks for your help.
Stefano.

Re: aCC 3.65 core dump

Peter Godron — Fri, 07 Apr 2006 04:00:42 GMT

Stefano,
did this code compile before on another compiler?
Signal 11 are normally SIGSEGV (Signal Segment Violation) were a process is attempting to overstep a resource constraint.
What is you maxssize set to ? ulimit -a should show it.

Re: aCC 3.65 core dump

StefanoIntroini — Fri, 07 Apr 2006 04:21:50 GMT

Hi,
the sorce code was developed with aCC 3.31 on HP 11.0, we are going to migrate it to HP11.11 32 bits.

the Ulimit -a shows:
time(seconds) unlimited
file(blocks) unlimited
data(kbytes) 2097152
stack(kbytes) 131072
memory(kbytes) unlimited
coredump(blocks) 4194303
nofiles(descriptors) 8192

From SAM:
maxssiz 134217728
massiz_64 1073741824

Re: aCC 3.65 core dump

Arunvijai_4 — Fri, 07 Apr 2006 04:32:07 GMT

Hi Stefano,

What is your debug level set to ? If you have anything like -g0, you can remove and try compiling it.

-Arun

Re: aCC 3.65 core dump

StefanoIntroini — Fri, 07 Apr 2006 04:37:32 GMT

I've removed the -g flag but it doesen't work.

Re: aCC 3.65 core dump

Arunvijai_4 — Fri, 07 Apr 2006 04:44:45 GMT

OK, You can take a look at these threads, it could be helpful

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=825891&admit=-682735245+1144403033043+28353475
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=103917

-Arun

Re: aCC 3.65 core dump

StefanoIntroini — Fri, 07 Apr 2006 05:02:02 GMT

Hi,
i saw the threads listes,but nothing changed.

Stefano.

Re: aCC 3.65 core dump

StefanoIntroini — Fri, 07 Apr 2006 05:32:25 GMT

The only way to get rid of such errror, is to specify the O2 optimization level.

Re: aCC 3.65 core dump

StefanoIntroini — Fri, 07 Apr 2006 05:39:13 GMT

I'm sorry, ignore the previous post, i wrongly changed the source code.

Stefano

Re: aCC 3.65 core dump

Dennis Handly — Fri, 07 Apr 2006 21:32:12 GMT

This appears to be a new compiler bug. Are you using +nrv? If not, try adding it. If so, remove it.

For aC++ issues, you may want to join the CXX-DEV mailing list.
http://h21007.www2.hp.com/dspp/comm/comm_Community_Home_IDX/1,1273,,00.html
> tried to install the PHSS_34411 (in order to upgrade the compiler to 3.67)

You can't patch a trial version.

Peter Godron:
>Signal 11 are normally SIGSEGV (Signal Segment Violation) were a process is attempting to overstep a resource constraint.

This hasn't been my experience. This is just a NULL pointer dereference. If it is a stack overflow, you get a slightly different signal 11.

Re: aCC 3.65 core dump

StefanoIntroini — Mon, 10 Apr 2006 03:53:24 GMT

Hi,
i'm not using the nrv flag, i tried to add it but nothing changed.

Stefano.

Re: aCC 3.65 core dump

StefanoIntroini — Fri, 05 May 2006 08:39:46 GMT

Hi,
i've got the 3.67 version of the compiler (PHSS_34411) but the problem is still present.

Any ideas ???

Thanks
Stefano.

Re: aCC 3.65 core dump

Dennis Handly — Fri, 05 May 2006 17:24:49 GMT

> i've got the 3.67 version of the compiler (PHSS_34411) but the problem is still present.

You'll have to contact the Response Center and file a bug report.