03-03-2009
Aix xlc interesting SEGV on exit
Hello all,
One of the application we port to Aix from linux Segmentation faults when it exits. Here is part of backtrace of SEGV:
(dbx) where
splay(??, ??, ??) at
free_y(??, ??) at
free_common(??) at
....
exit(??) at
...
Application seem to perform everything expected well and while exiting it SEGV. By doing lot of googling learnt this could be due to heap corruption.
Tried the following:
1. setenv MALLOCDEBUG verbose
Result:
Malloc Report: The address passed to free, <x>, is
outside the valid range of addresses allocated by malloc (errno = 2).
2. setenv MALLOCTYPE debug
Run the application no SEGV at all. It exits normally.
3. setenv MALLOCTYPE watson
setenv MALLOCDEBUG catch_overflow
Run the application no SEGV at all. It exits normally.
Tried running same application same way with valgrind, totalview, MEMORY_CHECK_2 in linux to find potential coding error but found none.
Queries:
1. How can I find what is happening in memory address "x"? Who changes/disturbs it? Tried setting a watchpoint for that location but the error was:
Unreadable instruction at address 0x10000000
What else can i do? How can I find what happens in this location?
2. What difference does malloctype debug/watson does compared to default(yoktown)? Why isn't there a SEGV at that time?
3. Is there any tool to watch/examine heap?
4. Is there any other tool/option which could be used to debug this problem further?
Any pointers/suggestions/help is most welcome
Thanks in advance.
10 More Discussions You Might Find Interesting
1. Programming
when i compile with Xlc i sometimes get the following Error:
(S) Initialization between types "int" and "struct MskTtimerData*" is not allowed.
when i compile with cc with the same Flags, i only get a (W) with the same message, but it compiles fine.
How can i get the Xlc to 'ignore' the... (3 Replies)
Discussion started by: Lazzar
3 Replies
2. AIX
Hi guys,
I need to compile sap libraries, installed make, automake, gcc, libstc libgcc and everything I used to install on linux when compiling
when i run make, it aborts with the error msg
/usr/bin/sh: xlc: not found.
I see the binary xlc is missing
lslpp -l | grep -i xlc
... (3 Replies)
Discussion started by: funksen
3 Replies
3. AIX
Hello,
I'm getting "xlc: not found" error compiling code using 'make oracle' command on IBM AIX. C/C++ compiler is installed, but 'xlc' still missing. Could you please tell me which OS package contains the 'xlc' executable?
$ > make oracle
xlc -qmkshrobj -e MQStart -L/usr/mqm/lib -lmqmcs_r... (4 Replies)
Discussion started by: safir29
4 Replies
4. Programming
Executive summary:
Code (posted below) cores in AIX 5.3, despite being compiled and run successfully on several other operating systems. Code is attempting to verify that pthread_mutex_lock can be successfully aborted by siglongjmp. I do not believe this is an unreasonable requirement.
If... (1 Reply)
Discussion started by: DreamWarrior
1 Replies
5. AIX
Hi Team
I have a native applicaiton built on AIX using the xLC v8
Which could be the possible impacts from a code change point of view if I'll compile with the xLC v10?
Thanks
Marco (0 Replies)
Discussion started by: antcos
0 Replies
6. UNIX for Dummies Questions & Answers
Hello - I need to know the detail of HUP and SEGV.
I know HUP is Hangup and can be use to kill a Unix login session remotely by sending a hangup signal to the process running the login session.
Could someone tell me in detail prupose of HUP and SEGV (segmentation violation)?
I need to... (1 Reply)
Discussion started by: namasteall2000
1 Replies
7. AIX
Problem
when compiling c-source with xlc it will generate some other files .lst .call. This kind of files is generated on the current directory.
Question
Can I influence the directory where the .lst and .call is generated.
What i want is a dir stucture like this:
project
|---c-source... (0 Replies)
Discussion started by: corley
0 Replies
8. UNIX for Advanced & Expert Users
Who can explain in a few words the difference between XLC for AIX and XLC enterprise Edition for AIX (0 Replies)
Discussion started by: fvaltat1
0 Replies
9. Red Hat
Red Hat Enterprise 4.5 (32 bit)
In strace we see "unexpected reloc type 0x38" What does that code 0x38 mean?
mprotect(0x59a000, 42229760, PROT_READ|PROT_WRITE) = 0
writev(2, unexpected reloc type 0x38", 26}, {"", 0}, {"", 0}, {"\n", 1}], 1
0) = 113
exit_group(127) ... (6 Replies)
Discussion started by: Sunusernewbie
6 Replies
10. AIX
Hello,
We are migrating some applications from previously running on an AIX system to a Linux RedHat system.
I was tasked with recompiling some of the code. Unfortunatly I am a big novice on this.
So i have the commands used to compile the code with xlc compiler in the AIX environment and i am... (0 Replies)
Discussion started by: spooksman
0 Replies
LEARN ABOUT MOJAVE
sigtrap
sigtrap(3pm) Perl Programmers Reference Guide sigtrap(3pm)
NAME
sigtrap - Perl pragma to enable simple signal handling
SYNOPSIS
use sigtrap;
use sigtrap qw(stack-trace old-interface-signals); # equivalent
use sigtrap qw(BUS SEGV PIPE ABRT);
use sigtrap qw(die INT QUIT);
use sigtrap qw(die normal-signals);
use sigtrap qw(die untrapped normal-signals);
use sigtrap qw(die untrapped normal-signals
stack-trace any error-signals);
use sigtrap 'handler' => &my_handler, 'normal-signals';
use sigtrap qw(handler my_handler normal-signals
stack-trace error-signals);
DESCRIPTION
The sigtrap pragma is a simple interface to installing signal handlers. You can have it install one of two handlers supplied by sigtrap
itself (one which provides a Perl stack trace and one which simply "die()"s), or alternately you can supply your own handler for it to
install. It can be told only to install a handler for signals which are either untrapped or ignored. It has a couple of lists of signals
to trap, plus you can supply your own list of signals.
The arguments passed to the "use" statement which invokes sigtrap are processed in order. When a signal name or the name of one of
sigtrap's signal lists is encountered a handler is immediately installed, when an option is encountered it affects subsequently installed
handlers.
OPTIONS
SIGNAL HANDLERS
These options affect which handler will be used for subsequently installed signals.
stack-trace
The handler used for subsequently installed signals outputs a Perl stack trace to STDERR and then tries to dump core. This is the
default signal handler.
die The handler used for subsequently installed signals calls "die" (actually "croak") with a message indicating which signal was caught.
handler your-handler
your-handler will be used as the handler for subsequently installed signals. your-handler can be any value which is valid as an
assignment to an element of %SIG. See perlvar for examples of handler functions.
SIGNAL LISTS
sigtrap has a few built-in lists of signals to trap. They are:
normal-signals
These are the signals which a program might normally expect to encounter and which by default cause it to terminate. They are HUP,
INT, PIPE and TERM.
error-signals
These signals usually indicate a serious problem with the Perl interpreter or with your script. They are ABRT, BUS, EMT, FPE, ILL,
QUIT, SEGV, SYS and TRAP.
old-interface-signals
These are the signals which were trapped by default by the old sigtrap interface, they are ABRT, BUS, EMT, FPE, ILL, PIPE, QUIT, SEGV,
SYS, TERM, and TRAP. If no signals or signals lists are passed to sigtrap, this list is used.
For each of these three lists, the collection of signals set to be trapped is checked before trapping; if your architecture does not
implement a particular signal, it will not be trapped but rather silently ignored.
OTHER
untrapped
This token tells sigtrap to install handlers only for subsequently listed signals which aren't already trapped or ignored.
any This token tells sigtrap to install handlers for all subsequently listed signals. This is the default behavior.
signal
Any argument which looks like a signal name (that is, "/^[A-Z][A-Z0-9]*$/") indicates that sigtrap should install a handler for that
name.
number
Require that at least version number of sigtrap is being used.
EXAMPLES
Provide a stack trace for the old-interface-signals:
use sigtrap;
Ditto:
use sigtrap qw(stack-trace old-interface-signals);
Provide a stack trace on the 4 listed signals only:
use sigtrap qw(BUS SEGV PIPE ABRT);
Die on INT or QUIT:
use sigtrap qw(die INT QUIT);
Die on HUP, INT, PIPE or TERM:
use sigtrap qw(die normal-signals);
Die on HUP, INT, PIPE or TERM, except don't change the behavior for signals which are already trapped or ignored:
use sigtrap qw(die untrapped normal-signals);
Die on receipt one of an of the normal-signals which is currently untrapped, provide a stack trace on receipt of any of the error-signals:
use sigtrap qw(die untrapped normal-signals
stack-trace any error-signals);
Install my_handler() as the handler for the normal-signals:
use sigtrap 'handler', &my_handler, 'normal-signals';
Install my_handler() as the handler for the normal-signals, provide a Perl stack trace on receipt of one of the error-signals:
use sigtrap qw(handler my_handler normal-signals
stack-trace error-signals);
perl v5.18.2 2013-11-04 sigtrap(3pm)