More context on this from the bug report below.  Looks like
a problem with the shell codebase.

-----

ksh93 dumping core is resulting in inetd core dumping too; This is causing
inetd service not to come online, effectively causing the system to be stand
alone one:
jack@va64-v40za-blr03:/var/tmp# Feb  1 17:05:24 va64-v40za-blr03
svc.startd[100002]: network/inetd:default failed: transitioned to maintenance
(see 'svcs -xv' for details)

stack trace of inted core dump looks as follows :
 fee44640 strlen   (0, 80f6ebc, 24, 8047d2c) + 30
 0805c246 create_method_info (8102d20, 8047cec, 4, fee7bbb2) + 72
 0805c6ea create_method_infos (80f6f28, 8047d30, 80d2a34, 80d2a30) + 32
 0805cbc7 read_instance_cfg (80f6f28, fefbee44, 8047d98, 805d184) + d7
 08058703 handle_restarter_event (810ee48, 3, 1, 8058a5a) + 24b
 08058c45 process_restarter_event (8047df4, 8047df8, 8047e08, 805aff9, 8047e08, 
0) + 1f9
 0805ab48 event_loop (8047e08, 0, 3, 4, 8047e08, 8047f00) + cc
 0805aff9 start_method (8047e08, feffb93c, 8047e40, 805638d, 2, 8047e4c) + 89
 0805b28c main     (2, 8047e4c, 8047e58, feffb93c) + d0
 0805638d _start   (2, 8047eec, 8047f00, 0, 8047f06, 8047f24) + 7d

In create_method_info(), it calls wordexp() which forks a process and runs
ksh93 to get some work done. As this process cores, what wordexp() retuns to
create_method_info is junk and makes inetd core too.

-----

Studio 12.2 compilers are available for download here:
http://www.oracle.com/technetwork/server-storage/solarisstudio/overview/index.html

Venky.


On Tue, Feb 01, 2011 at 04:29:44PM +0100, Roland Mainz wrote:
> On Tue, Feb 1, 2011 at 4:21 PM,  <[email protected]> wrote:
> > *Synopsis*: ksh93 dumps core with ss12.2 compilers
> >
> > CR 7016258 changed on Feb 1 2011 by <User 1-5Q-1267>
> >
> > === Field ============ === New Value ============= === Old Value 
> > =============
> >
> > Keyword ? ? ? ? ? ? ? ?s11-cbe-aten
> > ====================== =========================== 
> > ===========================
> >
> >
> > *Change Request ID*: 7016258
> >
> > *Synopsis*: ksh93 dumps core with ss12.2 compilers
> >
> > ?Product: solaris
> > ?Category: shell
> > ?Subcategory: korn93
> > ?Type: Defect
> > ?Subtype:
> > ?Status: 3-Accepted
> > ?Substatus:
> > ?Priority: 1-Very High
> > ?Introduced In Release:
> > ?Introduced In Build:
> > ?Responsible Engineer: <User 1-5Q-8426>
> > ?Keywords: s11-cbe-aten
> >
> > === *Description* 
> > ============================================================
> > ksh93 is dumping core when compiled with ss12.2 compilers.
> > The stack trace looks as follows :
> >
> > fffffd7fffdf8330 libshell.so.1`extend+0x194()
> > fffffd7fffdfd5c0 libast.so.1`sfvprintf+0xa51()
> > fffffd7fffdfd6a0 libast.so.1`sfprintf+0x95()
> > fffffd7fffdfd7d0 libshell.so.1`b_print+0x67e()
> > fffffd7fffdfd810 libshell.so.1`b_printf+0x47()
> > fffffd7fffdfe2c0 libshell.so.1`sh_exec+0x3272()
> > fffffd7fffdfed70 libshell.so.1`sh_exec+0x1e26()
> > fffffd7fffdff060 libshell.so.1`sh_subshell+0x5f9()
> > fffffd7fffdff250 libshell.so.1`comsubst+0x861()
> > fffffd7fffdff580 libshell.so.1`varsub+0x31f8()
> > fffffd7fffdff650 libshell.so.1`copyto+0xb20()
> > fffffd7fffdff6e0 libshell.so.1`sh_mactrim+0x12a()
> > fffffd7fffdff7c0 libshell.so.1`sh_mactry+0xc2()
> > fffffd7fffdff820 libshell.so.1`io_prompt+0xf5()
> > fffffd7fffdff880 libshell.so.1`slowread+0x98()
> > fffffd7fffdff980 libast.so.1`sfrd+0x5f7()
> > fffffd7fffdff9e0 libast.so.1`_sffilbuf+0x297()
> > fffffd7fffdffa50 libast.so.1`sfreserve+0x235()
> > fffffd7fffdffb70 libshell.so.1`exfile+0x645()
> > fffffd7fffdffc60 libshell.so.1`sh_main+0x836()
> > fffffd7fffdffc90 main+0x57()
> >
> > *** (#1 of 1): 2011-02-01 09:24:50 GMT+00:00 <User 1-5Q-4611>
> 
> Erm... can anyone from Oracle please provide some more context, please
> ? Is this a compiler issue or a problem in our (shell) codebase ?
> 
> BTW: Where can I get Sun Studio 12.2 from ?
> 
> ----
> 
> Bye,
> Roland
_______________________________________________
tools-compilers mailing list
[email protected]

Reply via email to