*Synopsis*: ksh93 dumps core during fftw2 build in SFW
Due to a change requested by <User 1-5Q-12691>,
<User 1-5Q-12691> is now the responsible engineer for:
CR 6859442 changed on Aug 21 2009 by <User 1-5Q-12691>
=== Field ============ === New Value ============= === Old Value =============
Commit to Fix in Build snv_123
Evaluation New Note
Introduced in Release solaris_nevada
Responsible Engineer <User 1-5Q-12691>
Status 7-Fix in Progress 1-Dispatched
Targeted Release solaris_nevada
====================== =========================== ===========================
*Change Request ID*: 6859442
*Synopsis*: ksh93 dumps core during fftw2 build in SFW
Product: solaris
Category: shell
Subcategory: korn93
Type: Defect
Subtype:
Status: 7-Fix in Progress
Substatus:
Priority: 2-High
Introduced In Release: solaris_nevada
Introduced In Build:
Responsible Engineer: <User 1-5Q-12691>
Keywords:
=== *Description* ============================================================
When building SFW on OpenSolaris (where /bin/sh is ksh93), I consistently get a
core dump from ksh93 when it's running the configure script in the 32-bit,
double-precision build of fftw2. The dump happens when configure is "checking
for Fortran 77 libraries". The core file is attached, but the stack trace is
here:
libc_hwcap1.so.1`strlen+0xb(8061218, 809bdf5, 0, fedaa222)
libshell.so.1`copyto+0x9cc(8061218, 0)
libshell.so.1`sh_mactrim+0x11d(fee06d68, 809bdf5, ffffffff, fedb2cfd)
libshell.so.1`nv_setlist+0x272(809bdec, 20200, 809be10, 0)
libshell.so.1`sh_exec+0x2747(809be10, 4, 0, fedd425d)
libshell.so.1`sh_exec+0x1e23(809d900, 4, 0, fedd425d)
libshell.so.1`sh_exec+0x11c6(809b244, 4, 0, fedd425d)
libshell.so.1`sh_exec+0x11c6(809b10c, 4, 0, fedd425d)
libshell.so.1`sh_exec+0x1e23(80a03d4, 4, 808c7a4, fedd425d)
libshell.so.1`sh_exec+0x11f1(80673c4, 4, 0, fedb1b4d)
libshell.so.1`exfile+0x716(fee06d68, 80739b8)
libshell.so.1`sh_main+0x808(7, 8047160, 0, fee70fc2, 4, 0)
main+0x4a(7, 8047160, 8047180, 804711c)
_start+0x7d(7, 8047314, 804731c, 8047328, 8047334, 8047342)
While I get the dump consistently when doing a full nightly, a simple make in
the fftw2 directory fails to reproduce. Perhaps there's something load
related, though of the four nearly identical configures that happen for fftw2
(the same source is unpacked into four directories -- one each for single and
double precision, 32 and 64 bit), only the double-precision 32-bit version (the
first one to run) fails.
*** (#1 of 1): 2009-07-12 00:41:03 GMT+00:00 <User 1-5Q-2233>
=== *Public Comments* ========================================================
=== *Workaround* =============================================================
Roland says: "Use the 64bit version of ksh93, the bug appears to be specific
to the 32bit version."
*** (#1 of 1): 2009-07-12 03:16:31 GMT+00:00 <User 1-5Q-1267>
=== *Additional Details* =====================================================
Targeted Release: solaris_nevada
Commit To Fix In Build: snv_123
Fixed In Build:
Integrated In Build:
Verified In Build:
See Also: 6817601, 6859444
Duplicate of:
Hooks:
Hook1:
Hook2:
Hook3:
Hook4:
Hook5:
Hook6:
Program Management:
Root Cause:
Fix Affects Documentation: No
Fix Affects Localization: No
=== *History* ================================================================
Date Submitted: 2009-07-12 00:40:43 GMT+00:00
Submitted By: <User 1-5Q-2233>
Status Changed Date Updated Updated By
7-Fix in Progress 2009-08-21 04:47:00 GMT+00:00 <User 1-5Q-12691>
=== *Service Request* ========================================================
Impact: Critical
Functionality: Primary
Severity: 1
Product Name: solaris
Product Release: osol_2009.06
Product Build:
Operating System: solaris
Hardware: generic
Submitted Date: 2009-07-12 00:41:03 GMT+00:00
=== *Multiple Release (MR) Cluster* - 0 ======================================
