CREAD buggy?

[howarth at bromo dot msbb dot uc dot edu (Jack Howarth)]

    I have been able to build xplor-nih using the new cray pointer patches
and xplor-nih passes most of its testsuite. However, I see subset of tests
fail with segfaults that all seem to be in calls to CREAD. The gdb output
shows...

 COOR>ATOM      1  C4  DNP   250      -4.070  58.954 181.339  1.00  0.00 

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x00016ff0
0xffff8c94 in ___memcpy () at /System/Library/Frameworks/System.framework/PrivateHeaders/ppc/cpu_capabilities.h:189
189     /System/Library/Frameworks/System.framework/PrivateHeaders/ppc/cpu_capabilities.h: No such file or directory.
        in /System/Library/Frameworks/System.framework/PrivateHeaders/ppc/cpu_capabilities.h
(gdb) 

...a backtrace shows...

#0  0xffff8c94 in ___memcpy () at /System/Library/Frameworks/System.framework/PrivateHeaders/ppc/cpu_capabilities.h:189
#1  0x0122f604 in *__gfortrani_format_error (f=0xbfffd406, message=0xfffffeea <Address 0xfffffeea out of bounds>) at ../../../libgfortran/io/format.c:923
#2  0x0123e2dc in formatted_transfer (type=BT_CHARACTER, p=0x4fdc8, len=44) at ../../../libgfortran/io/transfer.c:431
#3  0x0123c874 in *__gfortran_transfer_character (p=0x2, len=-278) at ../../../libgfortran/io/transfer.c:884
#4  0x0003e00c in cread_ (x=0x4011400, y=0x4011c00, z=0x4012400, wmain=0x4010400, qmain=0x4010c00, flags=0x3c0f2e0, segid=0x4018000, res=0x4018400, resid=0x4018800, type=0x4018c00, _segid=-1073752064, _res=4, _resid=4, _type=1) at coorio.f:422
#5  0x019612c8 in corma2_ (islct=0x3c0f2e0, rmsd=0x400ac00, kcnstr=0x400b400, refx=0x400c400, refy=0x400cc00, refz=0x400d400, xcomp=0x400dc00, ycomp=0x400e400, zcomp=0x400ec00, wcomp=0x400f400, qcomp=0x400fc00, fbeta=0x400bc00, wmain=0x4010400, qmain=0x4010c00, x=0x4011400, y=0x4011c00, z=0x4012400, dx=0x4012c00, dy=0x4013400, dz=0x4013c00, amass=0x4016400) at corman.f:162
#6  0x019606a8 in corman_ () at corman.f:37
#7  0x00044f68 in xplor_parse_ (qsubshell=@0xbfffe328) at xplorFunc.f:446
#8  0x00002940 in main (argc=1, argv=0xbfffe440) at xplor.c:116

The CREAD call in question is of the form...

      CALL CREAD(X, Y, Z, WMAIN, QMAIN, ISLCT, heap(p_segid),
     &                 heap(p_res), heap(p_resid), heap(p_type))

where the types for the parameters in the subroutine making this
call are..

      double precision z(*)
      double precision y(*)
      double precision x(*)
      double precision wmain(*)
      double precision qmain(*)
      INTEGER ISLCT(*) 
      INTEGER HEAPDM
      PARAMETER (HEAPDM=2)
C dummy array to be able to use pointers in FORTRAN.
      INTEGER HEAP(HEAPDM)

I'll ponder this to see if I can come up with a test case. Unfortunately
its rather complex code and the CREAD call seems to make it through the
first time in this case, hence the...

COOR>ATOM      1  C4  DNP   250      -4.070  58.954 181.339  1.00  0.00

output. Its the second time it calls CREAD with the same parameters that
I get the segfault.
              Jack
ps I see this with both 4.0 and 4.1 branch and at -O0, -O2 and -O3
optimizations.