segfault allocating polymophic variable with polymorphic component with allocatable component

[Andrew Benson]

I forgot to include the link to the PR:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78356

On Monday, November 14, 2016 5:34:24 PM PST Andrew Benson wrote:
> I've opened a PR for an apparent regression in gfortran 7.0.0 when
> attempting to allocate a component of a polymorphic variable which itself
> contains an allocatable component. The code compiles and runs successfully
> with gfortran 6.0.0.
> 
> -Andrew
> 
> --
> 
> module m
>   implicit none
>   type ac
>   end type ac
>   type, extends(ac) :: a
>      integer, allocatable, dimension(:) :: b
>   end type a
>   type n
>      class(ac), allocatable, dimension(:) :: acr
>    contains
>      procedure :: ncp => nncp
>   end type n
> contains
>   subroutine nncp(self,tg)
>     implicit none
>     class(n), intent(in   ) :: self
>     class(n), intent(inout) :: tg
>     allocate(tg%acr(1),source=self%acr(1))
>     return
>   end subroutine nncp
> end module m
> program p
>   use m
>   implicit none
>   type(n) :: s,t
>   allocate(a :: s%acr(1))
>   call s%ncp(t)
> end program
> 
> $ gfortran -v
> Using built-in specs.
> COLLECT_GCC=gfortran
> COLLECT_LTO_WRAPPER=/home/abenson/Galacticus/Tools/libexec/gcc/x86_64-pc-
> linux-gnu/7.0.0/lto-wrapper
> Target: x86_64-pc-linux-gnu
> Configured with: ../gcc-trunk/configure --prefix=/home/abenson/Galacticus/
> Tools --enable-languages=c,c++,fortran --disable-multilib
> Thread model: posix
> gcc version 7.0.0 20161115 (experimental) (GCC)
> 
> $ a.out
> 
> Program received signal SIGSEGV: Segmentation fault - invalid memory
> reference.
> 
> Backtrace for this error:
> #0  0x7ff186f0d1ef in ???
>         at /data001/abenson/Galacticus/Tools/glibc-2.12.1/signal/../sysdeps/
> unix/sysv/linux/x86_64/sigaction.c:0
> #1  0x7ff186f5a460 in ???
>         at ../sysdeps/x86_64/memcpy.S:161
> #2  0x40167b in __copy_m_A
>         at /home/abenson/Galacticus/v0.9.4/c.F90:22
> #3  0x400f59 in __m_MOD_nncp
>         at /home/abenson/Galacticus/v0.9.4/c.F90:18
> #4  0x401593 in p
>         at /home/abenson/Galacticus/v0.9.4/c.F90:27
> #5  0x4016c4 in main
>         at /home/abenson/Galacticus/v0.9.4/c.F90:23
> Segmentation fault
> 
> $ valgrind a.out
> ==1544== Memcheck, a memory error detector
> ==1544== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
> ==1544== Using Valgrind-3.11.0 and LibVEX; rerun with -h for copyright info
> ==1544== Command: a.out
> ==1544==
> ==1544== Argument 'size' of function malloc has a fishy (possibly negative)
> value: -274793986988
> ==1544==    at 0x4A05B60: malloc (vg_replace_malloc.c:299)
> ==1544==    by 0x401647: __copy_m_A.3590 (c.F90:22)
> ==1544==    by 0x400F59: __m_MOD_nncp (c.F90:18)
> ==1544==    by 0x401593: MAIN__ (c.F90:27)
> ==1544==    by 0x4016C4: main (c.F90:23)
> ==1544==
> ==1544== Source and destination overlap in memcpy(0x0, 0x2,
> 18446743798915564628)
> ==1544==    at 0x4A0A4D3: memcpy (vg_replace_strmem.c:1019)
> ==1544==    by 0x40167B: __copy_m_A.3590 (c.F90:22)
> ==1544==    by 0x400F59: __m_MOD_nncp (c.F90:18)
> ==1544==    by 0x401593: MAIN__ (c.F90:27)
> ==1544==    by 0x4016C4: main (c.F90:23)
> ==1544==
> ==1544== Invalid read of size 2
> ==1544==    at 0x4A0A6E8: memcpy (vg_replace_strmem.c:1019)
> ==1544==    by 0x40167B: __copy_m_A.3590 (c.F90:22)
> ==1544==    by 0x400F59: __m_MOD_nncp (c.F90:18)
> ==1544==    by 0x401593: MAIN__ (c.F90:27)
> ==1544==    by 0x4016C4: main (c.F90:23)
> ==1544==  Address 0x2 is not stack'd, malloc'd or (recently) free'd
> ==1544==
> 
> Program received signal SIGSEGV: Segmentation fault - invalid memory
> reference.
> 
> Backtrace for this error:
> #0  0x56431ef in ???
>         at /data001/abenson/Galacticus/Tools/glibc-2.12.1/signal/../sysdeps/
> unix/sysv/linux/x86_64/sigaction.c:0
> #1  0x4a0a6e8 in _vgr20180ZZ_libcZdsoZa_memcpy
>         at ../shared/vg_replace_strmem.c:1019
> #2  0x40167b in ???
> #3  0x400f59 in ???
> #4  0x401593 in ???
> #5  0x4016c4 in ???
> #6  0x562fd0c in __libc_start_main
>         at
> /data001/abenson/Galacticus/Tools/glibc-2.12.1/csu/libc-start.c:226 #7 
> 0x400728 in ???
> ==1544==
> ==1544== Process terminating with default action of signal 11 (SIGSEGV)
> ==1544==    at 0x564313E: raise (raise.c:67)
> ==1544==    by 0x56431EF: ??? (in /home/abenson/Galacticus/Tools/lib/
> libc-2.12.1.so)
> ==1544==    by 0x4A0A6E7: memcpy (vg_replace_strmem.c:1019)
> ==1544==    by 0x40167B: __copy_m_A.3590 (c.F90:22)
> ==1544==    by 0x400F59: __m_MOD_nncp (c.F90:18)
> ==1544==    by 0x401593: MAIN__ (c.F90:27)
> ==1544==    by 0x4016C4: main (c.F90:23)
> ==1544==
> ==1544== HEAP SUMMARY:
> ==1544==     in use at exit: 3,964 bytes in 17 blocks
> ==1544==   total heap usage: 21 allocs, 4 frees, 12,104 bytes allocated
> ==1544==
> ==1544== LEAK SUMMARY:
> ==1544==    definitely lost: 0 bytes in 0 blocks
> ==1544==    indirectly lost: 0 bytes in 0 blocks
> ==1544==      possibly lost: 0 bytes in 0 blocks
> ==1544==    still reachable: 3,964 bytes in 17 blocks
> ==1544==         suppressed: 0 bytes in 0 blocks
> ==1544== Rerun with --leak-check=full to see details of leaked memory
> ==1544==
> ==1544== For counts of detected and suppressed errors, rerun with: -v
> ==1544== ERROR SUMMARY: 3 errors from 3 contexts (suppressed: 3 from 3)
> Segmentation fault
> 
> 
> Allocating the "b" component of "s%acr(1)" doesn't resolve the problem.


-- 

* Andrew Benson: http://users.obs.carnegiescience.edu/abenson/contact.html

* Galacticus: http://sites.google.com/site/galacticusmodel