[Bug middle-end/92952] [10 regression] gfortran.dg/lto/pr87689 FAILs at -O2

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92952

Martin Sebor <msebor at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2019-12-16
     Ever confirmed|0                           |1

--- Comment #1 from Martin Sebor <msebor at gcc dot gnu.org> ---
-Wstringop-overflow is enabled only for the C family of languages (and for LTO)
so it shouldn't be issued for Fortran code (see also pr80545).  I think the
problem might be making the distinction between languages during LTO builds.

I can reproduce the warning but have no idea how to actually step into the LTO
code.  My best guess is that the size computation doesn't account for the
non-zero low bound in Fortran arrays.  The following avoids the warning so let
me go with it.

Index: gcc/builtins.c
===================================================================
--- gcc/builtins.c      (revision 279443)
+++ gcc/builtins.c      (working copy)
@@ -3999,6 +3999,16 @@ compute_objsize (tree dest, int ostype, tree *pdec
             above.  */
          if (TREE_CODE (dest) == ARRAY_REF)
            {
+             tree lowbnd = array_ref_low_bound (dest);
+             if (!integer_zerop (lowbnd) && tree_fits_uhwi_p (lowbnd))
+               {
+                 /* Adjust the offset by the low bound of the array
+                    domain (normally zero but 1 in Fortran).  */
+                 unsigned HOST_WIDE_INT lb = tree_to_uhwi (lowbnd);
+                 offrng[0] -= lb;
+                 offrng[1] -= lb;
+               }
+
              /* Convert the array index into a byte offset.  */
              tree eltype = TREE_TYPE (dest);
              tree tpsize = TYPE_SIZE_UNIT (eltype);