[Bug lto/44533] New: [4.6 Regression] Revision 160679 miscompiles capacita.f90 with -O3 -finline-limit=600 -flto

[Bug lto/44533] New: [4.6 Regression] Revision 160679 miscompiles capacita.f90 with -O3 -finline-limit=600 -flto

[dominiq at lps dot ens dot fr]

On x86_64-apple-darwin10 the polyhedron test capacita.f90 is miscompiled with
'-O3 -finline-limit=600 -flto' at revision 160679 (revision 160678 is fine):

[macbook] lin/test% gfcp -O3 -finline-limit=600 -flto capacita.f90
[macbook] lin/test% a.out 
 Give: N_x, N_y, GridSize_X, GridSize_Y, DoCircle, UseFFT, UsePreco

   N_x and N_y give the grid division in x and y-directions and
   should be powers of 2, DoCircle and UseFFT are T or F (true
   or false). If DoCircle=T, the case of an ellipse is computed
   (circle: C = 8*eps0*R, P = 16/3*eps0*R^3), else that of the
   completely filled grid. UseFFT is for increased speed.
 Calculating G
 Transforming G
Segmentation fault

Backtrace is

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000404
Crashed Thread:  0  Dispatch queue: com.apple.main-thread

Thread 0 Crashed:  Dispatch queue: com.apple.main-thread
0   a.out                               0x0000000100005230
__solv_cap_MOD_fourir2d + 3136
1   a.out                               0x000000010000d580
__solv_cap_MOD_init_solve + 2512
2   a.out                               0x000000010000ddaf MAIN__.2138 + 1647
3   a.out                               0x000000010000e0b1 main + 33
4   a.out                               0x0000000100000f24 start + 52

Note that '-fno-tree-vectorize' does not remove the crash, while it does not
occur with '-O2 -ftree-vectorize' (nor on powerpc-apple-darwin9 at revision
160686).


-- 
           Summary: [4.6 Regression] Revision 160679 miscompiles
                    capacita.f90 with -O3 -finline-limit=600 -flto
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dominiq at lps dot ens dot fr
 GCC build triplet: x86_64-apple-darwin10
  GCC host triplet: x86_64-apple-darwin10
GCC target triplet: x86_64-apple-darwin10


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44533

[Bug lto/44533] [4.6 Regression] Revision 160679 miscompiles capacita.f90 with -O3 -finline-limit=600 -flto

[dominiq at lps dot ens dot fr]

------- Comment #1 from dominiq at lps dot ens dot fr  2010-06-16 09:21 -------
The miscompilation occurs between -finline-limit=300 (works) and =400
(Segmentation fault). This seems related to the inlining of the subroutine
'fourir2d' (called 4 times with =400) instead of  'fourir' (called 14 times
with =300).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44533

[Bug lto/44533] [4.6 Regression] Revision 160679 miscompiles capacita.f90 with -O3 -finline-limit=600 -flto

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44533

[Bug lto/44533] [4.6 Regression] Revision 160679 miscompiles capacita.f90 with -O3 -finline-limit=600 -flto

[dominiq at lps dot ens dot fr]

------- Comment #2 from dominiq at lps dot ens dot fr  2010-07-18 10:44 -------
Silently fixed since at least revision 161009.


-- 

dominiq at lps dot ens dot fr changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44533