[Bug middle-end/25443] New: -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[Bug middle-end/25443] New: -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[ghazi at gcc dot gnu dot org]

I'm getting a 4.1/mainline failure in gcc.dg/tree-ssa/loop-3.c on x86_64 and
i686:

http://gcc.gnu.org/ml/gcc-testresults/2005-12/msg00892.html
http://gcc.gnu.org/ml/gcc-testresults/2005-12/msg00889.html
http://gcc.gnu.org/ml/gcc-testresults/2005-12/msg00891.html
http://gcc.gnu.org/ml/gcc-testresults/2005-12/msg00888.html

The failures look like this:
FAIL: gcc.dg/tree-ssa/loop-3.c scan-tree-dump-times step: 1


-- 
           Summary: -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ghazi at gcc dot gnu dot org
OtherBugsDependingO 23224
             nThis:


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[ghazi at gcc dot gnu dot org]

------- Comment #1 from ghazi at gcc dot gnu dot org  2005-12-16 15:10 -------
Forgot to say this is with -fpic or -fPIC.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2005-12-25 01:30 -------
I have not looked enough into this one but IIRC -fPIC changes the number of
registers available to allocate with which causes different IV selection so the
it might be the testcase is wrong for -fPIC.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[pinskia at gcc dot gnu dot org]

------- Comment #3 from pinskia at gcc dot gnu dot org  2006-01-29 21:01 -------
Confimed.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2006-01-29 21:01:55
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[dalej at apple dot com]

------- Comment #4 from dalej at apple dot com  2007-01-12 20:13 -------
With pic, the load of the global's address is pulled outside the loop, so the
"base" is a register rather than the global.  This causes ivopts to represent
index*step all in the index field rather than as two separate fields, so the
dump looks like
  MEM[base: &arr_base, index: (int *) iter * 4B] = D.1877;
instead of
  MEM[symbol: arr_base, index: (int *) iter, step: 4B] = D.1877;
so the test fails, but the generated code is correct and optimal.  I suggest
adding -fno-pic to the test, does that look OK?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[ghazi at gcc dot gnu dot org]

------- Comment #5 from ghazi at gcc dot gnu dot org  2007-01-13 01:01 -------
(In reply to comment #4)
> so the test fails, but the generated code is correct and optimal.  I suggest
> adding -fno-pic to the test, does that look OK?

I no longer have access to the x86 boxes I was using for testing last year.  So
I can't test any fix.  But anyway, if I recall correctly adding -fno-pic won't
help because stuff in RUNTESTFLAGS is added *after* the dg-options flags.  So
pic will still be turned on if the user selects it.  Instead do something like
what's in gcc.dg/sibcall-6.c, namely:

/* { dg-skip-if "" { { i?86-*-* x86_64-*-* } && ilp32 } { "-fpic" "-fPIC" } {
"" } } */

I don't think the check for ilp32 is wanted in this case though.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[dalej at apple dot com]

------- Comment #6 from dalej at apple dot com  2007-01-13 01:32 -------
Stuff in --tool_opts from RUNTESTFLAGS goes before the dg-options on the
command line, I just tried it.  Is there some other way to do it?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[ghazi at gcc dot gnu dot org]

------- Comment #7 from ghazi at gcc dot gnu dot org  2007-01-13 05:01 -------
(In reply to comment #6)
> Stuff in --tool_opts from RUNTESTFLAGS goes before the dg-options on the
> command line, I just tried it.  Is there some other way to do it?

Yes, the GCC docs suggest using --target_board:
http://gcc.gnu.org/install/test.html

I could be wrong, but I'm pretty sure that one appends the flags you supply.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[dalej at apple dot com]

------- Comment #8 from dalej at apple dot com  2007-01-15 23:41 -------
You are right, thanks.  Test case fixed thus in mainline (to be 4.3).
http://gcc.gnu.org/ml/gcc-patches/2007-01/msg01266.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[echristo at apple dot com]

-- 

echristo at apple dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |dalej at apple dot com
                   |dot org                     |
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443

[Bug middle-end/25443] -fpic/-fPIC failure in gcc.dg/tree-ssa/loop-3.c

[dalej at apple dot com]

------- Comment #9 from dalej at apple dot com  2007-01-15 23:48 -------
as per previous comment


-- 

dalej at apple dot com changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25443