[Bug tree-optimization/65637] New: expand_omp_for_static_chunk ssa-code is untested

[Bug tree-optimization/65637] New: expand_omp_for_static_chunk ssa-code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

            Bug ID: 65637
           Summary: expand_omp_for_static_chunk ssa-code is untested
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: minor
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vries at gcc dot gnu.org

This is the more generic version of
https://gcc.gnu.org/ml/gcc-patches/2014-11/msg01149.html .

Currently, ssa-handling code in expand_omp_for_static_chunk is dead and not
exercised by testing.

Ssa-handling code in omp-low.c is only triggered by pass_parallelize_loops, and
that pass doesn't specify a chunk size on the GIMPLE_OMP_FOR it constructs, so
that will only call expand_omp_for_static_nochunk.

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #1 from vries at gcc dot gnu.org ---
A patch like this will excercise the code:
...
diff --git a/gcc/tree-parloops.c b/gcc/tree-parloops.c
index 62a6444..b600b2c 100644
--- a/gcc/tree-parloops.c
+++ b/gcc/tree-parloops.c
@@ -1719,6 +1719,7 @@ create_parallel_loop (struct loop *loop, tree loop_fn,
tree data,
   type = TREE_TYPE (cvar);
   t = build_omp_clause (loc, OMP_CLAUSE_SCHEDULE);
   OMP_CLAUSE_SCHEDULE_KIND (t) = OMP_CLAUSE_SCHEDULE_STATIC;
+  OMP_CLAUSE_SCHEDULE_CHUNK_EXPR(t) = integer_one_node;

   for_stmt = gimple_build_omp_for (NULL, GF_OMP_FOR_KIND_FOR, t, 1, NULL);
   gimple_set_location (for_stmt, loc);
...

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #2 from vries at gcc dot gnu.org ---
(In reply to vries from comment #1)
> A patch like this will excercise the code:

Indeed:
...
$ grep -i 'internal compiler error' build/gcc/testsuite/gcc/gcc.log
gcc/testsuite/gcc.dg/autopar/pr46099.c:17:3: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/pr46099.c:17:3: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/pr46099.c:17:3: internal compiler error:
Segmentation fault
FAIL: gcc.dg/autopar/pr46099.c (internal compiler error)
gcc/testsuite/gcc.dg/autopar/pr46099.c:17:3: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/pr49580.c:24:9: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/pr49580.c:24:9: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/pr49580.c:24:9: internal compiler error:
Segmentation fault
FAIL: gcc.dg/autopar/pr49580.c (internal compiler error)
gcc/testsuite/gcc.dg/autopar/pr49580.c:24:9: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/reduc-3.c:21:3: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/reduc-3.c:21:3: internal compiler error:
Segmentation fault
gcc/testsuite/gcc.dg/autopar/reduc-3.c:21:3: internal compiler error:
Segmentation fault
...
FAIL: gcc.dg/autopar/reduc-3.c (internal compiler error)
gcc/testsuite/gcc.dg/autopar/reduc-3.c:21:3: internal compiler error:
Segmentation fault

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #3 from vries at gcc dot gnu.org ---
Created attachment 35221
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=35221&action=edit
patch for autopar-1.exe

For autopar-1.c, the original loop has a loop phi:

  # s.5_20 = PHI <s.5_12(4), 0.0(20)>

After expand_omp_for_static_chunk, there's an inner and an outer loop. The
outer loop phi is:

  # s.5_11 = PHI <0.0(15), s.5_12(21)>

and the inner loop phi is:

  # s.5_20 = PHI <s.5_12(4), 0.0(20)>

The inner loop phi should not have 0.0 as argument, but the result of the outer
loop phi:

  # s.5_20 = PHI <s.5_12(4), s.5_11(20)>


This patch fixes the inner loop phi, and autopar-1.exe passes.

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #4 from vries at gcc dot gnu.org ---
Created attachment 35224
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=35224&action=edit
current state of patch series

Current state of patch series. Passes libgomp graphite.exp, libgomp c.exp and
gcc autopar.exp.

Patch series:
1. Set chunk_size to one for parloops parallel
2. Fix assert after loop over phis and edge_var_map
3. Use edge from iter_part_bb to fin_bb
4. Fix inner loop reduction phi argument
5. Handle case that fin_bb has more than one predecessor

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #5 from vries at gcc dot gnu.org ---
(In reply to vries from comment #4)
> Created attachment 35224 [details]
> current state of patch series
> 
> Current state of patch series. Passes libgomp graphite.exp, libgomp c.exp
> and gcc autopar.exp.
> 
> Patch series:
> 1. Set chunk_size to one for parloops parallel
> 2. Fix assert after loop over phis and edge_var_map
> 3. Use edge from iter_part_bb to fin_bb
> 4. Fix inner loop reduction phi argument
> 5. Handle case that fin_bb has more than one predecessor

Bootstrapped and reg-tested on x86_64, no issues found.

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

vries at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch

--- Comment #6 from vries at gcc dot gnu.org ---
https://gcc.gnu.org/ml/gcc-patches/2015-04/msg00757.html

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #7 from vries at gcc dot gnu.org ---
Author: vries
Date: Mon May 18 13:22:15 2015
New Revision: 223290

URL: https://gcc.gnu.org/viewcvs?rev=223290&root=gcc&view=rev
Log:
Fix gcc_assert in expand_omp_for_static_chunk

2015-05-18  Tom de Vries  <tom@codesourcery.com>

        PR tree-optimization/65637
        * omp-low.c (expand_omp_for_static_chunk): Fix gcc_assert for the case
        that head is NULL.

Modified:
    branches/gomp-4_0-branch/gcc/ChangeLog
    branches/gomp-4_0-branch/gcc/omp-low.c

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #8 from vries at gcc dot gnu.org ---
Author: vries
Date: Mon May 18 13:22:26 2015
New Revision: 223291

URL: https://gcc.gnu.org/viewcvs?rev=223291&root=gcc&view=rev
Log:
Fix inner loop phi in expand_omp_for_static_chunk

2015-05-18  Tom de Vries  <tom@codesourcery.com>

        PR tree-optimization/65637
        * omp-low.c (find_phi_with_arg_on_edge): New function.
        (expand_omp_for_static_chunk): Fix inner loop phi.

Modified:
    branches/gomp-4_0-branch/gcc/ChangeLog
    branches/gomp-4_0-branch/gcc/omp-low.c

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #9 from vries at gcc dot gnu.org ---
Author: vries
Date: Mon May 18 13:22:36 2015
New Revision: 223292

URL: https://gcc.gnu.org/viewcvs?rev=223292&root=gcc&view=rev
Log:
Handle 2 preds for fin_bb in expand_omp_for_static_chunk

2015-05-18  Tom de Vries  <tom@codesourcery.com>

        PR tree-optimization/65637
        * omp-low.c (expand_omp_for_static_chunk): Handle case that fin_bb has
2
        predecessors.

Modified:
    branches/gomp-4_0-branch/gcc/ChangeLog
    branches/gomp-4_0-branch/gcc/omp-low.c

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #10 from vries at gcc dot gnu.org ---
ping ^2: https://gcc.gnu.org/ml/gcc-patches/2015-06/msg00558.html

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #11 from vries at gcc dot gnu.org ---
Resubmitted at https://gcc.gnu.org/ml/gcc-patches/2015-08/msg01867.html

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #14 from vries at gcc dot gnu.org ---
Author: vries
Date: Thu Sep  3 11:01:22 2015
New Revision: 227438

URL: https://gcc.gnu.org/viewcvs?rev=227438&root=gcc&view=rev
Log:
Handle 2 preds for fin_bb in expand_omp_for_static_chunk

2015-09-03  Tom de Vries  <tom@codesourcery.com>

        PR tree-optimization/65637
        * omp-low.c (expand_omp_for_static_chunk): Handle case that fin_bb has
2
        predecessors.

        * gcc.dg/autopar/reduc-4.c: New test.

Added:
    trunk/gcc/testsuite/gcc.dg/autopar/reduc-4.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/omp-low.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #12 from vries at gcc dot gnu.org ---
Author: vries
Date: Thu Sep  3 11:01:05 2015
New Revision: 227436

URL: https://gcc.gnu.org/viewcvs?rev=227436&root=gcc&view=rev
Log:
Fix gcc_assert in expand_omp_for_static_chunk

2015-09-03  Tom de Vries  <tom@codesourcery.com>

        PR tree-optimization/65637
        * omp-low.c (expand_omp_for_static_chunk): Fix gcc_assert for the case
        that head is NULL.

        * gcc.dg/autopar/pr46099-2.c: New test.

Added:
    trunk/gcc/testsuite/gcc.dg/autopar/pr46099-2.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/omp-low.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

--- Comment #13 from vries at gcc dot gnu.org ---
Author: vries
Date: Thu Sep  3 11:01:14 2015
New Revision: 227437

URL: https://gcc.gnu.org/viewcvs?rev=227437&root=gcc&view=rev
Log:
Fix inner loop phi in expand_omp_for_static_chunk

2015-09-03  Tom de Vries  <tom@codesourcery.com>

        PR tree-optimization/65637
        * omp-low.c (find_phi_with_arg_on_edge): New function.
        (expand_omp_for_static_chunk): Fix inner loop phi.

        * testsuite/libgomp.c/autopar-2.c: New test.

Added:
    trunk/libgomp/testsuite/libgomp.c/autopar-2.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/omp-low.c
    trunk/libgomp/ChangeLog

[Bug tree-optimization/65637] expand_omp_for_static_chunk ssa-handling code is untested

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65637

vries at gcc dot gnu.org changed:

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

--- Comment #15 from vries at gcc dot gnu.org ---
patch series committed to trunk, with testcases (made possible by the
introduction of --param parloops-chunk-size)