[Bug middle-end/67790] [6 Regression] verify_ssa failed: definition in block 20 follows the use

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67790

Marek Polacek  changed:

   What|Removed |Added

 CC||mpolacek at gcc dot gnu.org
   Target Milestone|--- |6.0
Summary|verify_ssa failed:  |[6 Regression] verify_ssa
   |definition in block 20  |failed: definition in block
   |follows the use |20 follows the use

--- Comment #2 from Marek Polacek  ---
Started with r222354.

[Bug middle-end/67790] [6 Regression] verify_ssa failed: definition in block 20 follows the use

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67790

Richard Biener  changed:

   What|Removed |Added

 Status|NEW |ASSIGNED
   Assignee|unassigned at gcc dot gnu.org  |rguenth at gcc dot 
gnu.org

--- Comment #3 from Richard Biener  ---
Mine.

[Bug middle-end/67790] [6 Regression] verify_ssa failed: definition in block 20 follows the use

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67790

--- Comment #4 from Richard Biener  ---
The issue here is that

  /* If we detected "res -= x[i]" earlier, rewrite it into
 "res += -x[i]" now.  If this turns out to be useless reassoc
 will clean it up again.  */
  if (orig_code == MINUS_EXPR)
{
  tree rhs = gimple_assign_rhs2 (def_stmt);
  tree negrhs = make_ssa_name (TREE_TYPE (rhs));
  gimple *negate_stmt = gimple_build_assign (negrhs, NEGATE_EXPR, rhs);
  gimple_stmt_iterator gsi = gsi_for_stmt (def_stmt);
  set_vinfo_for_stmt (negate_stmt, new_stmt_vec_info (negate_stmt,
  loop_info));
  gsi_insert_before (&gsi, negate_stmt, GSI_NEW_STMT);
  gimple_assign_set_rhs2 (def_stmt, negrhs);
  gimple_assign_set_rhs_code (def_stmt, PLUS_EXPR);
  update_stmt (def_stmt);
}

actually inserts stmts into the IL which assigns UIDs to them (via
set_vinfo_for_stmt) which later confuses get_later/earlier_stmt.

The "proper" way of doing the above is by using a pattern which avoids
this issue.

[Bug middle-end/67790] [6 Regression] verify_ssa failed: definition in block 20 follows the use

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67790

--- Comment #5 from Richard Biener  ---
Aww.  And then I'm running into the issue of pattern recog / reduction analysis
ordering which has caused me headaches before...  maybe time to (finally) fix
that.

[Bug middle-end/67790] [6 Regression] verify_ssa failed: definition in block 20 follows the use

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67790

--- Comment #6 from Richard Biener  ---
Author: rguenth
Date: Wed Nov 18 08:31:32 2015
New Revision: 230518

URL: https://gcc.gnu.org/viewcvs?rev=230518&root=gcc&view=rev
Log:
2015-11-18  Richard Biener  

PR tree-optimization/67790
* tree-vect-loop.c (vect_is_simple_reduction_1): Remove
IL rewrite for MINUS_EXPR reductions, rename back to ...
(vect_is_simple_reduction): ... this, removing the wrapper.
(vect_force_simple_reduction): Adjust.
(vectorizable_reduction): Adjust reduc_index for MINUS_EXPR
reductions and make use if reduc_index in all places.  For
the final reduction of MINUS_EXPR use PLUS_EXPR.

* gcc.dg/vect/pr67790.c: New testcase.

Added:
trunk/gcc/testsuite/gcc.dg/vect/pr67790.c
Modified:
trunk/gcc/ChangeLog
trunk/gcc/testsuite/ChangeLog
trunk/gcc/tree-vect-loop.c

[Bug middle-end/67790] [6 Regression] verify_ssa failed: definition in block 20 follows the use

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67790

Richard Biener  changed:

   What|Removed |Added

 Status|ASSIGNED|RESOLVED
 Resolution|--- |FIXED

--- Comment #7 from Richard Biener  ---
Fixed.