This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

[PATCH] Don't create superfluous parm in expand_omp_taskreg

From: Tom de Vries <Tom_deVries at mentor dot com>
To: Richard Biener <rguenther at suse dot de>
Cc: Thomas Schwinge <thomas at codesourcery dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, Jakub Jelinek <jakub at redhat dot com>
Date: Tue, 11 Aug 2015 20:53:39 +0200
Subject: [PATCH] Don't create superfluous parm in expand_omp_taskreg
Authentication-results: sourceware.org; auth=none
References: <546743BC dot 5070804 at mentor dot com> <54678B3D dot 8020009 at mentor dot com> <54730EE7 dot 40000 at mentor dot com> <5474665A dot 4070101 at mentor dot com> <87iocp1d8o dot fsf at kepler dot schwinge dot homeip dot net> <alpine dot LSU dot 2 dot 11 dot 1504220932150 dot 20496 at zhemvz dot fhfr dot qr> <557076A5 dot 7050207 at mentor dot com> <55C1BA11 dot 70008 at mentor dot com> <alpine dot LSU dot 2 dot 11 dot 1508050928320 dot 19642 at zhemvz dot fhfr dot qr> <55C1CDAF dot 8050402 at mentor dot com> <alpine dot LSU dot 2 dot 11 dot 1508051116450 dot 19642 at zhemvz dot fhfr dot qr> <55C1EA05 dot 5070904 at mentor dot com> <alpine dot LSU dot 2 dot 11 dot 1508051310230 dot 19642 at zhemvz dot fhfr dot qr>

[ was: Re: [committed, gomp4] Fix release_dangling_ssa_names ]

On 05/08/15 13:13, Richard Biener wrote:

On Wed, 5 Aug 2015, Tom de Vries wrote:

On 05/08/15 11:30, Richard Biener wrote:

On Wed, 5 Aug 2015, Tom de Vries wrote:

On 05/08/15 09:29, Richard Biener wrote:

This patch fixes that by making sure we reset the def stmt to NULL.
This
means

we can simplify release_dangling_ssa_names to just test for NULL def

stmts.

Not sure if I understand the problem correctly but why are you not
simply
releasing the SSA name when you remove its definition?


In move_sese_region_to_fn we move a region of blocks from one function to
another, bit by bit.

When we encounter an ssa_name as def or use in the region, we:
- generate a new ssa_name,
- set the def stmt of the old name as def stmt of the new name, and
- add a mapping from the old to the new name.
The next time we encounter the same ssa_name in another statement, we find
it
in the map.

If we release the old ssa name, we effectively create statements with
operands
in the free-list. The first point where that cause breakage, is in
walk_gimple_op, which expects the TREE_TYPE of the lhs of an assign to be
defined, which is not the case if it's in the free-list:
...
case GIMPLE_ASSIGN:
    /* Walk the RHS operands.  If the LHS is of a non-renamable type or
       is a register variable, we may use a COMPONENT_REF on the RHS.*/
    if (wi)
      {
        tree lhs = gimple_assign_lhs (stmt);
        wi->val_only
          = (is_gimple_reg_type (TREE_TYPE (lhs)) && !is_gimple_reg (lhs))
             || gimple_assign_rhs_class (stmt) != GIMPLE_SINGLE_RHS;
      }
...


Hmm, ok, probably because the stmt moving doesn't happen in DOM
order (move defs before uses).  But


There seems to be similar code for the rhs, so I don't think changing the
order would fix anything.

+
+      if (!SSA_NAME_IS_DEFAULT_DEF (name))
+       /* The statement has been moved to the child function.  It no
longer
+          defines name in the original function.  Mark the def stmt NULL,
and
+          let release_dangling_ssa_names deal with it.  */
+       SSA_NAME_DEF_STMT (name) = NULL;

applies also to uses - I don't see why it couldn't happen that you
move a use but not its def (the def would be a parameter to the
split-out function).  You'd wreck the IL of the source function this way.


If you first move a use, you create a mapping. When you encounter the def, you
use the mapping. Indeed, if the def is a default def, we don't encounter the
def. Which is why we create a nop as defining def for those cases. The default
def in the source function still has a defining nop, and has no uses anymore.
I don't understand what is broken here.


If you never encounter the DEF then it's broken.  Say, if for

foo(int a)
{
   int b = a;
   if (b)
     {
       < code using b >
     }
}

you move < code using b > to a function.  Then the def is still in
foo but you create a mapping for its use(s).  Clearly the outlining
process in this case has to pass b as parameter to the outlined
function, something that may not happen currently.


Ah, I see. Indeed, this is a situation that is assumed not to happen.

It would probably be cleaner to separate the def and use remapping
to separate functions and record on whether we saw a def or not.

Right, or some other means to detect this situation, say when copyingthe def stmt in replace_ssa_name, check whether it's part of the seseregion.

I think that the whole dance of actually moving things instead of
just copying it isn't worth the extra maintainance (well, if we already
have a machinery duplicating a SESE region to another function - I
suppose gimple_duplicate_sese_region could be trivially changed to
support that).


I'll mention that as todo. For now, I think the fastest way to get a working
version is to fix move_sese_region_to_fn.


Sure.

Trunk doesn't have release_dangling_ssa_names it seems


Yep, I only ran into this trouble for the kernels region handling. But I don't
exclude the possibility it could happen for trunk as well.

but I think
it belongs to move_sese_region_to_fn and not to omp-low.c


Makes sense indeed.

and it
could also just walk the d->vars_map replace_ssa_name fills to
iterate over the removal candidates


Agreed, I suppose in general that's a win over iterating over all the ssa
names.

(and if the situation of
moving uses but not defs cannot happen you don't need any
SSA_NAME_DEF_STMT dance either).


I'd prefer to keep the SSA_NAME_DEF_STMT () = NULL bit. It makes sure a stmt
is the defining stmt of only one ssa-name at all times.

I'll prepare a patch for trunk then.


This patch fixes two problems with expand_omp_taskreg:
- it makes sure we don't generate a dummy default def in the original
  function (which we cannot get rid of afterwards, given that it's a
  default def).
- it releases ssa-names in the original function that have defining
  statements that have been moved to the split-off function.

Bootstrapped and reg-tested on x86_64.

OK for trunk?

Thanks,
- Tom

Don't create superfluous parm in expand_omp_taskreg

2015-08-11  Tom de Vries  <tom@codesourcery.com>

	* omp-low.c (expand_omp_taskreg): If in ssa, set rhs of parcopy stmt to
	parm_decl, rather than generating a dummy default def in cfun.
	* tree-cfg.c (replace_ssa_name): Assume no default defs.  Make sure
	ssa_name from cfun and child_fn do not share a stmt as def stmt.
	(move_stmt_op): Handle PARM_DECl.
	(gather_ssa_name_hash_map_from): New function.
	(move_sese_region_to_fn): Add default defs for function params, and add
	them to vars_map.  Release copied ssa names.
	* tree-cfg.h (gather_ssa_name_hash_map_from): Declare.
---
 gcc/omp-low.c  | 20 ++++++++++----------
 gcc/tree-cfg.c | 45 ++++++++++++++++++++++++++++++++++++++++++---
 gcc/tree-cfg.h |  1 +
 3 files changed, 53 insertions(+), 13 deletions(-)

diff --git a/gcc/omp-low.c b/gcc/omp-low.c
index c1dc919..6f32a4a 100644
--- a/gcc/omp-low.c
+++ b/gcc/omp-low.c
@@ -5417,7 +5417,7 @@ expand_omp_taskreg (struct omp_region *region)
 	  basic_block entry_succ_bb
 	    = single_succ_p (entry_bb) ? single_succ (entry_bb)
 				       : FALLTHRU_EDGE (entry_bb)->dest;
-	  tree arg, narg;
+	  tree arg;
 	  gimple parcopy_stmt = NULL;
 
 	  for (gsi = gsi_start_bb (entry_succ_bb); ; gsi_next (&gsi))
@@ -5462,15 +5462,15 @@ expand_omp_taskreg (struct omp_region *region)
 	    }
 	  else
 	    {
-	      /* If we are in ssa form, we must load the value from the default
-		 definition of the argument.  That should not be defined now,
-		 since the argument is not used uninitialized.  */
-	      gcc_assert (ssa_default_def (cfun, arg) == NULL);
-	      narg = make_ssa_name (arg, gimple_build_nop ());
-	      set_ssa_default_def (cfun, arg, narg);
-	      /* ?? Is setting the subcode really necessary ??  */
-	      gimple_omp_set_subcode (parcopy_stmt, TREE_CODE (narg));
-	      gimple_assign_set_rhs1 (parcopy_stmt, narg);
+	      tree lhs = gimple_assign_lhs (parcopy_stmt);
+	      gcc_assert (SSA_NAME_VAR (lhs) == arg);
+	      /* We'd like to set the rhs to the default def in the child_fn,
+		 but it's too early to create ssa names in the child_fn.
+		 Instead, we set the rhs to the parm.  In
+		 move_sese_region_to_fn, we introduce a default def for the
+		 parm, map the parm to it's default def, and once we encounter
+		 this stmt, replace the parm with the default def.  */
+	      gimple_assign_set_rhs1 (parcopy_stmt, arg);
 	      update_stmt (parcopy_stmt);
 	    }
 	}
diff --git a/gcc/tree-cfg.c b/gcc/tree-cfg.c
index 588ab69..8afa5fb 100644
--- a/gcc/tree-cfg.c
+++ b/gcc/tree-cfg.c
@@ -6422,17 +6422,19 @@ replace_ssa_name (tree name, hash_map<tree, tree> *vars_map,
       tree decl = SSA_NAME_VAR (name);
       if (decl)
 	{
+	  gcc_assert (!SSA_NAME_IS_DEFAULT_DEF (name));
 	  replace_by_duplicate_decl (&decl, vars_map, to_context);
 	  new_name = make_ssa_name_fn (DECL_STRUCT_FUNCTION (to_context),
 				       decl, SSA_NAME_DEF_STMT (name));
-	  if (SSA_NAME_IS_DEFAULT_DEF (name))
-	    set_ssa_default_def (DECL_STRUCT_FUNCTION (to_context),
-				 decl, new_name);
 	}
       else
 	new_name = copy_ssa_name_fn (DECL_STRUCT_FUNCTION (to_context),
 				     name, SSA_NAME_DEF_STMT (name));
 
+      /* Now that we've used the def stmt to define new_name, make sure it
+	 doesn't define name anymore.  */
+      SSA_NAME_DEF_STMT (name) = NULL;
+
       vars_map->put (name, new_name);
     }
   else
@@ -6484,6 +6486,9 @@ move_stmt_op (tree *tp, int *walk_subtrees, void *data)
     {
       if (TREE_CODE (t) == SSA_NAME)
 	*tp = replace_ssa_name (t, p->vars_map, p->to_context);
+      else if (TREE_CODE (t) == PARM_DECL
+	       && gimple_in_ssa_p (cfun))
+	*tp = *(p->vars_map->get (t));
       else if (TREE_CODE (t) == LABEL_DECL)
 	{
 	  if (p->new_label_map)
@@ -6994,6 +6999,19 @@ verify_sese (basic_block entry, basic_block exit, vec<basic_block> *bbs_p)
   BITMAP_FREE (bbs);
 }
 
+/* If FROM is an SSA_NAME, mark the version in bitmap DATA.  */
+
+bool
+gather_ssa_name_hash_map_from (tree const &from, tree const &, void *data)
+{
+  bitmap release_names = (bitmap)data;
+
+  if (TREE_CODE (from) != SSA_NAME)
+    return true;
+
+  bitmap_set_bit (release_names, SSA_NAME_VERSION (from));
+  return true;
+}
 
 /* Move a single-entry, single-exit region delimited by ENTRY_BB and
    EXIT_BB to function DEST_CFUN.  The whole region is replaced by a
@@ -7191,6 +7209,14 @@ move_sese_region_to_fn (struct function *dest_cfun, basic_block entry_bb,
   d.eh_map = eh_map;
   d.remap_decls_p = true;
 
+  if (gimple_in_ssa_p (cfun))
+    for (tree arg = DECL_ARGUMENTS (d.to_context); arg; arg = DECL_CHAIN (arg))
+      {
+	tree narg = make_ssa_name_fn (dest_cfun, arg, gimple_build_nop ());
+	set_ssa_default_def (dest_cfun, arg, narg);
+	vars_map.put (arg, narg);
+      }
+
   FOR_EACH_VEC_ELT (bbs, i, bb)
     {
       /* No need to update edge counts on the last block.  It has
@@ -7248,6 +7274,19 @@ move_sese_region_to_fn (struct function *dest_cfun, basic_block entry_bb,
   if (eh_map)
     delete eh_map;
 
+  if (gimple_in_ssa_p (cfun))
+    {
+      /* We need to release ssa-names in a defined order, so first find them,
+	 and then iterate in ascending version order.  */
+      bitmap release_names = BITMAP_ALLOC (NULL);
+      vars_map.traverse<void *, gather_ssa_name_hash_map_from> (release_names);
+      bitmap_iterator bi;
+      unsigned i;
+      EXECUTE_IF_SET_IN_BITMAP (release_names, 0, i, bi)
+	release_ssa_name (ssa_name (i));
+      BITMAP_FREE (release_names);
+    }
+
   /* Rewire the entry and exit blocks.  The successor to the entry
      block turns into the successor of DEST_FN's ENTRY_BLOCK_PTR in
      the child function.  Similarly, the predecessor of DEST_FN's
diff --git a/gcc/tree-cfg.h b/gcc/tree-cfg.h
index 6c4b1d9..4bd6fcf 100644
--- a/gcc/tree-cfg.h
+++ b/gcc/tree-cfg.h
@@ -75,6 +75,7 @@ extern bool gimple_duplicate_sese_tail (edge, edge, basic_block *, unsigned,
 extern void gather_blocks_in_sese_region (basic_block entry, basic_block exit,
 					  vec<basic_block> *bbs_p);
 extern void verify_sese (basic_block, basic_block, vec<basic_block> *);
+extern bool gather_ssa_name_hash_map_from (tree const &, tree const &, void *);
 extern basic_block move_sese_region_to_fn (struct function *, basic_block,
 				           basic_block, tree);
 extern void dump_function_to_file (tree, FILE *, int);
-- 
1.9.1

Follow-Ups:
- Re: [PATCH] Don't create superfluous parm in expand_omp_taskreg
  - From: Richard Biener

References:
- [committed, gomp4] Fix release_dangling_ssa_names
  - From: Tom de Vries
- Re: [committed, gomp4] Fix release_dangling_ssa_names
  - From: Richard Biener
- Re: [committed, gomp4] Fix release_dangling_ssa_names
  - From: Tom de Vries
- Re: [committed, gomp4] Fix release_dangling_ssa_names
  - From: Richard Biener
- Re: [committed, gomp4] Fix release_dangling_ssa_names
  - From: Tom de Vries
- Re: [committed, gomp4] Fix release_dangling_ssa_names
  - From: Richard Biener

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]