gcc: Upgrade to 4.5.3 and bring in latest linaro updates

We base now off release tarball of 4.5.3 instead of svn
Get relevant new arm linaro patches

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>

1 files changed, 784 insertions, 0 deletions

diff --git a/meta-oe/recipes-devtools/gcc/gcc-4.5/linaro/gcc-4.5-linaro-r99495.patch b/meta-oe/recipes-devtools/gcc/gcc-4.5/linaro/gcc-4.5-linaro-r99495.patch
new file mode 100644
index 000000000..bb866ce8d
--- /dev/null
+++ b/meta-oe/recipes-devtools/gcc/gcc-4.5/linaro/gcc-4.5-linaro-r99495.patch
@@ -0,0 +1,784 @@
+2011-03-24  Revital Eres  <revital.eres@linaro.org>
+
+        gcc/
+        * loop-doloop.c (doloop_condition_get): Support new form of
+        doloop pattern and use prev_nondebug_insn instead of PREV_INSN.
+        * config/arm/thumb2.md (*thumb2_addsi3_compare0): Remove "*".
+        (doloop_end): New.
+        * config/arm/arm.md (*addsi3_compare0): Remove "*".
+        * ddg.c (check_closing_branch_deps, get_node_of_insn_uid):
+        New functions.
+        (create_ddg): Pass sbitmap containing do-loop related
+        instructions instead of closing_branch_deps parameter and call
+        check_closing_branch_deps function.
+        * ddg.h (create_ddg): Adjust the function declaration.
+        * modulo-sched.c (PS_STAGE_COUNT): Rename to CALC_STAGE_COUNT
+        and redefine.
+        (doloop_register_get): Handle NONDEBUG_INSN_P.
+        (stage_count): New field in struct partial_schedule.
+        (mark_doloop_insns, calculate_stage_count): New functions.
+        (normalize_sched_times): Rename to reset_sched_times and handle
+        incrementing the sched time of the nodes by a constant value
+        passed as parameter.
+        (duplicate_insns_of_cycles): Skip closing branch.
+        (sms_schedule_by_order): Schedule closing branch when
+        closing_branch_deps is true.
+        (ps_insn_find_column): Handle closing branch.
+        (sms_schedule): Call reset_sched_times and handle case where
+        do-loop pattern is not decoupled from the other loop instructions.
+        Support new form of doloop pattern.
+        (ps_insert_empty_row): Update calls to normalize_sched_times
+        and rotate_partial_schedule functions.
+
+=== modified file 'gcc/config/arm/arm.md'
+--- old/gcc/config/arm/arm.md   2011-03-11 14:26:34 +0000
++++ new/gcc/config/arm/arm.md   2011-03-24 07:45:38 +0000
+@@ -734,7 +734,7 @@
+   ""
+ )
+ 
+-(define_insn "*addsi3_compare0"
++(define_insn "addsi3_compare0"
+   [(set (reg:CC_NOOV CC_REGNUM)
+        (compare:CC_NOOV
+         (plus:SI (match_operand:SI 1 "s_register_operand" "r, r")
+
+=== modified file 'gcc/config/arm/thumb2.md'
+--- old/gcc/config/arm/thumb2.md        2011-02-08 10:51:58 +0000
++++ new/gcc/config/arm/thumb2.md        2011-03-24 07:45:38 +0000
+@@ -1194,7 +1194,7 @@
+    (set_attr "length" "2")]
+ )
+ 
+-(define_insn "*thumb2_addsi3_compare0"
++(define_insn "thumb2_addsi3_compare0"
+   [(set (reg:CC_NOOV CC_REGNUM)
+        (compare:CC_NOOV
+          (plus:SI (match_operand:SI 1 "s_register_operand" "l,  0, r")
+@@ -1445,3 +1445,56 @@
+   [(set_attr "length" "4,4,16")
+    (set_attr "predicable" "yes")]
+ )
++
++
++;; Define the subtract-one-and-jump insns so loop.c
++;; knows what to generate.
++(define_expand "doloop_end"
++  [(use (match_operand 0 "" ""))      ; loop pseudo
++   (use (match_operand 1 "" ""))      ; iterations; zero if unknown
++   (use (match_operand 2 "" ""))      ; max iterations
++   (use (match_operand 3 "" ""))      ; loop level
++   (use (match_operand 4 "" ""))]     ; label
++  "TARGET_32BIT"
++  "
++ {
++   /* Currently SMS relies on the do-loop pattern to recognize loops
++      where (1) the control part consists of all insns defining and/or
++      using a certain 'count' register and (2) the loop count can be
++      adjusted by modifying this register prior to the loop.
++     ??? The possible introduction of a new block to initialize the
++      new IV can potentially affect branch optimizations.  */
++   if (optimize > 0 && flag_modulo_sched)
++   {
++     rtx s0;
++     rtx bcomp;
++     rtx loc_ref;
++     rtx cc_reg;
++     rtx insn;
++     rtx cmp;
++
++     /* Only use this on innermost loops.  */
++     if (INTVAL (operands[3]) > 1)
++       FAIL;
++
++     if (GET_MODE (operands[0]) != SImode)
++       FAIL;
++
++     s0 = operands [0];
++     if (TARGET_THUMB2)
++       insn = emit_insn (gen_thumb2_addsi3_compare0 (s0, s0, GEN_INT (-1)));
++     else
++       insn = emit_insn (gen_addsi3_compare0 (s0, s0, GEN_INT (-1)));
++
++     cmp = XVECEXP (PATTERN (insn), 0, 0);
++     cc_reg = SET_DEST (cmp);
++     bcomp = gen_rtx_NE (VOIDmode, cc_reg, const0_rtx);
++     loc_ref = gen_rtx_LABEL_REF (VOIDmode, operands [4]);
++     emit_jump_insn (gen_rtx_SET (VOIDmode, pc_rtx,
++                                  gen_rtx_IF_THEN_ELSE (VOIDmode, bcomp,
++                                                        loc_ref, pc_rtx)));
++     DONE;
++   }else
++      FAIL;
++ }")
++
+
+=== modified file 'gcc/ddg.c'
+--- old/gcc/ddg.c       2010-07-19 08:58:53 +0000
++++ new/gcc/ddg.c       2011-03-24 07:45:38 +0000
+@@ -60,6 +60,8 @@
+ static ddg_edge_ptr create_ddg_edge (ddg_node_ptr, ddg_node_ptr, dep_type,
+                                     dep_data_type, int, int);
+ static void add_edge_to_ddg (ddg_ptr g, ddg_edge_ptr);
++static ddg_node_ptr get_node_of_insn_uid (ddg_ptr, int);
++
+ 
+ /* Auxiliary variable for mem_read_insn_p/mem_write_insn_p.  */
+ static bool mem_ref_p;
+@@ -450,12 +452,65 @@
+   sched_free_deps (head, tail, false);
+ }
+ 
++/* Given DOLOOP_INSNS which holds the instructions that
++   belong to the do-loop part; mark closing_branch_deps field in ddg G
++   as TRUE if the do-loop part's instructions are dependent on the other
++   loop instructions.  Otherwise mark it as FALSE.  */
++static void
++check_closing_branch_deps (ddg_ptr g, sbitmap doloop_insns)
++{
++  sbitmap_iterator sbi;
++  unsigned int u = 0;
++
++  EXECUTE_IF_SET_IN_SBITMAP (doloop_insns, 0, u, sbi)
++  {
++    ddg_edge_ptr e;
++    ddg_node_ptr u_node = get_node_of_insn_uid (g, u);
++
++    gcc_assert (u_node);
++
++    for (e = u_node->in; e != 0; e = e->next_in)
++      {
++       ddg_node_ptr v_node = e->src;
++
++       if (((unsigned int) INSN_UID (v_node->insn) == u)
++           || DEBUG_INSN_P (v_node->insn))
++         continue;
++       
++       /* Ignore dependencies between memory writes and the
++          jump.  */
++       if (JUMP_P (u_node->insn)
++           && e->type == OUTPUT_DEP 
++            && mem_write_insn_p (v_node->insn))
++         continue;
++       if (!TEST_BIT (doloop_insns, INSN_UID (v_node->insn)))
++         {
++           g->closing_branch_deps = 1;
++           return;
++         }
++      }
++    for (e = u_node->out; e != 0; e = e->next_out)
++      {
++       ddg_node_ptr v_node = e->dest;
++
++       if (((unsigned int) INSN_UID (v_node->insn) == u)
++            || DEBUG_INSN_P (v_node->insn))
++         continue;
++       if (!TEST_BIT (doloop_insns, INSN_UID (v_node->insn)))
++         {
++           g->closing_branch_deps = 1;
++           return;
++         }
++      }
++  }
++  g->closing_branch_deps = 0;
++}
+ 
+ /* Given a basic block, create its DDG and return a pointer to a variable
+    of ddg type that represents it.
+    Initialize the ddg structure fields to the appropriate values.  */
+ ddg_ptr
+-create_ddg (basic_block bb, int closing_branch_deps)
++create_ddg (basic_block bb, sbitmap doloop_insns)
+ {
+   ddg_ptr g;
+   rtx insn, first_note;
+@@ -465,7 +520,6 @@
+   g = (ddg_ptr) xcalloc (1, sizeof (struct ddg));
+ 
+   g->bb = bb;
+-  g->closing_branch_deps = closing_branch_deps;
+ 
+   /* Count the number of insns in the BB.  */
+   for (insn = BB_HEAD (bb); insn != NEXT_INSN (BB_END (bb));
+@@ -538,6 +592,11 @@
+   /* Build the data dependency graph.  */
+   build_intra_loop_deps (g);
+   build_inter_loop_deps (g);
++
++  /* Check whether the do-loop part is decoupled from the other loop
++     instructions.  */
++  check_closing_branch_deps (g, doloop_insns);
++
+   return g;
+ }
+ 
+@@ -831,6 +890,18 @@
+   return NULL;
+ }
+ 
++/* Given the uid of an instruction UID return the node that represents it.  */
++static ddg_node_ptr
++get_node_of_insn_uid (ddg_ptr g, int uid)
++{
++  int i;
++
++  for (i = 0; i < g->num_nodes; i++)
++    if (uid == INSN_UID (g->nodes[i].insn))
++      return &g->nodes[i];
++  return NULL;
++}
++
+ /* Given a set OPS of nodes in the DDG, find the set of their successors
+    which are not in OPS, and set their bits in SUCC.  Bits corresponding to
+    OPS are cleared from SUCC.  Leaves the other bits in SUCC unchanged.  */
+
+=== modified file 'gcc/ddg.h'
+--- old/gcc/ddg.h       2009-11-25 10:55:54 +0000
++++ new/gcc/ddg.h       2011-03-24 07:45:38 +0000
+@@ -167,7 +167,7 @@
+ };
+ 
+ 
+-ddg_ptr create_ddg (basic_block, int closing_branch_deps);
++ddg_ptr create_ddg (basic_block, sbitmap);
+ void free_ddg (ddg_ptr);
+ 
+ void print_ddg (FILE *, ddg_ptr);
+
+=== modified file 'gcc/loop-doloop.c'
+--- old/gcc/loop-doloop.c       2010-07-19 08:58:53 +0000
++++ new/gcc/loop-doloop.c       2011-03-24 07:45:38 +0000
+@@ -78,6 +78,8 @@
+   rtx inc_src;
+   rtx condition;
+   rtx pattern;
++  rtx cc_reg = NULL_RTX;
++  rtx reg_orig = NULL_RTX;
+ 
+   /* The canonical doloop pattern we expect has one of the following
+      forms:
+@@ -96,7 +98,16 @@
+      2)  (set (reg) (plus (reg) (const_int -1))
+          (set (pc) (if_then_else (reg != 0)
+                                 (label_ref (label))
+-                                (pc))).  */
++                                (pc))).  
++
++     Some targets (ARM) do the comparison before the branch, as in the
++     following form:
++
++     3) (parallel [(set (cc) (compare ((plus (reg) (const_int -1), 0)))
++                   (set (reg) (plus (reg) (const_int -1)))])
++        (set (pc) (if_then_else (cc == NE)
++                                (label_ref (label))
++                                (pc))) */
+ 
+   pattern = PATTERN (doloop_pat);
+ 
+@@ -104,19 +115,47 @@
+     {
+       rtx cond;
+       rtx prev_insn = prev_nondebug_insn (doloop_pat);
++      rtx cmp_arg1, cmp_arg2;
++      rtx cmp_orig;
+ 
+-      /* We expect the decrement to immediately precede the branch.  */
++      /* In case the pattern is not PARALLEL we expect two forms
++        of doloop which are cases 2) and 3) above: in case 2) the
++        decrement immediately precedes the branch, while in case 3)
++        the compare and decrement instructions immediately precede
++        the branch.  */
+ 
+       if (prev_insn == NULL_RTX || !INSN_P (prev_insn))
+         return 0;
+ 
+       cmp = pattern;
+-      inc = PATTERN (PREV_INSN (doloop_pat));
++      if (GET_CODE (PATTERN (prev_insn)) == PARALLEL)
++        {
++         /* The third case: the compare and decrement instructions
++            immediately precede the branch.  */
++         cmp_orig = XVECEXP (PATTERN (prev_insn), 0, 0);
++         if (GET_CODE (cmp_orig) != SET)
++           return 0;
++         if (GET_CODE (SET_SRC (cmp_orig)) != COMPARE)
++           return 0;
++         cmp_arg1 = XEXP (SET_SRC (cmp_orig), 0);
++          cmp_arg2 = XEXP (SET_SRC (cmp_orig), 1);
++         if (cmp_arg2 != const0_rtx 
++             || GET_CODE (cmp_arg1) != PLUS)
++           return 0;
++         reg_orig = XEXP (cmp_arg1, 0);
++         if (XEXP (cmp_arg1, 1) != GEN_INT (-1) 
++             || !REG_P (reg_orig))
++           return 0;
++         cc_reg = SET_DEST (cmp_orig);
++         
++         inc = XVECEXP (PATTERN (prev_insn), 0, 1);
++       }
++      else
++        inc = PATTERN (prev_insn);
+       /* We expect the condition to be of the form (reg != 0)  */
+       cond = XEXP (SET_SRC (cmp), 0);
+       if (GET_CODE (cond) != NE || XEXP (cond, 1) != const0_rtx)
+         return 0;
+-
+     }
+   else
+     {
+@@ -162,11 +201,15 @@
+     return 0;
+ 
+   if ((XEXP (condition, 0) == reg)
++      /* For the third case:  */  
++      || ((cc_reg != NULL_RTX)
++         && (XEXP (condition, 0) == cc_reg)
++         && (reg_orig == reg))
+       || (GET_CODE (XEXP (condition, 0)) == PLUS
+-                  && XEXP (XEXP (condition, 0), 0) == reg))
++         && XEXP (XEXP (condition, 0), 0) == reg))
+    {
+      if (GET_CODE (pattern) != PARALLEL)
+-     /*  The second form we expect:
++     /*  For the second form we expect:
+ 
+          (set (reg) (plus (reg) (const_int -1))
+          (set (pc) (if_then_else (reg != 0)
+@@ -181,7 +224,24 @@
+                      (set (reg) (plus (reg) (const_int -1)))
+                      (additional clobbers and uses)])
+ 
+-         So we return that form instead.
++        For the third form we expect:
++
++        (parallel [(set (cc) (compare ((plus (reg) (const_int -1)), 0))
++                   (set (reg) (plus (reg) (const_int -1)))])
++        (set (pc) (if_then_else (cc == NE)
++                                (label_ref (label))
++                                (pc))) 
++
++        which is equivalent to the following:
++
++        (parallel [(set (cc) (compare (reg,  1))
++                   (set (reg) (plus (reg) (const_int -1)))
++                   (set (pc) (if_then_else (NE == cc)
++                                           (label_ref (label))
++                                           (pc))))])
++
++        So we return the second form instead for the two cases.
++
+      */
+         condition = gen_rtx_fmt_ee (NE, VOIDmode, inc_src, const1_rtx);
+ 
+
+=== modified file 'gcc/modulo-sched.c'
+--- old/gcc/modulo-sched.c      2009-11-25 10:55:54 +0000
++++ new/gcc/modulo-sched.c      2011-03-24 07:45:38 +0000
+@@ -116,8 +116,10 @@
+ 
+ /* The number of different iterations the nodes in ps span, assuming
+    the stage boundaries are placed efficiently.  */
+-#define PS_STAGE_COUNT(ps) ((PS_MAX_CYCLE (ps) - PS_MIN_CYCLE (ps) \
+-                            + 1 + (ps)->ii - 1) / (ps)->ii)
++#define CALC_STAGE_COUNT(min_cycle,max_cycle,ii) ((max_cycle - min_cycle \
++                         + 1 + ii - 1) / ii)
++/* The stage count of ps.  */
++#define PS_STAGE_COUNT(ps) (((partial_schedule_ptr)(ps))->stage_count)
+ 
+ /* A single instruction in the partial schedule.  */
+ struct ps_insn
+@@ -155,6 +157,8 @@
+   int max_cycle;
+ 
+   ddg_ptr g;   /* The DDG of the insns in the partial schedule.  */
++
++  int stage_count;  /* The stage count of the partial schedule.  */
+ };
+ 
+ /* We use this to record all the register replacements we do in
+@@ -195,6 +199,7 @@
+                                     rtx, rtx);
+ static void duplicate_insns_of_cycles (partial_schedule_ptr,
+                                       int, int, int, rtx);
++static int calculate_stage_count (partial_schedule_ptr ps);
+ 
+ #define SCHED_ASAP(x) (((node_sched_params_ptr)(x)->aux.info)->asap)
+ #define SCHED_TIME(x) (((node_sched_params_ptr)(x)->aux.info)->time)
+@@ -310,10 +315,10 @@
+      either a single (parallel) branch-on-count or a (non-parallel)
+      branch immediately preceded by a single (decrement) insn.  */
+   first_insn_not_to_check = (GET_CODE (PATTERN (tail)) == PARALLEL ? tail
+-                             : PREV_INSN (tail));
++                             : prev_nondebug_insn (tail));
+ 
+   for (insn = head; insn != first_insn_not_to_check; insn = NEXT_INSN (insn))
+-    if (reg_mentioned_p (reg, insn))
++    if (reg_mentioned_p (reg, insn) && NONDEBUG_INSN_P (insn))
+       {
+         if (dump_file)
+         {
+@@ -332,6 +337,24 @@
+ #endif
+ }
+ 
++/* Mark in DOLOOP_INSNS the instructions that belong to the do-loop part.
++  Use TAIL to recognize that part.  */
++static void
++mark_doloop_insns (sbitmap doloop_insns, rtx tail)
++{
++  rtx first_insn_not_to_check, insn;
++
++  /* This is the first instruction which belongs the doloop part.  */
++  first_insn_not_to_check = (GET_CODE (PATTERN (tail)) == PARALLEL ? tail
++                            : prev_nondebug_insn (tail));
++
++  sbitmap_zero (doloop_insns);
++  for (insn = first_insn_not_to_check; insn != NEXT_INSN (tail);
++       insn = NEXT_INSN (insn))
++    if (NONDEBUG_INSN_P (insn))
++      SET_BIT (doloop_insns, INSN_UID (insn));
++}
++
+ /* Check if COUNT_REG is set to a constant in the PRE_HEADER block, so
+    that the number of iterations is a compile-time constant.  If so,
+    return the rtx that sets COUNT_REG to a constant, and set COUNT to
+@@ -569,13 +592,12 @@
+     }
+ }
+ 
+-/* Bump the SCHED_TIMEs of all nodes to start from zero.  Set the values
+-   of SCHED_ROW and SCHED_STAGE.  */
++/* Bump the SCHED_TIMEs of all nodes by AMOUNT.  Set the values of
++   SCHED_ROW and SCHED_STAGE.  */
+ static void
+-normalize_sched_times (partial_schedule_ptr ps)
++reset_sched_times (partial_schedule_ptr ps, int amount)
+ {
+   int row;
+-  int amount = PS_MIN_CYCLE (ps);
+   int ii = ps->ii;
+   ps_insn_ptr crr_insn;
+ 
+@@ -584,6 +606,10 @@
+       {
+        ddg_node_ptr u = crr_insn->node;
+        int normalized_time = SCHED_TIME (u) - amount;
++       int new_min_cycle = PS_MIN_CYCLE (ps) - amount;
++        /* The first cycle in row zero after the rotation.  */
++       int new_first_cycle_in_row_zero = 
++         new_min_cycle + ii - SMODULO (new_min_cycle, ii);
+ 
+        if (dump_file)
+          fprintf (dump_file, "crr_insn->node=%d, crr_insn->cycle=%d,\
+@@ -592,8 +618,30 @@
+        gcc_assert (SCHED_TIME (u) >= ps->min_cycle);
+        gcc_assert (SCHED_TIME (u) <= ps->max_cycle);
+        SCHED_TIME (u) = normalized_time;
+-       SCHED_ROW (u) = normalized_time % ii;
+-       SCHED_STAGE (u) = normalized_time / ii;
++       crr_insn->cycle = normalized_time;
++       SCHED_ROW (u) = SMODULO (normalized_time, ii);
++
++       /* If min_cycle is in row zero after the rotation then
++          the stage count can be calculated by dividing the cycle
++          with ii.  Otherwise, the calculation is done by dividing the
++          SMSed kernel into two intervals:
++
++          1) min_cycle               <= interval 0 < first_cycle_in_row_zero
++          2) first_cycle_in_row_zero <= interval 1 < max_cycle
++          
++          Cycles in interval 0 are in stage 0. The stage of cycles
++          in interval 1 should be added by 1 to take interval 0 into
++          account.  */
++       if (SMODULO (new_min_cycle, ii) == 0)
++          SCHED_STAGE (u) = normalized_time / ii;
++        else
++         {
++            if (crr_insn->cycle < new_first_cycle_in_row_zero)
++             SCHED_STAGE (u) = 0;
++           else
++              SCHED_STAGE (u) = 
++               ((SCHED_TIME (u) - new_first_cycle_in_row_zero) / ii) + 1;
++         }
+       }
+ }
+ 
+@@ -646,9 +694,12 @@
+ 
+         /* Do not duplicate any insn which refers to count_reg as it
+            belongs to the control part.
++           If closing_branch_deps is true the closing branch is scheduled
++           as well and thus should be ignored.
+            TODO: This should be done by analyzing the control part of
+            the loop.  */
+-        if (reg_mentioned_p (count_reg, u_node->insn))
++        if (reg_mentioned_p (count_reg, u_node->insn)
++            || JUMP_P (ps_ij->node->insn))
+           continue;
+ 
+        if (for_prolog)
+@@ -894,7 +945,8 @@
+   basic_block condition_bb = NULL;
+   edge latch_edge;
+   gcov_type trip_count = 0;
+-
++  sbitmap doloop_insns;
++  
+   loop_optimizer_init (LOOPS_HAVE_PREHEADERS
+                       | LOOPS_HAVE_RECORDED_EXITS);
+   if (number_of_loops () <= 1)
+@@ -919,6 +971,7 @@
+   setup_sched_infos ();
+   haifa_sched_init ();
+ 
++  doloop_insns = sbitmap_alloc (get_max_uid () + 1);
+   /* Allocate memory to hold the DDG array one entry for each loop.
+      We use loop->num as index into this array.  */
+   g_arr = XCNEWVEC (ddg_ptr, number_of_loops ());
+@@ -1009,9 +1062,11 @@
+        continue;
+       }
+ 
+-      /* Don't handle BBs with calls or barriers, or !single_set insns,
+-         or auto-increment insns (to avoid creating invalid reg-moves
+-         for the auto-increment insns).
++      /* Don't handle BBs with calls or barriers or auto-increment insns 
++        (to avoid creating invalid reg-moves for the auto-increment insns),
++        or !single_set with the exception of instructions that include
++        count_reg---these instructions are part of the control part
++        that do-loop recognizes.
+          ??? Should handle auto-increment insns.
+          ??? Should handle insns defining subregs.  */
+      for (insn = head; insn != NEXT_INSN (tail); insn = NEXT_INSN (insn))
+@@ -1021,7 +1076,8 @@
+         if (CALL_P (insn)
+             || BARRIER_P (insn)
+             || (NONDEBUG_INSN_P (insn) && !JUMP_P (insn)
+-                && !single_set (insn) && GET_CODE (PATTERN (insn)) != USE)
++                && !single_set (insn) && GET_CODE (PATTERN (insn)) != USE
++                && !reg_mentioned_p (count_reg, insn))
+             || (FIND_REG_INC_NOTE (insn, NULL_RTX) != 0)
+             || (INSN_P (insn) && (set = single_set (insn))
+                 && GET_CODE (SET_DEST (set)) == SUBREG))
+@@ -1048,14 +1104,16 @@
+ 
+          continue;
+        }
+-
+-      if (! (g = create_ddg (bb, 0)))
++      mark_doloop_insns (doloop_insns, tail);
++      if (! (g = create_ddg (bb, doloop_insns)))
+         {
+           if (dump_file)
+            fprintf (dump_file, "SMS create_ddg failed\n");
+          continue;
+         }
+-
++      if (dump_file)
++        fprintf (dump_file, "SMS closing_branch_deps: %d\n", 
++                 g->closing_branch_deps); 
+       g_arr[loop->num] = g;
+       if (dump_file)
+         fprintf (dump_file, "...OK\n");
+@@ -1157,11 +1215,13 @@
+ 
+       ps = sms_schedule_by_order (g, mii, maxii, node_order);
+ 
+-      if (ps){
+-       stage_count = PS_STAGE_COUNT (ps);
+-        gcc_assert(stage_count >= 1);
+-      }
+-
++      if (ps)
++       {
++         stage_count = calculate_stage_count (ps);
++         gcc_assert(stage_count >= 1);
++         PS_STAGE_COUNT(ps) = stage_count;
++       }
++      
+       /* Stage count of 1 means that there is no interleaving between
+          iterations, let the scheduling passes do the job.  */
+       if (stage_count <= 1
+@@ -1182,17 +1242,7 @@
+       else
+        {
+          struct undo_replace_buff_elem *reg_move_replaces;
+-
+-         if (dump_file)
+-           {
+-             fprintf (dump_file,
+-                      "SMS succeeded %d %d (with ii, sc)\n", ps->ii,
+-                      stage_count);
+-             print_partial_schedule (ps, dump_file);
+-             fprintf (dump_file,
+-                      "SMS Branch (%d) will later be scheduled at cycle %d.\n",
+-                      g->closing_branch->cuid, PS_MIN_CYCLE (ps) - 1);
+-           }
++         int amount;
+ 
+          /* Set the stage boundaries.  If the DDG is built with closing_branch_deps,
+             the closing_branch was scheduled and should appear in the last (ii-1)
+@@ -1202,12 +1252,28 @@
+              TODO: Revisit the issue of scheduling the insns of the
+              control part relative to the branch when the control part
+              has more than one insn.  */
+-         normalize_sched_times (ps);
+-         rotate_partial_schedule (ps, PS_MIN_CYCLE (ps));
++         amount = (g->closing_branch_deps)? SCHED_TIME (g->closing_branch) + 1: 
++           PS_MIN_CYCLE (ps);
++         reset_sched_times (ps, amount);
++         rotate_partial_schedule (ps, amount);
++         
+          set_columns_for_ps (ps);
+ 
+          canon_loop (loop);
+ 
++          if (dump_file)
++            {
++             fprintf (dump_file,
++                      "SMS succeeded %d %d (with ii, sc)\n", ps->ii,
++                      stage_count);
++             print_partial_schedule (ps, dump_file);
++             if (!g->closing_branch_deps)
++               fprintf (dump_file,
++                        "SMS Branch (%d) will later be scheduled at \
++                        cycle %d.\n",
++                        g->closing_branch->cuid, PS_MIN_CYCLE (ps) - 1);
++            }
++
+           /* case the BCT count is not known , Do loop-versioning */
+          if (count_reg && ! count_init)
+             {
+@@ -1252,6 +1318,7 @@
+     }
+ 
+   free (g_arr);
++  sbitmap_free (doloop_insns);
+ 
+   /* Release scheduler data, needed until now because of DFA.  */
+   haifa_sched_finish ();
+@@ -1759,8 +1826,9 @@
+              RESET_BIT (tobe_scheduled, u);
+              continue;
+            }
+-
+-         if (JUMP_P (insn)) /* Closing branch handled later.  */
++         /* Closing branch handled later unless closing_branch_deps
++            is true.  */
++         if (JUMP_P (insn) && !g->closing_branch_deps) 
+            {
+              RESET_BIT (tobe_scheduled, u);
+              continue;
+@@ -1893,8 +1961,8 @@
+   if (dump_file)
+     fprintf (dump_file, "split_row=%d\n", split_row);
+ 
+-  normalize_sched_times (ps);
+-  rotate_partial_schedule (ps, ps->min_cycle);
++  reset_sched_times (ps, PS_MIN_CYCLE (ps));
++  rotate_partial_schedule (ps, PS_MIN_CYCLE (ps));
+ 
+   rows_new = (ps_insn_ptr *) xcalloc (new_ii, sizeof (ps_insn_ptr));
+   for (row = 0; row < split_row; row++)
+@@ -2571,6 +2639,7 @@
+   ps_insn_ptr next_ps_i;
+   ps_insn_ptr first_must_follow = NULL;
+   ps_insn_ptr last_must_precede = NULL;
++  ps_insn_ptr last_in_row = NULL;
+   int row;
+ 
+   if (! ps_i)
+@@ -2597,8 +2666,37 @@
+          else
+             last_must_precede = next_ps_i;
+         }
++      /* The closing branch must be the last in the row.  */
++      if (must_precede 
++         && TEST_BIT (must_precede, next_ps_i->node->cuid) 
++         && JUMP_P (next_ps_i->node->insn))     
++       return false;
++             
++       last_in_row = next_ps_i;
+     }
+ 
++  /* If closing_branch_deps is true we are scheduling the closing
++     branch as well.  Make sure there is no dependent instruction after
++     it as the branch should be the last instruction.  */
++  if (JUMP_P (ps_i->node->insn)) 
++    {
++      if (first_must_follow)
++       return false;
++      if (last_in_row)
++       {
++         /* Make the branch the last in the row.  New instructions
++            will be inserted at the beginning of the row or after the
++            last must_precede instruction thus the branch is guaranteed
++            to remain the last instruction in the row.  */
++         last_in_row->next_in_row = ps_i;
++         ps_i->prev_in_row = last_in_row;
++         ps_i->next_in_row = NULL;
++       }
++      else
++       ps->rows[row] = ps_i;
++      return true;
++    }
++  
+   /* Now insert the node after INSERT_AFTER_PSI.  */
+ 
+   if (! last_must_precede)
+@@ -2820,6 +2918,54 @@
+   return ps_i;
+ }
+ 
++/* Calculate the stage count of the partial schedule PS.  */
++int
++calculate_stage_count (partial_schedule_ptr ps)
++{
++  int stage_count;
++
++  /* If closing_branch_deps is false then the stage
++     boundaries are placed efficiently, meaning that min_cycle will be
++     placed at row 0. Otherwise, the closing branch will be placed in
++     row ii-1. For the later case we assume the final SMSed kernel can
++     be divided into two intervals. This assumption is used for the
++     stage count calculation:
++
++     1) min_cycle               <= interval 0 < first_cycle_in_row_zero
++     2) first_cycle_in_row_zero <= interval 1 < max_cycle
++   */
++  stage_count =
++    CALC_STAGE_COUNT (PS_MIN_CYCLE (ps), PS_MAX_CYCLE (ps), ps->ii);
++  if (ps->g->closing_branch_deps)
++    {
++      int new_min_cycle;
++      int new_min_cycle_row;
++      int rotation_amount = SCHED_TIME (ps->g->closing_branch) + 1;
++
++      /* This is the new value of min_cycle after the final rotation to
++         bring closing branch into row ii-1.  */
++      new_min_cycle = PS_MIN_CYCLE (ps) - rotation_amount;
++      /* This is the row which the the new min_cycle will be placed in.  */
++      new_min_cycle_row = SMODULO (new_min_cycle, ps->ii);
++      /* If the row of min_cycle is zero then interval 0 is empty.
++         Otherwise, we need to calculate interval 1 and add it by one
++         to take interval 0 into account.  */
++      if (new_min_cycle_row != 0)
++       {
++         int new_max_cycle, first_cycle_in_row_zero;
++
++         new_max_cycle = PS_MAX_CYCLE (ps) - rotation_amount;
++         first_cycle_in_row_zero =
++           new_min_cycle + ps->ii - new_min_cycle_row;
++
++         stage_count =
++           CALC_STAGE_COUNT (first_cycle_in_row_zero, new_max_cycle,
++                             ps->ii) + 1;
++       }
++    }
++  return stage_count;
++}
++
+ /* Rotate the rows of PS such that insns scheduled at time
+    START_CYCLE will appear in row 0.  Updates max/min_cycles.  */
+ void
+