final.c (compute_alignments): Compute/free loop info all the time.

	* final.c (compute_alignments): Compute/free loop info all the time.
	* config/rs6000/rs6000.h (LOOP_ALIGN): Define.
	* config/rs6000/rs6000-protos.h (rs6000_loop_align): Declare.
	* config/rs6000/t-rs6000 (rs6000.o): Add cfgloop.h.
	* config/rs6000/rs6000.c (cfgloop.h): Include.
	(can_override_loop_align): New.
	(rs6000_option_override_internal): Set it.
	(TARGET_ASM_LOOP_ALIGN_MAX_SKIP): Define target hook.
	(rs6000_loop_align): New function.
	(rs6000_loop_align_max_skip): Likewise.

	* gcc.target/powerpc/loop_align.c: New.

From-SVN: r166329

8 files changed, 79 insertions, 8 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 141d8eb..cf794db 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,16 @@
+2010-11-04  Pat Haugen  <pthaugen@us.ibm.com>
+
+	* final.c (compute_alignments): Compute/free loop info all the time.
+	* config/rs6000/rs6000.h (LOOP_ALIGN): Define.
+	* config/rs6000/rs6000-protos.h (rs6000_loop_align): Declare.
+	* config/rs6000/t-rs6000 (rs6000.o): Add cfgloop.h.
+	* config/rs6000/rs6000.c (cfgloop.h): Include.
+	(can_override_loop_align): New.
+	(rs6000_option_override_internal): Set it.
+	(TARGET_ASM_LOOP_ALIGN_MAX_SKIP): Define target hook.
+	(rs6000_loop_align): New function.
+	(rs6000_loop_align_max_skip): Likewise.
+
 2010-11-04  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 
 	PR driver/45703
diff --git a/gcc/config/rs6000/rs6000-protos.h b/gcc/config/rs6000/rs6000-protos.h
index 0678fc0..937f95b 100644
--- a/gcc/config/rs6000/rs6000-protos.h
+++ b/gcc/config/rs6000/rs6000-protos.h
@@ -131,6 +131,7 @@ extern rtx rs6000_machopic_legitimize_pic_address (rtx, enum machine_mode,
 extern rtx rs6000_address_for_fpconvert (rtx);
 extern rtx rs6000_allocate_stack_temp (enum machine_mode, bool, bool);
 extern void rs6000_expand_convert_si_to_sfdf (rtx, rtx, bool);
+extern int rs6000_loop_align (rtx);
 #endif /* RTX_CODE */
 
 #ifdef TREE_CODE
diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index 3014664..854f6d0 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6000.c
@@ -51,6 +51,7 @@
 #include "langhooks.h"
 #include "reload.h"
 #include "cfglayout.h"
+#include "cfgloop.h"
 #include "sched-int.h"
 #include "gimple.h"
 #include "tree-flow.h"
@@ -156,6 +157,9 @@ static GTY(()) bool rs6000_sched_groups;
 /* Align branch targets.  */
 static GTY(()) bool rs6000_align_branch_targets;
 
+/* Non-zero to allow overriding loop alignment. */
+static int can_override_loop_align = 0;
+
 /* Support for -msched-costly-dep option.  */
 const char *rs6000_sched_costly_dep_str;
 enum rs6000_dependence_cost rs6000_sched_costly_dep;
@@ -1140,6 +1144,7 @@ static void rs6000_option_override (void);
 static void rs6000_option_init_struct (struct gcc_options *);
 static void rs6000_option_default_params (void);
 static bool rs6000_handle_option (size_t, const char *, int);
+static int rs6000_loop_align_max_skip (rtx);
 static void rs6000_parse_tls_size_option (void);
 static void rs6000_parse_yes_no_option (const char *, const char *, int *);
 static int first_altivec_reg_to_save (void);
@@ -1607,6 +1612,9 @@ static const struct default_options rs6000_option_optimization_table[] =
 #undef TARGET_HANDLE_OPTION
 #define TARGET_HANDLE_OPTION rs6000_handle_option
 
+#undef TARGET_ASM_LOOP_ALIGN_MAX_SKIP
+#define TARGET_ASM_LOOP_ALIGN_MAX_SKIP rs6000_loop_align_max_skip
+
 #undef TARGET_OPTION_OVERRIDE
 #define TARGET_OPTION_OVERRIDE rs6000_option_override
 
@@ -3026,7 +3034,10 @@ rs6000_option_override_internal (const char *default_cpu)
 	  if (align_jumps <= 0)
 	    align_jumps = 16;
 	  if (align_loops <= 0)
-	    align_loops = 16;
+	    {
+	      can_override_loop_align = 1;
+	      align_loops = 16;
+	    }
 	}
       if (align_jumps_max_skip <= 0)
 	align_jumps_max_skip = 15;
@@ -3271,6 +3282,38 @@ rs6000_builtin_mask_for_load (void)
     return 0;
 }
 
+/* Implement LOOP_ALIGN. */
+int
+rs6000_loop_align (rtx label)
+{
+  basic_block bb;
+  int ninsns;
+
+  /* Don't override loop alignment if -falign-loops was specified. */
+  if (!can_override_loop_align)
+    return align_loops_log;
+
+  bb = BLOCK_FOR_INSN (label);
+  ninsns = num_loop_insns(bb->loop_father);
+
+  /* Align small loops to 32 bytes to fit in an icache sector, otherwise return default. */
+  if (ninsns > 4 && ninsns <= 8
+      && (rs6000_cpu == PROCESSOR_POWER4
+	  || rs6000_cpu == PROCESSOR_POWER5
+	  || rs6000_cpu == PROCESSOR_POWER6
+	  || rs6000_cpu == PROCESSOR_POWER7))
+    return 5;
+  else
+    return align_loops_log;
+}
+
+/* Implement TARGET_LOOP_ALIGN_MAX_SKIP. */
+static int
+rs6000_loop_align_max_skip (rtx label)
+{
+  return (1 << rs6000_loop_align (label)) - 1;
+}
+
 /* Implement targetm.vectorize.builtin_conversion.
    Returns a decl of a function that implements conversion of an integer vector
    into a floating-point vector, or vice-versa.  DEST_TYPE is the
diff --git a/gcc/config/rs6000/rs6000.h b/gcc/config/rs6000/rs6000.h
index c317025..e68936c 100644
--- a/gcc/config/rs6000/rs6000.h
+++ b/gcc/config/rs6000/rs6000.h
@@ -2385,6 +2385,9 @@ extern char rs6000_reg_names[][8];	/* register names (0 vs. %r0).  */
   if ((LOG) != 0)			\
     fprintf (FILE, "\t.align %d\n", (LOG))
 
+/* How to align the given loop. */
+#define LOOP_ALIGN(LABEL)  rs6000_loop_align(LABEL)
+
 /* Pick up the return address upon entry to a procedure. Used for
    dwarf2 unwind information.  This also enables the table driven
    mechanism.  */
diff --git a/gcc/config/rs6000/t-rs6000 b/gcc/config/rs6000/t-rs6000
index 8066c89..79ab4f5 100644
--- a/gcc/config/rs6000/t-rs6000
+++ b/gcc/config/rs6000/t-rs6000
@@ -27,7 +27,7 @@ rs6000.o: $(CONFIG_H) $(SYSTEM_H) coretypes.h $(TM_H) \
   $(OBSTACK_H) $(TREE_H) $(EXPR_H) $(OPTABS_H) except.h function.h \
   output.h $(BASIC_BLOCK_H) $(INTEGRATE_H) toplev.h $(GGC_H) $(HASHTAB_H) \
   $(TM_P_H) $(TARGET_H) $(TARGET_DEF_H) langhooks.h reload.h gt-rs6000.h \
-  cfglayout.h
+  cfglayout.h cfgloop.h
 
 rs6000-c.o: $(srcdir)/config/rs6000/rs6000-c.c \
     $(srcdir)/config/rs6000/rs6000-protos.h \
diff --git a/gcc/final.c b/gcc/final.c
index fb4ab26..060a245 100644
--- a/gcc/final.c
+++ b/gcc/final.c
@@ -715,8 +715,8 @@ compute_alignments (void)
     {
       dump_flow_info (dump_file, TDF_DETAILS);
       flow_loops_dump (dump_file, NULL, 1);
-      loop_optimizer_init (AVOID_CFG_MODIFICATIONS);
     }
+  loop_optimizer_init (AVOID_CFG_MODIFICATIONS);
   FOR_EACH_BB (bb)
     if (bb->frequency > freq_max)
       freq_max = bb->frequency;
@@ -808,11 +808,8 @@ compute_alignments (void)
       LABEL_TO_MAX_SKIP (label) = max_skip;
     }
 
-  if (dump_file)
-    {
-      loop_optimizer_finalize ();
-      free_dominance_info (CDI_DOMINATORS);
-    }
+  loop_optimizer_finalize ();
+  free_dominance_info (CDI_DOMINATORS);
   return 0;
 }
 
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 7eabeb4..7661634 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,7 @@
+2010-11-04  Pat Haugen  <pthaugen@us.ibm.com>
+
+	* gcc.target/powerpc/loop_align.c: New.
+
 2010-11-04  Iain Sandoe  <iains@gcc.gnu.org>
 
 	* gcc.target/powerpc/ppc64-abi-1.c: Skip for Darwin.
diff --git a/gcc/testsuite/gcc.target/powerpc/loop_align.c b/gcc/testsuite/gcc.target/powerpc/loop_align.c
new file mode 100644
index 0000000..489380f
--- /dev/null
+++ b/gcc/testsuite/gcc.target/powerpc/loop_align.c
@@ -0,0 +1,10 @@
+/* { dg-do compile { target { powerpc*-*-* } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-options "-O2 -mcpu=power7 -falign-functions=16" } */
+/* { dg-final { scan-assembler ".p2align 5,,31" } } */
+
+void f(double *a, double *b, double *c, int n) {
+  int i;
+  for (i=0; i < n; i++)
+    a[i] = b[i] + c[i];
+}