[AArch64] Optimize epilogue when there is no frame pointer.

From-SVN: r213000

10 files changed, 75 insertions, 10 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index cf3a2d9..1e827b5 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,5 +1,10 @@
 2014-07-24  Jiong Wang  <jiong.wang@arm.com>
 
+	* config/aarch64/aarch64.c (aarch64_popwb_single_reg): New function.
+	(aarch64_expand_epilogue): Optimize epilogue when !frame_pointer_needed.
+
+2014-07-24  Jiong Wang  <jiong.wang@arm.com>
+
 	* config/aarch64/aarch64.c (aarch64_pushwb_single_reg): New function.
 	(aarch64_expand_prologue): Optimize prologue when !frame_pointer_needed.
 
diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index 76e6027..ed80269 100644
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -1946,6 +1946,23 @@ aarch64_pushwb_single_reg (enum machine_mode mode, unsigned regno,
   RTX_FRAME_RELATED_P (insn) = 1;
 }
 
+static void
+aarch64_popwb_single_reg (enum machine_mode mode, unsigned regno,
+			  HOST_WIDE_INT adjustment)
+{
+  rtx base_rtx = stack_pointer_rtx;
+  rtx insn, reg, mem;
+
+  reg = gen_rtx_REG (mode, regno);
+  mem = gen_rtx_POST_MODIFY (Pmode, base_rtx,
+			     plus_constant (Pmode, base_rtx, adjustment));
+  mem = gen_rtx_MEM (mode, mem);
+
+  insn = emit_move_insn (reg, mem);
+  add_reg_note (insn, REG_CFA_RESTORE, reg);
+  RTX_FRAME_RELATED_P (insn) = 1;
+}
+
 static rtx
 aarch64_gen_storewb_pair (enum machine_mode mode, rtx base, rtx reg, rtx reg2,
 			  HOST_WIDE_INT adjustment)
@@ -2384,7 +2401,6 @@ aarch64_expand_epilogue (bool for_sibcall)
   HOST_WIDE_INT fp_offset;
   rtx insn;
   rtx cfa_reg;
-  bool skip_wb = false;
 
   aarch64_layout_frame ();
 
@@ -2432,22 +2448,41 @@ aarch64_expand_epilogue (bool for_sibcall)
       cfa_reg = stack_pointer_rtx;
     }
 
-  aarch64_restore_callee_saves (DFmode, frame_pointer_needed ? 0 : fp_offset,
-				V0_REGNUM, V31_REGNUM, skip_wb);
-
   if (offset > 0)
     {
+      unsigned reg1 = cfun->machine->frame.wb_candidate1;
+      unsigned reg2 = cfun->machine->frame.wb_candidate2;
+      bool skip_wb = true;
+
       if (frame_pointer_needed)
+	fp_offset = 0;
+      else if (fp_offset
+	       || reg1 == FIRST_PSEUDO_REGISTER
+	       || (reg2 == FIRST_PSEUDO_REGISTER
+		   && offset >= 256))
+	skip_wb = false;
+
+      aarch64_restore_callee_saves (DImode, fp_offset, R0_REGNUM, R30_REGNUM,
+				    skip_wb);
+      aarch64_restore_callee_saves (DFmode, fp_offset, V0_REGNUM, V31_REGNUM,
+				    skip_wb);
+
+      if (skip_wb)
 	{
-	  aarch64_restore_callee_saves (DImode, 0, R0_REGNUM, R28_REGNUM,
-					skip_wb);
-	  aarch64_popwb_pair_reg (DImode, R29_REGNUM, R30_REGNUM, offset,
-				  cfa_reg);
+	  enum machine_mode mode1 = (reg1 <= R30_REGNUM) ? DImode : DFmode;
+
+	  if (reg2 == FIRST_PSEUDO_REGISTER)
+	    aarch64_popwb_single_reg (mode1, reg1, offset);
+	  else
+	    {
+	      if (reg1 != HARD_FRAME_POINTER_REGNUM)
+		cfa_reg = NULL;
+
+	      aarch64_popwb_pair_reg (mode1, reg1, reg2, offset, cfa_reg);
+	    }
 	}
       else
 	{
-	  aarch64_restore_callee_saves (DImode, fp_offset, R0_REGNUM,
-					R30_REGNUM, skip_wb);
 	  insn = emit_insn (gen_add2_insn (stack_pointer_rtx,
 					   GEN_INT (offset)));
 	  RTX_FRAME_RELATED_P (insn) = 1;
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 901343c..1aeb7ab 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -2,6 +2,17 @@
 
 	* gcc.target/aarch64/test_frame_1.c: Match optimized instruction
 	sequences.
+	* gcc.target/aarch64/test_frame_2.c: Likewise.
+	* gcc.target/aarch64/test_frame_4.c: Likewise.
+	* gcc.target/aarch64/test_frame_6.c: Likewise.
+	* gcc.target/aarch64/test_frame_7.c: Likewise.
+	* gcc.target/aarch64/test_frame_8.c: Likewise.
+	* gcc.target/aarch64/test_frame_10.c: Likewise.
+
+2014-07-24  Jiong Wang  <jiong.wang@arm.com>
+
+	* gcc.target/aarch64/test_frame_1.c: Match optimized instruction
+	sequences.
 	* gcc.target/aarch64/test_frame_10.c: Likewise.
 	* gcc.target/aarch64/test_frame_2.c: Likewise.
 	* gcc.target/aarch64/test_frame_4.c: Likewise.
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_1.c b/gcc/testsuite/gcc.target/aarch64/test_frame_1.c
index e9d04aa..5b3c0ab 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_1.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_1.c
@@ -14,4 +14,6 @@ t_frame_pattern (test1, 200, )
 t_frame_run (test1)
 
 /* { dg-final { scan-assembler-times "str\tx30, \\\[sp, -\[0-9\]+\\\]!" 2 } } */
+/* { dg-final { scan-assembler-times "ldr\tx30, \\\[sp\\\], \[0-9\]+" 3 } } */
+
 /* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_10.c b/gcc/testsuite/gcc.target/aarch64/test_frame_10.c
index b646a71..525b49e 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_10.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_10.c
@@ -16,4 +16,6 @@ t_frame_pattern_outgoing (test10, 480, "x19", 24, a[8], a[9], a[10])
 t_frame_run (test10)
 
 /* { dg-final { scan-assembler-times "stp\tx19, x30, \\\[sp, -\[0-9\]+\\\]!" 1 } } */
+/* { dg-final { scan-assembler-times "ldp\tx19, x30, \\\[sp\\\], \[0-9\]+" 1 } } */
+
 /* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_2.c b/gcc/testsuite/gcc.target/aarch64/test_frame_2.c
index b972664..6ec4088 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_2.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_2.c
@@ -15,4 +15,6 @@ t_frame_run (test2)
 
 
 /* { dg-final { scan-assembler-times "stp\tx19, x30, \\\[sp, -\[0-9\]+\\\]!" 1 } } */
+/* { dg-final { scan-assembler-times "ldp\tx19, x30, \\\[sp\\\], \[0-9\]+" 2 } } */
+
 /* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_4.c b/gcc/testsuite/gcc.target/aarch64/test_frame_4.c
index 5a9a919..ebfb290 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_4.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_4.c
@@ -14,4 +14,6 @@ t_frame_pattern (test4, 400, "x19")
 t_frame_run (test4)
 
 /* { dg-final { scan-assembler-times "stp\tx19, x30, \\\[sp, -\[0-9\]+\\\]!" 1 } } */
+/* { dg-final { scan-assembler-times "ldp\tx19, x30, \\\[sp\\\], \[0-9\]+" 2 } } */
+
 /* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_6.c b/gcc/testsuite/gcc.target/aarch64/test_frame_6.c
index 6056f57..b5ea7ee 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_6.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_6.c
@@ -15,4 +15,6 @@ t_frame_pattern (test6, 700, )
 t_frame_run (test6)
 
 /* { dg-final { scan-assembler-times "str\tx30, \\\[sp, -\[0-9\]+\\\]!" 2 } } */
+/* { dg-final { scan-assembler-times "ldr\tx30, \\\[sp\\\], \[0-9\]+" 3 } } */
+
 /* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_7.c b/gcc/testsuite/gcc.target/aarch64/test_frame_7.c
index 991860c..daa1f42 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_7.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_7.c
@@ -15,4 +15,6 @@ t_frame_pattern (test7, 700, "x19")
 t_frame_run (test7)
 
 /* { dg-final { scan-assembler-times "stp\tx19, x30, \\\[sp, -\[0-9\]+\\\]!" 1 } } */
+/* { dg-final { scan-assembler-times "ldp\tx19, x30, \\\[sp\\\], \[0-9\]+" 2 } } */
+
 /* { dg-final { cleanup-saved-temps } } */
diff --git a/gcc/testsuite/gcc.target/aarch64/test_frame_8.c b/gcc/testsuite/gcc.target/aarch64/test_frame_8.c
index 4a4d93b..4ce630c 100644
--- a/gcc/testsuite/gcc.target/aarch64/test_frame_8.c
+++ b/gcc/testsuite/gcc.target/aarch64/test_frame_8.c
@@ -13,4 +13,6 @@ t_frame_pattern_outgoing (test8, 700, , 8, a[8])
 t_frame_run (test8)
 
 /* { dg-final { scan-assembler-times "str\tx30, \\\[sp, -\[0-9\]+\\\]!" 3 } } */
+/* { dg-final { scan-assembler-times "ldr\tx30, \\\[sp\\\], \[0-9\]+" 3 } } */
+
 /* { dg-final { cleanup-saved-temps } } */