--- testsuite/gcc.target/i386/incoming-7.c	(revision 0)
+++ testsuite/gcc.target/i386/incoming-7.c	(revision 0)
@@ -0,0 +1,16 @@ 
+/* PR target/40838 */
+/* { dg-do compile { target { { ! *-*-darwin* } && ilp32 } } } */
+/* { dg-options "-w -O2 -msse2 -mpreferred-stack-boundary=4" } */
+
+typedef int v4si __attribute__ ((vector_size (16)));
+
+extern v4si y(v4si, v4si, v4si, v4si, v4si);
+
+extern v4si s1, s2;
+
+v4si x(void)
+{
+  return y(s1, s2, s1, s2, s2);
+}
+
+/* { dg-final { scan-assembler "andl\[\\t \]*\\$-16,\[\\t \]*%esp" } } */
--- a/gcc/testsuite/gcc.target/i386/incoming-9.c	(revision 0)
+++ a/gcc/testsuite/gcc.target/i386/incoming-9.c	(revision 0)
@@ -0,0 +1,18 @@ 
+/* PR target/40838 */
+/* { dg-do compile { target { { ! *-*-darwin* } && ilp32 } } } */
+/* { dg-options "-w -O3 -mno-sse -mpreferred-stack-boundary=4" } */
+
+float
+foo (float f)
+{
+  float array[128];
+  float x;
+  int i;
+  for (i = 0; i < sizeof(array) / sizeof(*array); i++)
+    array[i] = f;
+  for (i = 0; i < sizeof(array) / sizeof(*array); i++)
+    x += array[i];
+  return x;
+}
+
+/* { dg-final { scan-assembler-not "andl\[\\t \]*\\$-16,\[\\t \]*%esp" } } */
--- a/gcc/testsuite/gcc.target/i386/incoming-6.c	(revision 0)
+++ a/gcc/testsuite/gcc.target/i386/incoming-6.c	(revision 0)
@@ -0,0 +1,17 @@ 
+/* PR target/40838 */
+/* { dg-do compile { target { { ! *-*-darwin* } && ilp32 } } } */
+/* { dg-options "-w -O2 -msse2 -mpreferred-stack-boundary=4" } */
+
+typedef int v4si __attribute__ ((vector_size (16)));
+
+extern v4si y(v4si *s3);
+
+extern v4si s1, s2;
+
+v4si x(void)
+{
+  v4si s3 = s1 + s2;
+  return y(&s3);
+}
+
+/* { dg-final { scan-assembler "andl\[\\t \]*\\$-16,\[\\t \]*%esp" } } */
--- a/gcc/testsuite/gcc.target/i386/incoming-8.c	(revision 0)
+++ a/gcc/testsuite/gcc.target/i386/incoming-8.c	(revision 0)
@@ -0,0 +1,18 @@ 
+/* PR target/40838 */
+/* { dg-do compile { target { { ! *-*-darwin* } && ilp32 } } } */
+/* { dg-options "-w -O3 -msse2 -mpreferred-stack-boundary=4" } */
+
+float
+foo (float f)
+{
+  float array[128];
+  float x;
+  int i;
+  for (i = 0; i < sizeof(array) / sizeof(*array); i++)
+    array[i] = f;
+  for (i = 0; i < sizeof(array) / sizeof(*array); i++)
+    x += array[i];
+  return x;
+}
+
+/* { dg-final { scan-assembler "andl\[\\t \]*\\$-16,\[\\t \]*%esp" } } */
--- a/gcc/config/i386/i386.h	(revision 150821)
+++ a/gcc/config/i386/i386.h	(working copy)
@@ -2404,6 +2404,8 @@ struct machine_function GTY(())
   /* This value is used for amd64 targets and specifies the current abi
      to be used. MS_ABI means ms abi. Otherwise SYSV_ABI means sysv abi.  */
   int call_abi;
+  /* Use STACK_BOUNDARY for incoming stack boundary.  */
+  int use_stack_boundary_for_incoming_stack_boundary;
 };
 
 #define ix86_stack_locals (cfun->machine->stack_locals)
--- a/gcc/config/i386/i386.c	(revision 150821)
+++ a/gcc/config/i386/i386.c	(working copy)
@@ -8038,11 +8038,19 @@ find_drap_reg (void)
 static void
 ix86_update_stack_boundary (void)
 {
+  /* Should we use STACK_BOUNDARY for incoming stack boundary?  */
+  unsigned int incoming_stack_boundary;
+
+  if (cfun->machine->use_stack_boundary_for_incoming_stack_boundary)
+    incoming_stack_boundary = STACK_BOUNDARY;
+  else
+    incoming_stack_boundary = ix86_default_incoming_stack_boundary;
+
   /* Prefer the one specified at command line. */
   ix86_incoming_stack_boundary 
     = (ix86_user_incoming_stack_boundary
        ? ix86_user_incoming_stack_boundary
-       : ix86_default_incoming_stack_boundary);
+       : incoming_stack_boundary);
 
   /* Incoming stack alignment can be changed on individual functions
      via force_align_arg_pointer attribute.  We use the smallest
@@ -19560,6 +19568,10 @@ ix86_local_alignment (tree exp, enum mac
   return align;
 }
 
+#define VALID_SSE_VECTOR_MODE(MODE) \
+  ((MODE) == V4SFmode || (MODE) == V4SImode || (MODE) == V2DFmode \
+   || (MODE) == V16QImode || (MODE) == V8HImode || (MODE) == V2DImode)
+
 /* Compute the minimum required alignment for dynamic stack realignment
    purposes for a local variable, parameter or a stack slot.  EXP is
    the data type or decl itself, MODE is its mode and ALIGN is the
@@ -19571,7 +19583,7 @@ ix86_minimum_alignment (tree exp, enum m
 {
   tree type, decl;
 
-  if (TARGET_64BIT || align != 64 || ix86_preferred_stack_boundary >= 64)
+  if (TARGET_64BIT)
     return align;
 
   if (exp && DECL_P (exp))
@@ -19585,6 +19597,23 @@ ix86_minimum_alignment (tree exp, enum m
       decl = NULL;
     }
 
+  /* In 32bit, use STACK_BOUNDARY for incoming stack boundary if any
+     SSE variables are put on stack.  Since gcc 4.4 may not generate
+     local vector variables for vectorizer, we also check array with
+     128bit user alignment set by vectorizer.  */
+  if (VALID_SSE_VECTOR_MODE (mode)
+      || (type
+	  && (VALID_SSE_VECTOR_MODE (TYPE_MODE (type))
+	      || (flag_tree_vectorize != 0
+		  && TREE_CODE (type) == ARRAY_TYPE
+		  && decl
+		  && DECL_USER_ALIGN (decl)
+		  && DECL_ALIGN (decl) == 128))))
+     cfun->machine->use_stack_boundary_for_incoming_stack_boundary = 1;
+
+  if (align != 64 || ix86_preferred_stack_boundary >= 64)
+    return align;
+
   /* Don't do dynamic stack realignment for long long objects with
      -mpreferred-stack-boundary=2.  */
   if ((mode == DImode || (type && TYPE_MODE (type) == DImode))