diff --git a/configure b/configure
index 4d0e1163a2..c989fece9a 100755
--- a/configure
+++ b/configure
@@ -278,7 +278,7 @@ QEMU_CFLAGS="-fno-strict-aliasing $QEMU_CFLAGS"
 QEMU_CFLAGS="-Wall -Wundef -Wwrite-strings -Wmissing-prototypes $QEMU_CFLAGS"
 QEMU_CFLAGS="-Wstrict-prototypes -Wredundant-decls $QEMU_CFLAGS"
 QEMU_CFLAGS="-D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE $QEMU_CFLAGS"
-QEMU_INCLUDES="-I. -I\$(SRC_PATH) -I\$(SRC_PATH)/include -I\$(SRC_PATH)/fpu"
+QEMU_INCLUDES="-I. -I\$(SRC_PATH) -I\$(SRC_PATH)/include"
 if test "$debug_info" = "yes"; then
     CFLAGS="-g $CFLAGS"
     LDFLAGS="-g $LDFLAGS"
diff --git a/fpu/softfloat.c b/fpu/softfloat.c
index 841314686b..0cfa6b4831 100644
--- a/fpu/softfloat.c
+++ b/fpu/softfloat.c
@@ -40,7 +40,7 @@ these four paragraphs for those parts of this code that are retained.
  */
 #include "config.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 /*----------------------------------------------------------------------------
 | Primitive arithmetic functions, including multi-word arithmetic, and
diff --git a/fpu/softfloat.h b/include/fpu/softfloat.h
similarity index 100%
rename from fpu/softfloat.h
rename to include/fpu/softfloat.h
diff --git a/include/qemu/bswap.h b/include/qemu/bswap.h
index cc7f84d30f..2006fcd621 100644
--- a/include/qemu/bswap.h
+++ b/include/qemu/bswap.h
@@ -4,7 +4,7 @@
 #include "config-host.h"
 
 #include <inttypes.h>
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #ifdef CONFIG_MACHINE_BSWAP_H
 #include <sys/endian.h>
diff --git a/linux-user/arm/nwfpe/double_cpdo.c b/linux-user/arm/nwfpe/double_cpdo.c
index 8e9b28fa68..41c28f3096 100644
--- a/linux-user/arm/nwfpe/double_cpdo.c
+++ b/linux-user/arm/nwfpe/double_cpdo.c
@@ -19,7 +19,7 @@
 */
 
 #include "fpa11.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "fpopcode.h"
 
 float64 float64_exp(float64 Fm);
diff --git a/linux-user/arm/nwfpe/extended_cpdo.c b/linux-user/arm/nwfpe/extended_cpdo.c
index 880ce030a5..48eca3b4a6 100644
--- a/linux-user/arm/nwfpe/extended_cpdo.c
+++ b/linux-user/arm/nwfpe/extended_cpdo.c
@@ -19,7 +19,7 @@
 */
 
 #include "fpa11.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "fpopcode.h"
 
 floatx80 floatx80_exp(floatx80 Fm);
diff --git a/linux-user/arm/nwfpe/fpa11.h b/linux-user/arm/nwfpe/fpa11.h
index 002b3cbb82..bb9ac6554e 100644
--- a/linux-user/arm/nwfpe/fpa11.h
+++ b/linux-user/arm/nwfpe/fpa11.h
@@ -43,7 +43,7 @@ extern CPUARMState *user_registers;
 
 /* includes */
 #include "fpsr.h"		/* FP control and status register definitions */
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define		typeNone		0x00
 #define		typeSingle		0x01
diff --git a/linux-user/arm/nwfpe/fpa11_cpdt.c b/linux-user/arm/nwfpe/fpa11_cpdt.c
index 3e7a938253..007a3d6505 100644
--- a/linux-user/arm/nwfpe/fpa11_cpdt.c
+++ b/linux-user/arm/nwfpe/fpa11_cpdt.c
@@ -20,7 +20,7 @@
 */
 
 #include "fpa11.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "fpopcode.h"
 //#include "fpmodule.h"
 //#include "fpmodule.inl"
diff --git a/linux-user/arm/nwfpe/fpa11_cprt.c b/linux-user/arm/nwfpe/fpa11_cprt.c
index 801189798b..7be93fa54f 100644
--- a/linux-user/arm/nwfpe/fpa11_cprt.c
+++ b/linux-user/arm/nwfpe/fpa11_cprt.c
@@ -20,7 +20,7 @@
 */
 
 #include "fpa11.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "fpopcode.h"
 #include "fpa11.inl"
 //#include "fpmodule.h"
diff --git a/linux-user/arm/nwfpe/fpopcode.c b/linux-user/arm/nwfpe/fpopcode.c
index 82ac92f0ce..0dc5c9cd5d 100644
--- a/linux-user/arm/nwfpe/fpopcode.c
+++ b/linux-user/arm/nwfpe/fpopcode.c
@@ -19,7 +19,7 @@
 */
 
 #include "fpa11.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "fpopcode.h"
 #include "fpsr.h"
 //#include "fpmodule.h"
diff --git a/linux-user/arm/nwfpe/single_cpdo.c b/linux-user/arm/nwfpe/single_cpdo.c
index 26168e2201..2bfb359eb6 100644
--- a/linux-user/arm/nwfpe/single_cpdo.c
+++ b/linux-user/arm/nwfpe/single_cpdo.c
@@ -19,7 +19,7 @@
 */
 
 #include "fpa11.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "fpopcode.h"
 
 float32 float32_exp(float32 Fm);
diff --git a/target-alpha/cpu.h b/target-alpha/cpu.h
index 137703f6d7..b75c856f97 100644
--- a/target-alpha/cpu.h
+++ b/target-alpha/cpu.h
@@ -29,7 +29,7 @@
 
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define TARGET_HAS_ICE 1
 
diff --git a/target-alpha/fpu_helper.c b/target-alpha/fpu_helper.c
index fe988ec459..fad3575549 100644
--- a/target-alpha/fpu_helper.c
+++ b/target-alpha/fpu_helper.c
@@ -19,7 +19,7 @@
 
 #include "cpu.h"
 #include "helper.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define FP_STATUS (env->fp_status)
 
diff --git a/target-alpha/helper.c b/target-alpha/helper.c
index 2430f70aca..22c9c6ec8c 100644
--- a/target-alpha/helper.c
+++ b/target-alpha/helper.c
@@ -22,7 +22,7 @@
 #include <stdio.h>
 
 #include "cpu.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "helper.h"
 
 uint64_t cpu_alpha_load_fpcr (CPUAlphaState *env)
diff --git a/target-alpha/translate.c b/target-alpha/translate.c
index c94126737f..3afc3c63f9 100644
--- a/target-alpha/translate.c
+++ b/target-alpha/translate.c
@@ -611,7 +611,7 @@ static void gen_qual_roundmode(DisasContext *ctx, int fn11)
     }
 
 #if defined(CONFIG_SOFTFLOAT_INLINE)
-    /* ??? The "softfloat.h" interface is to call set_float_rounding_mode.
+    /* ??? The "fpu/softfloat.h" interface is to call set_float_rounding_mode.
        With CONFIG_SOFTFLOAT that expands to an out-of-line call that just
        sets the one field.  */
     tcg_gen_st8_i32(tmp, cpu_env,
diff --git a/target-arm/cpu.h b/target-arm/cpu.h
index 7f87efa7d9..ffddfcbc0d 100644
--- a/target-arm/cpu.h
+++ b/target-arm/cpu.h
@@ -29,7 +29,7 @@
 #include "qemu-common.h"
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define TARGET_HAS_ICE 1
 
diff --git a/target-i386/cpu.h b/target-i386/cpu.h
index f3f50a0499..0709780356 100644
--- a/target-i386/cpu.h
+++ b/target-i386/cpu.h
@@ -46,7 +46,7 @@
 
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define R_EAX 0
 #define R_ECX 1
diff --git a/target-m68k/cpu.h b/target-m68k/cpu.h
index b37af1fffd..adaf56c471 100644
--- a/target-m68k/cpu.h
+++ b/target-m68k/cpu.h
@@ -28,7 +28,7 @@
 #include "qemu-common.h"
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define MAX_QREGS 32
 
diff --git a/target-microblaze/cpu.h b/target-microblaze/cpu.h
index 5358941ae4..4de22266ef 100644
--- a/target-microblaze/cpu.h
+++ b/target-microblaze/cpu.h
@@ -27,7 +27,7 @@
 #define CPUArchState struct CPUMBState
 
 #include "exec/cpu-defs.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 struct CPUMBState;
 typedef struct CPUMBState CPUMBState;
 #if !defined(CONFIG_USER_ONLY)
diff --git a/target-mips/cpu.h b/target-mips/cpu.h
index 183ba9fbd8..31602ac8ed 100644
--- a/target-mips/cpu.h
+++ b/target-mips/cpu.h
@@ -13,7 +13,7 @@
 #include "qemu-common.h"
 #include "mips-defs.h"
 #include "exec/cpu-defs.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 struct CPUMIPSState;
 
diff --git a/target-openrisc/cpu.h b/target-openrisc/cpu.h
index 876b386a3c..3beab45c3c 100644
--- a/target-openrisc/cpu.h
+++ b/target-openrisc/cpu.h
@@ -31,7 +31,7 @@ struct OpenRISCCPU;
 #include "config.h"
 #include "qemu-common.h"
 #include "exec/cpu-defs.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 #include "qom/cpu.h"
 #include "qapi/error.h"
 
diff --git a/target-ppc/cpu.h b/target-ppc/cpu.h
index 610bcd544b..e88ebe00d4 100644
--- a/target-ppc/cpu.h
+++ b/target-ppc/cpu.h
@@ -75,7 +75,7 @@
 
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define TARGET_HAS_ICE 1
 
diff --git a/target-s390x/cpu.h b/target-s390x/cpu.h
index acb8c73174..dda0b9af66 100644
--- a/target-s390x/cpu.h
+++ b/target-s390x/cpu.h
@@ -36,7 +36,7 @@
 
 #include "exec/cpu-all.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define NB_MMU_MODES 3
 
diff --git a/target-sh4/cpu.h b/target-sh4/cpu.h
index 7c50c798b8..34e9b0acf7 100644
--- a/target-sh4/cpu.h
+++ b/target-sh4/cpu.h
@@ -41,7 +41,7 @@
 
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define TARGET_PAGE_BITS 12	/* 4k XXXXX */
 
@@ -230,8 +230,6 @@ static inline void cpu_set_tls(CPUSH4State *env, target_ulong newtls)
 
 void cpu_load_tlb(CPUSH4State * env);
 
-#include "softfloat.h"
-
 static inline CPUSH4State *cpu_init(const char *cpu_model)
 {
     SuperHCPU *cpu = cpu_sh4_init(cpu_model);
diff --git a/target-sparc/cpu.h b/target-sparc/cpu.h
index 04b6659909..7389b03514 100644
--- a/target-sparc/cpu.h
+++ b/target-sparc/cpu.h
@@ -27,7 +27,7 @@
 
 #include "exec/cpu-defs.h"
 
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define TARGET_HAS_ICE 1
 
diff --git a/target-unicore32/cpu.h b/target-unicore32/cpu.h
index cd8f730756..509ce7c69d 100644
--- a/target-unicore32/cpu.h
+++ b/target-unicore32/cpu.h
@@ -24,7 +24,7 @@
 #include "config.h"
 #include "qemu-common.h"
 #include "exec/cpu-defs.h"
-#include "softfloat.h"
+#include "fpu/softfloat.h"
 
 #define NB_MMU_MODES            2