[Buildroot] [PATCH 0/8] ARM floating point improvements and EABIhf support

[Buildroot] [PATCH 0/8] ARM floating point improvements and EABIhf support

[Thomas Petazzoni]

Hello,

Here is a patch series that improves the support for the various
floating point possibilities on ARM, and adds support for the new
EABIhf ARM ABI. Until now, we only had a global BR2_SOFT_FLOAT option
to enable/disable soft-float support, and on ARM a BR2_VFP_FLOAT
option to enable -mfpu=vfp. This clearly isn't sufficient to take into
the various possibilities offered by the Cortex-A ARM cores.

In summary:

 * PATCH 1 introduces the BR2_GCC_TARGET_FLOAT and BR2_GCC_TARGET_FPU
   hidden options, that will be used by the Config.in.<arch> files to
   define which -mfloat-abi= and -mfpu= values they want.

 * PATCH 2 moves BR2_SOFT_FLOAT as global hidden option, and adds
   per-architecture visible options. This will allow each architecture
   to do its own business in terms of floating point strategy
   selection, while still providing a global BR2_SOFT_FLOAT for
   Buildroot code that needs to know whether floating point is
   emulated in software or not.

 * PATCH 3 is the core of the patch set, adding many new options to
   select various VFP versions for various ARM cores, and adding
   support for the EABIhf ARM ABI. A special attention has been made
   on the Config.in help texts, to help the user making the right
   choice in terms of VFP selection.

 * PATCH 4 and 5 adjust gcc and binutils version availabilities
   according to the selected ABI or VFP version.

 * PATCH 6 removes the BR2_VFP_FLOAT option that no longer makes any
   sense.

 * PATCH 7 updates some defconfig that were using BR2_VFP_FLOAT.

 * PATCH 8 updates the external toolchain Config.in options after the
   introduction of the VFP selection and EABIhf ABI.

I have tested the following combinations:
 * ARM926 soft-float EABI internal toolchain
 * Cortex-A8 softfp EABI internal toolchain, using VFPv3-D16
 * Cortex-A8 EABIhf toolchain, using VFPv3-D16
 * Cortex-A8 soft-float with the Arago toolchain (the libc/libm itself
   uses VFP instructions, but programs are properly built soft-float)
 * Cortex-A8 softfp with the Arago toolchain, using VFPv3-D16
 * Cortex-A8 EABIhf with the Linaro toolchain, using VFPv3-D16

For each case, it was run time tested on a Cortex-A8 real platform
(not Qemu), using a small demo program that makes a few floating point
computations, and the assembly code was checked to verify that we're
correctly in the soft-float, softfp or hard cases.

Thanks,

Thomas

Thomas Petazzoni (8):
  arch: introduce BR2_GCC_TARGET_{FPU,FLOAT}
  arch: Refactor BR2_SOFT_FLOAT into per-architecture options
  arch: improve ARM floating point support and add support for EABIhf
  binutils: exclude binutils versions that don't support EABIhf
  gcc: take into account ARM floating point capabilities
  toolchain: remove the old BR2_VFP_FLOAT option
  configs: update defconfigs after VFP option changes
  toolchain-external: update config options after EABIhf introduction

 Config.in.legacy                                   |   9 +
 arch/Config.in                                     |   9 +
 arch/Config.in.arm                                 | 221 ++++++++++++++++++++-
 arch/Config.in.avr32                               |  10 +
 arch/Config.in.mips                                |  10 +
 arch/Config.in.powerpc                             |  10 +
 configs/calao_snowball_defconfig                   |   3 +-
 configs/mx53loco_defconfig                         |   3 +-
 configs/qemu_arm_nuri_defconfig                    |   3 +-
 configs/qemu_arm_vexpress_defconfig                |   3 +-
 package/Makefile.in                                |   6 +
 package/binutils/Config.in.host                    |   3 +
 package/gcc/Config.in.host                         |  10 +
 package/gcc/gcc.mk                                 |  10 +
 toolchain/toolchain-buildroot/Config.in.2          |  12 --
 toolchain/toolchain-common.in                      |  21 --
 toolchain/toolchain-external/Config.in             |  24 ++-
 toolchain/toolchain-external/ext-tool.mk           |  15 +-
 .../toolchain-external/ext-toolchain-wrapper.c     |   9 +-
 19 files changed, 330 insertions(+), 61 deletions(-)

-- 
1.8.1.2

[Buildroot] [PATCH 1/8] arch: introduce BR2_GCC_TARGET_{FPU, FLOAT}

[Thomas Petazzoni]

Buildroot already has the BR2_GCC_TARGET_{TUNE,ARCH,ABI,CPU} hidden
kconfig strings that allow per-architecture Config.in files to feed
the appropriate values of --with-{tune,arch,abi-cpu} when building
gcc, or the appropriate flags for the external toolchain wrapper.

This commit has two additional options: BR2_GCC_TARGET_{FPU,FLOAT},
that allows to define the --with-{fpu,float} gcc configure options for
the internal backend, or the -mfpu/-mfloat-abi options for the flags
of the external toolchain wrapper.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 arch/Config.in                                       |  6 ++++++
 package/gcc/gcc.mk                                   | 10 ++++++++++
 toolchain/toolchain-external/ext-tool.mk             | 10 ++++++++++
 toolchain/toolchain-external/ext-toolchain-wrapper.c |  6 ++++++
 4 files changed, 32 insertions(+)

diff --git a/arch/Config.in b/arch/Config.in
index 5ca05cd..d95d4ef 100644
--- a/arch/Config.in
+++ b/arch/Config.in
@@ -192,6 +192,12 @@ config BR2_GCC_TARGET_CPU
 config BR2_GCC_TARGET_CPU_REVISION
 	string
 
+config BR2_GCC_TARGET_FPU
+	string
+
+config BR2_GCC_TARGET_FLOAT_ABI
+	string
+
 # Set up target binary format
 choice
 	prompt "Target Binary Format"
diff --git a/package/gcc/gcc.mk b/package/gcc/gcc.mk
index 968354b..c5ca857 100644
--- a/package/gcc/gcc.mk
+++ b/package/gcc/gcc.mk
@@ -165,6 +165,16 @@ HOST_GCC_COMMON_CONF_OPT += --with-cpu=$(call qstrip,$(BR2_GCC_TARGET_CPU))
 endif
 endif
 
+GCC_TARGET_FPU = $(call qstrip,$(BR2_GCC_TARGET_FPU))
+ifneq ($(GCC_TARGET_FPU),)
+HOST_GCC_COMMON_CONF_OPT += --with-fpu=$(GCC_TARGET_FPU)
+endif
+
+GCC_TARGET_FLOAT_ABI = $(call qstrip,$(BR2_GCC_TARGET_FLOAT_ABI))
+ifneq ($(GCC_TARGET_FLOAT_ABI),)
+HOST_GCC_COMMON_CONF_OPT += --with-float=$(GCC_TARGET_FLOAT_ABI)
+endif
+
 # Branding works on >= 4.3
 ifneq ($(findstring x4.2.,x$(GCC_VERSION)),x4.2.)
 HOST_GCC_COMMON_CONF_OPT += \
diff --git a/toolchain/toolchain-external/ext-tool.mk b/toolchain/toolchain-external/ext-tool.mk
index b8d77ad..5c51e06 100644
--- a/toolchain/toolchain-external/ext-tool.mk
+++ b/toolchain/toolchain-external/ext-tool.mk
@@ -145,6 +145,8 @@ CC_TARGET_CPU_:=$(call qstrip,$(BR2_GCC_TARGET_CPU)-$(BR2_GCC_TARGET_CPU_REVISIO
 endif
 CC_TARGET_ARCH_:=$(call qstrip,$(BR2_GCC_TARGET_ARCH))
 CC_TARGET_ABI_:=$(call qstrip,$(BR2_GCC_TARGET_ABI))
+CC_TARGET_FLOAT_:=$(call qstrip,$(BR2_GCC_TARGET_FLOAT))
+CC_TARGET_FPU_:=$(call qstrip,$(BR2_GCC_TARGET_FPU))
 
 # march/mtune/floating point mode needs to be passed to the external toolchain
 # to select the right multilib variant
@@ -168,6 +170,14 @@ ifneq ($(CC_TARGET_ABI_),)
 TOOLCHAIN_EXTERNAL_CFLAGS += -mabi=$(CC_TARGET_ABI_)
 TOOLCHAIN_EXTERNAL_WRAPPER_ARGS += -DBR_ABI='"$(CC_TARGET_ABI_)"'
 endif
+ifneq ($(CC_TARGET_FLOAT_),)
+TOOLCHAIN_EXTERNAL_CFLAGS += -mfloat-abi=$(CC_TARGET_FLOAT_)
+TOOLCHAIN_EXTERNAL_WRAPPER_ARGS += -DBR_FLOAT='"$(CC_TARGET_FLOAT_)"'
+endif
+ifneq ($(CC_TARGET_FPU_),)
+TOOLCHAIN_EXTERNAL_CFLAGS += -mfpu=$(CC_TARGET_FPU_)
+TOOLCHAIN_EXTERNAL_WRAPPER_ARGS += -DBR_FPU='"$(CC_TARGET_FPU_)"'
+endif
 ifeq ($(BR2_BINFMT_FLAT),y)
 TOOLCHAIN_EXTERNAL_CFLAGS += -Wl,-elf2flt
 TOOLCHAIN_EXTERNAL_WRAPPER_ARGS += -DBR_BINFMT_FLAT
diff --git a/toolchain/toolchain-external/ext-toolchain-wrapper.c b/toolchain/toolchain-external/ext-toolchain-wrapper.c
index 9d79d68..e504ff2 100644
--- a/toolchain/toolchain-external/ext-toolchain-wrapper.c
+++ b/toolchain/toolchain-external/ext-toolchain-wrapper.c
@@ -38,6 +38,12 @@ static char *predef_args[] = {
 #ifdef BR_ABI
 	"-mabi=" BR_ABI,
 #endif
+#ifdef BR_FLOAT
+	"-mfloat-abi=" BR_FLOAT,
+#endif
+#ifdef BR2_FPU
+	"-mfpu=" BR_FPU,
+#endif
 #ifdef BR_SOFTFLOAT
 	"-msoft-float",
 #endif /* BR_SOFTFLOAT */
-- 
1.8.1.2

[Buildroot] [PATCH 2/8] arch: Refactor BR2_SOFT_FLOAT into per-architecture options

[Thomas Petazzoni]

As we are going to introduced a more advanced support of floating
point options for the ARM architecture, we need to adjust how the
soft-float option is handled. We replace the current hidden option
BR2_PREFER_SOFT_FLOAT option and the visible BR2_SOFT_FLOAT option by:

 * A global hidden BR2_SOFT_FLOAT option, defined in arch/Config.in,
   that tells whether the architecture-specific code is using software
   emulated floating point. This hidden option can be used throughout
   Buildroot to determine whether soft float is used or not.

 * Per-architecture visible BR2_<arch>_SOFT_FLOAT options, for the
   architecture for which it makes sense, which allows users to select
   soft float emulation when needed.

This change will allow each architecture to have a different way of
presenting its floating point capabilities.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 arch/Config.in                |  3 +++
 arch/Config.in.arm            | 10 ++++++++++
 arch/Config.in.avr32          | 10 ++++++++++
 arch/Config.in.mips           | 10 ++++++++++
 arch/Config.in.powerpc        | 10 ++++++++++
 toolchain/toolchain-common.in | 21 ---------------------
 6 files changed, 43 insertions(+), 21 deletions(-)

diff --git a/arch/Config.in b/arch/Config.in
index d95d4ef..8daeee3 100644
--- a/arch/Config.in
+++ b/arch/Config.in
@@ -1,6 +1,9 @@
 config BR2_ARCH_IS_64
        bool
 
+config BR2_SOFT_FLOAT
+	bool
+
 choice
 	prompt "Target Architecture"
 	default BR2_i386
diff --git a/arch/Config.in.arm b/arch/Config.in.arm
index 983cac4..a0c7228 100644
--- a/arch/Config.in.arm
+++ b/arch/Config.in.arm
@@ -86,6 +86,16 @@ config BR2_ARM_EABI
 	def_bool y
 	depends on !BR2_ARM_OABI
 
+config BR2_ARM_SOFT_FLOAT
+	bool "Use soft-float"
+	default y
+	select BR2_SOFT_FLOAT
+	help
+	  If your target CPU does not have a Floating Point Unit (FPU)
+	  or a kernel FPU emulator, but you still wish to support
+	  floating point functions, then everything will need to be
+	  compiled with soft floating point support (-msoft-float).
+
 config BR2_ARM_ENABLE_NEON
 	bool "Enable NEON SIMD extension support"
 	depends on BR2_ARM_CPU_MAYBE_HAS_NEON
diff --git a/arch/Config.in.avr32 b/arch/Config.in.avr32
index ebf8454..34055d9 100644
--- a/arch/Config.in.avr32
+++ b/arch/Config.in.avr32
@@ -3,3 +3,13 @@ config BR2_ARCH
 
 config BR2_ENDIAN
        default "BIG"
+
+config BR2_AVR32_SOFT_FLOAT
+	bool "Use soft-float"
+	default y
+	select BR2_SOFT_FLOAT
+	help
+	  If your target CPU does not have a Floating Point Unit (FPU)
+	  or a kernel FPU emulator, but you still wish to support
+	  floating point functions, then everything will need to be
+	  compiled with soft floating point support (-msoft-float).
diff --git a/arch/Config.in.mips b/arch/Config.in.mips
index 1454fb4..6242bcc 100644
--- a/arch/Config.in.mips
+++ b/arch/Config.in.mips
@@ -52,6 +52,16 @@ config BR2_MIPS_NABI64
 	depends on BR2_ARCH_IS_64
 endchoice
 
+config BR2_MIPS_SOFT_FLOAT
+	bool "Use soft-float"
+	default y
+	select BR2_SOFT_FLOAT
+	help
+	  If your target CPU does not have a Floating Point Unit (FPU)
+	  or a kernel FPU emulator, but you still wish to support
+	  floating point functions, then everything will need to be
+	  compiled with soft floating point support (-msoft-float).
+
 config BR2_ARCH
 	default "mips"		if BR2_mips
 	default "mipsel"	if BR2_mipsel
diff --git a/arch/Config.in.powerpc b/arch/Config.in.powerpc
index 8643efc..721d40a 100644
--- a/arch/Config.in.powerpc
+++ b/arch/Config.in.powerpc
@@ -88,6 +88,16 @@ config BR2_powerpc_SPE
 	depends on BR2_powerpc_8540 || BR2_powerpc_8548
 endchoice
 
+config BR2_POWERPC_SOFT_FLOAT
+	bool "Use soft-float"
+	default y
+	select BR2_SOFT_FLOAT
+	help
+	  If your target CPU does not have a Floating Point Unit (FPU)
+	  or a kernel FPU emulator, but you still wish to support
+	  floating point functions, then everything will need to be
+	  compiled with soft floating point support (-msoft-float).
+
 config BR2_ARCH
 	default "powerpc"	if BR2_powerpc
 
diff --git a/toolchain/toolchain-common.in b/toolchain/toolchain-common.in
index 7c9b842..dc3bd2a 100644
--- a/toolchain/toolchain-common.in
+++ b/toolchain/toolchain-common.in
@@ -90,27 +90,6 @@ config BR2_USE_MMU
 	  If your target has a MMU, you should say Y here.  If you
 	  are unsure, just say Y.
 
-config BR2_PREFER_SOFT_FLOAT
-	bool
-	default y if BR2_arm
-	default y if BR2_armeb
-	default y if BR2_avr32
-	default y if BR2_mips
-	default y if BR2_mipsel
-
-config BR2_SOFT_FLOAT
-	bool "Use software floating point by default"
-	depends on BR2_arm || BR2_armeb || BR2_avr32 || BR2_mips || BR2_mipsel || BR2_powerpc
-	# External toolchains will automatically do soft float or hard
-	# float depending on their configuration
-	depends on BR2_TOOLCHAIN_BUILDROOT || BR2_TOOLCHAIN_CTNG
-	default $(BR2_PREFER_SOFT_FLOAT)
-	help
-	  If your target CPU does not have a Floating Point Unit (FPU) or a
-	  kernel FPU emulator, but you still wish to support floating point
-	  functions, then everything will need to be compiled with soft
-	  floating point support (-msoft-float).
-
 config BR2_TARGET_OPTIMIZATION
 	string "Target Optimizations"
 	default "-pipe"
-- 
1.8.1.2

[Buildroot] [PATCH 3/8] arch: improve ARM floating point support and add support for EABIhf

[Thomas Petazzoni]

This commit introduces the support for the EABIhf ABI, next to the
existing support we have for EABI and OABI (even though OABI support
is deprecated). EABIhf allows to improve performance of floating point
workload by using floating point registers to transfer floating point
arguments when calling functions, instead of using integer registers
to do, as is done in the 'softfp' floating point model of EABI.

In addition to this, this commit introduces a list of options for the
floating point support:
 * Software floating point
 * VFP
 * VFPv3
 * VFPv3-D16
 * VFPv4
 * VFPv4-D16

and it introduces some logic to make sure the options are only visible
when it makes sense, depending on the ARM core being selected. This is
however made complicated by the fact that certain VFP capabilities are
mandatory on some cores, but optional on some other cores. The kconfig
logic tries to achieve the following goals:

 * Hide options that are definitely not possible.

 * Use safe default values (i.e for Cortex-A5 and A7, the presence of
   the VFPv4 unit is optional, so we default on software floating
   point on these cores)..

 * Show the available possibilities, even if some of them are not
   necessarily working on a particular core (again, for the Cortex-A5
   and A7 cores, there is no way of knowing whether the particular
   variant used by the user has VFPv4 or not, so we select software
   floating point by default, but still show VFP/VFPv3/VFPv4 options).

It is worth noting that this commit doesn't add support for all
possible -mfpu= values on ARM. We haven't added support for fpa, fpe2,
fpe3, maverick (those four are only used on very old ARM cores), for
vfpv3-fp16, vfpv3-d16-fp16, vfpv3xd, vfpv3xd-fp16, neon-fp16,
vfpv4-sp-d16. They can be added quite easily if needed thanks to the
new organization of the Config.in options.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 arch/Config.in.arm  | 231 ++++++++++++++++++++++++++++++++++++++++++++++++----
 package/Makefile.in |   6 ++
 2 files changed, 219 insertions(+), 18 deletions(-)

diff --git a/arch/Config.in.arm b/arch/Config.in.arm
index a0c7228..785462b 100644
--- a/arch/Config.in.arm
+++ b/arch/Config.in.arm
@@ -6,6 +6,31 @@ config BR2_ARM_CPU_HAS_NEON
 config BR2_ARM_CPU_MAYBE_HAS_NEON
 	bool
 
+# for some cores, VFPv2 is optional
+config BR2_ARM_CPU_MAYBE_HAS_VFPV2
+	bool
+
+config BR2_ARM_CPU_HAS_VFPV2
+	bool
+
+# for some cores, VFPv3 is optional
+config BR2_ARM_CPU_MAYBE_HAS_VFPV3
+	bool
+	select BR2_ARM_CPU_MAYBE_HAS_VFPV2
+
+config BR2_ARM_CPU_HAS_VFPV3
+	bool
+	select BR2_ARM_CPU_HAS_VFPV2
+
+# for some cores, VFPv4 is optional
+config BR2_ARM_CPU_MAYBE_HAS_VFPV4
+	bool
+	select BR2_ARM_CPU_MAYBE_HAS_VFPV3
+
+config BR2_ARM_CPU_HAS_VFPV4
+	bool
+	select BR2_ARM_CPU_HAS_VFPV3
+
 choice
 	prompt "Target Architecture Variant"
 	depends on BR2_arm || BR2_armeb
@@ -27,31 +52,40 @@ config BR2_arm10t
 	bool "arm10t"
 config BR2_arm1136jf_s_r0
 	bool "arm1136jf_s rev0"
+	select BR2_ARM_CPU_HAS_VFPV2
 config BR2_arm1136jf_s_r1
 	bool "arm1136jf_s rev1"
+	select BR2_ARM_CPU_HAS_VFPV2
 config BR2_arm1176jz_s
 	bool "arm1176jz-s"
 config BR2_arm1176jzf_s
 	bool "arm1176jzf-s"
+	select BR2_ARM_CPU_HAS_VFPV2
 config BR2_cortex_a5
 	bool "cortex-A5"
 	select BR2_ARM_CPU_MAYBE_HAS_NEON
+	select BR2_ARM_CPU_MAYBE_HAS_VFPV4
 config BR2_cortex_a7
 	bool "cortex-A7"
 	select BR2_ARM_CPU_HAS_NEON
+	select BR2_ARM_CPU_HAS_VFPV4
 config BR2_cortex_a8
 	bool "cortex-A8"
 	select BR2_ARM_CPU_HAS_NEON
+	select BR2_ARM_CPU_HAS_VFPV3
 config BR2_cortex_a9
 	bool "cortex-A9"
 	select BR2_ARM_CPU_MAYBE_HAS_NEON
+	select BR2_ARM_CPU_MAYBE_HAS_VFPV3
 config BR2_cortex_a15
 	bool "cortex-A15"
 	select BR2_ARM_CPU_HAS_NEON
+	select BR2_ARM_CPU_HAS_VFPV4
 config BR2_fa526
 	bool "fa526/626"
 config BR2_pj4
 	bool "pj4"
+	select BR2_ARM_CPU_HAS_VFPV3
 config BR2_strongarm
 	bool "strongarm sa110/sa1100"
 config BR2_xscale
@@ -67,34 +101,67 @@ config BR2_arm1136jf_s
 choice
 	prompt "Target ABI"
 	depends on BR2_arm || BR2_armeb
-	depends on BR2_DEPRECATED
 	default BR2_ARM_EABI
 	help
-	  Application Binary Interface to use
+	  Application Binary Interface to use. The Application Binary
+	  Interface describes the calling conventions (how arguments
+	  are passed to functions, how the return value is passed, how
+	  system calls are made, etc.).
 
 	  Note:
 	    Using OABI is discouraged.
 
-config BR2_ARM_EABI_CHOICE
+config BR2_ARM_EABI
 	bool "EABI"
-config BR2_ARM_OABI
-	bool "OABI"
-	depends on !BR2_GCC_VERSION_4_7_X
-endchoice
+	help
+	  The EABI is currently the standard ARM ABI, which is used in
+	  most projects. It supports both the 'soft' floating point
+	  model (in which floating point instructions are emulated in
+	  software) and the 'softfp' floating point model (in which
+	  floating point instructions are executed using an hardware
+	  floating point unit, but floating point arguments to
+	  functions are passed in integer registers).
 
-config BR2_ARM_EABI
-	def_bool y
-	depends on !BR2_ARM_OABI
+	  The 'softfp' floating point model is link-compatible with
+	  the 'soft' floating point model, i.e you can link a library
+	  built 'soft' with some other code built 'softfp'.
 
-config BR2_ARM_SOFT_FLOAT
-	bool "Use soft-float"
-	default y
-	select BR2_SOFT_FLOAT
+	  However, passing the floating point arguments in integer
+	  registers is a bit inefficient, so if your ARM processor has
+	  a floating point unit, and you don't have pre-compiled
+	  'soft' or 'softfp' code, using the EABIhf ABI will provide
+	  better floating point performances.
+
+	  If your processor does not have a floating point unit, then
+	  you must use this ABI.
+
+config BR2_ARM_EABIHF
+	bool "EABIhf"
+	depends on BR2_ARM_CPU_MAYBE_HAS_VFPV2 || BR2_ARM_CPU_HAS_VFPV2
 	help
-	  If your target CPU does not have a Floating Point Unit (FPU)
-	  or a kernel FPU emulator, but you still wish to support
-	  floating point functions, then everything will need to be
-	  compiled with soft floating point support (-msoft-float).
+	  The EABIhf is an extension of EABI which supports the 'hard'
+	  floating point model. This model uses the floating point
+	  unit to execute floating point instructions, and passes
+	  floating point arguments in floating point registers.
+
+	  It is more efficient than EABI for floating point related
+	  workload. However, it does not allow to link against code
+	  that has been pre-built for the 'soft' or 'softfp' floating
+	  point models.
+
+	  If your processor has a floating point unit, and you don't
+	  depend on existing pre-compiled code, this option is most
+	  likely the best choice.
+
+config BR2_ARM_OABI
+	bool "OABI"
+	depends on BR2_DEPRECATED
+	help
+	  The OABI is a deprecated ABI that Linux was using before ARM
+	  standardized the EABI. It should no longer be used for new
+	  projects.
+
+endchoice
 
 config BR2_ARM_ENABLE_NEON
 	bool "Enable NEON SIMD extension support"
@@ -105,6 +172,120 @@ config BR2_ARM_ENABLE_NEON
 	  Select this option if you are certain your particular
 	  implementation has NEON support and you want to use it.
 
+choice
+	prompt "Floating point strategy"
+	depends on BR2_ARM_EABI || BR2_ARM_EABIHF
+	default BR2_ARM_FPU_VFPV4D16 if BR2_ARM_CPU_HAS_VFPV4
+	default BR2_ARM_FPU_VFPV3D16 if BR2_ARM_CPU_HAS_VFPV3
+	default BR2_ARM_FPU_VFPV2 if BR2_ARM_CPU_HAS_VFPV2
+	default BR2_ARM_SOFT_FLOAT if !BR2_ARM_CPU_HAS_VFPV2
+
+config BR2_ARM_SOFT_FLOAT
+	bool "Soft float"
+	depends on BR2_ARM_EABI
+	select BR2_SOFT_FLOAT
+	help
+	  This option allows to use software emulated floating
+	  point. It should be used for ARM cores that do not include a
+	  Vector Floating Point unit, such as ARMv5 cores (ARM926 for
+	  example) or certain ARMv6 cores.
+
+config BR2_ARM_FPU_VFPV2
+	bool "VFPv2"
+	depends on BR2_ARM_CPU_HAS_VFPV2 || BR2_ARM_CPU_MAYBE_HAS_VFPV2
+	help
+	  This option allows to use the VFPv2 floating point unit, as
+	  available in some ARMv6 processors (ARM1136JF-S,
+	  ARM1176JZF-S and ARM11 MPCore).
+
+	  Note that this option is also safe to use for newer cores
+	  such as Cortex-A, because the VFPv3 and VFPv4 units are
+	  backward compatible with VFPv2.
+
+config BR2_ARM_FPU_VFPV3
+	bool "VFPv3"
+	depends on BR2_ARM_CPU_HAS_VFPV3 || BR2_ARM_CPU_MAYBE_HAS_VFPV3
+	help
+	  This option allows to use the VFPv3 floating point unit, as
+	  available in some ARMv7 processors (Cortex-A{8, 9}). This
+	  option requires a VFPv3 unit that has 32 double-precision
+	  registers, which is not necessarily the case in all SOCs
+	  based on Cortex-A{8, 9}. If you're unsure, use VFPv3-D16
+	  instead, which is guaranteed to work on all Cortex-A{8, 9}.
+
+	  Note that this option is also safe to use for newer cores
+	  that have a VFPv4 unit, because VFPv4 is backward compatible
+	  with VFPv3. They must of course also have 32
+	  double-precision registers.
+
+config BR2_ARM_FPU_VFPV3D16
+	bool "VFPv3-D16"
+	depends on BR2_ARM_CPU_HAS_VFPV3 || BR2_ARM_CPU_MAYBE_HAS_VFPV3
+	help
+	  This option allows to use the VFPv3 floating point unit, as
+	  available in some ARMv7 processors (Cortex-A{8, 9}). This
+	  option requires a VFPv3 unit that has 16 double-precision
+	  registers, which is generally the case in all SOCs based on
+	  Cortex-A{8, 9}, even though VFPv3 is technically optional on
+	  Cortex-A9. This is the safest option for those cores.
+
+	  Note that this option is also safe to use for newer cores
+	  such that have a VFPv4 unit, because the VFPv4 is backward
+	  compatible with VFPv3.
+
+config BR2_ARM_FPU_VFPV4
+	bool "VFPv4"
+	depends on BR2_ARM_CPU_HAS_VFPV4 || BR2_ARM_CPU_MAYBE_HAS_VFPV4
+	help
+	  This option allows to use the VFPv4 floating point unit, as
+	  available in some ARMv7 processors (Cortex-A{5, 7, 12,
+	  15}). This option requires a VFPv4 unit that has 32
+	  double-precision registers, which is not necessarily the
+	  case in all SOCs based on Cortex-A{5, 7, 12, 15}. If you're
+	  unsure, you should probably use VFPv4-D16 instead.
+
+	  Note that if you want binary code that works on all ARMv7
+	  cores, including the earlier Cortex-A{8, 9}, you should
+	  instead select VFPv3.
+
+config BR2_ARM_FPU_VFPV4D16
+	bool "VFPv4-D16"
+	depends on BR2_ARM_CPU_HAS_VFPV4 || BR2_ARM_CPU_MAYBE_HAS_VFPV4
+	help
+	  This option allows to use the VFPv4 floating point unit, as
+	  available in some ARMv7 processors (Cortex-A{5, 7, 12,
+	  15}). This option requires a VFPv4 unit that has 16
+	  double-precision registers, which is always available on
+	  Cortex-A12 and Cortex-A15, but optional on Cortex-A5 and
+	  Cortex-A7.
+
+	  Note that if you want binary code that works on all ARMv7
+	  cores, including the earlier Cortex-A{8, 9}, you should
+	  instead select VFPv3-D16.
+
+config BR2_ARM_FPU_NEON
+	bool "NEON"
+	depends on BR2_ARM_CPU_HAS_NEON
+	help
+	  This option allows to use the NEON SIMD unit, as available
+	  in some ARMv7 processors, as a floating-point unit. It
+	  should however be noted that using NEON for floating point
+	  operations doesn't provide a complete compatibility with the
+	  IEEE 754.
+
+config BR2_ARM_FPU_NEON_VFPV4
+	bool "NEON/VFPv4"
+	depends on BR2_ARM_CPU_HAS_VFPV4 || BR2_ARM_CPU_MAYBE_HAS_VFPV4
+	depends on BR2_ARM_CPU_HAS_NEON
+	help
+	  This option allows to use both the VFPv4 and the NEON SIMD
+	  units for floating point operations. Note that some ARMv7
+	  cores do not necessarily have VFPv4 and/or NEON support, for
+	  example on Cortex-A5 and Cortex-A7, support for VFPv4 and
+	  NEON is optional.
+
+endchoice
+
 config BR2_ARCH
 	default "arm"	if BR2_arm
 	default "armeb"	if BR2_armeb
@@ -161,3 +342,17 @@ config BR2_GCC_TARGET_ARCH
 config BR2_GCC_TARGET_ABI
 	default "apcs-gnu"	if BR2_ARM_OABI
 	default "aapcs-linux"	if BR2_ARM_EABI
+
+config BR2_GCC_TARGET_FPU
+	default "vfp"		if BR2_ARM_FPU_VFPV2
+	default "vfpv3"		if BR2_ARM_FPU_VFPV3
+	default "vfpv3-d16" 	if BR2_ARM_FPU_VFPV3D16
+	default "vfpv4" 	if BR2_ARM_FPU_VFPV4
+	default "vfpv4-d16" 	if BR2_ARM_FPU_VFPV4D16
+	default "neon" 		if BR2_ARM_FPU_NEON
+	default "neon-vfpv4" 	if BR2_ARM_FPU_NEON_VFPV4
+
+config BR2_GCC_TARGET_FLOAT_ABI
+	default "soft"		if BR2_ARM_SOFT_FLOAT
+	default "softfp"	if !BR2_ARM_SOFT_FLOAT && BR2_ARM_EABI
+	default "hard"		if !BR2_ARM_SOFT_FLOAT && BR2_ARM_EABIHF
diff --git a/package/Makefile.in b/package/Makefile.in
index 405b3e4..f570d67 100644
--- a/package/Makefile.in
+++ b/package/Makefile.in
@@ -40,6 +40,12 @@ ABI=gnueabi
 else
 ABI=eabi
 endif
+else ifeq ($(BR2_ARM_EABIHF),y)
+ifeq ($(LIBC),uclibc)
+ABI=gnueabihf
+else
+ABI=eabihf
+endif
 endif
 
 # For FSL PowerPC there's SPE
-- 
1.8.1.2

[Buildroot] [PATCH 4/8] binutils: exclude binutils versions that don't support EABIhf

[Thomas Petazzoni]

The ARM EABIhf support was introduced in Binutils 2.22, so earlier
versions should not be selected when EABIhf is used.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/binutils/Config.in.host | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/package/binutils/Config.in.host b/package/binutils/Config.in.host
index 3d687fe..9ddaed7 100644
--- a/package/binutils/Config.in.host
+++ b/package/binutils/Config.in.host
@@ -14,14 +14,17 @@ choice
 
 	config BR2_BINUTILS_VERSION_2_20_1
 		depends on !BR2_avr32
+		depends on !BR2_ARM_EABIHF
 		bool "binutils 2.20.1"
 
 	config BR2_BINUTILS_VERSION_2_21
 		depends on !BR2_avr32
+		depends on !BR2_ARM_EABIHF
 		bool "binutils 2.21"
 
 	config BR2_BINUTILS_VERSION_2_21_1
 		depends on !BR2_avr32
+		depends on !BR2_ARM_EABIHF
 		bool "binutils 2.21.1"
 
 	config BR2_BINUTILS_VERSION_2_22
-- 
1.8.1.2

[Buildroot] [PATCH 5/8] gcc: take into account ARM floating point capabilities

[Thomas Petazzoni]

The ARM EABIhf support was introduced in gcc 4.6.x, so versions
earlier than this one should not be used when EABIhf is selected.

The ARM VFPv4 support was introduced in gcc 4.5.x, so versions earlier
than this one should not be used when VFPv4 is used.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/gcc/Config.in.host | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/package/gcc/Config.in.host b/package/gcc/Config.in.host
index 1b19128..0a893b2 100644
--- a/package/gcc/Config.in.host
+++ b/package/gcc/Config.in.host
@@ -22,15 +22,22 @@ choice
 
 	config BR2_GCC_VERSION_4_3_X
 		depends on !BR2_arc && !BR2_avr32 && !BR2_sparc_sparchfleon && !BR2_sparc_sparchfleonv8 && !BR2_sparc_sparcsfleon && !BR2_sparc_sparcsfleonv8 && !BR2_cortex_a5 && !BR2_cortex_a7 && !BR2_cortex_a8 && !BR2_cortex_a9 && !BR2_cortex_a15 && !BR2_x86_atom && !BR2_powerpc_e300c2 && !BR2_powerpc_e300c3 && !BR2_powerpc_e500mc && !BR2_powerpc_464 && !BR2_powerpc_464fp && !BR2_powerpc_476 && !BR2_powerpc_476fp && !BR2_fa526 && !BR2_pj4
+		depends on !BR2_ARM_EABIHF
 		bool "gcc 4.3.x"
 
 	config BR2_GCC_VERSION_4_4_X
 		depends on !BR2_arc && !BR2_avr32 && !BR2_cortex_a5 && !BR2_cortex_a7 && !BR2_cortex_a15 && !BR2_x86_atom && !BR2_powerpc_476 && !BR2_powerpc_476fp && !BR2_fa526 && !BR2_pj4
 		bool "gcc 4.4.x"
+		# ARM EABIhf support appeared in gcc 4.6
+		depends on !BR2_ARM_EABIHF
+		# VFPv4 support appeared in gcc 4.5
+		depends on !BR2_ARM_FPU_VFPV4 && !BR2_ARM_FPU_VFPV4D16
 
 	config BR2_GCC_VERSION_4_5_X
 		depends on !BR2_arc && !BR2_avr32 && !BR2_cortex_a7 && !BR2_cortex_a15 && !BR2_sparc_sparchfleon && !BR2_sparc_sparchfleonv8 && !BR2_sparc_sparcsfleon && !BR2_sparc_sparcsfleonv8 && !BR2_fa526 && !BR2_pj4
 		select BR2_GCC_NEEDS_MPC
+		# ARM EABIhf support appeared in gcc 4.6
+		depends on !BR2_ARM_EABIHF
 		bool "gcc 4.5.x"
 
 	config BR2_GCC_VERSION_4_6_X
@@ -40,17 +47,20 @@ choice
 
 	config BR2_GCC_VERSION_4_7_X
 		depends on !BR2_arc && !BR2_avr32 && !BR2_sparc_sparchfleon && !BR2_sparc_sparchfleonv8 && !BR2_sparc_sparcsfleon && !BR2_sparc_sparcsfleonv8 && !BR2_pj4
+		depends on !BR2_ARM_OABI
 		select BR2_GCC_NEEDS_MPC
 		bool "gcc 4.7.x"
 
 	config BR2_GCC_VERSION_4_8_X
 		depends on !BR2_arc && !BR2_avr32 && !BR2_sparc_sparchfleon && !BR2_sparc_sparchfleonv8 && !BR2_sparc_sparcsfleon && !BR2_sparc_sparcsfleonv8
 		select BR2_GCC_NEEDS_MPC
+		depends on !BR2_ARM_OABI
 		bool "gcc 4.8.x"
 
 	config BR2_GCC_VERSION_SNAP
 		depends on !BR2_arc && !BR2_avr32 && !BR2_sparc_sparchfleon && !BR2_sparc_sparchfleonv8 && !BR2_sparc_sparcsfleon && !BR2_sparc_sparcsfleonv8
 		select BR2_GCC_NEEDS_MPC
+		depends on !BR2_ARM_OABI
 		bool "gcc snapshot"
 endchoice
 
-- 
1.8.1.2

[Buildroot] [PATCH 6/8] toolchain: remove the old BR2_VFP_FLOAT option

[Thomas Petazzoni]

Now that we have a much better way of selecting between the various
VFP versions and capabilities, the BR2_VFP_FLOAT version no longer
makes sense. This commit gets rid of it, and adds the appropriate
Config.in.legacy code.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 Config.in.legacy                                     |  9 +++++++++
 toolchain/toolchain-buildroot/Config.in.2            | 12 ------------
 toolchain/toolchain-external/ext-tool.mk             |  5 -----
 toolchain/toolchain-external/ext-toolchain-wrapper.c |  3 ---
 4 files changed, 9 insertions(+), 20 deletions(-)

diff --git a/Config.in.legacy b/Config.in.legacy
index 01bf900..b92b656 100644
--- a/Config.in.legacy
+++ b/Config.in.legacy
@@ -76,6 +76,15 @@ config BR2_ELF2FLT
 	  BR2_PACKAGE_HOST_ELF2FLT due to the conversion of elf2flt to
 	  the package infrastructure.
 
+config BR2_VFP_FLOAT
+	bool "the ARM VFP floating point option has been renamed"
+	select BR2_LEGACY
+	help
+	  Due to a major refactoring of the floating-point handling of
+	  the ARM architecture support, the BR2_VFP_FLOAT option has
+	  been replaced with a choice of options that allows to select
+	  between various VFP versions/capabilities.
+
 ###############################################################################
 comment "Legacy options removed in 2013.05"
 
diff --git a/toolchain/toolchain-buildroot/Config.in.2 b/toolchain/toolchain-buildroot/Config.in.2
index 9bbf016..5a191f3 100644
--- a/toolchain/toolchain-buildroot/Config.in.2
+++ b/toolchain/toolchain-buildroot/Config.in.2
@@ -23,16 +23,4 @@ config BR2_TOOLCHAIN_BUILDROOT_USE_SSP
 
 source "package/elf2flt/Config.in.host"
 
-config BR2_VFP_FLOAT
-	bool "Use ARM Vector Floating Point unit"
-	depends on !BR2_SOFT_FLOAT
-	depends on BR2_arm || BR2_armeb
-	help
-	  Setting this option will enable the "-mfpu=vfp" option.
-	  If your ARM CPU has a Vector Floating Point Unit (VFP)
-	  and the toolchain supports the option, then the
-	  code can be optimized.
-
-	  Most people will answer N.
-
 endif
diff --git a/toolchain/toolchain-external/ext-tool.mk b/toolchain/toolchain-external/ext-tool.mk
index 5c51e06..98a4c59 100644
--- a/toolchain/toolchain-external/ext-tool.mk
+++ b/toolchain/toolchain-external/ext-tool.mk
@@ -196,11 +196,6 @@ TOOLCHAIN_EXTERNAL_CFLAGS += -msoft-float
 TOOLCHAIN_EXTERNAL_WRAPPER_ARGS += -DBR_SOFTFLOAT=1
 endif
 
-ifeq ($(BR2_VFP_FLOAT),y)
-TOOLCHAIN_EXTERNAL_CFLAGS += -mfpu=vfp
-TOOLCHAIN_EXTERNAL_WRAPPER_ARGS += -DBR_VFPFLOAT=1
-endif
-
 ifeq ($(BR2_TOOLCHAIN_EXTERNAL_DOWNLOAD),y)
 TOOLCHAIN_EXTERNAL_DEPENDENCIES = $(TOOLCHAIN_EXTERNAL_DIR)/.extracted
 endif
diff --git a/toolchain/toolchain-external/ext-toolchain-wrapper.c b/toolchain/toolchain-external/ext-toolchain-wrapper.c
index e504ff2..d3319e9 100644
--- a/toolchain/toolchain-external/ext-toolchain-wrapper.c
+++ b/toolchain/toolchain-external/ext-toolchain-wrapper.c
@@ -47,9 +47,6 @@ static char *predef_args[] = {
 #ifdef BR_SOFTFLOAT
 	"-msoft-float",
 #endif /* BR_SOFTFLOAT */
-#ifdef BR_VFPFLOAT
-	"-mfpu=vfp",
-#endif /* BR_VFPFLOAT */
 #ifdef BR_64
 	"-m64",
 #endif
-- 
1.8.1.2

[Buildroot] [PATCH 7/8] configs: update defconfigs after VFP option changes

[Thomas Petazzoni]

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 configs/calao_snowball_defconfig    | 3 +--
 configs/mx53loco_defconfig          | 3 +--
 configs/qemu_arm_nuri_defconfig     | 3 +--
 configs/qemu_arm_vexpress_defconfig | 3 +--
 4 files changed, 4 insertions(+), 8 deletions(-)

diff --git a/configs/calao_snowball_defconfig b/configs/calao_snowball_defconfig
index 360ed51..fd9e1b6 100644
--- a/configs/calao_snowball_defconfig
+++ b/configs/calao_snowball_defconfig
@@ -1,10 +1,9 @@
 BR2_arm=y
 BR2_cortex_a9=y
 BR2_GCC_VERSION_4_7_X=y
-# BR2_SOFT_FLOAT is not set
+BR2_ARM_FPU_VFPV3D16=y
 BR2_TOOLCHAIN_BUILDROOT_WCHAR=y
 BR2_TOOLCHAIN_BUILDROOT_CXX=y
-BR2_VFP_FLOAT=y
 BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_MDEV=y
 BR2_TARGET_GENERIC_GETTY_PORT="ttyAMA2"
 BR2_PACKAGE_BUSYBOX_SHOW_OTHERS=y
diff --git a/configs/mx53loco_defconfig b/configs/mx53loco_defconfig
index ff3948b..08de3da 100644
--- a/configs/mx53loco_defconfig
+++ b/configs/mx53loco_defconfig
@@ -1,8 +1,7 @@
 # Architecture
 BR2_arm=y
 BR2_cortex_a8=y
-# BR2_SOFT_FLOAT is not set
-BR2_VFP_FLOAT=y
+BR2_ARM_FPU_VFPV3D16=y
 
 # System
 BR2_TARGET_GENERIC_GETTY_PORT="ttymxc0"
diff --git a/configs/qemu_arm_nuri_defconfig b/configs/qemu_arm_nuri_defconfig
index 27dac27..57c0470 100644
--- a/configs/qemu_arm_nuri_defconfig
+++ b/configs/qemu_arm_nuri_defconfig
@@ -2,8 +2,7 @@
 BR2_arm=y
 BR2_cortex_a9=y
 BR2_ARM_ENABLE_NEON=y
-BR2_VFP_FLOAT=y
-# BR2_SOFT_FLOAT is not set
+BR2_ARM_FPU_VFPV3D16=y
 
 # System
 BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_MDEV=y
diff --git a/configs/qemu_arm_vexpress_defconfig b/configs/qemu_arm_vexpress_defconfig
index 07b629e..4184da4 100644
--- a/configs/qemu_arm_vexpress_defconfig
+++ b/configs/qemu_arm_vexpress_defconfig
@@ -2,8 +2,7 @@
 BR2_arm=y
 BR2_cortex_a9=y
 BR2_ARM_ENABLE_NEON=y
-BR2_VFP_FLOAT=y
-# BR2_SOFT_FLOAT is not set
+BR2_ARM_FPU_VFPV3D16=y
 
 # System
 BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_MDEV=y
-- 
1.8.1.2

[Buildroot] [PATCH 8/8] toolchain-external: update config options after EABIhf introduction

[Thomas Petazzoni]

The introduction of the EABIhf ABI requires a few updates to the
configuration options for external toolchains, in order to ensure that
the user doesn't do any invalid selection. In detail:

 * The Linaro ARM toolchains now depend on BR2_ARM_EABIHF, because
   that's the ABI they use, and it is incompatible with EABI. The
   comment about the availability of Linaro toolchains is updated to
   inform users selecting EABI that they should select EABIhf if they
   want to see Linaro toolchains.

 * The Sourcery CodeBench toolchains now depend on BR2_ARM_EABI,
   because that's the ABI they use. A comment is added to inform users
   that have selected EABIhf that Sourcery CodeBench are only
   available when EABI is used.

 * The Arago toolchains now depend on BR2_ARM_EABI, because that's the
   ABI they use. The description of the ARMv7 Arago toolchain is also
   slightly improved.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 toolchain/toolchain-external/Config.in | 24 ++++++++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)

diff --git a/toolchain/toolchain-external/Config.in b/toolchain/toolchain-external/Config.in
index 946063f..6fd8f1f 100644
--- a/toolchain/toolchain-external/Config.in
+++ b/toolchain/toolchain-external/Config.in
@@ -8,6 +8,7 @@ config BR2_TOOLCHAIN_EXTERNAL_LINARO_2013_06
 	depends on BR2_arm
 	depends on BR2_GCC_TARGET_ARCH = "armv7-a"
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
+	depends on BR2_ARM_EABIHF
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -27,6 +28,7 @@ config BR2_TOOLCHAIN_EXTERNAL_LINARO_2013_05
 	depends on BR2_arm
 	depends on BR2_GCC_TARGET_ARCH = "armv7-a"
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
+	depends on BR2_ARM_EABIHF
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -46,6 +48,7 @@ config BR2_TOOLCHAIN_EXTERNAL_LINARO_2013_04
 	depends on BR2_arm
 	depends on BR2_GCC_TARGET_ARCH = "armv7-a"
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
+	depends on BR2_ARM_EABIHF
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -60,14 +63,16 @@ config BR2_TOOLCHAIN_EXTERNAL_LINARO_2013_04
 
 	  To use this toolchain, you must disable soft float usage.
 
-comment "Linaro toolchains available for Cortex-A{5,8,9,15}"
+comment "Linaro toolchains available for Cortex-A{5,7,8,9,15} and the EABIhf ABI"
 	depends on BR2_arm
 	depends on BR2_GCC_TARGET_ARCH != "armv7-a"
+	depends on !BR2_ARM_EABIHF
 
 config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_ARM201305
 	bool "Sourcery CodeBench ARM 2013.05"
 	depends on BR2_arm
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
+	depends on BR2_ARM_EABI
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -92,6 +97,7 @@ config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_ARM201203
 	bool "Sourcery CodeBench ARM 2012.03"
 	depends on BR2_arm
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
+	depends on BR2_ARM_EABI
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -116,6 +122,7 @@ config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_ARM201109
 	bool "Sourcery CodeBench ARM 2011.09"
 	depends on BR2_arm
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
+	depends on BR2_ARM_EABI
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -136,11 +143,17 @@ config BR2_TOOLCHAIN_EXTERNAL_CODESOURCERY_ARM201109
 	      Select BR2_SOFT_FLOAT
 	      Set BR2_TARGET_OPTIMIZATION to -mthumb
 
+comment "Sourcery CodeBench toolchains available for the EABI ABI"
+	depends on BR2_arm
+	depends on !BR2_ARM_EABI
+
 config BR2_TOOLCHAIN_EXTERNAL_ARAGO_ARMV7A_201109
 	bool "Arago ARMv7 2011.09"
 	depends on BR2_arm
 	depends on BR2_HOSTARCH = "x86_64" || BR2_HOSTARCH = "x86"
 	depends on BR2_GCC_TARGET_ARCH = "armv7-a"
+	depends on BR2_ARM_EABI
+	depends on BR2_ARM_CPU_HAS_VFPV3 || BR2_ARM_CPU_MAYBE_HAS_VFPV3
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
@@ -149,9 +162,11 @@ config BR2_TOOLCHAIN_EXTERNAL_ARAGO_ARMV7A_201109
 	  Texas Instruments Arago 2011.09 toolchain, with gcc 4.5.3,
 	  binutils 2.20.1, glibc 2.12, gdb 7.2.
 
-	  This toolchain uses -mfloat-abi=softfp (i.e use FPU
-	  instructions, but pass floating point function arguments in
-	  non-floating point registers).
+	  This toolchain uses -mfloat-abi=softfp (i.e can use FPU
+	  instructions, but passes floating point function arguments
+	  in integer registers), and requires a VFPv3 floating point
+	  unit to work properly. This unit is available on most
+	  Cortex-A ARM processors, but not all.
 
 config BR2_TOOLCHAIN_EXTERNAL_ARAGO_ARMV5TE_201109
 	bool "Arago ARMv5 2011.09"
@@ -160,6 +175,7 @@ config BR2_TOOLCHAIN_EXTERNAL_ARAGO_ARMV5TE_201109
 	depends on BR2_GCC_TARGET_ARCH != "armv4t" && \
 		BR2_GCC_TARGET_ARCH != "armv4" && \
 		BR2_GCC_TARGET_ARCH != "armv5t"
+	depends on BR2_ARM_EABI
 	select BR2_TOOLCHAIN_EXTERNAL_GLIBC
 	select BR2_TOOLCHAIN_HAS_NATIVE_RPC
 	select BR2_INSTALL_LIBSTDCPP
-- 
1.8.1.2

[Buildroot] [PATCH 2/8] arch: Refactor BR2_SOFT_FLOAT into per-architecture options

[Peter Korsgaard]

>>>>> "Thomas" == Thomas Petazzoni <thomas.petazzoni@free-electrons.com> writes:

 Thomas> As we are going to introduced a more advanced support of floating
 Thomas> point options for the ARM architecture, we need to adjust how the
 Thomas> soft-float option is handled. We replace the current hidden option
 Thomas> BR2_PREFER_SOFT_FLOAT option and the visible BR2_SOFT_FLOAT option by:

 Thomas>  * A global hidden BR2_SOFT_FLOAT option, defined in arch/Config.in,
 Thomas>    that tells whether the architecture-specific code is using software
 Thomas>    emulated floating point. This hidden option can be used throughout
 Thomas>    Buildroot to determine whether soft float is used or not.

 Thomas>  * Per-architecture visible BR2_<arch>_SOFT_FLOAT options, for the
 Thomas>    architecture for which it makes sense, which allows users to select
 Thomas>    soft float emulation when needed.

 Thomas> This change will allow each architecture to have a different way of
 Thomas> presenting its floating point capabilities.

 Thomas> diff --git a/arch/Config.in.avr32 b/arch/Config.in.avr32
 Thomas> index ebf8454..34055d9 100644
 Thomas> --- a/arch/Config.in.avr32
 Thomas> +++ b/arch/Config.in.avr32
 Thomas> @@ -3,3 +3,13 @@ config BR2_ARCH
 
 Thomas>  config BR2_ENDIAN
 Thomas>         default "BIG"
 Thomas> +
 Thomas> +config BR2_AVR32_SOFT_FLOAT
 Thomas> +	bool "Use soft-float"
 Thomas> +	default y
 Thomas> +	select BR2_SOFT_FLOAT
 Thomas> +	help
 Thomas> +	  If your target CPU does not have a Floating Point Unit (FPU)
 Thomas> +	  or a kernel FPU emulator, but you still wish to support
 Thomas> +	  floating point functions, then everything will need to be
 Thomas> +	  compiled with soft floating point support (-msoft-float).

Does avr32 SoCs with hard float even exist? Does
uClibc/kernel/.. support it? If not, we could just make this a hidden
option.
 
 Thomas> +config BR2_POWERPC_SOFT_FLOAT
 Thomas> +	bool "Use soft-float"
 Thomas> +	default y
 Thomas> +	select BR2_SOFT_FLOAT
 Thomas> +	help
 Thomas> +	  If your target CPU does not have a Floating Point Unit (FPU)
 Thomas> +	  or a kernel FPU emulator, but you still wish to support
 Thomas> +	  floating point functions, then everything will need to be
 Thomas> +	  compiled with soft floating point support (-msoft-float).

We didn't use to default to soft float for powerpc, as most cores have a
FPU.
 
 Thomas> -config BR2_PREFER_SOFT_FLOAT
 Thomas> -	bool
 Thomas> -	default y if BR2_arm
 Thomas> -	default y if BR2_armeb
 Thomas> -	default y if BR2_avr32
 Thomas> -	default y if BR2_mips
 Thomas> -	default y if BR2_mipsel

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 3/8] arch: improve ARM floating point support and add support for EABIhf

[Peter Korsgaard]

>>>>> "Thomas" == Thomas Petazzoni <thomas.petazzoni@free-electrons.com> writes:

 Thomas> This commit introduces the support for the EABIhf ABI, next to the
 Thomas> existing support we have for EABI and OABI (even though OABI support
 Thomas> is deprecated). EABIhf allows to improve performance of floating point
 Thomas> workload by using floating point registers to transfer floating point
 Thomas> arguments when calling functions, instead of using integer registers
 Thomas> to do, as is done in the 'softfp' floating point model of EABI.

 Thomas> In addition to this, this commit introduces a list of options for the
 Thomas> floating point support:
 Thomas>  * Software floating point
 Thomas>  * VFP
 Thomas>  * VFPv3
 Thomas>  * VFPv3-D16
 Thomas>  * VFPv4
 Thomas>  * VFPv4-D16

 Thomas> and it introduces some logic to make sure the options are only visible
 Thomas> when it makes sense, depending on the ARM core being selected. This is
 Thomas> however made complicated by the fact that certain VFP capabilities are
 Thomas> mandatory on some cores, but optional on some other cores. The kconfig
 Thomas> logic tries to achieve the following goals:

 Thomas>  * Hide options that are definitely not possible.

 Thomas>  * Use safe default values (i.e for Cortex-A5 and A7, the presence of
 Thomas>    the VFPv4 unit is optional, so we default on software floating
 Thomas>    point on these cores)..

 Thomas>  * Show the available possibilities, even if some of them are not
 Thomas>    necessarily working on a particular core (again, for the Cortex-A5
 Thomas>    and A7 cores, there is no way of knowing whether the particular
 Thomas>    variant used by the user has VFPv4 or not, so we select software
 Thomas>    floating point by default, but still show VFP/VFPv3/VFPv4 options).

 Thomas> It is worth noting that this commit doesn't add support for all
 Thomas> possible -mfpu= values on ARM. We haven't added support for fpa, fpe2,
 Thomas> fpe3, maverick (those four are only used on very old ARM cores), for
 Thomas> vfpv3-fp16, vfpv3-d16-fp16, vfpv3xd, vfpv3xd-fp16, neon-fp16,
 Thomas> vfpv4-sp-d16. They can be added quite easily if needed thanks to the
 Thomas> new organization of the Config.in options.

 Thomas> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>


 Thomas>  config BR2_cortex_a15
 Thomas>  	bool "cortex-A15"
 Thomas>  	select BR2_ARM_CPU_HAS_NEON
 Thomas> +	select BR2_ARM_CPU_HAS_VFPV4
 Thomas>  config BR2_fa526
 Thomas>  	bool "fa526/626"
 Thomas>  config BR2_pj4
 Thomas>  	bool "pj4"
 Thomas> +	select BR2_ARM_CPU_HAS_VFPV3
 Thomas>  config BR2_strongarm
 Thomas>  	bool "strongarm sa110/sa1100"
 Thomas>  config BR2_xscale
 Thomas> @@ -67,34 +101,67 @@ config BR2_arm1136jf_s
 Thomas>  choice
 Thomas>  	prompt "Target ABI"
 Thomas>  	depends on BR2_arm || BR2_armeb
 Thomas> -	depends on BR2_DEPRECATED
 Thomas>  	default BR2_ARM_EABI

Looks good. You could argue that we should do:

      default BR2_ARM_EABIHF

instead (E.G. default to eabihf if available, otherwise fall back to
eabi), but with all the optional vfp units it's perhaps better to play
it safe..

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 0/8] ARM floating point improvements and EABIhf support

[Peter Korsgaard]

>>>>> "Thomas" == Thomas Petazzoni <thomas.petazzoni@free-electrons.com> writes:

 Thomas> Hello,
 Thomas> Here is a patch series that improves the support for the various
 Thomas> floating point possibilities on ARM, and adds support for the new
 Thomas> EABIhf ARM ABI. Until now, we only had a global BR2_SOFT_FLOAT option
 Thomas> to enable/disable soft-float support, and on ARM a BR2_VFP_FLOAT
 Thomas> option to enable -mfpu=vfp. This clearly isn't sufficient to take into
 Thomas> the various possibilities offered by the Cortex-A ARM cores.

Series looks good (+/- the minor issues I mentioned), but I will wait
until you send an update with the oabi stuff from Yann integrated.

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 2/8] arch: Refactor BR2_SOFT_FLOAT into per-architecture options

[Thomas Petazzoni]

Dear Peter Korsgaard,

On Sun, 07 Jul 2013 22:24:05 +0200, Peter Korsgaard wrote:

> Does avr32 SoCs with hard float even exist? Does
> uClibc/kernel/.. support it? If not, we could just make this a hidden
> option.

According to http://en.wikipedia.org/wiki/AVR32, the AVR32 AP7 never
had a floating point unit, only the UC3, but I believe Linux was only
running on AP7 cores, but I'm not sure. Anyway, I'll make AVR32
soft-float only.

>  Thomas> +config BR2_POWERPC_SOFT_FLOAT
>  Thomas> +	bool "Use soft-float"
>  Thomas> +	default y
>  Thomas> +	select BR2_SOFT_FLOAT
>  Thomas> +	help
>  Thomas> +	  If your target CPU does not have a Floating Point Unit (FPU)
>  Thomas> +	  or a kernel FPU emulator, but you still wish to support
>  Thomas> +	  floating point functions, then everything will need to be
>  Thomas> +	  compiled with soft floating point support (-msoft-float).
> 
> We didn't use to default to soft float for powerpc, as most cores have a
> FPU.

Correct, will fix. As a followup to this patch set, it'd be nice if
someone knowledgeable in PPC stuff did a similar handling of FPU stuff.

Thanks,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] [PATCH 3/8] arch: improve ARM floating point support and add support for EABIhf

[Thomas Petazzoni]

Dear Peter Korsgaard,

On Sun, 07 Jul 2013 22:44:08 +0200, Peter Korsgaard wrote:

> Looks good. You could argue that we should do:
> 
>       default BR2_ARM_EABIHF
> 
> instead (E.G. default to eabihf if available, otherwise fall back to
> eabi), but with all the optional vfp units it's perhaps better to play
> it safe..

Yeah, I thought that defaulting on eabihf today (for ARMv7) was a bit
too much. I played conservative, and remained on eabi by default.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com