crosstool-ng/config/target.in
Cody Schafer 13fcbed86e arch: add option to let gccdecide what floating point to use
Add an ARCH_FLOAT_AUTO option to avoid passing float options
to gcc and allow it to choose

Signed-off-by: Cody P Schafer <dev@codyps.com>
[yann.morin.1998@free.fr: slightly reword the commit log]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Message-Id: <088489641f4790262685.1399687813@localhost>
Patchwork-Id: 347584
2014-05-09 19:10:08 -07:00

384 lines
10 KiB
Plaintext

# Target definition: architecture, optimisations, etc...
menu "Target options"
config ARCH
string
# Pre-declare target optimisation variables
config ARCH_SUPPORTS_BOTH_MMU
config ARCH_SUPPORTS_BOTH_ENDIAN
config ARCH_SUPPORTS_32
config ARCH_SUPPORTS_64
config ARCH_SUPPORTS_WITH_ARCH
config ARCH_SUPPORTS_WITH_ABI
config ARCH_SUPPORTS_WITH_CPU
config ARCH_SUPPORTS_WITH_TUNE
config ARCH_SUPPORTS_WITH_FLOAT
config ARCH_SUPPORTS_WITH_FPU
config ARCH_SUPPORTS_SOFTFP
config ARCH_DEFAULT_HAS_MMU
config ARCH_DEFAULT_BE
config ARCH_DEFAULT_LE
config ARCH_DEFAULT_32
config ARCH_DEFAULT_64
config ARCH_ARCH
config ARCH_ABI
config ARCH_CPU
config ARCH_TUNE
config ARCH_FPU
config ARCH_BE
config ARCH_LE
config ARCH_32
config ARCH_64
config ARCH_BITNESS
config ARCH_FLOAT_HW
config ARCH_FLOAT_SW
config TARGET_CFLAGS
config TARGET_LDFLAGS
source "config.gen/arch.in"
config ARCH_SUFFIX
string
prompt "Suffix to the arch-part"
help
Some architectures have multiple variants and being able to specify
the variant instead of the arch is quite convenient. This is commonly
seen for instance when "armv5tel-" is used as a prefix instead of the
more generic "arm-", or with "alphaev6-" instead of "alpha-".
Whatever you enter here will be appended to the architecture-part of the
tuple, just before the first '-'. It will override any architecture-
specific suffix that crosstool-NG may compute.
If you are not sure about what this is, leave it blank.
#--------------------------------------
comment "Generic target options"
#--------------------------------------
config MULTILIB
bool
prompt "Build a multilib toolchain (READ HELP!!!)"
help
If you say 'y' here, then the toolchain will also contain the C library
optimised for some variants of the selected architecture, besides the
default settings.
This means the build time of the C library will be in O(nb_variants).
The list of variants is dependent on the architecture, and is hard-coded
in gcc, so it is not possible to say what variants to support, only
whether hard-coded variants should be supported or not.
NOTE: The multilib feature in crosstool-NG is not well-tested.
Use at your own risk, and report success and/or failure.
#--------------------------------------
config ARCH_SUPPORTS_BOTH_MMU
bool
config ARCH_DEFAULT_HAS_MMU
bool
config ARCH_USE_MMU
bool
prompt "Use the MMU" if ARCH_SUPPORTS_BOTH_MMU
default y if ARCH_DEFAULT_HAS_MMU
help
If your architecture has an MMU and you want to use it,
say 'Y' here.
OTOH, if you don't want to use the MMU, or your arch
lacks an MMU, say 'N' here.
Note that some architectures (eg. ARM) has variants that
lacks an MMU (eg. ARM Cortex-M3), while other variants
have one (eg. ARM Cortex-A8).
#--------------------------------------
config ARCH_SUPPORTS_BOTH_ENDIAN
bool
config ARCH_DEFAULT_BE
bool
config ARCH_DEFAULT_LE
bool
choice
bool
prompt "Endianness:"
depends on ARCH_SUPPORTS_BOTH_ENDIAN
default ARCH_BE if ARCH_DEFAULT_BE
default ARCH_LE if ARCH_DEFAULT_LE
config ARCH_BE
bool
prompt "Big endian"
config ARCH_LE
bool
prompt "Little endian"
endchoice
config ARCH_ENDIAN
string
depends on ARCH_SUPPORTS_BOTH_ENDIAN
default "big" if ARCH_BE
default "little" if ARCH_LE
#--------------------------------------
config ARCH_SUPPORTS_32
bool
config ARCH_SUPPORTS_64
bool
config ARCH_DEFAULT_32
bool
config ARCH_DEFAULT_64
bool
config ARCH_BITNESS
int
default "32" if ARCH_32
default "64" if ARCH_64
choice
bool
prompt "Bitness:"
default ARCH_32 if ARCH_DEFAULT_32
default ARCH_64 if ARCH_DEFAULT_64
config ARCH_32
bool
prompt "32-bit"
depends on ARCH_SUPPORTS_32
config ARCH_64
bool
prompt "64-bit"
depends on ARCH_SUPPORTS_64
endchoice
#--------------------------------------
comment "Target optimisations"
config ARCH_SUPPORTS_WITH_ARCH
bool
config ARCH_SUPPORTS_WITH_ABI
bool
config ARCH_SUPPORTS_WITH_CPU
bool
config ARCH_SUPPORTS_WITH_TUNE
bool
config ARCH_SUPPORTS_WITH_FLOAT
bool
config ARCH_SUPPORTS_WITH_FPU
bool
config ARCH_SUPPORTS_SOFTFP
bool
config ARCH_ARCH
string
prompt "Architecture level"
depends on ARCH_SUPPORTS_WITH_ARCH
default ""
help
GCC uses this name to determine what kind of instructions it can emit
when generating assembly code. This option can be used in conjunction
with or instead of the ARCH_CPU option (above), or a (command-line)
-mcpu= option.
This is the configuration flag --with-arch=XXXX, and the runtime flag
-march=XXX.
Pick a value from the gcc manual for your choosen gcc version and your
target CPU.
Leave blank if you don't know, or if your target architecture does not
offer this option.
config ARCH_ABI
string
prompt "Generate code for the specific ABI"
depends on ARCH_SUPPORTS_WITH_ABI
default ""
help
Generate code for the given ABI.
This is the configuration flag --with-abi=XXXX, and the runtime flag
-mabi=XXX.
Pick a value from the gcc manual for your choosen gcc version and your
target CPU.
Leave blank if you don't know, or if your target architecture does not
offer this option.
config ARCH_CPU
string
prompt "Emit assembly for CPU"
depends on ARCH_SUPPORTS_WITH_CPU
default ""
help
This specifies the name of the target processor. GCC uses this name
to determine what kind of instructions it can emit when generating
assembly code.
This is the configuration flag --with-cpu=XXXX, and the runtime flag
-mcpu=XXX.
Pick a value from the gcc manual for your choosen gcc version and your
target CPU.
Leave blank if you don't know, or if your target architecture does not
offer this option.
config ARCH_TUNE
string
prompt "Tune for CPU"
depends on ARCH_SUPPORTS_WITH_TUNE
default ""
help
This option is very similar to the ARCH_CPU option (above), except
that instead of specifying the actual target processor type, and hence
restricting which instructions can be used, it specifies that GCC should
tune the performance of the code as if the target were of the type
specified in this option, but still choosing the instructions that it
will generate based on the cpu specified by the ARCH_CPU option
(above), or a (command-line) -mcpu= option.
This is the configuration flag --with-tune=XXXX, and the runtime flag
-mtune=XXX.
Pick a value from the gcc manual for your choosen gcc version and your
target CPU.
Leave blank if you don't know, or if your target architecture does not
offer this option.
config ARCH_FPU
string
prompt "Use specific FPU"
depends on ARCH_SUPPORTS_WITH_FPU
default ""
help
On some targets (eg. ARM), you can specify the kind of FPU to emit
code for.
This is the configuration flag --with-fpu=XXX, and the runtime flag
-mfpu=XXX.
See below wether to actually emit FP opcodes, or to emulate them.
Pick a value from the gcc manual for your choosen gcc version and your
target CPU.
Leave blank if you don't know, or if your target architecture does not
offer this option.
choice
bool
prompt "Floating point:"
depends on ARCH_SUPPORTS_WITH_FLOAT
config ARCH_FLOAT_AUTO
bool
prompt "auto (let gcc decide)"
help
Instead of explicitly passing a float option, don't
pass any float options and let gcc figure it out.
For multilib configurations, this may help.
config ARCH_FLOAT_HW
bool
prompt "hardware (FPU)"
help
Emit hardware floating point opcodes.
If you've got a processor with a FPU, then you want that.
If your hardware has no FPU, you still can use HW floating point, but
need to compile support for FPU emulation in your kernel. Needless to
say that emulating the FPU is /slooowwwww/...
One situation you'd want HW floating point without a FPU is if you get
binary blobs from different vendors that are compiling this way and
can't (don't wan't to) change.
config ARCH_FLOAT_SOFTFP
bool
prompt "softfp (FPU)"
depends on ARCH_SUPPORTS_SOFTFP
help
Emit hardware floating point opcodes but use the software
floating point calling convention.
Architectures such as ARM use different registers for passing
floating point values depending on if they're in software mode
or hardware mode. softfp emits FPU instructions but uses the
software FP calling convention allowing softfp code to
interoperate with legacy software only code.
If in doubt, use 'software' or 'hardware' mode instead.
config ARCH_FLOAT_SW
bool
prompt "software (no FPU)"
help
Do not emit any hardware floating point opcode.
If your processor has no FPU, then you most probably want this, as it
is faster than emulating the FPU in the kernel.
endchoice
config TARGET_CFLAGS
string
prompt "Target CFLAGS"
default ""
help
Used to add specific options when compiling libraries of the toolchain,
that will run on the target (eg. libc.so).
Note that the options above for ARCH, ABI, CPU, TUNE and FPU will be
automatically used. You don't need to specify them here.
Leave blank if you don't know better.
config TARGET_LDFLAGS
string
prompt "Target LDFLAGS"
default ""
help
Used to add specific options when linking libraries of the toolchain,
that will run on your target.
Leave blank if you don't know better.
config ARCH_FLOAT
string
default "" if ! ARCH_SUPPORTS_WITH_FLOAT
default "auto" if ARCH_FLOAT_AUTO
default "hard" if ARCH_FLOAT_HW
default "soft" if ARCH_FLOAT_SW
default "softfp" if ARCH_FLOAT_SOFTFP
source "config.gen/arch.in.2"
endmenu