Half precision floating point numbers will be encoded using the IEEE 754
half precision floating point format - 16 bits in total, 1 for sign, 5
for exponent and 10 bits of mantissa.
This patch implements the float16 directive for both the IEEE 754 format
and the Arm alternative format for the Arm backend.
The syntax of the directive is:
.float16 <0-n decimal numbers>
e.g.
.float16 12.0
.float16 0.23, 433.1, 0.06
The Arm alternative format is almost identical to the IEEE 754 format,
except that it doesn't encode for NaNs or Infinity (instead an exponent
of 0x1F represents a normalized number in the range 65536 to 131008).
The alternative format is documented in the reference manual:
https://static.docs.arm.com/ddi0487/db/DDI0487D_b_armv8_arm.pdf?_ga=2.72318806.49764181.1561632697-999473562.1560847439
Which format is used is controlled by the .float16_format <format>
directive, where if <format> = ieee, then use the IEEE 754
half-precision format else if <format> = alternative, then use the
Arm alternative format
Or the format can be set on the command line via the -mfp16-format
option that has a similar syntax. -mfp16-format=<ieee|alternative>.
This also fixes the format and it cannot be changed by any directives.
Once the format has been set (either by the command line option or a directive) it cannot be changed,
and any attempts to change it (i.e. with the float16_format directive) will result in a warning and the
line being ignored.
For ELF targets the appropriate EABI attribute will be written out at the end of assembling
if the format has been explicitly specified. If no format has been explicitly specified then no
EABI attributes will be written.
If the format is not explicitly specified then any float16 directives are encoding using the IEEE 754-2008
format by default until the format is fixed or changed with the float16_format directive.
gas * config/tc-arm.c (enum fp_16bit_format): Add enum to represent the 2 float16 encodings.
(md_atof): Set precision for float16 type.
(arm_is_largest_exponent_ok): Check for whether to encode with the IEEE or alternative
format.
(set_fp16_format): Parse a float16_format directive.
(arm_parse_fp16_opt): Parse the fp16-format command line option.
(aeabi_set_public_attributes): For ELF encode the FP16 format EABI attribute.
* config/tc-arm.h (TC_LARGEST_EXPONENT_IS_NORMAL): Macro that expands to
arm_is_largest_exponent_ok.
(arm_is_largest_exponent_ok): Add prototype for arm_is_largest_exponent_ok function.
* doc/c-arm.texi: Add documentation for .float16, .float16_format and -mfp16-format=
* testsuite/gas/arm/float16-bad.d: New test.
* testsuite/gas/arm/float16-bad.l: New test.
* testsuite/gas/arm/float16-bad.s: New test.
* testsuite/gas/arm/float16-be.d: New test.
* testsuite/gas/arm/float16-format-bad.d: New test.
* testsuite/gas/arm/float16-format-bad.l: New test.
* testsuite/gas/arm/float16-format-bad.s: New test.
* testsuite/gas/arm/float16-format-opt-bad.d: New test.
* testsuite/gas/arm/float16-format-opt-bad.l: New test.
* testsuite/gas/arm/float16-le.d: New test.
* testsuite/gas/arm/float16.s: New test.
* testsuite/gas/arm/float16-eabi-alternative-format.d: New test.
* testsuite/gas/arm/float16-eabi-ieee-format.d: New test.
* testsuite/gas/arm/float16-eabi-no-format.d: New test.
* testsuite/gas/arm/float16-eabi.s: New test.
* config/atof-ieee.c (H_PRECISION): Macro for precision of float16
type.
(atof_ieee): Set precision and exponent bits for encoding float16
types.
(gen_to_words): NaN and Infinity encoding for float16.
(ieee_md_atof): Set precision for encoding float16 type.
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
README for GAS
A number of things have changed since version 1 and the wonderful
world of gas looks very different. There's still a lot of irrelevant
garbage lying around that will be cleaned up in time. Documentation
is scarce, as are logs of the changes made since the last gas release.
My apologies, and I'll try to get something useful.
Unpacking and Installation - Summary
====================================
See ../binutils/README.
To build just the assembler, make the target all-gas.
Documentation
=============
The GAS release includes texinfo source for its manual, which can be processed
into `info' or `dvi' forms.
The DVI form is suitable for printing or displaying; the commands for doing
this vary from system to system. On many systems, `lpr -d' will print a DVI
file. On others, you may need to run a program such as `dvips' to convert the
DVI file into a form your system can print.
If you wish to build the DVI file, you will need to have TeX installed on your
system. You can rebuild it by typing:
cd gas/doc
make as.dvi
The Info form is viewable with the GNU Emacs `info' subsystem, or the
stand-alone `info' program, available as part of the GNU Texinfo distribution.
To build the info files, you will need the `makeinfo' program. Type:
cd gas/doc
make info
Specifying names for hosts and targets
======================================
The specifications used for hosts and targets in the `configure'
script are based on a three-part naming scheme, but some short
predefined aliases are also supported. The full naming scheme encodes
three pieces of information in the following pattern:
ARCHITECTURE-VENDOR-OS
For example, you can use the alias `sun4' as a HOST argument or in a
`--target=TARGET' option. The equivalent full name is
`sparc-sun-sunos4'.
The `configure' script accompanying GAS does not provide any query
facility to list all supported host and target names or aliases.
`configure' calls the Bourne shell script `config.sub' to map
abbreviations to full names; you can read the script, if you wish, or
you can use it to test your guesses on abbreviations--for example:
% sh config.sub i386v
i386-unknown-sysv
% sh config.sub i786v
Invalid configuration `i786v': machine `i786v' not recognized
`configure' options
===================
Here is a summary of the `configure' options and arguments that are
most often useful for building GAS. `configure' also has several other
options not listed here.
configure [--help]
[--prefix=DIR]
[--srcdir=PATH]
[--host=HOST]
[--target=TARGET]
[--with-OPTION]
[--enable-OPTION]
You may introduce options with a single `-' rather than `--' if you
prefer; but you may abbreviate option names if you use `--'.
`--help'
Print a summary of the options to `configure', and exit.
`-prefix=DIR'
Configure the source to install programs and files under directory
`DIR'.
`--srcdir=PATH'
Look for the package's source code in directory DIR. Usually
`configure' can determine that directory automatically.
`--host=HOST'
Configure GAS to run on the specified HOST. Normally the
configure script can figure this out automatically.
There is no convenient way to generate a list of all available
hosts.
`--target=TARGET'
Configure GAS for cross-assembling programs for the specified
TARGET. Without this option, GAS is configured to assemble .o files
that run on the same machine (HOST) as GAS itself.
There is no convenient way to generate a list of all available
targets.
`--enable-OPTION'
These flags tell the program or library being configured to
configure itself differently from the default for the specified
host/target combination. See below for a list of `--enable'
options recognized in the gas distribution.
`configure' accepts other options, for compatibility with configuring
other GNU tools recursively; but these are the only options that affect
GAS or its supporting libraries.
The `--enable' options recognized by software in the gas distribution are:
`--enable-targets=...'
This causes one or more specified configurations to be added to those for
which BFD support is compiled. Currently gas cannot use any format other
than its compiled-in default, so this option is not very useful.
`--enable-bfd-assembler'
This causes the assembler to use the new code being merged into it to use
BFD data structures internally, and use BFD for writing object files.
For most targets, this isn't supported yet. For most targets where it has
been done, it's already the default. So generally you won't need to use
this option.
Compiler Support Hacks
======================
On a few targets, the assembler has been modified to support a feature
that is potentially useful when assembling compiler output, but which
may confuse assembly language programmers. If assembler encounters a
.word pseudo-op of the form symbol1-symbol2 (the difference of two
symbols), and the difference of those two symbols will not fit in 16
bits, the assembler will create a branch around a long jump to
symbol1, and insert this into the output directly before the next
label: The .word will (instead of containing garbage, or giving an
error message) contain (the address of the long jump)-symbol2. This
allows the assembler to assemble jump tables that jump to locations
very far away into code that works properly. If the next label is
more than 32K away from the .word, you lose (silently); RMS claims
this will never happen. If the -K option is given, you will get a
warning message when this happens.
REPORTING BUGS IN GAS
=====================
Bugs in gas should be reported to:
bug-binutils@gnu.org.
They may be cross-posted to gcc-bugs@gnu.org if they affect the use of
gas with gcc. They should not be reported just to gcc-bugs, since not
all of the maintainers read that list.
See ../binutils/README for what we need in a bug report.
Copyright (C) 2012-2019 Free Software Foundation, Inc.
Copying and distribution of this file, with or without modification,
are permitted in any medium without royalty provided the copyright
notice and this notice are preserved.
5312fe52e9