Go to file
Jose E. Marchesi 75ac3a7f57 gas: fix bumping to architectures >v9 in sparc64-* targets.
This patch fixes two related problems:

- By default gas is supposed to bump the current architecture
  (starting with v6) as it finds "higher" instructions as the
  assembling progresses.  There are four possible cases depending on
  the usage of the -A and -bump options:

  (a) No -A and -bump are specified.  In this case max_architecture
      must be the highest architecture not conflicting with the
      default architecture.  The default opcode architecture is
      indirectly set in configure.tgt and is "v9" in sparc64 systems
      (from "v9-64").  Thus the maximum architecture in sparc64
      systems must be "v9b".  No warnings are echoed when the assembly
      of an instruction bumps the current architecture.

  (b) Only -bump is specified.  This is like (a) but warnings are
      always issued when the assembly of an instruction bumps the
      current architecture.

  (c) Only -A is specified.  In this case bumping to a new
      architecture is an error.

  (d) Both -A and -bump are specified.  In this case max_architecture
      must be the highest architecture not conflicting with the
      default architecture, but warnings are only to be issued when
      bumping to an architecture higher than the architecture selected
      in the -A option.

  `max_architecture' is a global variable defined in tc-sparc.c which
  is initialized to the opcode architecture corresponding to the
  default architecture ("sparclite" for sparc-* targets and "v9" for
  sparc64-* targets).  Then in `md_begin' it is set to the highest
  non-conflicting architecture, but only when both -A and -bump are
  specified.

  Thus (a) does not work:

    $ echo "fzero %f0" | as
    {standard input}: Assembler messages:
    {standard input}:1: Error: Architecture mismatch on "fzero".
    {standard input}:1:  (Requires v9a|v9b; requested architecture is v9.)

  Neither (b):

    $ echo "fzero %f0" | as -bump
    {standard input}: Assembler messages:
    {standard input}:1: Error: Architecture mismatch on "fzero".
    {standard input}:1:  (Requires v9a|v9b; requested architecture is v9.)

  Only (d) does:

    $ echo "fzero %f0" | as -Av9 -bump
    {standard input}: Assembler messages:
    {standard input}:1: Warning: architecture bumped from "v6" to "v9a" on "fzero"

  This patch fixes that function to "upgrade" `max_architecture' also
  in the (a) and (b) cases.

  Note that this problem becomes apparent only in sparc64-* targets
  because in sparc-* targets the default architecture is the "higher"
  among the 32bit architectures ("sparclite").

- Gas maintains a set of hardware capabilities associated with each
  gas architecture, in `sparc_arch_table'.  On the other hand
  libopcodes maintains a set of hardware capabilities needed by each
  individual sparc instruction.

  When an instruction is assembled in `sparc_ip' gas checks for the
  presence of the hardware capabilities required by the instruction,
  emitting an error if some capability is missing.

  However, this mechanism does not work properly if the current
  architecture is bumped due to an instruction requiring new hw
  capabilities not present on either the default architecture or an
  architecture specified with -A:

  $ echo "fzero %f0" | as -bump
  {standard input}: Assembler messages:
  {standard input}:1: Warning: architecture bumped from "v6" to "v9a" on "fzero"
  {standard input}:1: Error: Hardware capability "vis" not enabled for "fzero".

  This patch fixes this by adding the set of required hw caps of an
  instruction if it triggers an architecture bump.

The patch has been tested in sparc64-unknown-linux-gnu.

gas/ChangeLog:

2014-09-12  Jose E. Marchesi  <jose.marchesi@oracle.com>

	* config/tc-sparc.c (sparc_ip): Update the set of allowed hwcaps
	when bumping the current architecture.
	(md_begin): Adjust the highetst architecture level also when a
	specific architecture is not requested.
2014-09-12 15:38:21 +02:00
bfd Add mips*-img-elf* target triple. 2014-09-12 14:25:15 +01:00
binutils Add mips*-img-elf* target triple. 2014-09-12 14:25:15 +01:00
config
cpu
elfcpp Make Elf_file::section_name() a const function, so that it can be used in 2014-09-02 15:44:58 -07:00
etc
gas gas: fix bumping to architectures >v9 in sparc64-* targets. 2014-09-12 15:38:21 +02:00
gdb PR tdep/17379: Fix internal-error when stack pointer is invalid. 2014-09-12 09:20:25 -03:00
gold Add NaCl (NativeClient) specific classes Target_mips_nacl and 2014-09-03 09:44:11 -07:00
gprof
include [PATCH/AArch64] Implement LSE feature 2014-09-03 14:53:53 +01:00
intl
ld Add mips*-img-elf* target triple. 2014-09-12 14:25:15 +01:00
libdecnumber
libiberty
opcodes Properly handle suffix for iret and sysret 2014-09-10 09:39:24 -07:00
readline
sim Shadow SIM's debug_printf function 2014-08-28 11:59:09 +01:00
texinfo
.cvsignore
.gitattributes
.gitignore
ChangeLog Add mips*-img-elf* target triple. 2014-09-12 14:25:15 +01:00
compile
config-ml.in
config.guess
config.rpath
config.sub
configure Add mips*-img-elf* target triple. 2014-09-12 14:25:15 +01:00
configure.ac Add mips*-img-elf* target triple. 2014-09-12 14:25:15 +01:00
COPYING
COPYING3
COPYING3.LIB
COPYING.LIB
COPYING.LIBGLOSS
COPYING.NEWLIB
depcomp
djunpack.bat
install-sh
libtool.m4
lt~obsolete.m4
ltgcc.m4
ltmain.sh
ltoptions.m4
ltsugar.m4
ltversion.m4
MAINTAINERS
Makefile.def
Makefile.in
Makefile.tpl
makefile.vms
missing
mkdep
mkinstalldirs
move-if-change
README
README-maintainer-mode
setup.com
src-release.sh src-release.sh: Convert src-release to a shell script 2014-08-27 13:50:06 +01:00
symlink-tree
ylwrap

		   README for GNU development tools

This directory contains various GNU compilers, assemblers, linkers, 
debuggers, etc., plus their support routines, definitions, and documentation.

If you are receiving this as part of a GDB release, see the file gdb/README.
If with a binutils release, see binutils/README;  if with a libg++ release,
see libg++/README, etc.  That'll give you info about this
package -- supported targets, how to use it, how to report bugs, etc.

It is now possible to automatically configure and build a variety of
tools with one command.  To build all of the tools contained herein,
run the ``configure'' script here, e.g.:

	./configure 
	make

To install them (by default in /usr/local/bin, /usr/local/lib, etc),
then do:
	make install

(If the configure script can't determine your type of computer, give it
the name as an argument, for instance ``./configure sun4''.  You can
use the script ``config.sub'' to test whether a name is recognized; if
it is, config.sub translates it to a triplet specifying CPU, vendor,
and OS.)

If you have more than one compiler on your system, it is often best to
explicitly set CC in the environment before running configure, and to
also set CC when running make.  For example (assuming sh/bash/ksh):

	CC=gcc ./configure
	make

A similar example using csh:

	setenv CC gcc
	./configure
	make

Much of the code and documentation enclosed is copyright by
the Free Software Foundation, Inc.  See the file COPYING or
COPYING.LIB in the various directories, for a description of the
GNU General Public License terms under which you can copy the files.

REPORTING BUGS: Again, see gdb/README, binutils/README, etc., for info
on where and how to report problems.