platform/upstream/nasm.git
12 years agoAdd support for warning on invalid LOCK prefixes
H. Peter Anvin [Sat, 25 Feb 2012 19:11:42 +0000 (11:11 -0800)]
Add support for warning on invalid LOCK prefixes

Add an LOCK flag to the instruction template, and make the presence of
a LOCK prefix trigger a warning if it is not set.  Simplify the LOCK
and HLE logic by hard-coding the knowledge that operand 0 has to be
memory.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoMove HLE byte codes to \264..\267
H. Peter Anvin [Sat, 25 Feb 2012 18:24:24 +0000 (10:24 -0800)]
Move HLE byte codes to \264..\267

Move the HLE byte codes to \264..\267 so as not to break up an unused
group of 8 (\240..\247).

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoAdd a "nohle" byte code to skip an instruction pattern
H. Peter Anvin [Sat, 25 Feb 2012 18:19:56 +0000 (10:19 -0800)]
Add a "nohle" byte code to skip an instruction pattern

The a2/a3 mem_offs MOV opcodes are invalid with XRELEASE; those
instructions instead have to use a modrm form.  Therefore give a way
to annotate those instruction patters so the pattern matcher will move
on to the next pattern, rather than selecting them and then issue a
warning.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoinsns.pl: Bundle all "plain bytecodes" into a hash
H. Peter Anvin [Sat, 25 Feb 2012 06:06:39 +0000 (22:06 -0800)]
insns.pl: Bundle all "plain bytecodes" into a hash

Instead of a long string of ifs, put all the plain opcodes which just
turn into single bytecodes with no other effect into a hash.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoHLE: Implement the basic mechanism for XACQUIRE/XRELEASE
H. Peter Anvin [Sat, 25 Feb 2012 05:51:46 +0000 (21:51 -0800)]
HLE: Implement the basic mechanism for XACQUIRE/XRELEASE

This implements the mechanism for XACQUIRE/XRELEASE.  It does not
include the necessary annotations in insns.dat.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agonasm.c: Fix is_suppressed_warning()
H. Peter Anvin [Sat, 25 Feb 2012 05:50:53 +0000 (21:50 -0800)]
nasm.c: Fix is_suppressed_warning()

The logic in is_suppressed_warning() was severely wrong, which would
cause a lot of legitimate warnings to be suppressed while some
warnings would be duplicated.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoHLE: Split the LOCK and REP prefix slots
H. Peter Anvin [Sat, 25 Feb 2012 04:57:04 +0000 (20:57 -0800)]
HLE: Split the LOCK and REP prefix slots

With HLE, the sequence REP LOCK actually makes sense, so support it.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoinsns: create a symbolic "wait" token for the \341 byte code
H. Peter Anvin [Fri, 10 Feb 2012 19:11:08 +0000 (11:11 -0800)]
insns: create a symbolic "wait" token for the \341 byte code

Create a symbolic name "wait" for the \341 byte code, so we don't have
to open-code it.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoinsns: fix IMUL patterns to get rid of open-coded \100 bytecodes
H. Peter Anvin [Fri, 10 Feb 2012 19:08:09 +0000 (11:08 -0800)]
insns: fix IMUL patterns to get rid of open-coded \100 bytecodes

There already is a standard encoding for "use this operand in multiple
slots"; no need to open-code it.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoinsns: correct the TSX opcodes
H. Peter Anvin [Fri, 10 Feb 2012 00:24:32 +0000 (16:24 -0800)]
insns: correct the TSX opcodes

All except XTEST are RTM, not HLE; XBEGIN is like a JMP or CALL.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoinsns: Add AVX2 transactional synchronization extensions
Cyrill Gorcunov [Thu, 9 Feb 2012 22:11:22 +0000 (02:11 +0400)]
insns: Add AVX2 transactional synchronization extensions

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoinsns: replace open-coded \322 opcode with odf (operand default)
H. Peter Anvin [Fri, 10 Feb 2012 00:20:21 +0000 (16:20 -0800)]
insns: replace open-coded \322 opcode with odf (operand default)

Create a mnemonic for the open-coded opcode \322.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agobackend: add support for x32 ELF
H.J. Lu [Fri, 3 Feb 2012 20:06:04 +0000 (12:06 -0800)]
backend: add support for x32 ELF

Add an x32 ELF (32-bit code with the CPU in 64-bit mode) backend.

12 years agoBR 3463230: Add VMFUNC instruction
H. Peter Anvin [Wed, 21 Dec 2011 16:26:48 +0000 (08:26 -0800)]
BR 3463230: Add VMFUNC instruction

Add VMFUNC instruction from the Intel SDM version 041.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>
12 years agoinsns: Fix up sizes for MOVSD and VMOVSS instructions
Cyrill Gorcunov [Sat, 17 Dec 2011 08:09:02 +0000 (12:09 +0400)]
insns: Fix up sizes for MOVSD and VMOVSS instructions

Reported-by: Jasper Neumann <jasper.neumann@scpsoftware.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoDon't forget to setup warning levels on preprocessor phase
Cyrill Gorcunov [Sun, 4 Dec 2011 21:56:40 +0000 (01:56 +0400)]
Don't forget to setup warning levels on preprocessor phase

http://bugzilla.nasm.us/show_bug.cgi?id=3143109

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoDrop tab/space mess from parse_cmdline
Cyrill Gorcunov [Sun, 4 Dec 2011 21:44:43 +0000 (01:44 +0400)]
Drop tab/space mess from parse_cmdline

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoUse proper bracing on setting warnings in comman line parsing
Cyrill Gorcunov [Sun, 4 Dec 2011 21:39:04 +0000 (01:39 +0400)]
Use proper bracing on setting warnings in comman line parsing

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoAdd WARN_IDX helper and ERR_WARN_TERM named constant
Cyrill Gorcunov [Sun, 4 Dec 2011 21:01:06 +0000 (01:01 +0400)]
Add WARN_IDX helper and ERR_WARN_TERM named constant

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoSimplify is_suppressed_warning helper
Cyrill Gorcunov [Sun, 4 Dec 2011 15:24:25 +0000 (19:24 +0400)]
Simplify is_suppressed_warning helper

The former is really hard to read.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3392199: Revert "insns: Add MOVD as aliases to MOVQ for compatibility with AMD"
Cyrill Gorcunov [Sun, 20 Nov 2011 13:16:28 +0000 (17:16 +0400)]
BR3392199: Revert "insns: Add MOVD as aliases to MOVQ for compatibility with AMD"

This reverts commit 70712c0df6c437c50452c4997aa2e3de5a0e0299.

Conflicts:

insns.dat

Our instructions matcher fuzzy logic fails to handle it at moment.

Reported-by: KO Myung-Hun <komh@chollian.net>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agopreproc: Proper bracing with list_for_each
Cyrill Gorcunov [Sun, 20 Nov 2011 06:42:13 +0000 (10:42 +0400)]
preproc: Proper bracing with list_for_each

It is a potential place for a bug if list_for_each
helper is used with multiple lines of code without
a proper bracing.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoinsns.dat: Fix VPCMPEQQ template
Cyrill Gorcunov [Mon, 14 Nov 2011 21:21:48 +0000 (01:21 +0400)]
insns.dat: Fix VPCMPEQQ template

http://bugzilla.nasm.us/show_bug.cgi?id=3392197

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3392198: Fix compilation warning on prefixes
Cyrill Gorcunov [Sat, 12 Nov 2011 07:41:51 +0000 (11:41 +0400)]
BR3392198: Fix compilation warning on prefixes

insn->prefixes might contain not only values from
'enum prefixes' but from 'enum reg_enum' as well so
make it generic 'int' instead.

This calms down the compiler about enum's mess and
eliminates a wrong assumption that we always have
values by particular type in this field.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3392195: insns: Drop MMX flag from MOVD
Cyrill Gorcunov [Sat, 12 Nov 2011 05:47:27 +0000 (09:47 +0400)]
BR3392195: insns: Drop MMX flag from MOVD

Typo in specification.

Reported-by: Jasper Neumann <sirrida@web.de>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoinsns: Fix typos for vcmpeq aliases
Cyrill Gorcunov [Fri, 11 Nov 2011 09:04:20 +0000 (13:04 +0400)]
insns: Fix typos for vcmpeq aliases

The patch came from herumi@nifty.com

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agopreproc: Drop NULL assignments for static variables
Cyrill Gorcunov [Sun, 9 Oct 2011 15:04:14 +0000 (19:04 +0400)]
preproc: Drop NULL assignments for static variables

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agopreproc: A few style fixups
Cyrill Gorcunov [Sun, 9 Oct 2011 14:58:46 +0000 (18:58 +0400)]
preproc: A few style fixups

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3376406: Update default optimization level in man
Cyrill Gorcunov [Wed, 5 Oct 2011 20:26:36 +0000 (00:26 +0400)]
BR3376406: Update default optimization level in man

Update default optimization level in man to reflect
the reality.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoinsns: Add MOVD as aliases to MOVQ for compatibility with AMD
Cyrill Gorcunov [Sun, 2 Oct 2011 06:53:37 +0000 (10:53 +0400)]
insns: Add MOVD as aliases to MOVQ for compatibility with AMD

AMD has MOVD for both 32bit and 64bit GPRs so in a sake of
compatibility bring them into insns.dat.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agopreproc: %ifenv should require environment variable name to exist
Cyrill Gorcunov [Sun, 2 Oct 2011 05:18:34 +0000 (09:18 +0400)]
preproc: %ifenv should require environment variable name to exist

Otherwise naked %ifenv/%endif passes without a notice.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3414012: Make %if[n]def to handle tail spaces
Cyrill Gorcunov [Sat, 1 Oct 2011 21:06:29 +0000 (01:06 +0400)]
BR3414012: Make %if[n]def to handle tail spaces

In case if there a trailing spaces after expansion
%if[n]def might produce a false alarms on macro id
being expected.

Note it also makes the constructions like bare

%ifdef
%endif

invalid, while before they were passed well without
a notice.

Reported-by: KO Myung-Hun
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoassemble: Drop redundant variable
Cyrill Gorcunov [Sun, 25 Sep 2011 14:01:45 +0000 (18:01 +0400)]
assemble: Drop redundant variable

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3411280: Fix incorrect compiler option for OS/2 target in Open Watcom makefile
Jiri Malak [Tue, 20 Sep 2011 17:37:06 +0000 (21:37 +0400)]
BR3411280: Fix incorrect compiler option for OS/2 target in Open Watcom makefile

For OS2 target the -bt option should be set accordingly.

Signed-off-by: Jiri Malak <malak.jiri@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoinsns.dat: Fixup VGATHERx instructions
Cyrill Gorcunov [Tue, 30 Aug 2011 21:34:43 +0000 (01:34 +0400)]
insns.dat: Fixup VGATHERx instructions

As being spotted by nasm64developer the memory
operands size is incorrect. Fix it.

Reported-by: nasm64developer
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoAdd AVX2 test file
Cyrill Gorcunov [Tue, 30 Aug 2011 21:22:43 +0000 (01:22 +0400)]
Add AVX2 test file

nasm64developer kindly provided a testfile converted
to nasm format.

http://sourceware.org/ml/binutils/2011-06/msg00150.html

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoA couple of simplifications to assemble.c
Cyrill Gorcunov [Sun, 28 Aug 2011 20:07:17 +0000 (00:07 +0400)]
A couple of simplifications to assemble.c

 - GEN_SIB and GEN_MODRM helpers added
 - a number of tabs vs space fixs
 - more use of is_class() helper

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoparser: No need to split line if it a bit longer than 80 chars
Cyrill Gorcunov [Sun, 28 Aug 2011 15:53:11 +0000 (19:53 +0400)]
parser: No need to split line if it a bit longer than 80 chars

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoAdd IS_SREG and IS_FSGS helpers
Cyrill Gorcunov [Sun, 28 Aug 2011 14:49:00 +0000 (18:49 +0400)]
Add IS_SREG and IS_FSGS helpers

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoA few style fixups in parser.c
Cyrill Gorcunov [Sun, 28 Aug 2011 14:02:31 +0000 (18:02 +0400)]
A few style fixups in parser.c

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoelf64: Make linelist to look the same as elf32
Cyrill Gorcunov [Sun, 28 Aug 2011 13:07:26 +0000 (17:07 +0400)]
elf64: Make linelist to look the same as elf32

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoelf32: Drop holes in linelist structure
Cyrill Gorcunov [Sun, 28 Aug 2011 13:05:17 +0000 (17:05 +0400)]
elf32: Drop holes in linelist structure

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agooutobj: Reorder Segment members to eliminate holes
Cyrill Gorcunov [Sun, 28 Aug 2011 13:03:21 +0000 (17:03 +0400)]
outobj: Reorder Segment members to eliminate holes

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agooutieee: Reorder ieeeSection members to eliminate holes
Cyrill Gorcunov [Sun, 28 Aug 2011 13:00:04 +0000 (17:00 +0400)]
outieee: Reorder ieeeSection members to eliminate holes

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoDrop empty line and bracket
Cyrill Gorcunov [Sun, 28 Aug 2011 12:33:39 +0000 (16:33 +0400)]
Drop empty line and bracket

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoStyle tuneup of nasm.h
Cyrill Gorcunov [Sun, 28 Aug 2011 11:00:26 +0000 (15:00 +0400)]
Style tuneup of nasm.h

 - a number of tabs to space conversion
 - redundant spaces removed
 - structure elements alignment

No functional changes.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agodisasm.c: ESP/RSP cannot be indicies, but XMM4/YMM4 can be
H. Peter Anvin [Mon, 22 Aug 2011 21:09:04 +0000 (14:09 -0700)]
disasm.c: ESP/RSP cannot be indicies, but XMM4/YMM4 can be

If we have XMM4/YMM4 as a VSIB, it does not mean no index; that is
only for scalar encoding.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
12 years agoassemble.c: remove stray debugging code
H. Peter Anvin [Mon, 22 Aug 2011 20:52:02 +0000 (13:52 -0700)]
assemble.c: remove stray debugging code

My bad for checking this in at all.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
12 years agoinsns, avx2: A couple of upper-case to lower-case conversion
Cyrill Gorcunov [Mon, 22 Aug 2011 20:35:32 +0000 (00:35 +0400)]
insns, avx2: A couple of upper-case to lower-case conversion

Just to be solid in style

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoinsns, avx2: A typo in VPERMPD
Cyrill Gorcunov [Mon, 22 Aug 2011 20:30:28 +0000 (00:30 +0400)]
insns, avx2: A typo in VPERMPD

The second VPERMD should be VPERMPD actually.

Thanks to nasm64developer for gas test file provided
which allowed to reveal this issue.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoassemble.c: Comment out debug printing
Cyrill Gorcunov [Mon, 22 Aug 2011 20:12:50 +0000 (00:12 +0400)]
assemble.c: Comment out debug printing

Probably we need some kind of pr_debug or
something like that instead.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
12 years agoBR3385573: Some AVX2 instructions fixups
Cyrill Gorcunov [Sun, 21 Aug 2011 21:38:40 +0000 (01:38 +0400)]
BR3385573: Some AVX2 instructions fixups

A few instruction templates for AVX2 set were wrong.

Reported-by: Agner Fog
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns: Change VPERMPQ to VPERMQ
Cyrill Gorcunov [Wed, 3 Aug 2011 18:16:29 +0000 (22:16 +0400)]
insns: Change VPERMPQ to VPERMQ

Was a typo

Reported-by: Agner Fog
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Drop emitting() rudimanet
Cyrill Gorcunov [Sun, 31 Jul 2011 10:51:49 +0000 (14:51 +0400)]
preproc: Drop emitting() rudimanet

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agodoc: update documentation for pending 2.10 release
Keith Kanios [Sat, 30 Jul 2011 18:14:18 +0000 (13:14 -0500)]
doc: update documentation for pending 2.10 release

13 years agoNASM 2.10rc8
H. Peter Anvin [Wed, 27 Jul 2011 21:32:08 +0000 (14:32 -0700)]
NASM 2.10rc8

13 years agoImplement insns.dat in human readable form
Anonymous [Wed, 20 Jul 2011 17:41:49 +0000 (21:41 +0400)]
Implement insns.dat in human readable form

 I converted almost all instructions in insns.dat (version
 7a6f97869840fc0b1b8677bd77c74754c5f024aa) to the more
 readable format that insns.pl has supported for years.

 I also made some changes to insns.pl. You can verify that the
 new insns.dat and insns.pl produce byte-identical output to
 the old insns.dat and insns.pl, so I think that this change
 is safe to check in, even though it is a large change to
 insns.dat.

The changes to insns.pl are:

 * fixed a bug: ib,u was not recognized
 * added support for a second immediate argument called "j" for
   instructions like ENTER imm,imm
 * added a "+r" syntax for \10..\13

[gorcunov: insns files remains the same, great job anonymous!]

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoNASM 2.10.rc7
Cyrill Gorcunov [Tue, 19 Jul 2011 21:30:31 +0000 (01:30 +0400)]
NASM 2.10.rc7

Intel's AVX2 instructions should be covered now.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns: A final pile of AVX2 instructions
Cyrill Gorcunov [Tue, 19 Jul 2011 21:16:11 +0000 (01:16 +0400)]
insns: A final pile of AVX2 instructions

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns: A few additional AVX2 templates
Cyrill Gorcunov [Tue, 19 Jul 2011 13:01:35 +0000 (17:01 +0400)]
insns: A few additional AVX2 templates

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns: One more small snippet of AVX2
Cyrill Gorcunov [Sun, 17 Jul 2011 21:15:25 +0000 (01:15 +0400)]
insns: One more small snippet of AVX2

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns: One more slab of AVX2 instructions
Cyrill Gorcunov [Sun, 17 Jul 2011 20:57:35 +0000 (00:57 +0400)]
insns: One more slab of AVX2 instructions

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns: Add a slab of AVX2 instructions
Cyrill Gorcunov [Sun, 17 Jul 2011 19:20:00 +0000 (23:20 +0400)]
insns: Add a slab of AVX2 instructions

Not all are covered yet, but still a step
forward I think.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns.dat: Move exsiting AVX2 insns to a separate section
Cyrill Gorcunov [Sun, 17 Jul 2011 16:24:16 +0000 (20:24 +0400)]
insns.dat: Move exsiting AVX2 insns to a separate section

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Align members in structures
Cyrill Gorcunov [Sun, 17 Jul 2011 16:20:14 +0000 (20:20 +0400)]
preproc: Align members in structures

For easier reading

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Reorder ExpInv to drop padding
Cyrill Gorcunov [Sun, 17 Jul 2011 16:11:08 +0000 (20:11 +0400)]
preproc: Reorder ExpInv to drop padding

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Reorder SMacro members to eliminate padding
Cyrill Gorcunov [Sun, 17 Jul 2011 16:06:20 +0000 (20:06 +0400)]
preproc: Reorder SMacro members to eliminate padding

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonasm.h: Reorder tokenval members to drop padding
Cyrill Gorcunov [Sun, 17 Jul 2011 16:00:30 +0000 (20:00 +0400)]
nasm.h: Reorder tokenval members to drop padding

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoSome cleanup on insns.h
Cyrill Gorcunov [Sun, 17 Jul 2011 13:52:03 +0000 (17:52 +0400)]
Some cleanup on insns.h

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoDrop unused 'opt' from nasm_opt_val
Cyrill Gorcunov [Sun, 17 Jul 2011 10:06:54 +0000 (14:06 +0400)]
Drop unused 'opt' from nasm_opt_val

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoelf64: Drop unused 'zero' variable
Cyrill Gorcunov [Sun, 17 Jul 2011 10:04:44 +0000 (14:04 +0400)]
elf64: Drop unused 'zero' variable

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoDrop unused 'type' from gencode
Cyrill Gorcunov [Sun, 17 Jul 2011 10:02:52 +0000 (14:02 +0400)]
Drop unused 'type' from gencode

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoFix saa_writeaddr for big-endian case
Cyrill Gorcunov [Sun, 17 Jul 2011 09:59:13 +0000 (13:59 +0400)]
Fix saa_writeaddr for big-endian case

saa_writeaddr on big-endian archs didn't write converted
argument, fix it.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoDrop unused 'arg' variable
Cyrill Gorcunov [Sun, 17 Jul 2011 09:55:25 +0000 (13:55 +0400)]
Drop unused 'arg' variable

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoMerge branch 'nasm-2.09.xx'
Cyrill Gorcunov [Fri, 15 Jul 2011 21:44:57 +0000 (01:44 +0400)]
Merge branch 'nasm-2.09.xx'

Conflicts:
version

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoNASM-2.09.10
Cyrill Gorcunov [Fri, 15 Jul 2011 21:41:28 +0000 (01:41 +0400)]
NASM-2.09.10

- Fix nasty problem with uninstaller in case of registry
  keys corruption, which might lead to arbitrary files
  removal.

  Just to be sure the proper files and directories are to
  be removed a few additional questions are brought into
  deinstallation procedure.

  The former issue reported here

http://forum.nasm.us/index.php?topic=1013.0

  Note: no functional changes are done inside nasm code,
  critical fixes will be ported to 2.09.xx branch, but
  mostly the (more or less) active development and various
  bug fixes are concentrated around 2.10 series.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agodoc: Describe changes for nasm-2.09.10
Cyrill Gorcunov [Fri, 15 Jul 2011 21:33:00 +0000 (01:33 +0400)]
doc: Describe changes for nasm-2.09.10

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonsis: Remove desktop bat link as well
Cyrill Gorcunov [Fri, 15 Jul 2011 19:10:46 +0000 (23:10 +0400)]
nsis: Remove desktop bat link as well

In case of installation procedure being used in
a per-user fasion the display link to a bat file
remain untouched at uninstallation time. Fix it.

I hope this time all things go smooth.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonsis: Inform a user about files being deleted
Cyrill Gorcunov [Fri, 15 Jul 2011 18:38:47 +0000 (22:38 +0400)]
nsis: Inform a user about files being deleted

Since a user might have some weird access rules on NTFS
such as allow to write files but not to delete them it
might happen that uninstallation procedure fails at
deleting files but deleting registry keys still success,
and in result a user might try to uninstall nasm for second
time with registry keys corrupted.

So prompt a user with files being deleted so he can double
check what is going on.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonsis: Make sure the registry keys exist
Cyrill Gorcunov [Wed, 6 Jul 2011 20:56:35 +0000 (00:56 +0400)]
nsis: Make sure the registry keys exist

It was reported on nasm forum (http://forum.nasm.us/index.php?topic=1013.0)
that if uninstallation procedure get stopped (for any reason) in a middle
leading to read empty registry values an attempt to remove arbitrary
files might happen. So make sure there are some sane values in registry.

N.B: People, if you find some bug in nasm, don't write to forum, file
bug directly in bugzilla, if it's not a bug we simply close it, otherwise
we might miss serious problems like this one!

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonsis: Remove desktop bat link as well
Cyrill Gorcunov [Fri, 15 Jul 2011 19:10:46 +0000 (23:10 +0400)]
nsis: Remove desktop bat link as well

In case of installation procedure being used in
a per-user fasion the display link to a bat file
remain untouched at uninstallation time. Fix it.

I hope this time all things go smooth.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonsis: Inform a user about files being deleted
Cyrill Gorcunov [Fri, 15 Jul 2011 18:38:47 +0000 (22:38 +0400)]
nsis: Inform a user about files being deleted

Since a user might have some weird access rules on NTFS
such as allow to write files but not to delete them it
might happen that uninstallation procedure fails at
deleting files but deleting registry keys still success,
and in result a user might try to uninstall nasm for second
time with registry keys corrupted.

So prompt a user with files being deleted so he can double
check what is going on.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoA few more AVX2 spec instructions
H. Peter Anvin [Fri, 8 Jul 2011 23:08:34 +0000 (16:08 -0700)]
A few more AVX2 spec instructions

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
13 years agoUse a normal quad-case for valueless /is4
H. Peter Anvin [Fri, 8 Jul 2011 00:21:24 +0000 (17:21 -0700)]
Use a normal quad-case for valueless /is4

When we don't have an immediate for the i-field in /is4, then use a
normal quad-bytecode encoding for it to save some small amount of
space and re-use existing machinery.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
13 years agoRemove support for DREX encoding
H. Peter Anvin [Thu, 7 Jul 2011 23:58:22 +0000 (16:58 -0700)]
Remove support for DREX encoding

The DREX encoding never hit production silicon, and has been replaced
by VEX/XOP encoding, so remove support for it.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
13 years agoinsns.dat: Clean up and fix the BMI instruction patterns
H. Peter Anvin [Thu, 7 Jul 2011 22:29:14 +0000 (15:29 -0700)]
insns.dat: Clean up and fix the BMI instruction patterns

Clean up the formatting of the BMI instruction patterns, and fix:

a) X64,FUTURE is wrong - it needs to be LONG,FUTURE
b) Fix the BLSI, BLSMSK, BLSR instruction patterns
c) Use a bracket pattern for TZCNT

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
13 years agoinsns: Fix up RORX template
Jasper Neuman [Thu, 7 Jul 2011 07:02:11 +0000 (11:02 +0400)]
insns: Fix up RORX template

Missed 64 bit case.

Signed-off-by: Jasper Neuman <jasper.neumann@web.de>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns.dat: Add some BMI1 and BMI2 instructions
Jasper Neuman [Wed, 6 Jul 2011 21:08:51 +0000 (01:08 +0400)]
insns.dat: Add some BMI1 and BMI2 instructions

Signed-off-by: Jasper Neuman <jasper.neumann@web.de>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agonsis: Make sure the registry keys exist
Cyrill Gorcunov [Wed, 6 Jul 2011 20:56:35 +0000 (00:56 +0400)]
nsis: Make sure the registry keys exist

It was reported on nasm forum (http://forum.nasm.us/index.php?topic=1013.0)
that if uninstallation procedure get stopped (for any reason) in a middle
leading to read empty registry values an attempt to remove arbitrary
files might happen. So make sure there are some sane values in registry.

N.B: People, if you find some bug in nasm, don't write to forum, file
bug directly in bugzilla, if it's not a bug we simply close it, otherwise
we might miss serious problems like this one!

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoinsns.h: Add BMI1 and BMI2 flags for further usage
Jasper Neuman [Wed, 6 Jul 2011 20:03:49 +0000 (00:03 +0400)]
insns.h: Add BMI1 and BMI2 flags for further usage

Signed-off-by: Jasper Neuman <jasper.neumann@web.de>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Drop useless assignments
Cyrill Gorcunov [Wed, 6 Jul 2011 19:12:55 +0000 (23:12 +0400)]
preproc: Drop useless assignments

new_ExpDef does zalloc'ate memory so no need
to assign zeros again.

Same time it should fix MSC complains on NULL assignment
on boolean variable.

Reported-by: Jasper Neuman <jasper.neumann@web.de>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoMerge branch 'nasm-2.09.xx'
Cyrill Gorcunov [Sun, 3 Jul 2011 20:51:53 +0000 (00:51 +0400)]
Merge branch 'nasm-2.09.xx'

Conflicts:
version

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoNASM 2.09.09
Cyrill Gorcunov [Sun, 3 Jul 2011 20:49:03 +0000 (00:49 +0400)]
NASM 2.09.09

Just a few fixes only but serious enough to yield
new stable series

 - Unitialized memory parsing in bin format section attibutes.
 - MachO NULL dereference

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agodoc: Describe changes for stable version
Cyrill Gorcunov [Sun, 3 Jul 2011 20:47:08 +0000 (00:47 +0400)]
doc: Describe changes for stable version

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agoBR3282788: Fix 64-bit Mach-O bug that crashes NASM due to NULL symbols
Keith Kanios [Tue, 12 Apr 2011 02:38:50 +0000 (21:38 -0500)]
BR3282788: Fix 64-bit Mach-O bug that crashes NASM due to NULL symbols

13 years agooutput/outbin.c: initialize section attribs upon creation
Cyrill Gorcunov [Sun, 3 Jul 2011 20:33:24 +0000 (00:33 +0400)]
output/outbin.c: initialize section attribs upon creation

Basically it's backport of commits

11db774a151e9d895fa05f980563a5cafb0f306a
3bc3ff2fb685a645698f9db9cfc903df30e4e555
c13deef255b621ace2130adf55530f3364a40458
e3f47806658de042af0eaccb1cc7896be388b397

They were missed to back-merge in a first place.

Reported-by: Keith Kanios <keith@kanios.net>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Move Preproc type to preproc_ops structure
Cyrill Gorcunov [Fri, 1 Jul 2011 06:38:25 +0000 (10:38 +0400)]
preproc: Move Preproc type to preproc_ops structure

There is no need to hide this structure into a type.
The former preproc_ops is a way more descriptive.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Refactor smacro paramters expansion
Cyrill Gorcunov [Wed, 29 Jun 2011 21:36:45 +0000 (01:36 +0400)]
preproc: Refactor smacro paramters expansion

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Add smacro paramter helpers
Cyrill Gorcunov [Wed, 29 Jun 2011 21:16:35 +0000 (01:16 +0400)]
preproc: Add smacro paramter helpers

I belive this should simply code a bit.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
13 years agopreproc: Alignment in loop
Cyrill Gorcunov [Wed, 29 Jun 2011 20:22:53 +0000 (00:22 +0400)]
preproc: Alignment in loop

Just for easier reading

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>