3 * Fix buffer overflow in the listing module.
4 * Fix the handling of hexadecimal escape codes in `...` strings.
8 * Add support for Intel AVX, CLMUL and FMA instructions,
9 including YMM registers.
10 * "dy", "resy" and "yword" for 32-byte operands.
11 * Fix some SSE5 instructions.
12 * Intel INVEPT, INVVPID and MOVBE instructions.
13 * Fix checking for critical expressions when the optimizer is enabled.
14 * Support the DWARF debugging format for ELF targets.
15 * Fix optimizations of signed bytes.
16 * Fix operation on bigendian machines.
17 * Fix buffer overflow in the preprocessor.
18 * SAFESEH support for Win32, IMAGEREL for Win64 (SEH).
19 * %? and %?? to refer to the name of a macro itself. In particular,
20 "%idefine keyword $%?" can be used to make a keyword "disappear".
21 * New options for dependency generation: -MD, -MF, -MP, -MT, -MQ.
22 * New preprocessor directives %pathsearch and %depend; INCBIN
23 reimplemented as a macro.
24 * %include now resolves macros in a sane manner.
25 * %substr can now be used to get other than one-character substrings.
26 * New type of character/string constants, using backquotes (`...`),
27 which support C-style escape sequences.
28 * %defstr and %idefstr to stringize macro definitions before creation.
29 * Fix forward references used in EQU statements.
33 * Additional fixes for MMX operands with explicit "qword", as well as
34 (hopefully) SSE operands with "oword".
35 * Fix handling of truncated strings with DO.
36 * Fix segfaults due to memory overwrites when floating-point constants
38 * Fix segfaults due to missing include files.
39 * Fix OpenWatcom Makefiles for DOS and OS/2.
40 * Add autogenerated instruction list back into the documentation.
41 * ELF: Fix segfault when generating stabs, and no symbols have been
43 * ELF: Experimental support for DWARF debugging information.
44 * New compile date and time standard macros.
45 * %ifnum now returns true for negative numbers.
46 * New %iftoken test for a single token.
47 * New %ifempty test for empty expansion.
48 * Add support for the XSAVE instruction group.
49 * Makefile for Netware/gcc.
50 * Fix issue with some warnings getting emitted way too many times.
51 * Autogenerated instruction list added to the documentation.
55 * Fix the handling of MMX registers with explicit "qword" tags on
56 memory (broken in 2.00 due to 64-bit changes.)
57 * Fix the PREFETCH instructions.
58 * Fix the documentation.
59 * Fix debugging info when using "-f elf" (backwards alias for "-f
61 * Man pages for rdoff tools (from the Debian project.)
62 * ELF: handle large numbers of sections.
63 * Fix corrupt output when the optimizer runs out of passes.
67 * Added c99 data-type compliance.
68 * Added general x86-64 support.
69 * Added win64 (x86-64 COFF) output format.
70 * Added __BITS__ standard macro.
71 * Renamed the elf output format to elf32 for clarity.
72 * Added elf64 and macho (MacOS X) output formats.
73 * Added Numeric constants in DQ directive.
74 * Added oword, do and reso pseudo operands.
75 * Allow underscores in numbers.
76 * Added 8-, 16- and 128-bit floating-point formats.
77 * Added binary, octal and hexadecimal floating-point.
78 * Correct the generation of floating-point constants.
79 * Added Floating-point option control.
80 * Added Infinity and NaN floating point support.
81 * Added ELF Symbol Visibility support.
82 * Added Setting OSABI value in ELF header directive.
83 * Added Generate Makefile Dependencies option.
84 * Added Unlimited Optimization Passes option.
85 * Added %IFN and %ELIFN support.
86 * Added Logical Negation Operator.
87 * Enhanced Stack Relative Preprocessor Directives.
88 * Enhanced ELF Debug Formats.
89 * Enhanced Send Errors to a File option.
90 * Added SSSE3, SSE4.1, SSE4.2, SSE5 support.
91 * Added a large number of additional instructions.
92 * Significant performance improvements.
97 * fix outas86's .bss handling
98 * "make spotless" no longer deletes config.h.in.
99 * %(el)if(n)idn insensitivity to string quotes difference (#809300).
100 * (nasm.c) __OUTPUT_FORMAT__ changed to string value instead of symbol.
104 * Add Makefile for 16-bit DOS binaries under OpenWatcom, and modify
105 mkdep.pl to be able to generate completely pathless dependencies, as
106 required by OpenWatcom wmake (it supports path searches, but not
108 * Fix the STR instruction.
109 * Fix the ELF output format, which was broken under certain
110 circumstances due to the addition of stabs support.
111 * Quick-fix Borland format debug-info for -f obj
112 * Fix for %rep with no arguments (#560568)
113 * Fix concatenation of preprocessor function call (#794686)
114 * Fix long label causes coredump (#677841)
115 * Use autoheader as well as autoconf to keep configure from generating
116 ridiculously long command lines.
117 * Make sure that all of the formats which support debugging output
118 actually will suppress debugging output when -g not specified.
122 * Paths given in "-I" switch searched for "incbin"ed as
123 well as "%include"ed files.
124 * Added stabs debugging for the ELF output format, patch from
126 * Fix output/outbin.c to allow origin > 80000000h.
127 * Make -U switch work.
128 * Fix the use of relative offsets with explicit prefixes, e.g. "a32 loop foo".
129 * Remove "backslash()".
130 * Fix the SMSW and SLDT instructions.
131 * -O2 and -O3 are no longer aliases for -O10 and -O15. If you mean the
132 latter, please say so! :)
136 * Update rdoff - librarian/archiver - common rec - docs!
137 * Fix signed/unsigned problems.
138 * Fix JMP FAR label and CALL FAR label.
139 * Add new multisection support - map files - fix align bug
140 * Fix sysexit, movhps/movlps reg,reg bugs in insns.dat
141 * "Q" or "O" suffixes indicate octal
142 * Support Prescott new instructions (PNI).
143 * Cyrix XSTORE instruction.
147 * Fix build failure on 16-bit DOS (Makefile.bc3 workaround for compiler bug.)
148 * Fix dependencies and compiler warnings.
149 * Add "const" in a number of places.
150 * Add -X option to specify error reporting format (use -Xvc to
151 integrate with Microsoft Visual Studio.)
152 * Minor changes for code legibility.
153 * Drop use of tmpnam() in rdoff (security fix.)
157 * Correct additional address-size vs. operand-size confusions.
158 * Generate dependencies for all Makefiles automatically.
159 * Add support for unimplemented (but theoretically available)
160 registers such as tr0 and cr5. Segment registers 6 and 7 are called
161 segr6 and segr7 for the operations which they can be represented.
162 * Correct some disassembler bugs related to redundant address-size prefixes.
163 Some work still remains in this area.
164 * Correctly generate an error for things like "SEG eax".
165 * Add the JMPE instruction, enabled by "CPU IA64".
166 * Correct compilation on newer gcc/glibc platforms.
167 * Issue an error on things like "jmp far eax".
173 * New __NASM_PATCHLEVEL__ and __NASM_VERSION_ID__ standard macros to
174 round out the version-query macros. version.pl now understands
175 X.YYplWW or X.YY.ZZplWW as a version number, equivalent to
176 X.YY.ZZ.WW (or X.YY.0.WW, as appropriate).
177 * New keyword "strict" to disable the optimization of specific
179 * Fix the handing of size overrides with JMP instructions
180 (instructions such as "jmp dword foo".)
181 * Fix the handling of "ABSOLUTE label", where "label" points into a
183 * Fix OBJ output format with lots of externs.
184 * More documentation updates.
185 * Add -Ov option to get verbose information about optimizations.
186 * Undo a braindead change which broke %elif directives.
193 * Fix NASM crashing when %macro directives were left unterminated.
194 * Lots of documentation updates.
195 * Complete rewrite of the PostScript/PDF documentation generator.
196 * The MS Visual C++ Makefile was updated and corrected.
197 * Recognize .rodata as a standard section name in ELF.
198 * Fix some obsolete Perl4-isms in Perl scripts.
199 * Fix configure.in to work with autoconf 2.5x.
200 * Fix a couple of "make cleaner" misses.
201 * Make the normal "./configure && make" work with Cygwin.
207 * Correctly build in a separate object directory again.
208 * Derive all references to the version number from the version file.
209 * New standard macros __NASM_SUBMINOR__ and __NASM_VER__ macros.
210 * Lots of Makefile updates and bug fixes.
211 * New %ifmacro directive to test for multiline macros.
212 * Documentation updates.
213 * Fixes for 16-bit OBJ format output.
214 * Changed the NASM environment variable to NASMENV.
220 * Changed doc files a lot: completely removed old READMExx and
221 Wishlist files, incorporating all information in CHANGES and TODO.
222 * I waited a long time to rename zoutieee.c to (original) outieee.c
223 * moved all output modules to output/ subdirectory.
224 * Added 'make strip' target to strip debug info from nasm & ndisasm.
225 * Added INSTALL file with installation instructions.
226 * Added -v option description to nasm man.
227 * Added dist makefile target to produce source distributions.
228 * 16-bit support for ELF output format (GNU extension, but useful.)
234 * Fastcooked this for Debian's Woody release:
235 Frank applied the INCBIN bug patch to 0.98.25alt and called
236 it 0.98.28 to not confuse poor little apt-get.
242 * Reorganised files even better from 0.98.25alt
248 * Prettified the source tree. Moved files to more reasonable places.
249 * Added findleak.pl script to misc/ directory.
250 * Attempted to fix doc.
255 * Line continuation character '\'
256 * Docs inadvertantly reverted - "dos packaging".
262 * FIXME: Someone, document this please.
268 * Documentation - Ndisasm doc added to Nasm.doc.
274 * Attempted to remove rdoff version1
275 * Lino Mastrodomenico's patches to preproc.c (%$$ bug?).
281 * Update rdoff2 - attempt to remove v1.
287 * Optimization fixes.
293 * Optimization fixes.
299 * H. J. Lu's patch back out.
305 * Added ".rdata" to "-f win32".
311 * H. J. Lu's "bogus elf" patch. (Red Hat problem?)
317 * Fix whitespace before "[section ..." bug.
323 * Fix fixes to memory leaks.
329 * (there was no '.13)
334 * Update optimization (new function of "-O1")
335 * Changes to test/bintest.asm (?).
340 * Optimization changes.
342 * (there was no '.10)
347 * Add multiple sections support to "-f bin".
348 * Changed GLOBAL_TEMP_BASE in outelf.c from 6 to 15.
349 * Add "-v" as an alias to the "-r" switch.
350 * Remove "#ifdef" from Tasm compatibility options.
351 * Remove redundant size-overrides on "mov ds, ex", etc.
352 * Fixes to SSE2, other insns.dat (?).
353 * Enable uppercase "I" and "P" switches.
354 * Case insinsitive "seg" and "wrt".
355 * Update install.sh (?).
356 * Allocate tokens in blocks.
357 * Improve "invalid effective address" messages.
362 * Add "%strlen" and "%substr" macro operators
363 * Fixed broken c16.mac.
364 * Unterminated string error reported.
365 * Fixed bugs as per 0.98bf
368 0.98.09b with John Coffman patches released 28-Oct-2001
369 -------------------------------------------------------
371 Changes from 0.98.07 release to 98.09b as of 28-Oct-2001
373 * More closely compatible with 0.98 when -O0 is implied
374 or specified. Not strictly identical, since backward
375 branches in range of short offsets are recognized, and signed
376 byte values with no explicit size specification will be
377 assembled as a single byte.
379 * More forgiving with the PUSH instruction. 0.98 requires
380 a size to be specified always. 0.98.09b will imply the size
381 from the current BITS setting (16 or 32).
383 * Changed definition of the optimization flag:
385 -O0 strict two-pass assembly, JMP and Jcc are
386 handled more like 0.98, except that back-
387 ward JMPs are short, if possible.
389 -O1 strict two-pass assembly, but forward
390 branches are assembled with code guaranteed
391 to reach; may produce larger code than
392 -O0, but will produce successful assembly
393 more often if branch offset sizes are not
396 -O2 multi-pass optimization, minimize branch
397 offsets; also will minimize signed immed-
398 iate bytes, overriding size specification.
400 -O3 like -O2, but more passes taken, if needed
403 0.98.07 released 01/28/01
404 -------------------------
406 * Added Stepane Denis' SSE2 instructions to a *working*
407 version of the code - some earlier versions were based on
408 broken code - sorry 'bout that. version "0.98.07"
414 * Cosmetic modifications to nasm.c, nasm.h,
418 0.98.06f released 01/18/01
419 --------------------------
421 * - Add "metalbrain"s jecxz bug fix in insns.dat
422 - alter nasmdoc.src to match - version "0.98.06f"
425 0.98.06e released 01/09/01
426 --------------------------
428 * Removed the "outforms.h" file - it appears to be
429 someone's old backup of "outform.h". version "0.98.06e"
435 * fbk - finally added the fix for the "multiple %includes bug",
436 known since 7/27/99 - reported originally (?) and sent to
437 us by Austin Lunnen - he reports that John Fine had a fix
438 within the day. Here it is...
440 * Nelson Rush resigns from the group. Big thanks to Nelson for
441 his leadership and enthusiasm in getting these changes
442 incorporated into Nasm!
444 * fbk - [list +], [list -] directives - ineptly implemented, should
445 be re-written or removed, perhaps.
447 * Brian Raiter / fbk - "elfso bug" fix - applied to aoutb format
448 as well - testing might be desirable...
454 * James Seter - -postfix, -prefix command line switches.
455 * Yuri Zaporogets - rdoff utility changes.
461 * GAS-like palign (Panos Minos)
462 * FIXME: Someone, fill this in with details
466 ----------------------------------
468 * Fixed - elf and aoutb bug - shared libraries
469 - multiple "%include" bug in "-f obj"
471 - unrecognized option bug in ndisasm
473 0.98.03 with John Coffman's changes released 27-Jul-2000
474 --------------------------------------------------------
476 * Added signed byte optimizations for the 0x81/0x83 class
477 of instructions: ADC, ADD, AND, CMP, OR, SBB, SUB, XOR:
478 when used as 'ADD reg16,imm' or 'ADD reg32,imm.' Also
479 optimization of signed byte form of 'PUSH imm' and 'IMUL
480 reg,imm'/'IMUL reg,reg,imm.' No size specification is needed.
482 * Added multi-pass JMP and Jcc offset optimization. Offsets
483 on forward references will preferentially use the short form,
484 without the need to code a specific size (short or near) for
485 the branch. Added instructions for 'Jcc label' to use the
486 form 'Jnotcc $+3/JMP label', in cases where a short offset
487 is out of bounds. If compiling for a 386 or higher CPU, then
488 the 386 form of Jcc will be used instead.
490 This feature is controlled by a new command-line switch: "O",
491 (upper case letter O). "-O0" reverts the assembler to no
492 extra optimization passes, "-O1" allows up to 5 extra passes,
493 and "-O2"(default), allows up to 10 extra optimization passes.
495 * Added a new directive: 'cpu XXX', where XXX is any of:
496 8086, 186, 286, 386, 486, 586, pentium, 686, PPro, P2, P3 or
497 Katmai. All are case insensitive. All instructions will
498 be selected only if they apply to the selected cpu or lower.
499 Corrected a couple of bugs in cpu-dependence in 'insns.dat'.
501 * Added to 'standard.mac', the "use16" and "use32" forms of
502 the "bits 16/32" directive. This is nothing new, just conforms
503 to a lot of other assemblers. (minor)
505 * Changed label allocation from 320/32 (10000 labels @ 200K+)
506 to 32/37 (1000 labels); makes running under DOS much easier.
507 Since additional label space is allocated dynamically, this
508 should have no effect on large programs with lots of labels.
509 The 37 is a prime, believed to be better for hashing. (minor)
511 * Integrated patchfile 0.98-0.98.01. I call this version
512 0.98.03, for historical reasons: 0.98.02 was trashed.
514 --John Coffman <johninsd@san.rr.com> 27-Jul-2000
517 Kendall Bennett's SciTech MGL changes
518 -------------------------------------
519 Note that you must define "TASM_COMPAT" at compile-time
520 to get the Tasm Ideal Mode compatibility.
522 All changes can be compiled in and out using the TASM_COMPAT macros,
523 and when compiled without TASM_COMPAT defined we get the exact same
524 binary as the unmodified 0.98 sources.
526 standard.mac, macros.c:
527 . Added macros to ignore TASM directives before first include
530 . Added extern declaration for tasm_compatible_mode
533 . Added global variable tasm_compatible_mode
534 . Added command line switch for TASM compatible mode (-t)
535 . Changed version command line to reflect when compiled with TASM additions
536 . Added response file processing to allow all arguments on a single
537 line (response file is @resp rather than -@resp for NASM format).
540 . Changes islocal() macro to support TASM style @@local labels.
541 . Added islocalchar() macro to support TASM style @@local labels.
544 . Added support for TASM style memory references (ie: mov [DWORD eax],10
545 rather than the NASM style mov DWORD [eax],10).
548 . Added new directives, %arg, %local, %stacksize to directives table
549 . Added support for TASM style directives without a leading % symbol.
551 Integrated a block of changes from Andrew Zabolotny <bit@eltech.ru>:
553 * A new keyword %xdefine and its case-insensitive counterpart %ixdefine.
554 They work almost the same way as %define and %idefine but expand
555 the definition immediately, not on the invocation. Something like a cross
556 between %define and %assign. The "x" suffix stands for "eXpand", so
557 "xdefine" can be deciphered as "expand-and-define". Thus you can do
563 %xdefine %1 dword [esp+ofs]
567 * Changed the place where the expansion of %$name macros are expanded.
568 Now they are converted into ..@ctxnum.name form when detokenizing, so
569 there are no quirks as before when using %$name arguments to macros,
570 in macros etc. For example:
579 Now last line will be expanded into "hello" as expected. This also allows
580 for lots of goodies, a good example are extended "proc" macros included
583 * Added a check for "cstk" in smacro_defined() before calling get_ctx() -
584 this allows for things like:
589 to work without warnings even in no context.
591 * Added a check for "cstk" in %if*ctx and %elif*ctx directives -
592 this allows to use %ifctx without excessive warnings. If there is
593 no active context, %ifctx goes through "false" branch.
595 * Removed "user error: " prefix with %error directive: it just clobbers the
596 output and has absolutely no functionality. Besides, this allows to write
597 macros that does not differ from built-in functions in any way.
599 * Added expansion of string that is output by %error directive. Now you
602 %define hello(x) Hello, x!
605 %error "hello(%$name)"
607 Same happened with %include directive.
609 * Now all directives that expect an identifier will try to expand and
610 concatenate everything without whitespaces in between before usage.
611 For example, with "unfixed" nasm the commands
614 %define __%$abc goodbye
617 would produce "incorrect" output: last line will expand to
621 Not quite what you expected, eh? :-) The answer is that preprocessor
622 treats the %define construct as if it would be
624 %define __ %$abc goodbye
626 (note the white space between __ and %$abc). After my "fix" it
627 will "correctly" expand into
631 as expected. Note that I use quotes around words "correct", "incorrect"
632 etc because this is rather a feature not a bug; however current behaviour
633 is more logical (and allows more advanced macro usage :-).
635 Same change was applied to:
636 %push,%macro,%imacro,%define,%idefine,%xdefine,%ixdefine,
637 %assign,%iassign,%undef
639 * A new directive [WARNING {+|-}warning-id] have been added. It works only
640 if the assembly phase is enabled (i.e. it doesn't work with nasm -e).
642 * A new warning type: macro-selfref. By default this warning is disabled;
643 when enabled NASM warns when a macro self-references itself; for example
644 the following source:
646 [WARNING macro-selfref]
657 will produce a warning, but if we remove the first line we won't see it
658 anymore (which is The Right Thing To Do {tm} IMHO since C preprocessor
659 eats such constructs without warnings at all).
661 * Added a "error" routine to preprocessor which always will set ERR_PASS1
662 bit in severity_code. This removes annoying repeated errors on first
663 and second passes from preprocessor.
665 * Added the %+ operator in single-line macros for concatenating two
666 identifiers. Usage example:
668 %define _myfunc _otherfunc
669 %define cextern(x) _ %+ x
672 After first expansion, third line will become "_myfunc". After this
673 expansion is performed again so it becomes "_otherunc".
675 * Now if preprocessor is in a non-emmitting state, no warning or error
676 will be emmitted. Example:
681 put anything you want between these two brackets,
682 even macro-parameter references %1 or local labels %$zz
683 or macro-local labels %%zz - no warning will be emmitted.
686 * Context-local variables on expansion as a last resort are looked up
687 in outer contexts. For example, the following piece:
697 will expand correctly the fourth line to [esp]; if we'll define another
698 %$a inside the "inner" context, it will take precedence over outer
699 definition. However, this modification has been applied only to
700 expand_smacro and not to smacro_define: as a consequence expansion
701 looks in outer contexts, but %ifdef won't look in outer contexts.
703 This behaviour is needed because we don't want nested contexts to
704 act on already defined local macros. Example:
706 %define %$arg1 [esp+4]
712 In this example the "if" mmacro enters into the "if" context, so %$arg1
713 is not valid anymore inside "if". Of course it could be worked around
714 by using explicitely %$$arg1 but this is ugly IMHO.
716 * Fixed memory leak in %undef. The origline wasn't freed before
719 * Fixed trap in preprocessor when line expanded to empty set of tokens.
720 This happens, for example, in the following case:
729 All changes since NASM 0.98p3 have been produced by H. Peter Anvin <hpa@zytor.com>.
731 * The documentation comment delimiter is \# not #.
732 * Allow EQU definitions to refer to external labels; reported by
734 * Re-enable support for RDOFF v1; reported by Pedro Gimeno.
735 * Updated License file per OK from Simon and Julian.
741 * Update documentation (although the instruction set reference will
742 have to wait; I don't want to hold up the 0.98 release for it.)
743 * Verified that the NASM implementation of the PEXTRW and PMOVMSKB
744 instructions is correct. The encoding differs from what the Intel
745 manuals document, but the Pentium III behaviour matches NASM, not
747 * Fix handling of implicit sizes in PSHUFW and PINSRW, reported by
749 * Resurrect the -s option, which was removed when changing the
750 diagnostic output to stdout.
756 * Fix for "DB" when NASM is running on a bigendian machine.
757 * Invoke insns.pl once for each output script, making Makefile.in
759 * Improve the Unix configure-based makefiles to make package
761 * Included an RPM .spec file for building RPM (RedHat Package Manager)
762 packages on Linux or Unix systems.
763 * Fix Makefile dependency problems.
764 * Change src/rdsrc.pl to include sectioning information in info
765 output; required for install-info to work.
766 * Updated the RDOFF distribution to version 2 from Jules; minor
767 massaging to make it compile in my environment.
768 * Split doc files that can be built by anyone with a Perl interpreter off
769 into a separate archive.
770 * "Dress rehearsal" release!
776 * Fixed opcodes with a third byte-sized immediate argument to not
777 complain if given "byte" on the immediate.
778 * Allow %undef to remove single-line macros with arguments. This
779 matches the behaviour of #undef in the C preprocessor.
780 * Allow -d, -u, -i and -p to be specified as -D, -U, -I and -P for
781 compatibility with most C compilers and preprocessors. This allows
782 Makefile options to be shared between cc and nasm, for example.
784 * Went through the list of Katmai instructions and hopefully fixed the
785 (rather few) mistakes in it.
786 * (Hopefully) fixed a number of disassembler bugs related to ambiguous
787 instructions (disambiguated by -p) and SSE instructions with REP.
788 * Fix for bug reported by Mark Junger: "call dword 0x12345678" should
789 work and may add an OSP (affected CALL, JMP, Jcc).
790 * Fix for environments when "stderr" isn't a compile-time constant.
796 * Took officially over coordination of the 0.98 release; so drop
797 the p3.x notation. Skipped p4 and p5 to avoid confusion with John
798 Fine's J4 and J5 releases.
799 * Update the documentation; however, it still doesn't include
800 documentation for the various new instructions. I somehow wonder if
801 it makes sense to have an instruction set reference in the assembler
802 manual when Intel et al have PDF versions of their manuals online.
803 * Recognize "idt" or "centaur" for the -p option to ndisasm.
804 * Changed error messages back to stderr where they belong, but add an
805 -E option to redirect them elsewhere (the DOS shell cannot redirect
807 * -M option to generate Makefile dependencies (based on code from Alex
809 * %undef preprocessor directive, and -u option, that undefines a
811 * OS/2 Makefile (Mkfiles/Makefile.os2) for Borland under OS/2; from
813 * Various minor bugfixes (reported by):
814 - Dangling %s in preproc.c (Martin Junker)
815 * THERE ARE KNOWN BUGS IN SSE AND THE OTHER KATMAI INSTRUCTIONS. I am
816 on a trip and didn't bring the Katmai instruction reference, so I
817 can't work on them right now.
818 * Updated the License file per agreement with Simon and Jules to
819 include a GPL distribution clause.
825 * (Hopefully) fixed the canned Makefiles to include the outrdf2 and
827 * Renamed changes.asm to changed.asm.
833 * Fixed a bunch of instructions that were added in 0.98p3.5 which had
834 memory operands, and the address-size prefix was missing from the
841 * Merged in changes from John S. Fine's 0.98-J5 release. John's based
842 0.98-J5 on my 0.98p3.3 release; this merges the changes.
843 * Expanded the instructions flag field to a long so we can fit more
844 flags; mark SSE (KNI) and AMD or Katmai-specific instructions as
846 * Fix the "PRIV" flag on a bunch of instructions, and create new
847 "PROT" flag for protected-mode-only instructions (orthogonal to if
848 the instruction is privileged!) and new "SMM" flag for SMM-only
850 * Added AMD-only SYSCALL and SYSRET instructions.
851 * Make SSE actually work, and add new Katmai MMX instructions.
852 * Added a -p (preferred vendor) option to ndisasm so that it can
853 distinguish e.g. Cyrix opcodes also used in SSE. For example:
855 ndisasm -p cyrix aliased.bin
856 00000000 670F514310 paddsiw mm0,[ebx+0x10]
857 00000005 670F514320 paddsiw mm0,[ebx+0x20]
858 ndisasm -p intel aliased.bin
859 00000000 670F514310 sqrtps xmm0,[ebx+0x10]
860 00000005 670F514320 sqrtps xmm0,[ebx+0x20]
861 * Added a bunch of Cyrix-specific instructions.
867 * Made at least an attempt to modify all the additional Makefiles (in
868 the Mkfiles directory). I can't test it, but this was the best I
870 * DOS DJGPP+"Opus Make" Makefile from John S. Fine.
871 * changes.asm changes from John S. Fine.
877 * Patch from Conan Brink to allow nesting of %rep directives.
878 * If we're going to allow INT01 as an alias for INT1/ICEBP (one of
879 Jules 0.98p3 changes), then we should allow INT03 as an alias for INT3
881 * Updated changes.asm to include the latest changes.
882 * Tried to clean up the <CR>s that had snuck in from a DOS/Windows
883 environment into my Unix environment, and try to make sure than
884 DOS/Windows users get them back.
885 * We would silently generate broken tools if insns.dat wasn't sorted
886 properly. Change insns.pl so that the order doesn't matter.
887 * Fix bug in insns.pl (introduced by me) which would cause conditional
888 instructions to have an extra "cc" in disassembly, e.g. "jnz"
889 disassembled as "jccnz".
895 * Merged in John S. Fine's changes from his 0.98-J4 prerelease; see
896 http://www.csoft.net/cz/johnfine/
897 * Changed previous "spotless" Makefile target (appropriate for distribution)
898 to "distclean", and added "cleaner" target which is same as "clean"
899 except deletes files generated by Perl scripts; "spotless" is union.
900 * Removed BASIC programs from distribution. Get a Perl interpreter
902 * Calling this "pre-release 3.2" rather than "p3-hpa2" because of
903 John's contributions.
904 * Actually link in the IEEE output format (zoutieee.c); fix a bunch of
905 compiler warnings in that file. Note I don't know what IEEE output
906 is supposed to look like, so these changes were made "blind".
912 * Merged nasm098p3.zip with nasm-0.97.tar.gz to create a fully
913 buildable version for Unix systems (Makefile.in updates, etc.)
914 * Changed insns.pl to create the instruction tables in nasm.h and
915 names.c, so that a new instruction can be added by adding it *only*
917 * Added the following new instructions: SYSENTER, SYSEXIT, FXSAVE,
918 FXRSTOR, UD1, UD2 (the latter two are two opcodes that Intel
919 guarantee will never be used; one of them is documented as UD2 in
920 Intel documentation, the other one just as "Undefined Opcode" --
921 calling it UD1 seemed to make sense.)
922 * MAX_SYMBOL was defined to be 9, but LOADALL286 and LOADALL386 are 10
923 characters long. Now MAX_SYMBOL is derived from insns.dat.
924 * A note on the BASIC programs included: forget them. insns.bas is
925 already out of date. Get yourself a Perl interpreter for your
926 platform of choice at:
928 http://www.cpan.org/ports/index.html
934 added response file support, improved command line handling, new layout
937 fixed limit checking bug, 'OUT byte nn, reg' bug, and a couple of rdoff
938 related bugs, updated Wishlist; 0.98 Prerelease 3.
944 fixed bug in outcoff.c to do with truncating section names longer
945 than 8 characters, referencing beyond end of string; 0.98 pre-release 2
948 0.98 pre-released May 1999
949 --------------------------
951 Fixed a bug whereby STRUC didn't work at all in RDF.
953 Fixed a problem with group specification in PUBDEFs in OBJ.
955 Improved ease of adding new output formats. Contribution due to
958 Fixed a bug in relocations in the `bin' format: was showing up when
959 a relocatable reference crossed an 8192-byte boundary in any output
962 Fixed a bug in local labels: local-label lookups were inconsistent
963 between passes one and two if an EQU occurred between the definition
964 of a global label and the subsequent use of a local label local to
967 Fixed a seg-fault in the preprocessor (again) which happened when
968 you use a blank line as the first line of a multi-line macro
969 definition and then defined a label on the same line as a call to
972 Fixed a stale-pointer bug in the handling of the NASM environment
973 variable. Thanks to Thomas McWilliams.
975 ELF had a hard limit on the number of sections which caused
976 segfaults when transgressed. Fixed.
978 Added ability for ndisasm to read from stdin by using `-' as the
981 ndisasm wasn't outputting the TO keyword. Fixed.
983 Fixed error cascade on bogus expression in %if - an error in
984 evaluation was causing the entire %if to be discarded, thus creating
985 trouble later when the %else or %endif was encountered.
987 Forward reference tracking was instruction-granular not operand-
988 granular, which was causing 286-specific code to be generated
989 needlessly on code of the form `shr word [forwardref],1'. Thanks to
990 Jim Hague for sending a patch.
992 All messages now appear on stdout, as sending them to stderr serves
993 no useful purpose other than to make redirection difficult.
995 Fixed the problem with EQUs pointing to an external symbol - this
996 now generates an error message.
998 Allowed multiple size prefixes to an operand, of which only the first
999 is taken into account.
1001 Incorporated John Fine's changes, including fixes of a large number
1002 of preprocessor bugs, some small problems in OBJ, and a reworking of
1003 label handling to define labels before their line is assembled, rather
1006 Reformatted a lot of the source code to be more readable. Included
1007 'coding.txt' as a guideline for how to format code for contributors.
1009 Stopped nested %reps causing a panic - they now cause a slightly more
1010 friendly error message instead.
1012 Fixed floating point constant problems (patch by Pedro Gimeno)
1014 Fixed the return value of insn_size() not being checked for -1, indicating
1017 Incorporated 3D now instructions.
1019 Fixed the 'mov eax, eax + ebx' bug.
1021 Fixed the GLOBAL EQU bug in ELF. Released developers release 3.
1023 Incorporated John Fine's command line parsing changes
1025 Incorporated David Lindauer's OMF debug support
1027 Made changes for LCC 4.0 support (__NASM_CDecl__, removed register size
1028 specification warning when sizes agree).
1030 Released NASM 0.98 Pre-release 1
1033 0.97 released December 1997
1034 ---------------------------
1036 This was entirely a bug-fix release to 0.96, which seems to have got
1039 Fixed stupid mistake in OBJ which caused `MOV EAX,<constant>' to
1040 fail. Caused by an error in the `MOV EAX,<segment>' support.
1042 ndisasm hung at EOF when compiled with lcc on Linux because lcc on
1043 Linux somehow breaks feof(). ndisasm now does not rely on feof().
1045 A heading in the documentation was missing due to a markup error in
1046 the indexing. Fixed.
1048 Fixed failure to update all pointers on realloc() within extended-
1049 operand code in parser.c. Was causing wrong behaviour and seg faults
1050 on lines such as `dd 0.0,0.0,0.0,0.0,...'
1052 Fixed a subtle preprocessor bug whereby invoking one multi-line
1053 macro on the first line of the expansion of another, when the second
1054 had been invoked with a label defined before it, didn't expand the
1057 Added internal.doc back in to the distribution archives - it was
1058 missing in 0.96 *blush*
1060 Fixed bug causing 0.96 to be unable to assemble its own test files,
1061 specifically objtest.asm. *blush again*
1063 Fixed seg-faults and bogus error messages caused by mismatching
1064 %rep and %endrep within multi-line macro definitions.
1066 Fixed a problem with buffer overrun in OBJ, which was causing
1067 corruption at ends of long PUBDEF records.
1069 Separated DOS archives into main-program and documentation to reduce
1073 0.96 released November 1997
1074 ---------------------------
1076 Fixed a bug whereby, if `nasm sourcefile' would cause a filename
1077 collision warning and put output into `nasm.out', then `nasm
1078 sourcefile -o outputfile' still gave the warning even though the
1081 Fixed name pollution under Digital UNIX: one of its header files
1082 defined R_SP, which broke the enum in nasm.h.
1084 Fixed minor instruction table problems: FUCOM and FUCOMP didn't have
1085 two-operand forms; NDISASM didn't recognise the longer register
1086 forms of PUSH and POP (eg FF F3 for PUSH BX); TEST mem,imm32 was
1087 flagged as undocumented; the 32-bit forms of CMOV had 16-bit operand
1088 size prefixes; `AAD imm' and `AAM imm' are no longer flagged as
1089 undocumented because the Intel Architecture reference documents
1092 Fixed a problem with the local-label mechanism, whereby strange
1093 types of symbol (EQUs, auto-defined OBJ segment base symbols)
1094 interfered with the `previous global label' value and screwed up
1097 Fixed a bug whereby the stub preprocessor didn't communicate with
1098 the listing file generator, so that the -a and -l options in
1099 conjunction would produce a useless listing file.
1101 Merged `os2' object file format back into `obj', after discovering
1102 that `obj' _also_ shouldn't have a link pass separator in a module
1103 containing a non-trivial MODEND. Flat segments are now declared
1104 using the FLAT attribute. `os2' is no longer a valid object format
1107 Removed the fixed-size temporary storage in the evaluator. Very very
1108 long expressions (like `mov ax,1+1+1+1+...' for two hundred 1s or
1109 so) should now no longer crash NASM.
1111 Fixed a bug involving segfaults on disassembly of MMX instructions,
1112 by changing the meaning of one of the operand-type flags in nasm.h.
1113 This may cause other apparently unrelated MMX problems; it needs to
1114 be tested thoroughly.
1116 Fixed some buffer overrun problems with large OBJ output files.
1117 Thanks to DJ Delorie for the bug report and fix.
1119 Made preprocess-only mode actually listen to the %line markers as it
1120 prints them, so that it can report errors more sanely.
1122 Re-designed the evaluator to keep more sensible track of expressions
1123 involving forward references: can now cope with previously-nightmare
1129 Added the ALIGN and ALIGNB standard macros.
1131 Added PIC support in ELF: use of WRT to obtain the four extra
1132 relocation types needed.
1134 Added the ability for output file formats to define their own
1135 extensions to the GLOBAL, COMMON and EXTERN directives.
1137 Implemented common-variable alignment, and global-symbol type and
1138 size declarations, in ELF.
1140 Implemented NEAR and FAR keywords for common variables, plus
1141 far-common element size specification, in OBJ.
1143 Added a feature whereby EXTERNs and COMMONs in OBJ can be given a
1144 default WRT specification (either a segment or a group).
1146 Transformed the Unix NASM archive into an auto-configuring package.
1148 Added a sanity-check for people applying SEG to things which are
1149 already segment bases: this previously went unnoticed by the SEG
1150 processing and caused OBJ-driver panics later.
1152 Added the ability, in OBJ format, to deal with `MOV EAX,<segment>'
1153 type references: OBJ doesn't directly support dword-size segment
1154 base fixups, but as long as the low two bytes of the constant term
1155 are zero, a word-size fixup can be generated instead and it will
1158 Added the ability to specify sections' alignment requirements in
1159 Win32 object files and pure binary files.
1161 Added preprocess-time expression evaluation: the %assign (and
1162 %iassign) directive and the bare %if (and %elif) conditional. Added
1163 relational operators to the evaluator, for use only in %if
1164 constructs: the standard relationals = < > <= >= <> (and C-like
1165 synonyms == and !=) plus low-precedence logical operators &&, ^^ and
1168 Added a preprocessor repeat construct: %rep / %exitrep / %endrep.
1170 Added the __FILE__ and __LINE__ standard macros.
1172 Added a sanity check for number constants being greater than
1173 0xFFFFFFFF. The warning can be disabled.
1175 Added the %0 token whereby a variadic multi-line macro can tell how
1176 many parameters it's been given in a specific invocation.
1178 Added %rotate, allowing multi-line macro parameters to be cycled.
1180 Added the `*' option for the maximum parameter count on multi-line
1181 macros, allowing them to take arbitrarily many parameters.
1183 Added the ability for the user-level forms of EXTERN, GLOBAL and
1184 COMMON to take more than one argument.
1186 Added the IMPORT and EXPORT directives in OBJ format, to deal with
1189 Added some more preprocessor %if constructs: %ifidn / %ifidni (exact
1190 textual identity), and %ifid / %ifnum / %ifstr (token type testing).
1192 Added the ability to distinguish SHL AX,1 (the 8086 version) from
1193 SHL AX,BYTE 1 (the 286-and-upwards version whose constant happens to
1196 Added NetBSD/FreeBSD/OpenBSD's variant of a.out format, complete
1197 with PIC shared library features.
1199 Changed NASM's idiosyncratic handling of FCLEX, FDISI, FENI, FINIT,
1200 FSAVE, FSTCW, FSTENV, and FSTSW to bring it into line with the
1201 otherwise accepted standard. The previous behaviour, though it was a
1202 deliberate feature, was a deliberate feature based on a
1203 misunderstanding. Apologies for the inconvenience.
1205 Improved the flexibility of ABSOLUTE: you can now give it an
1206 expression rather than being restricted to a constant, and it can
1207 take relocatable arguments as well.
1209 Added the ability for a variable to be declared as EXTERN multiple
1210 times, and the subsequent definitions are just ignored.
1212 We now allow instruction prefixes (CS, DS, LOCK, REPZ etc) to be
1213 alone on a line (without a following instruction).
1215 Improved sanity checks on whether the arguments to EXTERN, GLOBAL
1216 and COMMON are valid identifiers.
1218 Added misc/exebin.mac to allow direct generation of .EXE files by
1219 hacking up an EXE header using DB and DW; also added test/binexe.asm
1220 to demonstrate the use of this. Thanks to Yann Guidon for
1221 contributing the EXE header code.
1223 ndisasm forgot to check whether the input file had been successfully
1224 opened. Now it does. Doh!
1226 Added the Cyrix extensions to the MMX instruction set.
1228 Added a hinting mechanism to allow [EAX+EBX] and [EBX+EAX] to be
1229 assembled differently. This is important since [ESI+EBP] and
1230 [EBP+ESI] have different default base segment registers.
1232 Added support for the PharLap OMF extension for 4096-byte segment
1236 0.95 released July 1997
1237 -----------------------
1239 Fixed yet another ELF bug. This one manifested if the user relied on
1240 the default segment, and attempted to define global symbols without
1241 first explicitly declaring the target segment.
1243 Added makefiles (for NASM and the RDF tools) to build Win32 console
1244 apps under Symantec C++. Donated by Mark Junker.
1246 Added `macros.bas' and `insns.bas', QBasic versions of the Perl
1247 scripts that convert `standard.mac' to `macros.c' and convert
1248 `insns.dat' to `insnsa.c' and `insnsd.c'. Also thanks to Mark
1251 Changed the diassembled forms of the conditional instructions so
1252 that JB is now emitted as JC, and other similar changes. Suggested
1253 list by Ulrich Doewich.
1255 Added `@' to the list of valid characters to begin an identifier
1258 Documentary changes, notably the addition of the `Common Problems'
1259 section in nasm.doc.
1261 Fixed a bug relating to 32-bit PC-relative fixups in OBJ.
1263 Fixed a bug in perm_copy() in labels.c which was causing exceptions
1264 in cleanup_labels() on some systems.
1266 Positivity sanity check in TIMES argument changed from a warning to
1267 an error following a further complaint.
1269 Changed the acceptable limits on byte and word operands to allow
1270 things like `~10111001b' to work.
1272 Fixed a major problem in the preprocessor which caused seg-faults if
1273 macro definitions contained blank lines or comment-only lines.
1275 Fixed inadequate error checking on the commas separating the
1276 arguments to `db', `dw' etc.
1278 Fixed a crippling bug in the handling of macros with operand counts
1279 defined with a `+' modifier.
1281 Fixed a bug whereby object file formats which stored the input file
1282 name in the output file (such as OBJ and COFF) weren't doing so
1283 correctly when the output file name was specified on the command
1286 Removed [INC] and [INCLUDE] support for good, since they were
1289 Fixed a bug in OBJ which caused all fixups to be output in 16-bit
1290 (old-format) FIXUPP records, rather than putting the 32-bit ones in
1291 FIXUPP32 (new-format) records.
1293 Added, tentatively, OS/2 object file support (as a minor variant on
1296 Updates to Fox Cutter's Borland C makefile, Makefile.bc2.
1298 Removed a spurious second fclose() on the output file.
1300 Added the `-s' command line option to redirect all messages which
1301 would go to stderr (errors, help text) to stdout instead.
1303 Added the `-w' command line option to selectively suppress some
1304 classes of assembly warning messages.
1306 Added the `-p' pre-include and `-d' pre-define command-line options.
1308 Added an include file search path: the `-i' command line option.
1310 Fixed a silly little preprocessor bug whereby starting a line with a
1311 `%!' environment-variable reference caused an `unknown directive'
1314 Added the long-awaited listing file support: the `-l' command line
1317 Fixed a problem with OBJ format whereby, in the absence of any
1318 explicit segment definition, non-global symbols declared in the
1319 implicit default segment generated spurious EXTDEF records in the
1322 Added the NASM environment variable.
1324 From this version forward, Win32 console-mode binaries will be
1325 included in the DOS distribution in addition to the 16-bit binaries.
1326 Added Makefile.vc for this purpose.
1328 Added `return 0;' to test/objlink.c to prevent compiler warnings.
1330 Added the __NASM_MAJOR__ and __NASM_MINOR__ standard defines.
1332 Added an alternative memory-reference syntax in which prefixing an
1333 operand with `&' is equivalent to enclosing it in square brackets,
1334 at the request of Fox Cutter.
1336 Errors in pass two now cause the program to return a non-zero error
1337 code, which they didn't before.
1339 Fixed the single-line macro cycle detection, which didn't work at
1340 all on macros with no parameters (caused an infinite loop). Also
1341 changed the behaviour of single-line macro cycle detection to work
1342 like cpp, so that macros like `extrn' as given in the documentation
1345 Fixed the implementation of WRT, which was too restrictive in that
1346 you couldn't do `mov ax,[di+abc wrt dgroup]' because (di+abc) wasn't
1347 a relocatable reference.
1350 0.94 released April 1997
1351 ------------------------
1353 Major item: added the macro processor.
1355 Added undocumented instructions SMI, IBTS, XBTS and LOADALL286. Also
1356 reorganised CMPXCHG instruction into early-486 and Pentium forms.
1357 Thanks to Thobias Jones for the information.
1359 Fixed two more stupid bugs in ELF, which were causing `ld' to
1360 continue to seg-fault in a lot of non-trivial cases.
1362 Fixed a seg-fault in the label manager.
1364 Stopped FBLD and FBSTP from _requiring_ the TWORD keyword, which is
1365 the only option for BCD loads/stores in any case.
1367 Ensured FLDCW, FSTCW and FSTSW can cope with the WORD keyword, if
1368 anyone bothers to provide it. Previously they complained unless no
1369 keyword at all was present.
1371 Some forms of FDIV/FDIVR and FSUB/FSUBR were still inverted: a
1372 vestige of a bug that I thought had been fixed in 0.92. This was
1373 fixed, hopefully for good this time...
1375 Another minor phase error (insofar as a phase error can _ever_ be
1376 minor) fixed, this one occurring in code of the form
1377 rol ax,forward_reference
1378 forward_reference equ 1
1380 The number supplied to TIMES is now sanity-checked for positivity,
1381 and also may be greater than 64K (which previously didn't work on
1384 Added Watcom C makefiles, and misc/pmw.bat, donated by Dominik Behr.
1386 Added the INCBIN pseudo-opcode.
1388 Due to the advent of the preprocessor, the [INCLUDE] and [INC]
1389 directives have become obsolete. They are still supported in this
1390 version, with a warning, but won't be in the next.
1392 Fixed a bug in OBJ format, which caused incorrect object records to
1393 be output when absolute labels were made global.
1395 Updates to RDOFF subdirectory, and changes to outrdf.c.
1398 0.93 released January 1997
1399 --------------------------
1401 This release went out in a great hurry after semi-crippling bugs
1404 Really _did_ fix the stack overflows this time. *blush*
1406 Had problems with EA instruction sizes changing between passes, when
1407 an offset contained a forward reference and so 4 bytes were
1408 allocated for the offset in pass one; by pass two the symbol had
1409 been defined and happened to be a small absolute value, so only 1
1410 byte got allocated, causing instruction size mismatch between passes
1411 and hence incorrect address calculations. Fixed.
1413 Stupid bug in the revised ELF section generation fixed (associated
1414 string-table section for .symtab was hard-coded as 7, even when this
1415 didn't fit with the real section table). Was causing `ld' to
1416 seg-fault under Linux.
1418 Included a new Borland C makefile, Makefile.bc2, donated by Fox
1419 Cutter <lmb@comtch.iea.com>.
1422 0.92 released January 1997
1423 --------------------------
1425 The FDIVP/FDIVRP and FSUBP/FSUBRP pairs had been inverted: this was
1426 fixed. This also affected the LCC driver.
1428 Fixed a bug regarding 32-bit effective addresses of the form
1429 [other_register+ESP].
1431 Documentary changes, notably documentation of the fact that Borland
1432 Win32 compilers use `obj' rather than `win32' object format.
1434 Fixed the COMENT record in OBJ files, which was formatted
1437 Fixed a bug causing segfaults in large RDF files.
1439 OBJ format now strips initial periods from segment and group
1440 definitions, in order to avoid complications with the local label
1443 Fixed a bug in disassembling far calls and jumps in NDISASM.
1445 Added support for user-defined sections in COFF and ELF files.
1447 Compiled the DOS binaries with a sensible amount of stack, to
1448 prevent stack overflows on any arithmetic expression containing
1451 Fixed a bug in handling of files that do not terminate in a newline.
1454 0.91 released November 1996
1455 ---------------------------
1458 Support for RDF added.
1459 Support for DBG debugging format added.
1460 Support for 32-bit extensions to Microsoft OBJ format added.
1461 Revised for Borland C: some variable names changed, makefile added.
1462 LCC support revised to actually work.
1463 JMP/CALL NEAR/FAR notation added.
1464 `a16', `o16', `a32' and `o32' prefixes added.
1465 Range checking on short jumps implemented.
1466 MMX instruction support added.
1467 Negative floating point constant support added.
1468 Memory handling improved to bypass 64K barrier under DOS.
1469 $ prefix to force treatment of reserved words as identifiers added.
1470 Default-size mechanism for object formats added.
1471 Compile-time configurability added.
1472 `#', `@', `~' and `?' are now valid characters in labels.
1473 `-e' and `-k' options in NDISASM added.
1476 0.90 released October 1996
1477 --------------------------
1479 First release version. First support for object file output. Other
1480 changes from previous version (0.3x) too numerous to document.
projects / platform / upstream / nasm.git / blob