2 \# NASM revision history in nasmdoc format
5 \H{cl-2.xx} NASM 2 Series
7 The NASM 2 series supports x86-64, and is the production version of NASM
10 \S{cl-2.13} Version 2.13
12 \b Support the official forms of the \c{UD0} and \c{UD1} instructions.
14 \b Allow self-segment-relative expressions in immediates and
15 displacements, even when combined with an external or otherwise
16 out-of-segment special symbol, e.g.:
19 \c mov eax,[foo - $ + ebx] ; Now legal
21 \b Handle a 64-bit origin in NDISASM.
23 \b NASM can now generate sparse output files for relevant output
24 formats, if the underlying operating system supports them.
26 \b The \c{macho} object format now supports the \c{subsections_via_symbols}
27 and \c{no_dead_strip} directives, see \k{macho-ssvs}.
29 \b The \c{macho} object format now supports the \c{no_dead_strip},
30 \c{live_support} and \c{strip_static_syms} section flags, see
33 \b The \c{macho} object format now supports the \c{dwarf} debugging
34 format, as required by newer toolchains.
36 \b All warnings can now be suppressed if desired; warnings not
37 otherwise part of any warning class are now considered its own
38 warning class called \c{other} (e.g. \c{-w-other}). Furthermore,
39 warning-as-error can now be controlled on a per warning class
40 basis, using the syntax \c{-w+error=}\e{warning-class} and its
41 equivalent for all other warning control options. See \k{opt-w}
42 for the command-line options and warning classes and
43 \k{asmdir-warning} for the \c{[WARNING]} directive.
45 \b Fix a number of bugs related to AVX-512 decorators.
47 \b Significant improvements to building NASM with Microsoft Visual
48 Studio via \c{Mkfiles/msvc.mak}. It is now possible to build the
49 full Windows installer binary as long as the necessary
50 prerequisites are installed; see \c{Mkfiles/README}
52 \b To build NASM with custom modifications (table changes) or from the
53 git tree now requires Perl 5.8 at the very minimum, quite possibly
54 a higher version (Perl 5.24.1 tested.) There is no requirement to
55 have Perl on your system at all if all you want to do is build
56 unmodified NASM from source archives.
58 \b Fix the \c{\{z\}} decorator on AVX-512 \c{VMOVDQ*} instructions.
60 \b Add new warnings for certain dangerous constructs which never ought
61 to have been allowed. In particular, the \c{RESB} family of
62 instructions should have been taking a critical expression all
65 \b Fix the EVEX (AVX-512) versions of the \c{VPBROADCAST}, \c{VPEXTR},
66 and \c{VPINSR} instructions.
68 \b Support contracted forms of additional instructions. As a general
69 rule, if an instruction has a non-destructive source immediately
70 after a destination register that isn't used as an input, NASM
71 supports omitting that source register, using the destination
72 register as that value. This among other things makes it easier to
73 convert SSE code to the equivalent AVX code:
75 \c addps xmm1,xmm0 ; SSE instruction
76 \c vaddps ymm1,ymm1,ymm0 ; AVX official long form
77 \c vaddps ymm1,ymm0 ; AVX contracted form
79 \b Fix Codeview malformed compiler version record.
81 \b Add the \c{CLWB} and \c{PCOMMIT} instructions. Note that the
82 \c{PCOMMIT} instruction has been deprecated and will never be
83 included in a shipping product; it is included for completeness
86 \b Add the \c{%pragma} preprocessor directive for soft-error directives.
88 \b Add the \c{RDPID} instruction.
90 \S{cl-2.12.02} Version 2.12.02
92 \b Fix preprocessor errors, especially \c{%error} and \c{%warning},
93 inside \c{%if} statements.
95 \b Fix relative relocations in 32-bit Mach-O.
97 \b More Codeview debug format fixes.
99 \b If the MASM \c{PTR} keyword is encountered, issue a warning. This is
100 much more likely to indicate a MASM-ism encountered in NASM than it
101 is a valid label. This warning can be suppressed with \c{-w-ptr},
102 the \c{[warning]} directive (see \k{opt-w}) or by the macro
103 definition \c{%idefine ptr $%?} (see \k{selfref%?}).
105 \b When an error or a warning comes from the expansion of a multi-line
106 macro, display the file and line numbers for the expanded macros.
107 Macros defined with \c{.nolist} do not get displayed.
109 \b Add macros \c{ilog2fw()} and \c{ilog2cw()} to the \c{ifunc} macro
110 package. See \k{ilog2}.
113 \S{cl-2.12.01} Version 2.12.01
115 \b Portability fixes for some platforms.
117 \b Fix error when not specifying a list file.
119 \b Correct the handling of macro-local labels in the Codeview
122 \b Add \c{CLZERO}, \c{MONITORX} and \c{MWAITX} instructions.
125 \S{cl-2.12} Version 2.12
127 \b Major fixes to the \c{macho} backend (\k{machofmt}); earlier versions
128 would produce invalid symbols and relocations on a regular basis.
130 \b Support for thread-local storage in Mach-O.
132 \b Support for arbitrary sections in Mach-O.
134 \b Fix wrong negative size treated as a big positive value passed into
135 backend causing NASM to crash.
137 \b Fix handling of zero-extending unsigned relocations, we have been printing
138 wrong message and forgot to assign segment with predefined value before
139 passing it into output format.
141 \b Fix potential write of oversized (with size greater than allowed in
142 output format) relative relocations.
144 \b Portability fixes for building NASM with the LLVM compiler.
146 \b Add support of Codeview version 8 (\c{cv8}) debug format for
147 \c{win32} and \c{win64} formats in the \c{COFF} backend,
150 \b Allow 64-bit outputs in 16/32-bit only backends. Unsigned 64-bit
151 relocations are zero-extended from 32-bits with a warning
152 (suppressible via \c{-w-zext-reloc}); signed 64-bit relocations are
155 \b Line numbers in list files now correspond to the lines in the source
156 files, instead of simply being sequential.
158 \b There is now an official 64-bit (x64 a.k.a. x86-64) build for Windows.
161 \S{cl-2.11.09} Version 2.11.09
163 \b Fix potential stack overwrite in \c{macho32} backend.
165 \b Fix relocation records in \c{macho64} backend.
167 \b Fix symbol lookup computation in \c{macho64} backend.
169 \b Adjust \c{.symtab} and \c{.rela.text} sections alignments to 8 bytes
172 \b Fix section length computation in \c{bin} backend which leaded in incorrect
175 \S{cl-2.11.08} Version 2.11.08
177 \b Fix section length computation in \c{bin} backend which leaded in incorrect
180 \b Add a warning for numeric preprocessor definitions passed via command
181 line which might have unexpected results otherwise.
183 \b Add ability to specify a module name record in \c{rdoff} linker with
186 \b Increase label length capacity up to 256 bytes in \c{rdoff} backend for
187 FreePascal sake, which tends to generate very long labels for procedures.
189 \b Fix segmentation failure when rip addressing is used in \c{macho64} backend.
191 \b Fix access on out of memory when handling strings with a single
192 grave. We have sixed similar problem in previous release but not
193 all cases were covered.
195 \b Fix NULL dereference in disassembled on \c{BND} instruction.
197 \S{cl-2.11.07} Version 2.11.07
199 \b Fix 256 bit \c{VMOVNTPS} instruction.
201 \b Fix \c{-MD} option handling, which was rather broken in previous
202 release changing command line api.
204 \b Fix access to unitialized space when handling strings with
207 \b Fix nil dereference in handling memory reference parsing.
209 \S{cl-2.11.06} Version 2.11.06
211 \b Update AVX512 instructions based on the Extension Reference (319433-021 Sept
214 \b Fix the behavior of \c{-MF} and \c{-MD} options (Bugzilla 3392280)
216 \b Updated Win32 Makefile to fix issue with build
218 \S{cl-2.11.05} Version 2.11.05
220 \b Add \c{--v} as an alias for \c{-v} (see \k{opt-v}), for
221 command-line compatibility with Yasm.
223 \b Fix a bug introduced in 2.11.03 whereby certain instructions would
224 contain multiple REX prefixes, and thus be corrupt.
226 \S{cl-2.11.04} Version 2.11.04
228 \b Removed an invalid error checking code. Sometimes a memref only with
229 a displacement can also set an evex flag. For example:
231 \c vmovdqu32 [0xabcd]{k1}, zmm0
233 \b Fixed a bug in disassembler that EVEX.L'L vector length was not matched
234 when EVEX.b was set because it was simply considered as EVEC.RC.
235 Separated EVEX.L'L case from EVEX.RC which is ignored in matching.
237 \S{cl-2.11.03} Version 2.11.03
239 \b Fix a bug there REX prefixes were missing on instructions inside a
242 \S{cl-2.11.02} Version 2.11.02
244 \b Add the \c{XSAVEC}, \c{XSAVES} and \c{XRSTORS} family instructions.
246 \b Add the \c{CLFLUSHOPT} instruction.
248 \S{cl-2.11.01} Version 2.11.01
250 \b Allow instructions which implicitly uses \c{XMM0} (\c{VBLENDVPD},
251 \c{VBLENDVPS}, \c{PBLENDVB} and \c{SHA256RNDS2}) to be specified
252 without an explicit \c{xmm0} on the assembly line. In other words,
253 the following two lines produce the same output:
255 \c vblendvpd xmm2,xmm1,xmm0 ; Last operand is fixed xmm0
256 \c vblendvpd xmm2,xmm1 ; Implicit xmm0 omitted
258 \b In the ELF backends, don't crash the assembler if \c{section align}
259 is specified without a value.
261 \S{cl-2.11} Version 2.11
263 \b Add support for the Intel AVX-512 instruction set:
265 \b 16 new, 512-bit SIMD registers. Total 32 \c{(ZMM0 ~ ZMM31)}
267 \b 8 new opmask registers \c{(K0 ~ K7)}. One of 7 registers \c{(K1 ~ K7)} can
268 be used as an opmask for conditional execution.
270 \b A new EVEX encoding prefix. EVEX is based on VEX and provides more
271 capabilities: opmasks, broadcasting, embedded rounding and compressed
275 \c VDIVPD zmm0{k1}{z}, zmm1, zmm3 ; conditional vector operation
276 \c ; using opmask k1.
277 \c ; {z} is for zero-masking
279 \c VDIVPS zmm4, zmm5, [rbx]{1to16} ; load single-precision float and
280 \c ; replicate it 16 times. 32 * 16 = 512
281 \c - embedded rounding
282 \c VCVTSI2SD xmm6, xmm7, {rz-sae}, rax ; round toward zero. note that it
283 \c ; is used as if a separate operand.
284 \c ; it comes after the last SIMD operand
286 \b Add support for \c{ZWORD} (512 bits), \c{DZ} and \c{RESZ}.
288 \b Add support for the MPX and SHA instruction sets.
290 \b Better handling of section redefinition.
292 \b Generate manpages when running \c{'make dist'}.
294 \b Handle all token chains in mmacro params range.
296 \b Support split [base,index] effective address:
298 \c mov eax,[eax+8,ecx*4] ; eax=base, ecx=index, 4=scale, 8=disp
300 This is expected to be most useful for the MPX instructions.
302 \b Support \c{BND} prefix for branch instructions (for MPX).
304 \b The \c{DEFAULT} directive can now take \c{BND} and \c{NOBND}
305 options to indicate whether all relevant branches should be getting
306 \c{BND} prefixes. This is expected to be the normal for use in MPX
309 \b Add \c{{evex}}, \c{{vex3}} and \c{{vex2}} instruction prefixes to
310 have NASM encode the corresponding instruction, if possible, with an EVEX,
311 3-byte VEX, or 2-byte VEX prefix, respectively.
313 \b Support for section names longer than 8 bytes in Win32/Win64 COFF.
315 \b The \c{NOSPLIT} directive by itself no longer forces a single
316 register to become an index register, unless it has an explicit
319 \c mov eax,[nosplit eax] ; eax as base register
320 \c mov eax,[nosplit eax*1] ; eax as index register
322 \S{cl-2.10.09} Version 2.10.09
324 \b Pregenerate man pages.
326 \S{cl-2.10.08} Version 2.10.08
328 \b Fix \c{VMOVNTDQA}, \c{MOVNTDQA} and \c{MOVLPD} instructions.
330 \b Fix collision for \c{VGATHERQPS}, \c{VPGATHERQD} instructions.
332 \b Fix \c{VPMOVSXBQ}, \c{VGATHERQPD}, \c{VSPLLW} instructions.
334 \b Add a bunch of AMD TBM instructions.
336 \b Fix potential stack overwrite in numbers conversion.
338 \b Allow byte size in \c{PREFETCHTx} instructions.
340 \b Make manual pages up to date.
342 \b Make \c{F3} and \c{F2} SSE prefixes to override \c{66}.
344 \b Support of AMD SVM instructions in 32 bit mode.
346 \b Fix near offsets code generation for \c{JMP}, \c{CALL} instrictions
349 \b Fix preprocessor parse regression when id is expanding to a whitespace.
351 \S{cl-2.10.07} Version 2.10.07
353 \b Fix line continuation parsing being broken in previous version.
355 \S{cl-2.10.06} Version 2.10.06
357 \b Always quote the dependency source names when using the automatic
358 dependency generation options.
360 \b If no dependency target name is specified via the \c{-MT} or
361 \c{-MQ} options, quote the default output name.
363 \b Fix assembly of shift operations in \c{CPU 8086} mode.
365 \b Fix incorrect generation of explicit immediate byte for shift by 1
366 under certain circumstances.
368 \b Fix assembly of the \c{VPCMPGTQ} instruction.
370 \b Fix RIP-relative relocations in the \c{macho64} backend.
372 \S{cl-2.10.05} Version 2.10.05
374 \b Add the \c{CLAC} and \c{STAC} instructions.
376 \S{cl-2.10.04} Version 2.10.04
378 \b Add back the inadvertently deleted 256-bit version of the \c{VORPD}
381 \b Correct disassembly of instructions starting with byte \c{82} hex.
383 \b Fix corner cases in token pasting, for example:
385 \c %define N 1e%++%+ 5
388 \S{cl-2.10.03} Version 2.10.03
390 \b Correct the assembly of the instruction:
392 \c XRELEASE MOV [absolute],AL
394 \> Previous versions would incorrectly generate \c{F3 A2} for this
395 instruction and issue a warning; correct behavior is to emit \c{F3 88
398 \S{cl-2.10.02} Version 2.10.02
400 \b Add the \c{ifunc} macro package with integer functions, currently
401 only integer logarithms. See \k{pkg_ifunc}.
403 \b Add the \c{RDSEED}, \c{ADCX} and \c{ADOX} instructions.
405 \S{cl-2.10.01} Version 2.10.01
407 \b Add missing VPMOVMSKB instruction with reg32, ymmreg operands.
409 \S{cl-2.10} Version 2.10
411 \b When optimization is enabled, \c{mov r64,imm} now optimizes to the
412 shortest form possible between:
414 \c mov r32,imm32 ; 5 bytes
415 \c mov r64,imm32 ; 7 bytes
416 \c mov r64,imm64 ; 10 bytes
418 \> To force a specific form, use the \c{STRICT} keyword, see \k{strict}.
420 \b Add support for the Intel AVX2 instruction set.
422 \b Add support for Bit Manipulation Instructions 1 and 2.
424 \b Add support for Intel Transactional Synchronization Extensions (TSX).
426 \b Add support for x32 ELF (32-bit ELF with the CPU in 64-bit mode.)
429 \b Add support for bigendian UTF-16 and UTF-32. See \k{unicode}.
431 \S{cl-2.09.10} Version 2.09.10
433 \b Fix up NSIS script to protect uninstaller against registry keys
434 absence or corruption. It brings in a few additional questions
435 to a user during deinstallation procedure but still it is better
436 than unpredictable file removal.
438 \S{cl-2.09.09} Version 2.09.09
440 \b Fix initialization of section attributes of \c{bin} output format.
442 \b Fix \c{mach64} output format bug that crashes NASM due to NULL symbols.
445 \S{cl-2.09.08} Version 2.09.08
447 \b Fix \c{__OUTPUT_FORMAT__} assignment when output driver alias
448 is used. For example when \c{-f elf} is used \c{__OUTPUT_FORMAT__}
449 must be set to \c{elf}, if \c{-f elf32} is used \c{__OUTPUT_FORMAT__}
450 must be assigned accordingly, i.e. to \c{elf32}. The rule applies to
451 all output driver aliases. See \k{ofmtm}.
454 \S{cl-2.09.07} Version 2.09.07
456 \b Fix attempts to close same file several times
457 when \c{-a} option is used.
459 \b Fixes for VEXTRACTF128, VMASKMOVPS encoding.
462 \S{cl-2.09.06} Version 2.09.06
464 \b Fix missed section attribute initialization in \c{bin} output target.
467 \S{cl-2.09.05} Version 2.09.05
469 \b Fix arguments encoding for VPEXTRW instruction.
471 \b Remove invalid form of VPEXTRW instruction.
473 \b Add \c{VLDDQU} as alias for \c{VLDQQU} to
477 \S{cl-2.09.04} Version 2.09.04
479 \b Fix incorrect labels offset for VEX intructions.
481 \b Eliminate bogus warning on implicit operand size override.
483 \b \c{%if} term could not handle 64 bit numbers.
485 \b The COFF backend was limiting relocations number to 16 bits even if
486 in real there were a way more relocations.
489 \S{cl-2.09.03} Version 2.09.03
491 \b Print \c{%macro} name inside \c{%rep} blocks on error.
493 \b Fix preprocessor expansion behaviour. It happened sometime
494 too early and sometime simply wrong. Move behaviour back to
495 the origins (down to NASM 2.05.01).
497 \b Fix unitialized data dereference on OMF output format.
499 \b Issue warning on unterminated \c{%{} construct.
501 \b Fix for documentation typo.
504 \S{cl-2.09.02} Version 2.09.02
506 \b Fix reversed tokens when \c{%deftok} produces more than one output token.
508 \b Fix segmentation fault on disassembling some VEX instructions.
510 \b Missing \c{%endif} did not always cause error.
512 \b Fix typo in documentation.
514 \b Compound context local preprocessor single line macro identifiers
515 were not expanded early enough and as result lead to unresolved
519 \S{cl-2.09.01} Version 2.09.01
521 \b Fix NULL dereference on missed %deftok second parameter.
523 \b Fix NULL dereference on invalid %substr parameters.
526 \S{cl-2.09} Version 2.09
528 \b Fixed assignment the magnitude of \c{%rep} counter. It is limited
531 \b Fixed NULL dereference if argument of \c{%strlen} resolves
532 to whitespace. For example if nonexistent macro parameter is used.
534 \b \c{%ifenv}, \c{%elifenv}, \c{%ifnenv}, and \c{%elifnenv} directives
535 introduced. See \k{ifenv}.
537 \b Fixed NULL dereference if environment variable is missed.
539 \b Updates of new AVX v7 Intel instructions.
541 \b \c{PUSH imm32} is now officially documented.
543 \b Fix for encoding the LFS, LGS and LSS in 64-bit mode.
545 \b Fixes for compatibility with OpenWatcom compiler and DOS 8.3 file
548 \b Macros parameters range expansion introduced. See \k{mlmacrange}.
550 \b Backward compatibility on expanging of local sigle macros restored.
552 \b 8 bit relocations for \c{elf} and \c{bin} output formats are introduced.
554 \b Short intersegment jumps are permitted now.
556 \b An alignment more than 64 bytes are allowed for \c{win32},
557 \c{win64} output formats.
559 \b \c{SECTALIGN} directive introduced. See \k{sectalign}.
561 \b \c{nojmp} option introduced in \c{smartalign} package. See
564 \b Short aliases \c{win}, \c{elf} and \c{macho} for output formats are
565 introduced. Each stands for \c{win32}, \c{elf32} and \c{macho32}
568 \b Faster handling of missing directives implemented.
570 \b Various small improvements in documentation.
572 \b No hang anymore if unable to open malloc.log file.
574 \b The environments without vsnprintf function are able to build nasm again.
576 \b AMD LWP instructions updated.
578 \b Tighten EA checks. We warn a user if there overflow in EA addressing.
580 \b Make \c{-Ox} the default optimization level. For the legacy
581 behavior, specify \c{-O0} explicitly. See \k{opt-O}.
583 \b Environment variables read with \c{%!} or tested with \c{%ifenv}
584 can now contain non-identifier characters if surrounded by quotes.
587 \b Add a new standard macro package \c{%use fp} for floating-point
588 convenience macros. See \k{pkg_fp}.
591 \S{cl-2.08.02} Version 2.08.02
593 \b Fix crash under certain circumstances when using the \c{%+} operator.
596 \S{cl-2.08.01} Version 2.08.01
598 \b Fix the \c{%use} statement, which was broken in 2.08.
601 \S{cl-2.08} Version 2.08
603 \b A number of enhancements/fixes in macros area.
605 \b Support for converting strings to tokens. See \k{deftok}.
607 \b Fuzzy operand size logic introduced.
609 \b Fix COFF stack overrun on too long export identifiers.
611 \b Fix Macho-O alignment bug.
613 \b Fix crashes with -fwin32 on file with many exports.
615 \b Fix stack overrun for too long [DEBUG id].
617 \b Fix incorrect sbyte usage in IMUL (hit only if optimization
620 \b Append ending token for \c{.stabs} records in the ELF output format.
622 \b New NSIS script which uses ModernUI and MultiUser approach.
624 \b Visual Studio 2008 NASM integration (rules file).
626 \b Warn a user if a constant is too long (and as result will be stripped).
628 \b The obsoleted pre-XOP AMD SSE5 instruction set which was never actualized
631 \b Fix stack overrun on too long error file name passed from the command line.
633 \b Bind symbols to the .text section by default (ie in case if SECTION
634 directive was omitted) in the ELF output format.
636 \b Fix sync points array index wrapping.
638 \b A few fixes for FMA4 and XOP instruction templates.
640 \b Add AMD Lightweight Profiling (LWP) instructions.
642 \b Fix the offset for \c{%arg} in 64-bit mode.
644 \b An undefined local macro (\c{%$}) no longer matches a global macro
647 \b Fix NULL dereference on too long local labels.
650 \S{cl-2.07} Version 2.07
652 \b NASM is now under the 2-clause BSD license. See \k{legal}.
654 \b Fix the section type for the \c{.strtab} section in the \c{elf64}
657 \b Fix the handling of \c{COMMON} directives in the \c{obj} output format.
659 \b New \c{ith} and \c{srec} output formats; these are variants of the
660 \c{bin} output format which output Intel hex and Motorola S-records,
661 respectively. See \k{ithfmt} and \k{srecfmt}.
663 \b \c{rdf2ihx} replaced with an enhanced \c{rdf2bin}, which can output
664 binary, COM, Intel hex or Motorola S-records.
666 \b The Windows installer now puts the NASM directory first in the
667 \c{PATH} of the "NASM Shell".
669 \b Revert the early expansion behavior of \c{%+} to pre-2.06 behavior:
670 \c{%+} is only expanded late.
672 \b Yet another Mach-O alignment fix.
674 \b Don't delete the list file on errors. Also, include error and
675 warning information in the list file.
677 \b Support for 64-bit Mach-O output, see \k{machofmt}.
679 \b Fix assert failure on certain operations that involve strings with
683 \S{cl-2.06} Version 2.06
685 \b This release is dedicated to the memory of Charles A. Crayne, long
686 time NASM developer as well as moderator of \c{comp.lang.asm.x86} and
687 author of the book \e{Serious Assembler}. We miss you, Chuck.
689 \b Support for indirect macro expansion (\c{%[...]}). See \k{indmacro}.
691 \b \c{%pop} can now take an argument, see \k{pushpop}.
693 \b The argument to \c{%use} is no longer macro-expanded. Use
694 \c{%[...]} if macro expansion is desired.
696 \b Support for thread-local storage in ELF32 and ELF64. See \k{elftls}.
698 \b Fix crash on \c{%ifmacro} without an argument.
700 \b Correct the arguments to the \c{POPCNT} instruction.
702 \b Fix section alignment in the Mach-O format.
704 \b Update AVX support to version 5 of the Intel specification.
706 \b Fix the handling of accesses to context-local macros from higher
707 levels in the context stack.
709 \b Treat \c{WAIT} as a prefix rather than as an instruction, thereby
710 allowing constructs like \c{O16 FSAVE} to work correctly.
712 \b Support for structures with a non-zero base offset. See \k{struc}.
714 \b Correctly handle preprocessor token concatenation (see \k{concat})
715 involving floating-point numbers.
717 \b The \c{PINSR} series of instructions have been corrected and
720 \b Removed AMD SSE5, replaced with the new XOP/FMA4/CVT16 (rev 3.03)
723 \b The ELF backends no longer automatically generate a \c{.comment} section.
725 \b Add additional "well-known" ELF sections with default attributes. See
729 \S{cl-2.05.01} Version 2.05.01
731 \b Fix the \c{-w}/\c{-W} option parsing, which was broken in NASM 2.05.
734 \S{cl-2.05} Version 2.05
736 \b Fix redundant REX.W prefix on \c{JMP reg64}.
738 \b Make the behaviour of \c{-O0} match NASM 0.98 legacy behavior.
741 \b \c{-w-user} can be used to suppress the output of \c{%warning} directives.
744 \b Fix bug where \c{ALIGN} would issue a full alignment datum instead of
747 \b Fix offsets in list files.
749 \b Fix \c{%include} inside multi-line macros or loops.
751 \b Fix error where NASM would generate a spurious warning on valid
752 optimizations of immediate values.
754 \b Fix arguments to a number of the \c{CVT} SSE instructions.
756 \b Fix RIP-relative offsets when the instruction carries an immediate.
758 \b Massive overhaul of the ELF64 backend for spec compliance.
760 \b Fix the Geode \c{PFRCPV} and \c{PFRSQRTV} instruction.
762 \b Fix the SSE 4.2 \c{CRC32} instruction.
765 \S{cl-2.04} Version 2.04
767 \b Sanitize macro handing in the \c{%error} directive.
769 \b New \c{%warning} directive to issue user-controlled warnings.
771 \b \c{%error} directives are now deferred to the final assembly phase.
773 \b New \c{%fatal} directive to immediately terminate assembly.
775 \b New \c{%strcat} directive to join quoted strings together.
777 \b New \c{%use} macro directive to support standard macro directives. See
780 \b Excess default parameters to \c{%macro} now issues a warning by default.
783 \b Fix \c{%ifn} and \c{%elifn}.
785 \b Fix nested \c{%else} clauses.
787 \b Correct the handling of nested \c{%rep}s.
789 \b New \c{%unmacro} directive to undeclare a multi-line macro.
792 \b Builtin macro \c{__PASS__} which expands to the current assembly pass.
795 \b \c{__utf16__} and \c{__utf32__} operators to generate UTF-16 and UTF-32
796 strings. See \k{unicode}.
798 \b Fix bug in case-insensitive matching when compiled on platforms that
799 don't use the \c{configure} script. Of the official release binaries,
800 that only affected the OS/2 binary.
802 \b Support for x87 packed BCD constants. See \k{bcdconst}.
804 \b Correct the \c{LTR} and \c{SLDT} instructions in 64-bit mode.
806 \b Fix unnecessary REX.W prefix on indirect jumps in 64-bit mode.
808 \b Add AVX versions of the AES instructions (\c{VAES}...).
810 \b Fix the 256-bit FMA instructions.
812 \b Add 256-bit AVX stores per the latest AVX spec.
814 \b VIA XCRYPT instructions can now be written either with or without
815 \c{REP}, apparently different versions of the VIA spec wrote them
818 \b Add missing 64-bit \c{MOVNTI} instruction.
820 \b Fix the operand size of \c{VMREAD} and \c{VMWRITE}.
822 \b Numerous bug fixes, especially to the AES, AVX and VTX instructions.
824 \b The optimizer now always runs until it converges. It also runs even
825 when disabled, but doesn't optimize. This allows most forward references
826 to be resolved properly.
828 \b \c{%push} no longer needs a context identifier; omitting the context
829 identifier results in an anonymous context.
832 \S{cl-2.03.01} Version 2.03.01
834 \b Fix buffer overflow in the listing module.
836 \b Fix the handling of hexadecimal escape codes in `...` strings.
838 \b The Postscript/PDF documentation has been reformatted.
840 \b The \c{-F} option now implies \c{-g}.
843 \S{cl-2.03} Version 2.03
845 \b Add support for Intel AVX, CLMUL and FMA instructions,
846 including YMM registers.
848 \b \c{dy}, \c{resy} and \c{yword} for 32-byte operands.
850 \b Fix some SSE5 instructions.
852 \b Intel \c{INVEPT}, \c{INVVPID} and \c{MOVBE} instructions.
854 \b Fix checking for critical expressions when the optimizer is enabled.
856 \b Support the DWARF debugging format for ELF targets.
858 \b Fix optimizations of signed bytes.
860 \b Fix operation on bigendian machines.
862 \b Fix buffer overflow in the preprocessor.
864 \b \c{SAFESEH} support for Win32, \c{IMAGEREL} for Win64 (SEH).
866 \b \c{%?} and \c{%??} to refer to the name of a macro itself. In particular,
867 \c{%idefine keyword $%?} can be used to make a keyword "disappear".
869 \b New options for dependency generation: \c{-MD}, \c{-MF},
870 \c{-MP}, \c{-MT}, \c{-MQ}.
872 \b New preprocessor directives \c{%pathsearch} and \c{%depend}; INCBIN
873 reimplemented as a macro.
875 \b \c{%include} now resolves macros in a sane manner.
877 \b \c{%substr} can now be used to get other than one-character substrings.
879 \b New type of character/string constants, using backquotes (\c{`...`}),
880 which support C-style escape sequences.
882 \b \c{%defstr} and \c{%idefstr} to stringize macro definitions before
885 \b Fix forward references used in \c{EQU} statements.
888 \S{cl-2.02} Version 2.02
890 \b Additional fixes for MMX operands with explicit \c{qword}, as well as
891 (hopefully) SSE operands with \c{oword}.
893 \b Fix handling of truncated strings with \c{DO}.
895 \b Fix segfaults due to memory overwrites when floating-point constants
898 \b Fix segfaults due to missing include files.
900 \b Fix OpenWatcom Makefiles for DOS and OS/2.
902 \b Add autogenerated instruction list back into the documentation.
904 \b ELF: Fix segfault when generating stabs, and no symbols have been
907 \b ELF: Experimental support for DWARF debugging information.
909 \b New compile date and time standard macros.
911 \b \c{%ifnum} now returns true for negative numbers.
913 \b New \c{%iftoken} test for a single token.
915 \b New \c{%ifempty} test for empty expansion.
917 \b Add support for the \c{XSAVE} instruction group.
919 \b Makefile for Netware/gcc.
921 \b Fix issue with some warnings getting emitted way too many times.
923 \b Autogenerated instruction list added to the documentation.
926 \S{cl-2.01} Version 2.01
928 \b Fix the handling of MMX registers with explicit \c{qword} tags on
929 memory (broken in 2.00 due to 64-bit changes.)
931 \b Fix the PREFETCH instructions.
933 \b Fix the documentation.
935 \b Fix debugging info when using \c{-f elf}
936 (backwards compatibility alias for \c{-f elf32}).
938 \b Man pages for rdoff tools (from the Debian project.)
940 \b ELF: handle large numbers of sections.
942 \b Fix corrupt output when the optimizer runs out of passes.
945 \S{cl-2.00} Version 2.00
947 \b Added c99 data-type compliance.
949 \b Added general x86-64 support.
951 \b Added win64 (x86-64 COFF) output format.
953 \b Added \c{__BITS__} standard macro.
955 \b Renamed the \c{elf} output format to \c{elf32} for clarity.
957 \b Added \c{elf64} and \c{macho} (MacOS X) output formats.
959 \b Added Numeric constants in \c{dq} directive.
961 \b Added \c{oword}, \c{do} and \c{reso} pseudo operands.
963 \b Allow underscores in numbers.
965 \b Added 8-, 16- and 128-bit floating-point formats.
967 \b Added binary, octal and hexadecimal floating-point.
969 \b Correct the generation of floating-point constants.
971 \b Added floating-point option control.
973 \b Added Infinity and NaN floating point support.
975 \b Added ELF Symbol Visibility support.
977 \b Added setting OSABI value in ELF header directive.
979 \b Added Generate Makefile Dependencies option.
981 \b Added Unlimited Optimization Passes option.
983 \b Added \c{%IFN} and \c{%ELIFN} support.
985 \b Added Logical Negation Operator.
987 \b Enhanced Stack Relative Preprocessor Directives.
989 \b Enhanced ELF Debug Formats.
991 \b Enhanced Send Errors to a File option.
993 \b Added SSSE3, SSE4.1, SSE4.2, SSE5 support.
995 \b Added a large number of additional instructions.
997 \b Significant performance improvements.
999 \b \c{-w+warning} and \c{-w-warning} can now be written as -Wwarning and
1000 -Wno-warning, respectively. See \k{opt-w}.
1002 \b Add \c{-w+error} to treat warnings as errors. See \k{opt-w}.
1004 \b Add \c{-w+all} and \c{-w-all} to enable or disable all suppressible
1005 warnings. See \k{opt-w}.
1008 \H{cl-0.98.xx} NASM 0.98 Series
1010 The 0.98 series was the production versions of NASM from 1999 to 2007.
1013 \S{cl-0.98.39} Version 0.98.39
1015 \b fix buffer overflow
1017 \b fix outas86's \c{.bss} handling
1019 \b "make spotless" no longer deletes config.h.in.
1021 \b \c{%(el)if(n)idn} insensitivity to string quotes difference (#809300).
1023 \b (nasm.c)\c{__OUTPUT_FORMAT__} changed to string value instead of symbol.
1025 \S{cl-0.98.38} Version 0.98.38
1028 \b Add Makefile for 16-bit DOS binaries under OpenWatcom, and modify
1029 \c{mkdep.pl} to be able to generate completely pathless dependencies, as
1030 required by OpenWatcom wmake (it supports path searches, but not
1033 \b Fix the \c{STR} instruction.
1035 \b Fix the ELF output format, which was broken under certain
1036 circumstances due to the addition of stabs support.
1038 \b Quick-fix Borland format debug-info for \c{-f obj}
1040 \b Fix for \c{%rep} with no arguments (#560568)
1042 \b Fix concatenation of preprocessor function call (#794686)
1044 \b Fix long label causes coredump (#677841)
1046 \b Use autoheader as well as autoconf to keep configure from generating
1047 ridiculously long command lines.
1049 \b Make sure that all of the formats which support debugging output
1050 actually will suppress debugging output when \c{-g} not specified.
1052 \S{cl-0.98.37} Version 0.98.37
1055 \b Paths given in \c{-I} switch searched for \c{incbin}-ed as
1056 well as \c{%include}-ed files.
1058 \b Added stabs debugging for the ELF output format, patch from
1061 \b Fix \c{output/outbin.c} to allow origin > 80000000h.
1063 \b Make \c{-U} switch work.
1065 \b Fix the use of relative offsets with explicit prefixes, e.g.
1068 \b Remove \c{backslash()}.
1070 \b Fix the \c{SMSW} and \c{SLDT} instructions.
1072 \b \c{-O2} and \c{-O3} are no longer aliases for \c{-O10} and \c{-O15}.
1073 If you mean the latter, please say so! :)
1075 \S{cl-0.98.36} Version 0.98.36
1078 \b Update rdoff - librarian/archiver - common rec - docs!
1080 \b Fix signed/unsigned problems.
1082 \b Fix \c{JMP FAR label} and \c{CALL FAR label}.
1084 \b Add new multisection support - map files - fix align bug
1086 \b Fix sysexit, movhps/movlps reg,reg bugs in insns.dat
1088 \b \c{Q} or \c{O} suffixes indicate octal
1090 \b Support Prescott new instructions (PNI).
1092 \b Cyrix \c{XSTORE} instruction.
1095 \S{cl-0.98.35} Version 0.98.35
1097 \b Fix build failure on 16-bit DOS (Makefile.bc3 workaround for compiler bug.)
1099 \b Fix dependencies and compiler warnings.
1101 \b Add "const" in a number of places.
1103 \b Add -X option to specify error reporting format (use -Xvc to
1104 integrate with Microsoft Visual Studio.)
1106 \b Minor changes for code legibility.
1108 \b Drop use of tmpnam() in rdoff (security fix.)
1111 \S{cl-0.98.34} Version 0.98.34
1113 \b Correct additional address-size vs. operand-size confusions.
1115 \b Generate dependencies for all Makefiles automatically.
1117 \b Add support for unimplemented (but theoretically available)
1118 registers such as tr0 and cr5. Segment registers 6 and 7 are called
1119 segr6 and segr7 for the operations which they can be represented.
1121 \b Correct some disassembler bugs related to redundant address-size prefixes.
1122 Some work still remains in this area.
1124 \b Correctly generate an error for things like "SEG eax".
1126 \b Add the JMPE instruction, enabled by "CPU IA64".
1128 \b Correct compilation on newer gcc/glibc platforms.
1130 \b Issue an error on things like "jmp far eax".
1133 \S{cl-0.98.33} Version 0.98.33
1135 \b New __NASM_PATCHLEVEL__ and __NASM_VERSION_ID__ standard macros to
1136 round out the version-query macros. version.pl now understands
1137 X.YYplWW or X.YY.ZZplWW as a version number, equivalent to
1138 X.YY.ZZ.WW (or X.YY.0.WW, as appropriate).
1140 \b New keyword "strict" to disable the optimization of specific
1143 \b Fix the handing of size overrides with JMP instructions
1144 (instructions such as "jmp dword foo".)
1146 \b Fix the handling of "ABSOLUTE label", where "label" points into a
1147 relocatable segment.
1149 \b Fix OBJ output format with lots of externs.
1151 \b More documentation updates.
1153 \b Add -Ov option to get verbose information about optimizations.
1155 \b Undo a braindead change which broke \c{%elif} directives.
1157 \b Makefile updates.
1160 \S{cl-0.98.32} Version 0.98.32
1162 \b Fix NASM crashing when \c{%macro} directives were left unterminated.
1164 \b Lots of documentation updates.
1166 \b Complete rewrite of the PostScript/PDF documentation generator.
1168 \b The MS Visual C++ Makefile was updated and corrected.
1170 \b Recognize .rodata as a standard section name in ELF.
1172 \b Fix some obsolete Perl4-isms in Perl scripts.
1174 \b Fix configure.in to work with autoconf 2.5x.
1176 \b Fix a couple of "make cleaner" misses.
1178 \b Make the normal "./configure && make" work with Cygwin.
1181 \S{cl-0.98.31} Version 0.98.31
1183 \b Correctly build in a separate object directory again.
1185 \b Derive all references to the version number from the version file.
1187 \b New standard macros __NASM_SUBMINOR__ and __NASM_VER__ macros.
1189 \b Lots of Makefile updates and bug fixes.
1191 \b New \c{%ifmacro} directive to test for multiline macros.
1193 \b Documentation updates.
1195 \b Fixes for 16-bit OBJ format output.
1197 \b Changed the NASM environment variable to NASMENV.
1200 \S{cl-0.98.30} Version 0.98.30
1202 \b Changed doc files a lot: completely removed old READMExx and
1203 Wishlist files, incorporating all information in CHANGES and TODO.
1205 \b I waited a long time to rename zoutieee.c to (original) outieee.c
1207 \b moved all output modules to output/ subdirectory.
1209 \b Added 'make strip' target to strip debug info from nasm & ndisasm.
1211 \b Added INSTALL file with installation instructions.
1213 \b Added -v option description to nasm man.
1215 \b Added dist makefile target to produce source distributions.
1217 \b 16-bit support for ELF output format (GNU extension, but useful.)
1220 \S{cl-0.98.28} Version 0.98.28
1222 \b Fastcooked this for Debian's Woody release:
1223 Frank applied the INCBIN bug patch to 0.98.25alt and called
1224 it 0.98.28 to not confuse poor little apt-get.
1227 \S{cl-0.98.26} Version 0.98.26
1229 \b Reorganised files even better from 0.98.25alt
1232 \S{cl-0.98.25alt} Version 0.98.25alt
1234 \b Prettified the source tree. Moved files to more reasonable places.
1236 \b Added findleak.pl script to misc/ directory.
1238 \b Attempted to fix doc.
1241 \S{cl-0.98.25} Version 0.98.25
1243 \b Line continuation character \c{\\}.
1245 \b Docs inadvertantly reverted - "dos packaging".
1248 \S{cl-0.98.24p1} Version 0.98.24p1
1250 \b FIXME: Someone, document this please.
1253 \S{cl-0.98.24} Version 0.98.24
1255 \b Documentation - Ndisasm doc added to Nasm.doc.
1258 \S{cl-0.98.23} Version 0.98.23
1260 \b Attempted to remove rdoff version1
1262 \b Lino Mastrodomenico's patches to preproc.c (%$$ bug?).
1265 \S{cl-0.98.22} Version 0.98.22
1267 \b Update rdoff2 - attempt to remove v1.
1270 \S{cl-0.98.21} Version 0.98.21
1272 \b Optimization fixes.
1275 \S{cl-0.98.20} Version 0.98.20
1277 \b Optimization fixes.
1280 \S{cl-0.98.19} Version 0.98.19
1282 \b H. J. Lu's patch back out.
1285 \S{cl-0.98.18} Version 0.98.18
1287 \b Added ".rdata" to "-f win32".
1290 \S{cl-0.98.17} Version 0.98.17
1292 \b H. J. Lu's "bogus elf" patch. (Red Hat problem?)
1295 \S{cl-0.98.16} Version 0.98.16
1297 \b Fix whitespace before "[section ..." bug.
1300 \S{cl-0.98.15} Version 0.98.15
1302 \b Rdoff changes (?).
1304 \b Fix fixes to memory leaks.
1307 \S{cl-0.98.14} Version 0.98.14
1309 \b Fix memory leaks.
1312 \S{cl-0.98.13} Version 0.98.13
1314 \b There was no 0.98.13
1317 \S{cl-0.98.12} Version 0.98.12
1319 \b Update optimization (new function of "-O1")
1321 \b Changes to test/bintest.asm (?).
1324 \S{cl-0.98.11} Version 0.98.11
1326 \b Optimization changes.
1331 \S{cl-0.98.10} Version 0.98.10
1333 \b There was no 0.98.10
1336 \S{cl-0.98.09} Version 0.98.09
1338 \b Add multiple sections support to "-f bin".
1340 \b Changed GLOBAL_TEMP_BASE in outelf.c from 6 to 15.
1342 \b Add "-v" as an alias to the "-r" switch.
1344 \b Remove "#ifdef" from Tasm compatibility options.
1346 \b Remove redundant size-overrides on "mov ds, ex", etc.
1348 \b Fixes to SSE2, other insns.dat (?).
1350 \b Enable uppercase "I" and "P" switches.
1352 \b Case insinsitive "seg" and "wrt".
1354 \b Update install.sh (?).
1356 \b Allocate tokens in blocks.
1358 \b Improve "invalid effective address" messages.
1361 \S{cl-0.98.08} Version 0.98.08
1363 \b Add "\c{%strlen}" and "\c{%substr}" macro operators
1365 \b Fixed broken c16.mac.
1367 \b Unterminated string error reported.
1369 \b Fixed bugs as per 0.98bf
1372 \S{cl-0.98.09b with John Coffman patches released 28-Oct-2001} Version 0.98.09b with John Coffman patches released 28-Oct-2001
1374 Changes from 0.98.07 release to 98.09b as of 28-Oct-2001
1376 \b More closely compatible with 0.98 when -O0 is implied
1377 or specified. Not strictly identical, since backward
1378 branches in range of short offsets are recognized, and signed
1379 byte values with no explicit size specification will be
1380 assembled as a single byte.
1382 \b More forgiving with the PUSH instruction. 0.98 requires
1383 a size to be specified always. 0.98.09b will imply the size
1384 from the current BITS setting (16 or 32).
1386 \b Changed definition of the optimization flag:
1388 \c -O0 strict two-pass assembly, JMP and Jcc are
1389 \c handled more like 0.98, except that back-
1390 \c ward JMPs are short, if possible.
1392 \c -O1 strict two-pass assembly, but forward
1393 \c branches are assembled with code guaranteed
1394 \c to reach; may produce larger code than
1395 \c -O0, but will produce successful assembly
1396 \c more often if branch offset sizes are not
1399 \c -O2 multi-pass optimization, minimize branch
1400 \c offsets; also will minimize signed immed-
1401 \c iate bytes, overriding size specification.
1403 \c -O3 like -O2, but more passes taken, if needed
1406 \S{cl-0.98.07 released 01/28/01} Version 0.98.07 released 01/28/01
1408 \b Added Stepane Denis' SSE2 instructions to a *working*
1409 version of the code - some earlier versions were based on
1410 broken code - sorry 'bout that. version "0.98.07"
1412 \b Cosmetic modifications to nasm.c, nasm.h,
1416 \S{cl-0.98.06f released 01/18/01} Version 0.98.06f released 01/18/01
1419 \b Add "metalbrain"s jecxz bug fix in insns.dat
1421 \b Alter nasmdoc.src to match - version "0.98.06f"
1424 \S{cl-0.98.06e released 01/09/01} Version 0.98.06e released 01/09/01
1427 \b Removed the "outforms.h" file - it appears to be
1428 someone's old backup of "outform.h". version "0.98.06e"
1430 \b fbk - finally added the fix for the "multiple %includes bug",
1431 known since 7/27/99 - reported originally (?) and sent to
1432 us by Austin Lunnen - he reports that John Fine had a fix
1433 within the day. Here it is...
1435 \b Nelson Rush resigns from the group. Big thanks to Nelson for
1436 his leadership and enthusiasm in getting these changes
1437 incorporated into Nasm!
1439 \b fbk - [list +], [list -] directives - ineptly implemented, should
1440 be re-written or removed, perhaps.
1442 \b Brian Raiter / fbk - "elfso bug" fix - applied to aoutb format
1443 as well - testing might be desirable...
1445 \b James Seter - -postfix, -prefix command line switches.
1447 \b Yuri Zaporozhets - rdoff utility changes.
1450 \S{cl-0.98p1} Version 0.98p1
1452 \b GAS-like palign (Panos Minos)
1454 \b FIXME: Someone, fill this in with details
1457 \S{cl-0.98bf (bug-fixed)} Version 0.98bf (bug-fixed)
1459 \b Fixed - elf and aoutb bug - shared libraries
1460 - multiple "%include" bug in "-f obj"
1462 - unrecognized option bug in ndisasm
1464 \S{cl-0.98.03 with John Coffman's changes released 27-Jul-2000} Version 0.98.03 with John Coffman's changes released 27-Jul-2000
1466 \b Added signed byte optimizations for the 0x81/0x83 class
1467 of instructions: ADC, ADD, AND, CMP, OR, SBB, SUB, XOR:
1468 when used as 'ADD reg16,imm' or 'ADD reg32,imm.' Also
1469 optimization of signed byte form of 'PUSH imm' and 'IMUL
1470 reg,imm'/'IMUL reg,reg,imm.' No size specification is needed.
1472 \b Added multi-pass JMP and Jcc offset optimization. Offsets
1473 on forward references will preferentially use the short form,
1474 without the need to code a specific size (short or near) for
1475 the branch. Added instructions for 'Jcc label' to use the
1476 form 'Jnotcc $+3/JMP label', in cases where a short offset
1477 is out of bounds. If compiling for a 386 or higher CPU, then
1478 the 386 form of Jcc will be used instead.
1480 \> This feature is controlled by a new command-line switch: "O",
1481 (upper case letter O). "-O0" reverts the assembler to no
1482 extra optimization passes, "-O1" allows up to 5 extra passes,
1483 and "-O2"(default), allows up to 10 extra optimization passes.
1485 \b Added a new directive: 'cpu XXX', where XXX is any of:
1486 8086, 186, 286, 386, 486, 586, pentium, 686, PPro, P2, P3 or
1487 Katmai. All are case insensitive. All instructions will
1488 be selected only if they apply to the selected cpu or lower.
1489 Corrected a couple of bugs in cpu-dependence in 'insns.dat'.
1491 \b Added to 'standard.mac', the "use16" and "use32" forms of
1492 the "bits 16/32" directive. This is nothing new, just conforms
1493 to a lot of other assemblers. (minor)
1495 \b Changed label allocation from 320/32 (10000 labels @ 200K+)
1496 to 32/37 (1000 labels); makes running under DOS much easier.
1497 Since additional label space is allocated dynamically, this
1498 should have no effect on large programs with lots of labels.
1499 The 37 is a prime, believed to be better for hashing. (minor)
1502 \S{cl-0.98.03} Version 0.98.03
1504 "Integrated patchfile 0.98-0.98.01. I call this version 0.98.03 for
1505 historical reasons: 0.98.02 was trashed." --John Coffman
1506 <johninsd@san.rr.com>, 27-Jul-2000
1508 \b Kendall Bennett's SciTech MGL changes
1510 \b Note that you must define "TASM_COMPAT" at compile-time
1511 to get the Tasm Ideal Mode compatibility.
1513 \b All changes can be compiled in and out using the TASM_COMPAT macros,
1514 and when compiled without TASM_COMPAT defined we get the exact same
1515 binary as the unmodified 0.98 sources.
1517 \b standard.mac, macros.c: Added macros to ignore TASM directives before
1520 \b nasm.h: Added extern declaration for tasm_compatible_mode
1522 \b nasm.c: Added global variable tasm_compatible_mode
1524 \b Added command line switch for TASM compatible mode (-t)
1526 \b Changed version command line to reflect when compiled with TASM additions
1528 \b Added response file processing to allow all arguments on a single
1529 line (response file is @resp rather than -@resp for NASM format).
1531 \b labels.c: Changes islocal() macro to support TASM style @@local labels.
1533 \b Added islocalchar() macro to support TASM style @@local labels.
1535 \b parser.c: Added support for TASM style memory references (ie: mov
1536 [DWORD eax],10 rather than the NASM style mov DWORD [eax],10).
1538 \b preproc.c: Added new directives, \c{%arg}, \c{%local}, \c{%stacksize} to directives
1541 \b Added support for TASM style directives without a leading % symbol.
1543 \b Integrated a block of changes from Andrew Zabolotny <bit@eltech.ru>:
1545 \b A new keyword \c{%xdefine} and its case-insensitive counterpart \c{%ixdefine}.
1546 They work almost the same way as \c{%define} and \c{%idefine} but expand
1547 the definition immediately, not on the invocation. Something like a cross
1548 between \c{%define} and \c{%assign}. The "x" suffix stands for "eXpand", so
1549 "xdefine" can be deciphered as "expand-and-define". Thus you can do
1555 \c %xdefine %1 dword [esp+ofs]
1556 \c %assign ofs ofs+4
1559 \b Changed the place where the expansion of %$name macros are expanded.
1560 Now they are converted into ..@ctxnum.name form when detokenizing, so
1561 there are no quirks as before when using %$name arguments to macros,
1562 in macros etc. For example:
1571 \> Now last line will be expanded into "hello" as expected. This also allows
1572 for lots of goodies, a good example are extended "proc" macros included
1575 \b Added a check for "cstk" in smacro_defined() before calling get_ctx() -
1576 this allows for things like:
1581 \> to work without warnings even in no context.
1583 \b Added a check for "cstk" in %if*ctx and %elif*ctx directives -
1584 this allows to use \c{%ifctx} without excessive warnings. If there is
1585 no active context, \c{%ifctx} goes through "false" branch.
1587 \b Removed "user error: " prefix with \c{%error} directive: it just clobbers the
1588 output and has absolutely no functionality. Besides, this allows to write
1589 macros that does not differ from built-in functions in any way.
1591 \b Added expansion of string that is output by \c{%error} directive. Now you
1594 \c %define hello(x) Hello, x!
1596 \c %define %$name andy
1597 \c %error "hello(%$name)"
1599 \> Same happened with \c{%include} directive.
1601 \b Now all directives that expect an identifier will try to expand and
1602 concatenate everything without whitespaces in between before usage.
1603 For example, with "unfixed" nasm the commands
1605 \c %define %$abc hello
1606 \c %define __%$abc goodbye
1609 \> would produce "incorrect" output: last line will expand to
1611 \c hello goodbyehello
1613 \> Not quite what you expected, eh? :-) The answer is that preprocessor
1614 treats the \c{%define} construct as if it would be
1616 \c %define __ %$abc goodbye
1618 \> (note the white space between __ and %$abc). After my "fix" it
1619 will "correctly" expand into
1623 \> as expected. Note that I use quotes around words "correct", "incorrect"
1624 etc because this is rather a feature not a bug; however current behaviour
1625 is more logical (and allows more advanced macro usage :-).
1627 Same change was applied to:
1628 \c{%push},\c{%macro},\c{%imacro},\c{%define},\c{%idefine},\c{%xdefine},\c{%ixdefine},
1629 \c{%assign},\c{%iassign},\c{%undef}
1631 \b A new directive [WARNING {+|-}warning-id] have been added. It works only
1632 if the assembly phase is enabled (i.e. it doesn't work with nasm -e).
1634 \b A new warning type: macro-selfref. By default this warning is disabled;
1635 when enabled NASM warns when a macro self-references itself; for example
1636 the following source:
1638 \c [WARNING macro-selfref]
1649 \> will produce a warning, but if we remove the first line we won't see it
1650 anymore (which is The Right Thing To Do {tm} IMHO since C preprocessor
1651 eats such constructs without warnings at all).
1653 \b Added a "error" routine to preprocessor which always will set ERR_PASS1
1654 bit in severity_code. This removes annoying repeated errors on first
1655 and second passes from preprocessor.
1657 \b Added the %+ operator in single-line macros for concatenating two
1658 identifiers. Usage example:
1660 \c %define _myfunc _otherfunc
1661 \c %define cextern(x) _ %+ x
1664 \> After first expansion, third line will become "_myfunc". After this
1665 expansion is performed again so it becomes "_otherunc".
1667 \b Now if preprocessor is in a non-emitting state, no warning or error
1668 will be emitted. Example:
1673 \c put anything you want between these two brackets,
1674 \c even macro-parameter references %1 or local
1675 \c labels %$zz or macro-local labels %%zz - no
1676 \c warning will be emitted.
1679 \b Context-local variables on expansion as a last resort are looked up
1680 in outer contexts. For example, the following piece:
1683 \c %define %$a [esp]
1690 \> will expand correctly the fourth line to [esp]; if we'll define another
1691 %$a inside the "inner" context, it will take precedence over outer
1692 definition. However, this modification has been applied only to
1693 expand_smacro and not to smacro_define: as a consequence expansion
1694 looks in outer contexts, but \c{%ifdef} won't look in outer contexts.
1696 \> This behaviour is needed because we don't want nested contexts to
1697 act on already defined local macros. Example:
1699 \c %define %$arg1 [esp+4]
1705 \> In this example the "if" mmacro enters into the "if" context, so %$arg1
1706 is not valid anymore inside "if". Of course it could be worked around
1707 by using explicitely %$$arg1 but this is ugly IMHO.
1709 \b Fixed memory leak in \c{%undef}. The origline wasn't freed before
1712 \b Fixed trap in preprocessor when line expanded to empty set of tokens.
1713 This happens, for example, in the following case:
1715 \c #define SOMETHING
1719 \S{cl-0.98} Version 0.98
1721 All changes since NASM 0.98p3 have been produced by H. Peter Anvin <hpa@zytor.com>.
1723 \b The documentation comment delimiter is \# not #.
1725 \b Allow EQU definitions to refer to external labels; reported by
1728 \b Re-enable support for RDOFF v1; reported by Pedro Gimeno.
1730 \b Updated License file per OK from Simon and Julian.
1733 \S{cl-0.98p9} Version 0.98p9
1735 \b Update documentation (although the instruction set reference will
1736 have to wait; I don't want to hold up the 0.98 release for it.)
1738 \b Verified that the NASM implementation of the PEXTRW and PMOVMSKB
1739 instructions is correct. The encoding differs from what the Intel
1740 manuals document, but the Pentium III behaviour matches NASM, not
1743 \b Fix handling of implicit sizes in PSHUFW and PINSRW, reported by
1746 \b Resurrect the -s option, which was removed when changing the
1747 diagnostic output to stdout.
1750 \S{cl-0.98p8} Version 0.98p8
1752 \b Fix for "DB" when NASM is running on a bigendian machine.
1754 \b Invoke insns.pl once for each output script, making Makefile.in
1755 legal for "make -j".
1757 \b Improve the Unix configure-based makefiles to make package
1760 \b Included an RPM .spec file for building RPM (RedHat Package Manager)
1761 packages on Linux or Unix systems.
1763 \b Fix Makefile dependency problems.
1765 \b Change src/rdsrc.pl to include sectioning information in info
1766 output; required for install-info to work.
1768 \b Updated the RDOFF distribution to version 2 from Jules; minor
1769 massaging to make it compile in my environment.
1771 \b Split doc files that can be built by anyone with a Perl interpreter off
1772 into a separate archive.
1774 \b "Dress rehearsal" release!
1777 \S{cl-0.98p7} Version 0.98p7
1779 \b Fixed opcodes with a third byte-sized immediate argument to not
1780 complain if given "byte" on the immediate.
1782 \b Allow \c{%undef} to remove single-line macros with arguments. This
1783 matches the behaviour of #undef in the C preprocessor.
1785 \b Allow -d, -u, -i and -p to be specified as -D, -U, -I and -P for
1786 compatibility with most C compilers and preprocessors. This allows
1787 Makefile options to be shared between cc and nasm, for example.
1791 \b Went through the list of Katmai instructions and hopefully fixed the
1792 (rather few) mistakes in it.
1794 \b (Hopefully) fixed a number of disassembler bugs related to ambiguous
1795 instructions (disambiguated by -p) and SSE instructions with REP.
1797 \b Fix for bug reported by Mark Junger: "call dword 0x12345678" should
1798 work and may add an OSP (affected CALL, JMP, Jcc).
1800 \b Fix for environments when "stderr" isn't a compile-time constant.
1803 \S{cl-0.98p6} Version 0.98p6
1806 \b Took officially over coordination of the 0.98 release; so drop
1807 the p3.x notation. Skipped p4 and p5 to avoid confusion with John
1808 Fine's J4 and J5 releases.
1810 \b Update the documentation; however, it still doesn't include
1811 documentation for the various new instructions. I somehow wonder if
1812 it makes sense to have an instruction set reference in the assembler
1813 manual when Intel et al have PDF versions of their manuals online.
1815 \b Recognize "idt" or "centaur" for the -p option to ndisasm.
1817 \b Changed error messages back to stderr where they belong, but add an
1818 -E option to redirect them elsewhere (the DOS shell cannot redirect
1821 \b -M option to generate Makefile dependencies (based on code from Alex
1824 \b \c{%undef} preprocessor directive, and -u option, that undefines a
1827 \b OS/2 Makefile (Mkfiles/Makefile.os2) for Borland under OS/2; from
1830 \b Various minor bugfixes (reported by):
1831 - Dangling \c{%s} in preproc.c (Martin Junker)
1833 \b THERE ARE KNOWN BUGS IN SSE AND THE OTHER KATMAI INSTRUCTIONS. I am
1834 on a trip and didn't bring the Katmai instruction reference, so I
1835 can't work on them right now.
1837 \b Updated the License file per agreement with Simon and Jules to
1838 include a GPL distribution clause.
1841 \S{cl-0.98p3.7} Version 0.98p3.7
1843 \b (Hopefully) fixed the canned Makefiles to include the outrdf2 and
1846 \b Renamed changes.asm to changed.asm.
1849 \S{cl-0.98p3.6} Version 0.98p3.6
1851 \b Fixed a bunch of instructions that were added in 0.98p3.5 which had
1852 memory operands, and the address-size prefix was missing from the
1853 instruction pattern.
1856 \S{cl-0.98p3.5} Version 0.98p3.5
1858 \b Merged in changes from John S. Fine's 0.98-J5 release. John's based
1859 0.98-J5 on my 0.98p3.3 release; this merges the changes.
1861 \b Expanded the instructions flag field to a long so we can fit more
1862 flags; mark SSE (KNI) and AMD or Katmai-specific instructions as
1865 \b Fix the "PRIV" flag on a bunch of instructions, and create new
1866 "PROT" flag for protected-mode-only instructions (orthogonal to if
1867 the instruction is privileged!) and new "SMM" flag for SMM-only
1870 \b Added AMD-only SYSCALL and SYSRET instructions.
1872 \b Make SSE actually work, and add new Katmai MMX instructions.
1874 \b Added a -p (preferred vendor) option to ndisasm so that it can
1875 distinguish e.g. Cyrix opcodes also used in SSE. For example:
1877 \c ndisasm -p cyrix aliased.bin
1878 \c 00000000 670F514310 paddsiw mm0,[ebx+0x10]
1879 \c 00000005 670F514320 paddsiw mm0,[ebx+0x20]
1880 \c ndisasm -p intel aliased.bin
1881 \c 00000000 670F514310 sqrtps xmm0,[ebx+0x10]
1882 \c 00000005 670F514320 sqrtps xmm0,[ebx+0x20]
1884 \b Added a bunch of Cyrix-specific instructions.
1887 \S{cl-0.98p3.4} Version 0.98p3.4
1889 \b Made at least an attempt to modify all the additional Makefiles (in
1890 the Mkfiles directory). I can't test it, but this was the best I
1893 \b DOS DJGPP+"Opus Make" Makefile from John S. Fine.
1895 \b changes.asm changes from John S. Fine.
1898 \S{cl-0.98p3.3} Version 0.98p3.3
1900 \b Patch from Conan Brink to allow nesting of \c{%rep} directives.
1902 \b If we're going to allow INT01 as an alias for INT1/ICEBP (one of
1903 Jules 0.98p3 changes), then we should allow INT03 as an alias for INT3
1906 \b Updated changes.asm to include the latest changes.
1908 \b Tried to clean up the <CR>s that had snuck in from a DOS/Windows
1909 environment into my Unix environment, and try to make sure than
1910 DOS/Windows users get them back.
1912 \b We would silently generate broken tools if insns.dat wasn't sorted
1913 properly. Change insns.pl so that the order doesn't matter.
1915 \b Fix bug in insns.pl (introduced by me) which would cause conditional
1916 instructions to have an extra "cc" in disassembly, e.g. "jnz"
1917 disassembled as "jccnz".
1920 \S{cl-0.98p3.2} Version 0.98p3.2
1922 \b Merged in John S. Fine's changes from his 0.98-J4 prerelease; see
1923 http://www.csoft.net/cz/johnfine/
1925 \b Changed previous "spotless" Makefile target (appropriate for distribution)
1926 to "distclean", and added "cleaner" target which is same as "clean"
1927 except deletes files generated by Perl scripts; "spotless" is union.
1929 \b Removed BASIC programs from distribution. Get a Perl interpreter
1930 instead (see below.)
1932 \b Calling this "pre-release 3.2" rather than "p3-hpa2" because of
1933 John's contributions.
1935 \b Actually link in the IEEE output format (zoutieee.c); fix a bunch of
1936 compiler warnings in that file. Note I don't know what IEEE output
1937 is supposed to look like, so these changes were made "blind".
1940 \S{cl-0.98p3-hpa} Version 0.98p3-hpa
1942 \b Merged nasm098p3.zip with nasm-0.97.tar.gz to create a fully
1943 buildable version for Unix systems (Makefile.in updates, etc.)
1945 \b Changed insns.pl to create the instruction tables in nasm.h and
1946 names.c, so that a new instruction can be added by adding it *only*
1949 \b Added the following new instructions: SYSENTER, SYSEXIT, FXSAVE,
1950 FXRSTOR, UD1, UD2 (the latter two are two opcodes that Intel
1951 guarantee will never be used; one of them is documented as UD2 in
1952 Intel documentation, the other one just as "Undefined Opcode" --
1953 calling it UD1 seemed to make sense.)
1955 \b MAX_SYMBOL was defined to be 9, but LOADALL286 and LOADALL386 are 10
1956 characters long. Now MAX_SYMBOL is derived from insns.dat.
1958 \b A note on the BASIC programs included: forget them. insns.bas is
1959 already out of date. Get yourself a Perl interpreter for your
1960 platform of choice at
1961 \W{http://www.cpan.org/ports/index.html}{http://www.cpan.org/ports/index.html}.
1964 \S{cl-0.98p3} Version 0.98 pre-release 3
1966 \b added response file support, improved command line handling, new layout
1969 \b fixed limit checking bug, 'OUT byte nn, reg' bug, and a couple of rdoff
1970 related bugs, updated Wishlist; 0.98 Prerelease 3.
1973 \S{cl-0.98p2} Version 0.98 pre-release 2
1975 \b fixed bug in outcoff.c to do with truncating section names longer
1976 than 8 characters, referencing beyond end of string; 0.98 pre-release 2
1979 \S{cl-0.98p1} Version 0.98 pre-release 1
1981 \b Fixed a bug whereby STRUC didn't work at all in RDF.
1983 \b Fixed a problem with group specification in PUBDEFs in OBJ.
1985 \b Improved ease of adding new output formats. Contribution due to
1988 \b Fixed a bug in relocations in the `bin' format: was showing up when
1989 a relocatable reference crossed an 8192-byte boundary in any output
1992 \b Fixed a bug in local labels: local-label lookups were inconsistent
1993 between passes one and two if an EQU occurred between the definition
1994 of a global label and the subsequent use of a local label local to
1997 \b Fixed a seg-fault in the preprocessor (again) which happened when
1998 you use a blank line as the first line of a multi-line macro
1999 definition and then defined a label on the same line as a call to
2002 \b Fixed a stale-pointer bug in the handling of the NASM environment
2003 variable. Thanks to Thomas McWilliams.
2005 \b ELF had a hard limit on the number of sections which caused
2006 segfaults when transgressed. Fixed.
2008 \b Added ability for ndisasm to read from stdin by using `-' as the
2011 \b ndisasm wasn't outputting the TO keyword. Fixed.
2013 \b Fixed error cascade on bogus expression in \c{%if} - an error in
2014 evaluation was causing the entire \c{%if} to be discarded, thus creating
2015 trouble later when the \c{%else} or \c{%endif} was encountered.
2017 \b Forward reference tracking was instruction-granular not operand-
2018 granular, which was causing 286-specific code to be generated
2019 needlessly on code of the form `shr word [forwardref],1'. Thanks to
2020 Jim Hague for sending a patch.
2022 \b All messages now appear on stdout, as sending them to stderr serves
2023 no useful purpose other than to make redirection difficult.
2025 \b Fixed the problem with EQUs pointing to an external symbol - this
2026 now generates an error message.
2028 \b Allowed multiple size prefixes to an operand, of which only the first
2029 is taken into account.
2031 \b Incorporated John Fine's changes, including fixes of a large number
2032 of preprocessor bugs, some small problems in OBJ, and a reworking of
2033 label handling to define labels before their line is assembled, rather
2036 \b Reformatted a lot of the source code to be more readable. Included
2037 'coding.txt' as a guideline for how to format code for contributors.
2039 \b Stopped nested \c{%reps} causing a panic - they now cause a slightly more
2040 friendly error message instead.
2042 \b Fixed floating point constant problems (patch by Pedro Gimeno)
2044 \b Fixed the return value of insn_size() not being checked for -1, indicating
2047 \b Incorporated 3Dnow! instructions.
2049 \b Fixed the 'mov eax, eax + ebx' bug.
2051 \b Fixed the GLOBAL EQU bug in ELF. Released developers release 3.
2053 \b Incorporated John Fine's command line parsing changes
2055 \b Incorporated David Lindauer's OMF debug support
2057 \b Made changes for LCC 4.0 support (\c{__NASM_CDecl__}, removed register size
2058 specification warning when sizes agree).
2061 \H{cl-0.9x} NASM 0.9 Series
2063 Revisions before 0.98.
2066 \S{cl-0.97} Version 0.97 released December 1997
2068 \b This was entirely a bug-fix release to 0.96, which seems to have got
2071 \b Fixed stupid mistake in OBJ which caused `MOV EAX,<constant>' to
2072 fail. Caused by an error in the `MOV EAX,<segment>' support.
2074 \b ndisasm hung at EOF when compiled with lcc on Linux because lcc on
2075 Linux somehow breaks feof(). ndisasm now does not rely on feof().
2077 \b A heading in the documentation was missing due to a markup error in
2078 the indexing. Fixed.
2080 \b Fixed failure to update all pointers on realloc() within extended-
2081 operand code in parser.c. Was causing wrong behaviour and seg faults
2082 on lines such as `dd 0.0,0.0,0.0,0.0,...'
2084 \b Fixed a subtle preprocessor bug whereby invoking one multi-line
2085 macro on the first line of the expansion of another, when the second
2086 had been invoked with a label defined before it, didn't expand the
2089 \b Added internal.doc back in to the distribution archives - it was
2090 missing in 0.96 *blush*
2092 \b Fixed bug causing 0.96 to be unable to assemble its own test files,
2093 specifically objtest.asm. *blush again*
2095 \b Fixed seg-faults and bogus error messages caused by mismatching
2096 \c{%rep} and \c{%endrep} within multi-line macro definitions.
2098 \b Fixed a problem with buffer overrun in OBJ, which was causing
2099 corruption at ends of long PUBDEF records.
2101 \b Separated DOS archives into main-program and documentation to reduce
2105 \S{cl-0.96} Version 0.96 released November 1997
2107 \b Fixed a bug whereby, if `nasm sourcefile' would cause a filename
2108 collision warning and put output into `nasm.out', then `nasm
2109 sourcefile -o outputfile' still gave the warning even though the
2111 Fixed name pollution under Digital UNIX: one of its header files
2112 defined R_SP, which broke the enum in nasm.h.
2114 \b Fixed minor instruction table problems: FUCOM and FUCOMP didn't have
2115 two-operand forms; NDISASM didn't recognise the longer register
2116 forms of PUSH and POP (eg FF F3 for PUSH BX); TEST mem,imm32 was
2117 flagged as undocumented; the 32-bit forms of CMOV had 16-bit operand
2118 size prefixes; `AAD imm' and `AAM imm' are no longer flagged as
2119 undocumented because the Intel Architecture reference documents
2122 \b Fixed a problem with the local-label mechanism, whereby strange
2123 types of symbol (EQUs, auto-defined OBJ segment base symbols)
2124 interfered with the `previous global label' value and screwed up
2127 \b Fixed a bug whereby the stub preprocessor didn't communicate with
2128 the listing file generator, so that the -a and -l options in
2129 conjunction would produce a useless listing file.
2131 \b Merged `os2' object file format back into `obj', after discovering
2132 that `obj' _also_ shouldn't have a link pass separator in a module
2133 containing a non-trivial MODEND. Flat segments are now declared
2134 using the FLAT attribute. `os2' is no longer a valid object format
2137 \b Removed the fixed-size temporary storage in the evaluator. Very very
2138 long expressions (like `mov ax,1+1+1+1+...' for two hundred 1s or
2139 so) should now no longer crash NASM.
2141 \b Fixed a bug involving segfaults on disassembly of MMX instructions,
2142 by changing the meaning of one of the operand-type flags in nasm.h.
2143 This may cause other apparently unrelated MMX problems; it needs to
2144 be tested thoroughly.
2146 \b Fixed some buffer overrun problems with large OBJ output files.
2147 Thanks to DJ Delorie for the bug report and fix.
2149 \b Made preprocess-only mode actually listen to the \c{%line} markers as it
2150 prints them, so that it can report errors more sanely.
2152 \b Re-designed the evaluator to keep more sensible track of expressions
2153 involving forward references: can now cope with previously-nightmare
2160 \b Added the ALIGN and ALIGNB standard macros.
2162 \b Added PIC support in ELF: use of WRT to obtain the four extra
2163 relocation types needed.
2165 \b Added the ability for output file formats to define their own
2166 extensions to the GLOBAL, COMMON and EXTERN directives.
2168 \b Implemented common-variable alignment, and global-symbol type and
2169 size declarations, in ELF.
2171 \b Implemented NEAR and FAR keywords for common variables, plus
2172 far-common element size specification, in OBJ.
2174 \b Added a feature whereby EXTERNs and COMMONs in OBJ can be given a
2175 default WRT specification (either a segment or a group).
2177 \b Transformed the Unix NASM archive into an auto-configuring package.
2179 \b Added a sanity-check for people applying SEG to things which are
2180 already segment bases: this previously went unnoticed by the SEG
2181 processing and caused OBJ-driver panics later.
2183 \b Added the ability, in OBJ format, to deal with `MOV EAX,<segment>'
2184 type references: OBJ doesn't directly support dword-size segment
2185 base fixups, but as long as the low two bytes of the constant term
2186 are zero, a word-size fixup can be generated instead and it will
2189 \b Added the ability to specify sections' alignment requirements in
2190 Win32 object files and pure binary files.
2192 \b Added preprocess-time expression evaluation: the \c{%assign} (and
2193 \c{%iassign}) directive and the bare \c{%if} (and \c{%elif}) conditional. Added
2194 relational operators to the evaluator, for use only in \c{%if}
2195 constructs: the standard relationals = < > <= >= <> (and C-like
2196 synonyms == and !=) plus low-precedence logical operators &&, ^^ and
2199 \b Added a preprocessor repeat construct: \c{%rep} / \c{%exitrep} / \c{%endrep}.
2201 \b Added the __FILE__ and __LINE__ standard macros.
2203 \b Added a sanity check for number constants being greater than
2204 0xFFFFFFFF. The warning can be disabled.
2206 \b Added the %0 token whereby a variadic multi-line macro can tell how
2207 many parameters it's been given in a specific invocation.
2209 \b Added \c{%rotate}, allowing multi-line macro parameters to be cycled.
2211 \b Added the `*' option for the maximum parameter count on multi-line
2212 macros, allowing them to take arbitrarily many parameters.
2214 \b Added the ability for the user-level forms of EXTERN, GLOBAL and
2215 COMMON to take more than one argument.
2217 \b Added the IMPORT and EXPORT directives in OBJ format, to deal with
2220 \b Added some more preprocessor \c{%if} constructs: \c{%ifidn} / \c{%ifidni} (exact
2221 textual identity), and \c{%ifid} / \c{%ifnum} / \c{%ifstr} (token type testing).
2223 \b Added the ability to distinguish SHL AX,1 (the 8086 version) from
2224 SHL AX,BYTE 1 (the 286-and-upwards version whose constant happens to
2227 \b Added NetBSD/FreeBSD/OpenBSD's variant of a.out format, complete
2228 with PIC shared library features.
2230 \b Changed NASM's idiosyncratic handling of FCLEX, FDISI, FENI, FINIT,
2231 FSAVE, FSTCW, FSTENV, and FSTSW to bring it into line with the
2232 otherwise accepted standard. The previous behaviour, though it was a
2233 deliberate feature, was a deliberate feature based on a
2234 misunderstanding. Apologies for the inconvenience.
2236 \b Improved the flexibility of ABSOLUTE: you can now give it an
2237 expression rather than being restricted to a constant, and it can
2238 take relocatable arguments as well.
2240 \b Added the ability for a variable to be declared as EXTERN multiple
2241 times, and the subsequent definitions are just ignored.
2243 \b We now allow instruction prefixes (CS, DS, LOCK, REPZ etc) to be
2244 alone on a line (without a following instruction).
2246 \b Improved sanity checks on whether the arguments to EXTERN, GLOBAL
2247 and COMMON are valid identifiers.
2249 \b Added misc/exebin.mac to allow direct generation of .EXE files by
2250 hacking up an EXE header using DB and DW; also added test/binexe.asm
2251 to demonstrate the use of this. Thanks to Yann Guidon for
2252 contributing the EXE header code.
2254 \b ndisasm forgot to check whether the input file had been successfully
2255 opened. Now it does. Doh!
2257 \b Added the Cyrix extensions to the MMX instruction set.
2259 \b Added a hinting mechanism to allow [EAX+EBX] and [EBX+EAX] to be
2260 assembled differently. This is important since [ESI+EBP] and
2261 [EBP+ESI] have different default base segment registers.
2263 \b Added support for the PharLap OMF extension for 4096-byte segment
2267 \S{cl-0.95 released July 1997} Version 0.95 released July 1997
2269 \b Fixed yet another ELF bug. This one manifested if the user relied on
2270 the default segment, and attempted to define global symbols without
2271 first explicitly declaring the target segment.
2273 \b Added makefiles (for NASM and the RDF tools) to build Win32 console
2274 apps under Symantec C++. Donated by Mark Junker.
2276 \b Added `macros.bas' and `insns.bas', QBasic versions of the Perl
2277 scripts that convert `standard.mac' to `macros.c' and convert
2278 `insns.dat' to `insnsa.c' and `insnsd.c'. Also thanks to Mark
2281 \b Changed the diassembled forms of the conditional instructions so
2282 that JB is now emitted as JC, and other similar changes. Suggested
2283 list by Ulrich Doewich.
2285 \b Added `@' to the list of valid characters to begin an identifier
2288 \b Documentary changes, notably the addition of the `Common Problems'
2289 section in nasm.doc.
2291 \b Fixed a bug relating to 32-bit PC-relative fixups in OBJ.
2293 \b Fixed a bug in perm_copy() in labels.c which was causing exceptions
2294 in cleanup_labels() on some systems.
2296 \b Positivity sanity check in TIMES argument changed from a warning to
2297 an error following a further complaint.
2299 \b Changed the acceptable limits on byte and word operands to allow
2300 things like `~10111001b' to work.
2302 \b Fixed a major problem in the preprocessor which caused seg-faults if
2303 macro definitions contained blank lines or comment-only lines.
2305 \b Fixed inadequate error checking on the commas separating the
2306 arguments to `db', `dw' etc.
2308 \b Fixed a crippling bug in the handling of macros with operand counts
2309 defined with a `+' modifier.
2311 \b Fixed a bug whereby object file formats which stored the input file
2312 name in the output file (such as OBJ and COFF) weren't doing so
2313 correctly when the output file name was specified on the command
2316 \b Removed [INC] and [INCLUDE] support for good, since they were
2319 \b Fixed a bug in OBJ which caused all fixups to be output in 16-bit
2320 (old-format) FIXUPP records, rather than putting the 32-bit ones in
2321 FIXUPP32 (new-format) records.
2323 \b Added, tentatively, OS/2 object file support (as a minor variant on
2326 \b Updates to Fox Cutter's Borland C makefile, Makefile.bc2.
2328 \b Removed a spurious second fclose() on the output file.
2330 \b Added the `-s' command line option to redirect all messages which
2331 would go to stderr (errors, help text) to stdout instead.
2333 \b Added the `-w' command line option to selectively suppress some
2334 classes of assembly warning messages.
2336 \b Added the `-p' pre-include and `-d' pre-define command-line options.
2338 \b Added an include file search path: the `-i' command line option.
2340 \b Fixed a silly little preprocessor bug whereby starting a line with a
2341 `%!' environment-variable reference caused an `unknown directive'
2344 \b Added the long-awaited listing file support: the `-l' command line
2347 \b Fixed a problem with OBJ format whereby, in the absence of any
2348 explicit segment definition, non-global symbols declared in the
2349 implicit default segment generated spurious EXTDEF records in the
2352 \b Added the NASM environment variable.
2354 \b From this version forward, Win32 console-mode binaries will be
2355 included in the DOS distribution in addition to the 16-bit binaries.
2356 Added Makefile.vc for this purpose.
2358 \b Added `return 0;' to test/objlink.c to prevent compiler warnings.
2360 \b Added the __NASM_MAJOR__ and __NASM_MINOR__ standard defines.
2362 \b Added an alternative memory-reference syntax in which prefixing an
2363 operand with `&' is equivalent to enclosing it in square brackets,
2364 at the request of Fox Cutter.
2366 \b Errors in pass two now cause the program to return a non-zero error
2367 code, which they didn't before.
2369 \b Fixed the single-line macro cycle detection, which didn't work at
2370 all on macros with no parameters (caused an infinite loop). Also
2371 changed the behaviour of single-line macro cycle detection to work
2372 like cpp, so that macros like `extrn' as given in the documentation
2375 \b Fixed the implementation of WRT, which was too restrictive in that
2376 you couldn't do `mov ax,[di+abc wrt dgroup]' because (di+abc) wasn't
2377 a relocatable reference.
2380 \S{cl-0.94 released April 1997} Version 0.94 released April 1997
2383 \b Major item: added the macro processor.
2385 \b Added undocumented instructions SMI, IBTS, XBTS and LOADALL286. Also
2386 reorganised CMPXCHG instruction into early-486 and Pentium forms.
2387 Thanks to Thobias Jones for the information.
2389 \b Fixed two more stupid bugs in ELF, which were causing `ld' to
2390 continue to seg-fault in a lot of non-trivial cases.
2392 \b Fixed a seg-fault in the label manager.
2394 \b Stopped FBLD and FBSTP from _requiring_ the TWORD keyword, which is
2395 the only option for BCD loads/stores in any case.
2397 \b Ensured FLDCW, FSTCW and FSTSW can cope with the WORD keyword, if
2398 anyone bothers to provide it. Previously they complained unless no
2399 keyword at all was present.
2401 \b Some forms of FDIV/FDIVR and FSUB/FSUBR were still inverted: a
2402 vestige of a bug that I thought had been fixed in 0.92. This was
2403 fixed, hopefully for good this time...
2405 \b Another minor phase error (insofar as a phase error can _ever_ be
2406 minor) fixed, this one occurring in code of the form
2408 \c rol ax,forward_reference
2409 \c forward_reference equ 1
2411 \b The number supplied to TIMES is now sanity-checked for positivity,
2412 and also may be greater than 64K (which previously didn't work on
2415 \b Added Watcom C makefiles, and misc/pmw.bat, donated by Dominik Behr.
2417 \b Added the INCBIN pseudo-opcode.
2419 \b Due to the advent of the preprocessor, the [INCLUDE] and [INC]
2420 directives have become obsolete. They are still supported in this
2421 version, with a warning, but won't be in the next.
2423 \b Fixed a bug in OBJ format, which caused incorrect object records to
2424 be output when absolute labels were made global.
2426 \b Updates to RDOFF subdirectory, and changes to outrdf.c.
2429 \S{cl-0.93 released January 1997} Version 0.93 released January 1997
2431 This release went out in a great hurry after semi-crippling bugs
2434 \b Really \e{did} fix the stack overflows this time. *blush*
2436 \b Had problems with EA instruction sizes changing between passes, when
2437 an offset contained a forward reference and so 4 bytes were
2438 allocated for the offset in pass one; by pass two the symbol had
2439 been defined and happened to be a small absolute value, so only 1
2440 byte got allocated, causing instruction size mismatch between passes
2441 and hence incorrect address calculations. Fixed.
2443 \b Stupid bug in the revised ELF section generation fixed (associated
2444 string-table section for .symtab was hard-coded as 7, even when this
2445 didn't fit with the real section table). Was causing `ld' to
2446 seg-fault under Linux.
2448 \b Included a new Borland C makefile, Makefile.bc2, donated by Fox
2449 Cutter <lmb@comtch.iea.com>.
2452 \S{cl-0.92 released January 1997} Version 0.92 released January 1997
2454 \b The FDIVP/FDIVRP and FSUBP/FSUBRP pairs had been inverted: this was
2455 fixed. This also affected the LCC driver.
2457 \b Fixed a bug regarding 32-bit effective addresses of the form
2458 \c{[other_register+ESP]}.
2460 \b Documentary changes, notably documentation of the fact that Borland
2461 Win32 compilers use `obj' rather than `win32' object format.
2463 \b Fixed the COMENT record in OBJ files, which was formatted
2466 \b Fixed a bug causing segfaults in large RDF files.
2468 \b OBJ format now strips initial periods from segment and group
2469 definitions, in order to avoid complications with the local label
2472 \b Fixed a bug in disassembling far calls and jumps in NDISASM.
2474 \b Added support for user-defined sections in COFF and ELF files.
2476 \b Compiled the DOS binaries with a sensible amount of stack, to
2477 prevent stack overflows on any arithmetic expression containing
2480 \b Fixed a bug in handling of files that do not terminate in a newline.
2483 \S{cl-0.91 released November 1996} Version 0.91 released November 1996
2485 \b Loads of bug fixes.
2487 \b Support for RDF added.
2489 \b Support for DBG debugging format added.
2491 \b Support for 32-bit extensions to Microsoft OBJ format added.
2493 \b Revised for Borland C: some variable names changed, makefile added.
2495 \b LCC support revised to actually work.
2497 \b JMP/CALL NEAR/FAR notation added.
2499 \b `a16', `o16', `a32' and `o32' prefixes added.
2501 \b Range checking on short jumps implemented.
2503 \b MMX instruction support added.
2505 \b Negative floating point constant support added.
2507 \b Memory handling improved to bypass 64K barrier under DOS.
2509 \b \c{$} prefix to force treatment of reserved words as identifiers added.
2511 \b Default-size mechanism for object formats added.
2513 \b Compile-time configurability added.
2515 \b \c{#}, \c{@}, \c{~} and c\{?} are now valid characters in labels.
2517 \b \c{-e} and \c{-k} options in NDISASM added.
2520 \S{cl-0.90 released October 1996} Version 0.90 released October 1996
2522 First release version. First support for object file output. Other
2523 changes from previous version (0.3x) too numerous to document.