gas/NEWS

   1 -*- text -*-
   2
   3 Changes in 2.39:
   4
   5 * Remove (rudimentary) support for the x86-64 sub-architectures Intel L1OM and
   6   Intel K1OM.
   7
   8 Changes in 2.38:
   9
  10 * Add support for AArch64 system registers that were missing in previous
  11   releases.
  12
  13 * Add support for the LoongArch instruction set.
  14
  15 * Add a command-line option, -muse-unaligned-vector-move, for x86 target
  16   to encode aligned vector move as unaligned vector move.
  17
  18 * Add support for Cortex-R52+ for Arm.
  19
  20 * Add support for Cortex-A510, Cortex-A710, Cortex-X2 for AArch64.
  21
  22 * Add support for Cortex-A710 for Arm.
  23
  24 * Add support for Scalable Matrix Extension (SME) for AArch64.
  25
  26 * The --multibyte-handling=[allow|warn|warn-sym-only] option tells the
  27   assembler what to when it encoutners multibyte characters in the input.  The
  28   default is to allow them.  Setting the option to "warn" will generate a
  29   warning message whenever any multibyte character is encountered.  Using the
  30   option to "warn-sym-only" will make the assembler generate a warning whenever a
  31   symbol is defined containing multibyte characters.  (References to undefined
  32   symbols will not generate warnings).
  33
  34 * Outputs of .ds.x directive and .tfloat directive with hex input from
  35   x86 assembler have been reduced from 12 bytes to 10 bytes to match the
  36   output of .tfloat directive.
  37
  38 * Add support for 'armv8.8-a', 'armv9-a', 'armv9.1-a', 'armv9.2-a' and
  39   'armv9.3-a' for -march in AArch64 GAS.
  40
  41 * Add support for 'armv8.7-a', 'armv8.8-a', 'armv9-a', 'armv9.1-a',
  42   'armv9.2-a' and 'armv9.3-a' for -march in Arm GAS.
  43
  44 * Add support for Intel AVX512_FP16 instructions.
  45
  46 Changes in 2.37:
  47
  48 * arm-symbianelf support removed.
  49
  50 * Add support for Realm Management Extension (RME) for AArch64.
  51
  52 Changes in 2.36:
  53
  54 * Add support for Intel AVX VNNI instructions.
  55
  56 * Add support for Intel HRESET instruction.
  57
  58 * Add support for Intel UINTR instructions.
  59
  60 * Support non-absolute segment values for i386 lcall and ljmp.
  61
  62 * When setting the link order attribute of ELF sections, it is now possible to
  63   use a numeric section index instead of symbol name.
  64
  65 * Add support for Cortex-A78, Cortex-A78AE, Cortex-A78C and Cortex-X1 for
  66   AArch64 and ARM.
  67   Add support for Cortex-R82, Neoverse V1, and Neoverse N2 for ARM.
  68
  69 * Add support for ETMv4 (Embedded Trace Macrocell), ETE (Embedded Trace
  70   Extension), TRBE (Trace Buffer Extension)and BRBE (Branch Record Buffer
  71   Extension) system registers for AArch64.
  72
  73 * Add support for Armv8-R and Armv8.7-A AArch64.
  74
  75 * Add support for DSB memory nXS barrier, WFET and WFIT instruction for Armv8.7
  76   AArch64.
  77
  78 * Add support for +flagm feature for -march in Armv8.4 AArch64.
  79
  80 * Add support for +ls64 feature for -march in Armv8.7 AArch64. Add atomic
  81   64-byte load/store instructions for this feature.
  82
  83 * Add support for +pauth (Pointer Authentication) feature for -march in
  84   AArch64.
  85
  86 * Add support for Intel TDX instructions.
  87
  88 * Add support for Intel Key Locker instructions.
  89
  90 * Added a .nop directive to generate a single no-op instruction in a target
  91   neutral manner.  This instruction does have an effect on DWARF line number
  92   generation, if that is active.
  93
  94 * Removed --reduce-memory-overheads and --hash-size as gas now
  95   uses hash tables that can be expand and shrink automatically.
  96
  97 * Add {disp16} pseudo prefix to x86 assembler.
  98
  99 * Add support for Intel AMX instructions.
 100
 101 * Configure with --enable-x86-used-note by default for Linux/x86.
 102
 103 * Add support for the SHF_GNU_RETAIN flag, which can be applied to
 104   sections using the 'R' flag in the .section directive.
 105   SHF_GNU_RETAIN specifies that the section should not be garbage
 106   collected by the linker.  It requires the GNU or FreeBSD ELF OSABIs.
 107
 108 Changes in 2.35:
 109
 110 * X86 NaCl target support is removed.
 111
 112 * Extend .symver directive to update visibility of the original symbol
 113   and assign one original symbol to different versioned symbols.
 114
 115 * Add support for Intel SERIALIZE and TSXLDTRK instructions.
 116
 117 * Add -mlfence-after-load=, -mlfence-before-indirect-branch= and
 118   -mlfence-before-ret= options to x86 assembler to help mitigate
 119   CVE-2020-0551.
 120
 121 * Add --gdwarf-5 option to the assembler to generate DWARF 5 debug output
 122   (if such output is being generated).  Added the ability to generate
 123   version 5 .debug_line sections.
 124
 125 * Add -mbig-obj support to i386 MingW targets.
 126
 127 Changes in 2.34:
 128
 129 * Add -malign-branch-boundary=NUM, -malign-branch=TYPE[+TYPE...],
 130   -malign-branch-prefix-size=NUM and -mbranches-within-32B-boundaries
 131   options to x86 assembler to align branches within a fixed boundary
 132   with segment prefixes or NOPs.
 133
 134 * Add support for Zilog eZ80 and Zilog Z180 CPUs.
 135
 136 * Add support for z80-elf target.
 137
 138 * Add support for relocation of each byte or word of multibyte value to Z80
 139   targets (just use right shift to 0, 8, 16, or 24 bits or AND operation
 140   with 0xff/0xffff mask): ld a, label >> 16 \ ld hl, label & 0xffff
 141
 142 * Add SDCC support for Z80 targets.
 143
 144 Changes in 2.33:
 145
 146 * Add support for the Arm Scalable Vector Extension version 2 (SVE2)
 147   instructions.
 148
 149 * Add support for the Arm Transactional Memory Extension (TME)
 150   instructions.
 151
 152 * Add support for the Armv8.1-M Mainline and M-profile Vector Extension (MVE)
 153   instructions.
 154
 155 * For MIPS, Add -m[no-]fix-loongson3-llsc option to fix (or not) Loongson3
 156   LLSC Errata.  Add a --enable-mips-fix-loongson3-llsc=[yes|no] configure
 157   time option to set the default behavior. Set the default if the configure
 158   option is not used to "no".
 159
 160 * Add support for the Arm Cortex-A76AE, Cortex-A77 and Cortex-M35P
 161   processors.
 162
 163 * Add support for the AArch64 Cortex-A34, Cortex-A65, Cortex-A65AE,
 164   Cortex-A76AE, and Cortex-A77 processors.
 165
 166 * Add .float16 directive for both Arm and AArch64 to allow encoding of 16-bit
 167   floating point literals.  Add .float16_format directive and
 168   -mfp16-format=[ieee|alternative] option for Arm to control the format of the
 169   encoding.
 170
 171 * Add --gdwarf-cie-version command line flag.  This allows control over which
 172   version of DWARF CIE the assembler creates.
 173
 174 Changes in 2.32:
 175
 176 * Add -mvexwig=[0|1] option to x86 assembler to control encoding of
 177   VEX.W-ignored (WIG) VEX instructions.
 178
 179 * Add -mx86-used-note=[yes|no] option to generate (or not) x86 GNU property
 180   notes.  Add a --enable-x86-used-note configure time option to set the
 181   default behavior.  Set the default if the configure option is not used
 182   to "no".
 183
 184 * Add support for the MIPS Loongson EXTensions R2 (EXT2) instructions.
 185
 186 * Add support for the MIPS Loongson EXTensions (EXT) instructions.
 187
 188 * Add support for the MIPS Loongson Content Address Memory (CAM) ASE.
 189
 190 * Add support for the C-SKY processor series.
 191
 192 * Add support for the MIPS Loongson MultiMedia extensions Instructions (MMI)
 193   ASE.
 194
 195 Changes in 2.31:
 196
 197 * The ADR and ADRL pseudo-instructions supported by the ARM assembler
 198   now only set the bottom bit of the address of thumb function symbols
 199   if the -mthumb-interwork command line option is active.
 200
 201 * Add support for the MIPS Global INValidate (GINV) ASE.
 202
 203 * Add support for the MIPS Cyclic Redudancy Check (CRC) ASE.
 204
 205 * Add support for the Freescale S12Z architecture.
 206
 207 * Add --generate-missing-build-notes=[yes|no] option to create (or not) GNU
 208   Build Attribute notes if none are present in the input sources.  Add a
 209   --enable-generate-build-notes=[yes|no] configure time option to set the
 210   default behaviour.  Set the default if the configure option is not used
 211   to "no".
 212
 213 * Remove -mold-gcc command-line option for x86 targets.
 214
 215 * Add -O[2|s] command-line options to x86 assembler to enable alternate
 216   shorter instruction encoding.
 217
 218 * Add support for .nops directive.  It is currently supported only for
 219   x86 targets.
 220
 221 Changes in 2.30:
 222
 223 * Add support for loaction views in DWARF debug line information.
 224
 225 Changes in 2.29:
 226
 227 * Add support for ELF SHF_GNU_MBIND.
 228
 229 * Add support for the WebAssembly file format and wasm32 ELF conversion.
 230
 231 * PowerPC gas now checks that the correct register class is used in
 232   instructions.  For instance, "addi %f4,%cr3,%r31" warns three times
 233   that the registers are invalid.
 234
 235 * Add support for the Texas Instruments PRU processor.
 236
 237 * Support for the ARMv8-R architecture and Cortex-R52 processor has been
 238   added to the ARM port.
 239
 240 Changes in 2.28:
 241
 242 * Add support for the RISC-V architecture.
 243
 244 * Add support for the ARM Cortex-M23 and Cortex-M33 processors.
 245
 246 Changes in 2.27:
 247
 248 * Default to --enable-compressed-debug-sections=gas for Linux/x86 targets.
 249
 250 * Add --no-pad-sections to stop the assembler from padding the end of output
 251   sections up to their alignment boundary.
 252
 253 * Support for the ARMv8-M architecture has been added to the ARM port.  Support
 254   for the ARMv8-M Security and DSP Extensions has also been added to the ARM
 255   port.
 256
 257 * ARC backend accepts .extInstruction, .extCondCode, .extAuxRegister, and
 258   .extCoreRegister pseudo-ops that allow an user to define custom
 259   instructions, conditional codes, auxiliary and core registers.
 260
 261 * Add a configure option --enable-elf-stt-common to decide whether ELF
 262   assembler should generate common symbols with the STT_COMMON type by
 263   default.  Default to no.
 264
 265 * New command-line option --elf-stt-common= for ELF targets to control
 266   whether to generate common symbols with the STT_COMMON type.
 267
 268 * Add ability to set section flags and types via numeric values for ELF
 269   based targets.
 270
 271 * Add a configure option --enable-x86-relax-relocations to decide whether
 272   x86 assembler should generate relax relocations by default.  Default to
 273   yes, except for x86 Solaris targets older than Solaris 12.
 274
 275 * New command-line option -mrelax-relocations= for x86 target to control
 276   whether to generate relax relocations.
 277
 278 * New command-line option -mfence-as-lock-add=yes for x86 target to encode
 279   lfence, mfence and sfence as "lock addl $0x0, (%[re]sp)".
 280
 281 * Add assembly-time relaxation option for ARC cpus.
 282
 283 * Add --with-cpu=TYPE configure option for ARC gas.  This allows the default
 284   cpu type to be adjusted at configure time.
 285
 286 Changes in 2.26:
 287
 288 * Add a configure option --enable-compressed-debug-sections={all,gas} to
 289   decide whether DWARF debug sections should be compressed by default.
 290
 291 * Add support for the ARC EM/HS, and ARC600/700 architectures.  Remove
 292   assembler support for Argonaut RISC architectures.
 293
 294 * Symbol and label names can now be enclosed in double quotes (") which allows
 295   them to contain characters that are not part of valid symbol names in high
 296   level languages.
 297
 298 * Added the correctly spelled -march=armv6kz, for ARMv6KZ support.  The
 299   previous spelling, -march=armv6zk, is still accepted.
 300
 301 * Support for the ARMv8.1 architecture has been added to the Aarch64 port.
 302   Support for the individual ARMv8.1 Adv.SIMD, LOR and PAN architecture
 303   extensions has also been added to the Aarch64 port.
 304
 305 * Support for the ARMv8.1 architecture has been added to the ARM port.  Support
 306   for the individual ARMv8.1 Adv.SIMD and PAN architecture extensions has also
 307   been added to the ARM port.
 308
 309 * Extend --compress-debug-sections option to support
 310   --compress-debug-sections=[none|zlib|zlib-gnu|zlib-gabi] for ELF
 311   targets.
 312
 313 * --compress-debug-sections is turned on for Linux/x86 by default.
 314
 315 Changes in 2.25:
 316
 317 * Add support for the AVR Tiny microcontrollers.
 318
 319 * Replace support for openrisc and or32 with support for or1k.
 320
 321 * Enhanced the ARM port to accept the assembler output from the CodeComposer
 322   Studio tool.  Support is enabled via the new command-line option -mccs.
 323
 324 * Add support for the Andes NDS32.
 325
 326 Changes in 2.24:
 327
 328 * Add support for the Texas Instruments MSP430X processor.
 329
 330 * Add -gdwarf-sections command-line option to enable per-code-section
 331   generation of DWARF .debug_line sections.
 332
 333 * Add support for Altera Nios II.
 334
 335 * Add support for the Imagination Technologies Meta processor.
 336
 337 * Add support for the v850e3v5.
 338
 339 * Remove assembler support for MIPS ECOFF targets.
 340
 341 Changes in 2.23:
 342
 343 * Add support for the 64-bit ARM architecture: AArch64.
 344
 345 * Add support for S12X processor.
 346
 347 * Add support for the VLE extension to the PowerPC architecture.
 348
 349 * Add support for the Freescale XGATE architecture.
 350
 351 * Add support for .bundle_align_mode, .bundle_lock, and .bundle_unlock
 352   directives.  These are currently available only for x86 and ARM targets.
 353
 354 * Add support for the Renesas RL78 architecture.
 355
 356 * Add support for the Adapteva EPIPHANY architecture.
 357
 358 * For x86, allow 'rep bsf', 'rep bsr', and 'rep ret' syntax.
 359
 360 Changes in 2.22:
 361
 362 * Add support for the Tilera TILEPro and TILE-Gx architectures.
 363
 364 Changes in 2.21:
 365
 366 * Gas no longer requires doubling of ampersands in macros.
 367
 368 * Add support for the TMS320C6000 (TI C6X) processor family.
 369
 370 * GAS now understands an extended syntax in the .section directive flags
 371   for COFF targets that allows the section's alignment to be specified.  This
 372   feature has also been backported to the 2.20 release series, starting with
 373   2.20.1.
 374
 375 * Add support for the Renesas RX processor.
 376
 377 * New command-line option, --compress-debug-sections, which requests
 378   compression of DWARF debug information sections in the relocatable output
 379   file.  Compressed debug sections are supported by readelf, objdump, and
 380   gold, but not currently by Gnu ld.
 381
 382 Changes in 2.20:
 383
 384 * Added support for v850e2 and v850e2v3.
 385
 386 * GNU/Linux targets now supports "gnu_unique_object" as a value in the .type
 387   pseudo op.  It marks the symbol as being globally unique in the entire
 388   process.
 389
 390 * ARM assembler now supports .inst[.nw] pseudo-ops to insert opcodes specified
 391   in binary rather than text.
 392
 393 * Add support for common symbol alignment to PE formats.
 394
 395 * Add support for the new discriminator column in the DWARF line table,
 396   with a discriminator operand for the .loc directive.
 397
 398 * Add support for Sunplus score architecture.
 399
 400 * The .type pseudo-op now accepts a type of STT_GNU_IFUNC which can be used to
 401   indicate that if the symbol is the target of a relocation, its value should
 402   not be use.  Instead the function should be invoked and its result used as
 403   the value.
 404
 405 * Add support for Lattice Mico32 (lm32) architecture.
 406
 407 * Add support for Xilinx MicroBlaze architecture.
 408
 409 Changes in 2.19:
 410
 411 * New pseudo op .cfi_val_encoded_addr, to record constant addresses in unwind
 412   tables without runtime relocation.
 413
 414 * New command-line option, -h-tick-hex, for sh, m32c, and h8/300 targets, which
 415   adds compatibility with H'00 style hex constants.
 416
 417 * New command-line option, -msse-check=[none|error|warning], for x86
 418   targets.
 419
 420 * New sub-option added to the assembler's -a command-line switch to
 421   generate a listing output.  The 'g' sub-option will insert into the listing
 422   various information about the assembly, such as assembler version, the
 423   command-line options used, and a time stamp.
 424
 425 * New command-line option -msse2avx for x86 target to encode SSE
 426   instructions with VEX prefix.
 427
 428 * Add Intel XSAVE, EPT, MOVBE, AES, PCLMUL, AVX/FMA support for x86 target.
 429
 430 * New command-line options, -march=CPU[,+EXTENSION...], -mtune=CPU,
 431   -mmnemonic=[att|intel], -msyntax=[att|intel], -mindex-reg,
 432   -mnaked-reg and -mold-gcc, for x86 targets.
 433
 434 * Support for generating wide character strings has been added via the new
 435   pseudo ops: .string16, .string32 and .string64.
 436
 437 * Support for SSE5 has been added to the i386 port.
 438
 439 Changes in 2.18:
 440
 441 * The GAS sources are now released under the GPLv3.
 442
 443 * Support for the National Semiconductor CR16 target has been added.
 444
 445 * Added gas .reloc pseudo.  This is a low-level interface for creating
 446   relocations.
 447
 448 * Add support for x86_64 PE+ target.
 449
 450 * Add support for Score target.
 451
 452 Changes in 2.17:
 453
 454 * Support for the Infineon XC16X has been added by KPIT Cummins Infosystems.
 455
 456 * Support for ms2 architecture has been added.
 457
 458 * Support for the Z80 processor family has been added.
 459
 460 * Add support for the "@<file>" syntax to the command line, so that extra
 461   switches can be read from <file>.
 462
 463 * The SH target supports a new command-line switch --enable-reg-prefix which,
 464   if enabled, will allow register names to be optionally prefixed with a $
 465   character.  This allows register names to be distinguished from label names.
 466
 467 * Macros with a variable number of arguments are now supported. See the
 468   documentation for how this works.
 469
 470 * Added --reduce-memory-overheads switch to reduce the size of the hash
 471   tables used, at the expense of longer assembly times, and
 472   --hash-size=<NUMBER> to set the size of the hash tables used by gas.
 473
 474 * Macro names and macro parameter names can now be any identifier that would
 475   also be legal as a symbol elsewhere. For macro parameter names, this is
 476   known to cause problems in certain sources when the respective target uses
 477   characters inconsistently, and thus macro parameter references may no longer
 478   be recognized as such (see the documentation for details).
 479
 480 * Support the .f_floating, .d_floating, .g_floating and .h_floating directives
 481   for the VAX target in order to be more compatible with the VAX MACRO
 482   assembler.
 483
 484 * New command-line option -mtune=[itanium1|itanium2] for IA64 targets.
 485
 486 Changes in 2.16:
 487
 488 * Redefinition of macros now results in an error.
 489
 490 * New command-line option -mhint.b=[ok|warning|error] for IA64 targets.
 491
 492 * New command-line option -munwind-check=[warning|error] for IA64
 493   targets.
 494
 495 * The IA64 port now uses automatic dependency violation removal as its default
 496   mode.
 497
 498 * Port to MAXQ processor contributed by HCL Tech.
 499
 500 * Added support for generating unwind tables for ARM ELF targets.
 501
 502 * Add a -g command-line option to generate debug information in the target's
 503   preferred debug format.
 504
 505 * Support for the crx-elf target added.
 506
 507 * Support for the sh-symbianelf target added.
 508
 509 * Added a pseudo-op (.secrel32) to generate 32 bit section relative relocations
 510   on pe[i]-i386; required for this target's DWARF 2 support.
 511
 512 * Support for Motorola MCF521x/5249/547x/548x added.
 513
 514 * Support for ColdFire EMAC instructions added and Motorola syntax for MAC/EMAC
 515   instrucitons.
 516
 517 * New command-line option -mno-shared for MIPS ELF targets.
 518
 519 * New command-line option --alternate and pseudo-ops .altmacro and .noaltmacro
 520   added to enter (and leave) alternate macro syntax mode.
 521
 522 Changes in 2.15:
 523
 524 * The MIPS -membedded-pic option (Embedded-PIC code generation) is
 525   deprecated and will be removed in a future release.
 526
 527 * Added PIC m32r Linux (ELF) and support to M32R assembler.
 528
 529 * Added support for ARM V6.
 530
 531 * Added support for sh4a and variants.
 532
 533 * Support for Renesas M32R2 added.
 534
 535 * Limited support for Mapping Symbols as specified in the ARM ELF
 536   specification has been added to the arm assembler.
 537
 538 * On ARM architectures, added a new gas directive ".unreq" that undoes
 539   definitions created by ".req".
 540
 541 * Support for Motorola ColdFire MCF528x added.
 542
 543 * Added --gstabs+ switch to enable the generation of STABS debug format
 544   information with GNU extensions.
 545
 546 * Added support for MIPS64 Release 2.
 547
 548 * Added support for v850e1.
 549
 550 * Added -n switch for x86 assembler.  By default, x86 GAS replaces
 551   multiple nop instructions used for alignment within code sections
 552   with multi-byte nop instructions such as leal 0(%esi,1),%esi.  This
 553   switch disables the optimization.
 554
 555 * Removed -n option from MIPS assembler.  It was not useful, and confused the
 556   existing -non_shared option.
 557
 558 Changes in 2.14:
 559
 560 * Added support for MIPS32 Release 2.
 561
 562 * Added support for Xtensa architecture.
 563
 564 * Support for Intel's iWMMXt processor (an ARM variant) added.
 565
 566 * An assembler test generator has been contributed and an example file that
 567   uses it (gas/testsuite/gas/all/test-gen.c and test-exmaple.c).
 568
 569 * Support for SH2E added.
 570
 571 * GASP has now been removed.
 572
 573 * Support for Texas Instruments TMS320C4x and TMS320C3x series of
 574   DSP's contributed by Michael Hayes and Svein E. Seldal.
 575
 576 * Support for the Ubicom IP2xxx microcontroller added.
 577
 578 Changes in 2.13:
 579
 580 * Support for the Fujitsu FRV architecture added by Red Hat. Models for FR400
 581   and FR500 included.
 582
 583 * Support for DLX processor added.
 584
 585 * GASP has now been deprecated and will be removed in a future release.  Use
 586   the macro facilities in GAS instead.
 587
 588 * GASP now correctly parses floating point numbers. Unless the base is
 589   explicitly specified, they are interpreted as decimal numbers regardless of
 590   the currently specified base.
 591
 592 Changes in 2.12:
 593
 594 * Support for Don Knuth's MMIX, by Hans-Peter Nilsson.
 595
 596 * Support for the OpenRISC 32-bit embedded processor by OpenCores.
 597
 598 * The ARM assembler now accepts -march=..., -mcpu=... and -mfpu=... for
 599   specifying the target instruction set.  The old method of specifying the
 600   target processor has been deprecated, but is still accepted for
 601   compatibility.
 602
 603 * Support for the VFP floating-point instruction set has been added to
 604   the ARM assembler.
 605
 606 * New psuedo op: .incbin to include a set of binary data at a given point
 607   in the assembly.  Contributed by Anders Norlander.
 608
 609 * The MIPS assembler now accepts -march/-mtune. -mcpu has been deprecated
 610   but still works for compatability.
 611
 612 * The MIPS assembler no longer issues a warning by default when it
 613   generates a nop instruction from a macro. The new command-line option
 614   -n will turn on the warning.
 615
 616 Changes in 2.11:
 617
 618 * Support for PDP-11 and 2.11BSD a.out format, by Lars Brinkhoff.
 619
 620 * x86 gas now supports the full Pentium4 instruction set.
 621
 622 * Support for AMD x86-64 architecture, by Jan Hubicka, SuSE Labs.
 623
 624 * Support for Motorola 68HC11 and 68HC12.
 625
 626 * Support for Texas Instruments TMS320C54x (tic54x).
 627
 628 * Support for IA-64.
 629
 630 * Support for i860, by Jason Eckhardt.
 631
 632 * Support for CRIS (Axis Communications ETRAX series).
 633
 634 * x86 gas has a new .arch pseudo op to specify the target CPU architecture.
 635
 636 * x86 gas -q command-line option quietens warnings about register size changes
 637   due to suffix, indirect jmp/call without `*', stand-alone prefixes, and
 638   translating various deprecated floating point instructions.
 639
 640 Changes in 2.10:
 641
 642 * Support for the ARM msr instruction was changed to only allow an immediate
 643   operand when altering the flags field.
 644
 645 * Support for ATMEL AVR.
 646
 647 * Support for IBM 370 ELF.  Somewhat experimental.
 648
 649 * Support for numbers with suffixes.
 650
 651 * Added support for breaking to the end of repeat loops.
 652
 653 * Added support for parallel instruction syntax (DOUBLEBAR_PARALLEL).
 654
 655 * New .elseif pseudo-op added.
 656
 657 * New --fatal-warnings option.
 658
 659 * picoJava architecture support added.
 660
 661 * Motorola MCore 210 processor support added.
 662
 663 * A new pseudo-op .intel_syntax has been implemented to allow gas to parse i386
 664   assembly programs with intel syntax.
 665
 666 * New pseudo-ops .func,.endfunc to aid in debugging user-written assembler code.
 667
 668 * Added -gdwarf2 option to generate DWARF 2 debugging information.
 669
 670 * Full 16-bit mode support for i386.
 671
 672 * Greatly improved instruction operand checking for i386.  This change will
 673   produce errors or warnings on incorrect assembly code that previous versions
 674   of gas accepted.  If you get unexpected messages from code that worked with
 675   older versions of gas, please double check the code before reporting a bug.
 676
 677 * Weak symbol support added for COFF targets.
 678
 679 * Mitsubishi D30V support added.
 680
 681 * Texas Instruments c80 (tms320c80) support added.
 682
 683 * i960 ELF support added.
 684
 685 * ARM ELF support added.
 686
 687 Changes in 2.9:
 688
 689 * Texas Instruments c30 (tms320c30) support added.
 690
 691 * The assembler now optimizes the exception frame information generated by egcs
 692   and gcc 2.8.  The new --traditional-format option disables this optimization.
 693
 694 * Added --gstabs option to generate stabs debugging information.
 695
 696 * The -a option takes a new suboption, m (e.g., -alm) to expand macros in a
 697   listing.
 698
 699 * Added -MD option to print dependencies.
 700
 701 Changes in 2.8:
 702
 703 * BeOS support added.
 704
 705 * MIPS16 support added.
 706
 707 * Motorola ColdFire 5200 support added (configure for m68k and use -m5200).
 708
 709 * Alpha/VMS support added.
 710
 711 * m68k options --base-size-default-16, --base-size-default-32,
 712   --disp-size-default-16, and --disp-size-default-32 added.
 713
 714 * The alignment directives now take an optional third argument, which is the
 715   maximum number of bytes to skip.  If doing the alignment would require
 716   skipping more than the given number of bytes, the alignment is not done at
 717   all.
 718
 719 * The ELF assembler has a new pseudo-op, .symver, used for symbol versioning.
 720
 721 * The -a option takes a new suboption, c (e.g., -alc), to skip false
 722   conditionals in listings.
 723
 724 * Added new pseudo-op, .equiv; it's like .equ, except that it is an error if
 725   the symbol is already defined.
 726
 727 Changes in 2.7:
 728
 729 * The PowerPC assembler now allows the use of symbolic register names (r0,
 730   etc.) if -mregnames is used.  Symbolic names preceded by a '%' (%r0, etc.)
 731   can be used any time.  PowerPC 860 move to/from SPR instructions have been
 732   added.
 733
 734 * Alpha Linux (ELF) support added.
 735
 736 * PowerPC ELF support added.
 737
 738 * m68k Linux (ELF) support added.
 739
 740 * i960 Hx/Jx support added.
 741
 742 * i386/PowerPC gnu-win32 support added.
 743
 744 * SCO ELF support added.  For OpenServer 5 targets (i386-unknown-sco3.2v5) the
 745   default is to build COFF-only support.  To get a set of tools that generate
 746   ELF (they'll understand both COFF and ELF), you must configure with
 747   target=i386-unknown-sco3.2v5elf.
 748
 749 * m88k-motorola-sysv3* support added.
 750
 751 Changes in 2.6:
 752
 753 * Gas now directly supports macros, without requiring GASP.
 754
 755 * Gas now has an MRI assembler compatibility mode.  Use -M or --mri to select
 756   MRI mode.  The pseudo-op ``.mri 1'' will switch into the MRI mode until the
 757   ``.mri 0'' is seen; this can be convenient for inline assembler code.
 758
 759 * Added --defsym SYM=VALUE option.
 760
 761 * Added -mips4 support to MIPS assembler.
 762
 763 * Added PIC support to Solaris and SPARC SunOS 4 assembler.
 764
 765 Changes in 2.4:
 766
 767 * Converted this directory to use an autoconf-generated configure script.
 768
 769 * ARM support, from Richard Earnshaw.
 770
 771 * Updated VMS support, from Pat Rankin, including considerably improved
 772   debugging support.
 773
 774 * Support for the control registers in the 68060.
 775
 776 * Handles (ignores) a new directive ".this_GCC_requires_the_GNU_assembler", to
 777   provide for possible future gcc changes, for targets where gas provides some
 778   features not available in the native assembler.  If the native assembler is
 779   used, it should become obvious pretty quickly what the problem is.
 780
 781 * Usage message is available with "--help".
 782
 783 * The GNU Assembler Preprocessor (gasp) is included.  (Actually, it was in 2.3
 784   also, but didn't get into the NEWS file.)
 785
 786 * Weak symbol support for a.out.
 787
 788 * A bug in the listing code which could cause an infinite loop has been fixed.
 789   Bugs in listings when generating a COFF object file have also been fixed.
 790
 791 * Initial i386-svr4 PIC implementation from Eric Youngdale, based on code by
 792   Paul Kranenburg.
 793
 794 * Improved Alpha support.  Immediate constants can have a much larger range
 795   now. Support for the 21164 has been contributed by Digital.
 796
 797 * Updated ns32k (pc532-mach, netbsd532) support from Ian Dall.
 798
 799 Changes in 2.3:
 800
 801 * Mach i386 support, by David Mackenzie and Ken Raeburn.
 802
 803 * RS/6000 and PowerPC support by Ian Taylor.
 804
 805 * VMS command scripts (make-gas.com, config-gas.com) have been worked on a bit,
 806   based on mail received from various people.  The `-h#' option should work
 807   again too.
 808
 809 * HP-PA work, by Jeff Law.  Note, for the PA, gas-2.3 has been designed to work
 810   with gdb-4.12 and gcc-2.6.  As gcc-2.6 has not been released yet, a special
 811   version of gcc-2.5.8 has been patched to work with gas-2.3.  You can retrieve
 812   this special version of gcc-2.5.8 via anonymous ftp from jaguar.cs.utah.edu
 813   in the "dist" directory.
 814
 815 * Vax support in gas fixed for BSD, so it builds and seems to run a couple
 816   simple tests okay.  I haven't put it through extensive testing.  (GNU make is
 817   currently required for BSD 4.3 builds.)
 818
 819 * Support for the DEC Alpha, running OSF/1 (ECOFF format).  The gas support is
 820   based on code donated by CMU, which used an a.out-based format.  I'm afraid
 821   the alpha-a.out support is pretty badly mangled, and much of it removed;
 822   making it work will require rewriting it as BFD support for the format anyways.
 823
 824 * Irix 5 support.
 825
 826 * The test suites have been fixed up a bit, so that they should work with a
 827   couple different versions of expect and dejagnu.
 828
 829 * Symbols' values are now handled internally as expressions, permitting more
 830   flexibility in evaluating them in some cases.  Some details of relocation
 831   handling have also changed, and simple constant pool management has been
 832   added, to make the Alpha port easier.
 833
 834 * New option "--statistics" for printing out program run times.  This is
 835   intended to be used with the gcc "-Q" option, which prints out times spent in
 836   various phases of compilation.  (You should be able to get all of them
 837   printed out with "gcc -Q -Wa,--statistics", I think.)
 838
 839 Changes in 2.2:
 840
 841 * RS/6000 AIX and MIPS SGI Irix 5 support has been added.
 842
 843 * Configurations that are still in development (and therefore are convenient to
 844   have listed in configure.in) still get rejected without a minor change to
 845   gas/Makefile.in, so people not doing development work shouldn't get the
 846   impression that support for such configurations is actually believed to be
 847   reliable.
 848
 849 * The program name (usually "as") is printed when a fatal error message is
 850   displayed.  This should prevent some confusion about the source of occasional
 851   messages about "internal errors".
 852
 853 * ELF support is falling into place.  Support for the 386 should be working.
 854   Support for SPARC Solaris is in.  HPPA support from Utah is being integrated.
 855
 856 * Symbol values are maintained as expressions instead of being immediately
 857   boiled down to add-symbol, sub-symbol, and constant.  This permits slightly
 858   more complex calculations involving symbols whose values are not alreadey
 859   known.
 860
 861 * DBX-style debugging info ("stabs") is now supported for COFF formats.
 862   If any stabs directives are seen in the source, GAS will create two new
 863   sections: a ".stab" and a ".stabstr" section.  The format of the .stab
 864   section is nearly identical to the a.out symbol format, and .stabstr is
 865   its string table.  For this to be useful, you must have configured GCC
 866   to generate stabs (by defining DBX_DEBUGGING_INFO), and must have a GDB
 867   that can use the stab sections (4.11 or later).
 868
 869 * LynxOS, on i386 and m68k platforms, is now supported.  SPARC LynxOS
 870   support is in progress.
 871
 872 Changes in 2.1:
 873
 874 * Several small fixes for i386-aix (PS/2) support from Minh Tran-Le have been
 875   incorporated, but not well tested yet.
 876
 877 * Altered the opcode table split for m68k; it should require less VM to compile
 878   with gcc now.
 879
 880 * Some minor adjustments to add (Convergent Technologies') Miniframe support,
 881   suggested by Ronald Cole.
 882
 883 * HPPA support (running OSF only, not HPUX) has been contributed by Utah.  This
 884   includes improved ELF support, which I've started adapting for SPARC Solaris
 885   2.x.  Integration isn't completely, so it probably won't work.
 886
 887 * HP9000/300 support, donated by HP, has been merged in.
 888
 889 * Ian Taylor has finished the MIPS ECOFF (Ultrix, Irix) support.
 890
 891 * Better error messages for unsupported configurations (e.g., hppa-hpux).
 892
 893 * Test suite framework is starting to become reasonable.
 894
 895 Changes in 2.0:
 896
 897 * Mostly bug fixes.
 898
 899 * Some more merging of BFD and ELF code, but ELF still doesn't work.
 900
 901 Changes in 1.94:
 902
 903 * BFD merge is partly done.  Adventurous souls may try giving configure the
 904   "--with-bfd-assembler" option.  Currently, ELF format requires it, a.out
 905   format accepts it; SPARC CPU accepts it.  It's the default only for OS "elf"
 906   or "solaris".  (ELF isn't really supported yet.  It needs work.  I've got
 907   some code from Utah for HP-PA ELF, and from DG for m88k ELF, but they're not
 908   fully merged yet.)
 909
 910 * The 68K opcode table has been split in half.  It should now compile under gcc
 911   without consuming ridiculous amounts of memory.
 912
 913 * A couple data structures have been reduced in size.  This should result in
 914   saving a little bit of space at runtime.
 915
 916 * Support for MIPS, from OSF and Ralph Campbell, has been merged in.  The OSF
 917   code provided ROSE format support, which I haven't merged in yet.  (I can
 918   make it available, if anyone wants to try it out.)  Ralph's code, for BSD
 919   4.4, supports a.out format.  We don't have ECOFF support in just yet; it's
 920   coming.
 921
 922 * Support for the Hitachi H8/500 has been added.
 923
 924 * VMS host and target support should be working now, thanks chiefly to Eric
 925   Youngdale.
 926
 927 Changes in 1.93.01:
 928
 929 * For m68k, support for more processors has been added: 68040, CPU32, 68851.
 930
 931 * For i386, .align is now power-of-two; was number-of-bytes.
 932
 933 * For m68k, "%" is now accepted before register names.  For COFF format, which
 934   doesn't use underscore prefixes for C labels, it is required, so variable "a0"
 935   can be distinguished from the register.
 936
 937 * Last public release was 1.38.  Lots of configuration changes since then, lots
 938   of new CPUs and formats, lots of bugs fixed.
 939
 940 \f
 941 Copyright (C) 2012-2022 Free Software Foundation, Inc.
 942
 943 Copying and distribution of this file, with or without modification,
 944 are permitted in any medium without royalty provided the copyright
 945 notice and this notice are preserved.
 946
 947 Local variables:
 948 fill-column: 79
 949 End: