+
+ * Makefile.in (config-stamp): Make sure there is at least one
+ element in the for loop.
+
+
+ * config/tc-hppa.c: Misc trivial changes to make gcc -Wall happy.
+
+ * config/tc-hppa.h (elf_hppa_final_processing): Declare.
+
+
+ * as.c (parse_args): Cast xmalloc return value.
+
+ * Makefile.in (config-stamp): If $(defs) contains multiple words,
+ emit a #define line for each.
+ * configure.in: For sparc64 target, use sparc cpu files and add
+ sparcv9 to extra_defs. No longer treat sparc64-*-aout* specially.
+
+ * config/tc-sparc.c (membar_masks): Now static and const.
+ (md_show_usage) [!NO_V9]: Add -Av9 to usage message.
+ (current_architecture) [sparcv9]: Initialize to v9.
+ (md_begin) [sparcv9]: Don't bother changing it unconditionally
+ here.
+ (s_reserve): Don't pass unexpected argument to as_bad with
+ bad-segment message.
+
+ * as.h (bfd_alloc_by_size_t) [BFD_ASSEMBLER]: Declare.
+
+ * config/atof-ieee.c (int_to_gen): Commented out unused routine.
+
+ * config/tc-vax.c (md_assemble): Removed check of operand section.
+
+
+ * config/obj-vms.h (DBG_S_C_COMPLX4, DBG_S_C_COMPLX8): define
+ these new VMS symbol-type macros for `complex float' and
+ `complex double' support. Their values come from the existing
+ DSC$K_DTYPE_FC and DSC$K_DTYPE_DC macros in <descrip.h>.
+ (DBG_S_C_REAL8_G, DBG_S_C_COMPLX8_G): G_float versions of
+ REAL8 and COMPLX8; not used yet, because gcc outputs the same
+ .stabs for `double' regardless of whether `-mg' is used.
+ config/obj-vms.c (VMS_typedef_parse) [case 'r']: add entries
+ for gcc2's predefined types "complex float", "complex double",
+ and "complex long double" (identical to complex double).
+
+
+ * config/obj-coff.c (coff_frob_symbol): Use C_STAT for the .text
+ section symbol, not C_LABEL.
+
+ * config/tc-mips.c (mips_ip): Permit a modifier in 'o' case, and
+ permit non constant expressions in 'u' case. Lets ``lui
+ $8,%hi(foo); lw $8,%lo(foo)($8)'' work correctly.
+
+
+ * config/obj-aout.c (obj_aout_frob_symbol): Warn about an attempt
+ to put a common symbol in a set.
+
+
+ Add weak symbols as an extension to a.out.
+ * read.c (pseudo_set): Only preserve external bit for OBJ_AOUT and
+ OBJ_BOUT if not BFD_ASSEMBLER.
+ * config/aout_gnu.h (N_WEAKU, N_WEAKA, N_WEAKT, N_WEAKD, N_WEAKB):
+ Define as in ../include/aout/aout64.h.
+ * config/obj-aout.h (OBJ_SYMFIELD_TYPE): If not BFD_ASSEMBLER,
+ define as char.
+ (S_GET_WEAK, S_SET_WEAK): Define if not BFD_ASSEMBLER.
+ * config/obj-aout.c (obj_pseudo_table): Add "weak".
+ (obj_emit_symbols): Adjust type of weak symbols.
+ (obj_aout_weak): New static function.
+
+
+ * config/obj-elf.c (obj_elf_section): Don't set any flags based on
+ the type of a special section.
+
+ * config/ho-sunos.h: Include <stdlib.h>. Don't declare malloc,
+ realloc, free, or atol.
+
+
+ * Makefile.in (check): Delete as.new dependency, so that
+ regression test doesn't trigger an assembler build.
+
+
+ * Makefile.in (mostlyclean, realclean): New targets.
+ * doc/Makefile.in, testsuite/Makefile.in: Likewise.
+
+
+ * config/tc-mips.c (RDATA_SECTION_NAME): Define.
+ (macro): Correct M_LI_SS SVR4_PIC/EMBEDDED_PIC case. After M_LI_D
+ or M_L_DOB or label dob, force a new frag to avoid getting
+ confused in tc_gen_reloc.
+ (mips_ip): Use RDATA_SECTION_NAME, not .rdata.
+ (s_change_sec): Likewise.
+
+
+ * read.c (emit_expr): Use memset to zero out memory, rather than
+ going through md_number_to_chars. This permits handling symbolic
+ arguments when the size is larger than sizeof (valueT), if
+ TC_CONS_FIX_NEW is prepared to handle the case (as it is on MIPS).
+
+
+ * as.c (show_usage), config/tc-alpha.c (md_show_usage),
+ config/tc-mips.c (md_show_usage): Fix up messages.
+
+ * as.h: Replace flagseen with separate variables.
+ * as.c (parse_args): Set them. Don't accept -1 option, or -v
+ explicitly (it's a synonym for --version).
+ * as.c, input-scrub.c, messages.c, read.c, symbols.c, write.c,
+ config/obj-aout.c, config/obj-aout.h, config/obj-bout.c,
+ config/obj-bout.h, config/obj-coff.c, config/obj-coff.h,
+ config/obj-vms.c, config/tc-hppa.c, config/tc-i386.c,
+ config/tc-i960.c, config/tc-m68k.c, config/tc-mips.c,
+ config/tc-vax.c: Use the new flag variables instead of flagseen.
+ * config/tc-vax.c [OBJ_VMS]: Recognize -+, -1, -v, and document in
+ usage.
+
+ * as.c (show_usage): Remove target specific messages;
+ instead, call md_show_usage.
+ (parse_args): Use getopt_long_only. Take pointers to argc and
+ argv.
+ (main): Pass parse_args pointers.
+ * as.h: Remove 3 variables that are redundant with flagseen.
+ * as.c, messages.c: Change their users to use flagseen.
+ Define getopt stuff.
+ * tc.h: Update md_parse_option decl. Add md_show_usage decl.
+ * config/tc-*.c: Add md_shortopts, md_longopts,
+ md_longopts_size, md_show_usage. Change calling convention for
+ md_parse_option. Remove md_parse_long_option.
+ * config/tc-ns32k.c: Rename `struct option' to `struct ns32k_option'.
+ * config/tc-i386.h: Don't define md_parse_option.
+
+
+ * as.c (show_usage): New function.
+ (parse_args): Code moved from main.
+ Recognize --help and --version.
+ * config/tc-ns32k.h: Define TC_NS32K.
+ * doc/as.texinfo: Document all of the target-independent command
+ line options.
+
+
+ * gasp.c (hash_new_table): Clear newly allocated table.
+
+ * config/tc-m68k.c (enum _register): Add 68060 control registers
+ BUSCR and PCR.
+ (last_movec_reg): New macro.
+ (m68000_control_regs, m68010_control_regs, m68020_control_regs,
+ m68040_control_regs, m68060_control_regs): New arrays.
+ (control_regs): New pointer.
+ (m68k_ip): Use control_regs instead of testing CPU every time.
+ Use last_movec_reg too. In error messages, handle 68060, and
+ print 68060 for mfloat, too.
+ (m68k_init_after_args): Handle "68060". Use m68040up for making
+ m68851 choice. Set control_regs.
+ (md_parse_option): Handle "68060".
+ * configure.in: Setting cpu_type, recognize m68060 too.
+
+ * config/obj-coff.c (fixup_segment) [!BFD_ASSEMBLER]
+ [DIFF_EXPR_OK]: Do conversion to pc-relative for difference, even
+ if pcrel is already set.
+
+ * read.c (potable): Add this_gcc_requires_the_gnu_assembler in all
+ lower-case, in case we're ignoring case of opcodes in the input
+ file.
+
+ * doc/as.texinfo (.section): Document as unavailable for a.out
+ type formats.
+
+ * config/tc-alpha.c (machine): New variable.
+ (load_insn): New macro.
+ (load_insn_table): New function.
+ (md_begin): Call load_insn_table, once for basic instructions and
+ once for appropriate PAL instruction table.
+ (md_parse_option): Set `machine' based on -m##### arguments.
+ * config/alpha-opcode.h (alpha_pal21064_opcodes): Split out from
+ alpha_opcodes.
+ (alpha_pal21164_opcodes): New table.
+ (NUM21064OPCODES, NUM21164OPCODES): New macros.
+
+ * configure.in (target i386-*-netbsd0.8): Use 386bsd emulation.
+
+ * doc/Makefile.in (install-info-gasp): Use $$dir when installing
+ file.
+
+
+ * config/tc-mips.c (mips_ip): Force floating point values to be
+ aligned correctly.
+
+
+ Merged in changes from gas-2.3 net release:
+
+ * Makefile.in (VERSION): Updated to cygnus-2.3.1.
+
+ * config/obj-vms.c: Replaced unchecked uses of malloc with
+ xmalloc.
+
+ * listing.c (list_symbol_table): Only test BFD64, not
+ BFD_ASSEMBLER too.
+
+ * config/obj-coff.c (fixup_segment) [BFD_ASSEMBLER]
+ [DIFF_EXPR_OK]: Don't check pcrel, just convert it.
+
+ * config/obj-vms.c: Removed lots of extra semicolons after
+ compound statements.
+ (strchr): Don't declare here.
+
+ * config/ho-vax.h (realloc): Declare.
+
+ * config/ho-vms.h (strchr, strdup): Declare.
+
+ * config/tc-sparc.c (md_parse_option) [OBJ_ELF]: Accept and ignore
+ option `-q'.
+
+
+ * config/obj-vms.h (DBG_S_C_SQUAD, DBG_S_C_UQUAD): define these
+ new VMS symbol-type macros; signed and unsigned quadword integers,
+ for `long long' support. Their values come from the existing
+ DSC$K_DTYPE_QU and DSC$K_DTYPE_Q macros in <descrip.h>. The
+ VMS debugger now recognizes `long long' variables correctly.
+ config/obj-vms.c (VMS_typedef_parse) [case 'r']: add entries
+ for gcc2's predefined types "long double" (same as double, as
+ per gcc's current state), "long long int", "long long unsigned
+ int", and final `otherwise' case (to avoid uninitialized type
+ and size fields). [caveat: predefined types "complex int",
+ "complex float", "complex double", and "complex long double" are
+ still missing.]
+
+ * config/ho-vms.h (EXIT_FAILURE): define as 0x10000002 instead
+ of 0, because the latter indicates success rather than failure
+ when passed to `exit' or return from `main' compiled by gcc2.
+
+ * config/obj-vms.c (array_suffix, generate_suffix): replace two
+ hardcoded `0xa3's with macro DBG_S_C_ADVANCED_TYPE from obj-vms.h.
+ (VMS_typedef_parse): eliminate redundant if-then-else when
+ allocating new symbol entry and linking it to VMS_Symbol_type_list.
+
+
+ * config/obj-vms.c (Write_VMS_MHD_Records): don't try to interpret
+ the contents of the GAS_VERSION string when falling back to it for
+ language processor identification.
+
+ * make-gas.com, vmsconf.sh (ENVIRON): fix misspelling of
+ `psect_attr' in linker options.
+
+
+ * configure.bat: update to latest makefile.in
+ * config/te-go32.h: [new] go32's environment
+
+
+ * as.h: Don't declare parameters for strstr.
+
+
+ * config/tc-sparc.c (md_section_align): Don't change the size if
+ OBJ_ELF.
+
+
+ * Makefile.in (install): Redirect ln output to /dev/null. If ln
+ fails on gasp, install gasp.new, not gasp.
+
+
+ * config/tc-hppa.c (pa_ip): Use R_HPPA_ABS_CALL, not R_HPPA for
+ absolute calls.
+
+
+ * config/tc-hppa.c (pa_parse_fp_cmp_cond): Report an error
+ on a partial completer match.
+
+
+ * config/tc-hppa.c: Change .hppa_unwind to .PARISC.unwind
+ throughout code.
+ (is_complex): Delete definition and support for complex relocation
+ types.
+ (tc_gen_reloc): Delete special unwind crud for ELF. Simplify and
+ rewrite ELF code based on 94-02-02 PA ELF draft spec.
+ (pa_build_unwind_subspace): Use standard PARISC_DIR32 relocs for
+ the unwind descriptors.
+
+
+ * config/go32.mh: New makefile fragment for go32 crossing.
+ * configure.in (host==go32): Use new fragment.
+
+
+ * gasp.c: Include libiberty.h.
+ (main): Remove unused variable i.
+
+ * config/tc-ppc.c (md_begin): When using -many, permit comparison
+ instructions to appear multiple times in the opcode table.
+
+
+ * Makefile.in (VERSION): Update to 2.2.90.
+
+ * symbols.c (symbol_new) [BFD_ASSEMBLER]: Don't permit additions
+ to the symbol table if it's already been set in the output bfd.
+ (symbol_begin) [! EMIT_SECTION_SYMBOLS] [RELOC_REQUIRES_SYMBOL]:
+ Don't use bfd_abs_section.symbol for gas absolute symbol.
+
+ * doc/Makefile.in (distclean, clean-dvi, clean-info): Delete gasp
+ files too.
+
+
+ * config/tc-mips.c (mips_emit_delays): Make call to
+ mips_no_prev_insn unconditional.
+
+
+ * config/obj-coff*.*: Merged coffbfd versions into coff versions,
+ with a single "#ifdef BFD_ASSEMBLER" controlling most of it for
+ now. Deleted obj-coffbfd.* files.
+ * configure.in: Always use obj-coff.* for COFF targets.
+
+
+ * config/te-dpx2.h (TARGET_FORMAT, REGISTER_PREFIX_OPTIONAL):
+ Define.
+ * configure.in (m68k-bull-sysv3*): Enable.
+
+ * config/coff_gnu.h: Deleted.
+
+
+ * config/tc-m68k.h (TARGET_FORMAT): If not TE_SUN3, define as
+ "a.out-zero.big".
+
+ * config/obj-coffbfd.c (fixup_segment): Make common symbol and PC
+ relative adjustments when TE_LYNX is defined as well as when
+ TC_I386 is defined.
+
+
+ * as.h (strstr): Restore declaration.
+ (subseg_get) [BFD_ASSEMBLER]: Declare.
+
+ * write.c (write_object_file): If obj_adjust_symtab is defined,
+ invoke it. Then call set_symtab, and finally invoke *_frob_file
+ hooks.
+ * config/obj-coff.c (coff_adjust_symtab): Renamed from
+ coff_frob_file.
+ * config/obj-coff.h (coff_adjust_symtab): Changed declaration
+ accordingly.
+ (obj_adjust_symtab): Macro also changed.
+
+ * configure.in (i386-*-gnu*): New target, handled like i386-mach.
+
+
+ * config/obj-coff.h (TARGET_FORMAT) [TC_I960]: Select little
+ endian version.
+
+ * config/obj-coffbfd.h (TARGET_FORMAT) [TC_I960]: Ditto.
+
+ * config/obj-coff.c (coff_frob_section): Round up the size of
+ every section to a multiple of the alignment, so that BFD doesn't
+ surprise us.
+
+ Eliminate many simple differences between the two COFF back ends:
+
+ * config/obj-coffbfd.c: Removed all uses of DEFUN and DEFUN_VOID.
+ Made minor stylistic changes, deleted some register declarations.
+ (stack_top): Deleted.
+ (symbol_to_chars): Use absolute_section and reg_section instead of
+ the corresponding SEG_* symbols.
+ (obj_coff_endef, tag_find_or_make, fixup_segment): Likewise.
+ (stack typedef, stack_init, stack_delete, stack_push, stack_pop):
+ Moved to just after pseudo-op table. All functions now static.
+ (stack_delete): Removed declaration.
+ (tag_init, tag_insert, tag_find_or_make, tag_find): Moved to just
+ after stack functions.
+ * config/obj-coffbfd.h: Reordered some declarations and macros.
+ (stack_init, stack_delete, stack_push, stack_pop): Don't declare.
+ (stack typedef): Deleted.
+ (SYMBOLS_NEED_BACKPOINTERS): Always undef then define; don't test.
+ (SYM_AUXENT): New macro.
+ (SA_GET_*, SA_SET_*): Define in terms of SYM_AUXENT when feasible.
+ (SF_GET_*, SF_SET_*): Define in terms of SF_GET when feasible.
+ (SA_GET_SYM_TAGNDX, SA_GET_SYM_ENDNDX, SA_SET_SYM_TAGNDX,
+ SA_SET_SYM_ENDNDX, object_headers typedef, data_section_header,
+ text_section_header): Delete non-BFD_HEADERS versions, since we
+ always define that symbol now.
+
+ * config/obj-coff.c (stack_top): Deleted.
+ (obj_coff_endef, obj_coff_dim, obj_coff_line, obj_coff_size,
+ obj_coff_scl, obj_coff_tag, obj_coff_type, obj_coff_val): Change
+ argument name from "ignored" to "ignore".
+ (obj_coff_val): Use frag_now_fix.
+ (obj_pseudo_table): Removed IGNORE_DEBUG version, since it doesn't
+ get used.
+ (stack typedef, stack_init, stack_delete, stack_push, stack_pop):
+ Moved to just after pseudo-op table. All functions now static.
+ (tag_init, tag_insert, tag_find_or_make, tag_find): Moved to just
+ after stack functions.
+ * config/obj-coff.h: Reordered some declarations and macros.
+ Protected against multiple inclusions.
+ (stack_init, stack_delete, stack_push, stack_pop): Don't declare.
+ (stack typedef): Deleted.
+ (SYMBOLS_NEED_BACKPOINTERS): Always undef then define; don't test.
+ (stdoutput): Deleted declaration.
+ (TARGET_FORMAT) [TC_I386]: Don't define if already defined.
+
+
+ * subsegs.h (segment_info_type): Use fix_tail field even if
+ BFD_ASSEMBLER.
+ * subsegs.c (subseg_change): Initialize fix_tail field.
+ (subseg_get): Likewise.
+ * write.c (frags_chained): New static variable.
+ (fix_new_internal): If frags_chained is set, use fix_root and
+ fix_tail from seg_info (now_seg), rather than frchain_now.
+ (chain_frchains_together_1): Set fix_tail field.
+ (chain_frchains_together): Set frags_chained.
+
+
+ * config/tc-mips.c (s_option): Only refer to g_switch_* variables
+ if GPOPT is defined.
+ (s_abicalls): Ditto.
+ (md_apply_fix): Cast char* to unsigned char* to avoid pointer
+ mismatch.
+
+
+ * configure.in (i386-*-go32): Uses coff now.
+ * gasp.c (main): Now takes -D on command line.
+ (show_usage): Describe new options.
+
+
+ * listing.c (list_symbol_table): Print "NO DEFINED SYMBOLS" and
+ "NO UNDEFINED SYMBOLS" if there aren't any, instead of displaying
+ the header with an empty list.
+
+ * config/obj-coffbfd.c (fill_section): Check COFF_NOLOAD_PROBLEM
+ also before setting STYP_NOLOAD for .bss section.
+
+ * config/tc-m68k.c (flag_reg_prefix_optional): New variable.
+ Initialized to value of REGISTER_PREFIX_OPTIONAL, if defined, or
+ zero.
+ (m68k_reg_parse): If flag_reg_prefix_optional is set, permit
+ register prefix to be absent.
+ (m68k_ip_op): Accept `&' also for immediate constants.
+ (insert_reg): Don't bother with (two!?) sanity checks of the
+ symbol table when inserting each register.
+ (m68k_parse_long_option): New function. Set
+ flag_reg_prefix_optional if "register-prefix-optional" is passed.
+ * config/tc-m68k.h (REGISTER_PREFIX): Always define if not already
+ defined.
+ (OPTIONAL_REGISTER_PREFIX): Don't define.
+ (REGISTER_PREFIX_OPTIONAL): If not already defined, define as zero
+ or one depending on M68KCOFF.
+
+ * config/tc-m68k.c (DATA, ADDR, SP, FPREG, COPNUM, BAD, BAC):
+ Define as macros instead of enumerators, since the Apollo compiler
+ can't handle "enumVal1, enumVal2 = enumVal1" when defining an enum
+ type.
+ (make_pcrel_absolute) [NO_PCREL_RELOCS]: New function.
+ (tc_coff_fix2rtype) [NO_PCREL_RELOCS]: Generate only R_RELBYTE,
+ R_DIR16, and R_DIR32 relocs.
+ * config/tc-m68k.h [TE_APOLLO] (COFF_MAGIC, COFF_AOUTHDR_MAGIC):
+ Use Apollo versions.
+ [TE_APOLLO] (OBJ_COFF_OMIT_OPTIONAL_HEADER): Undefine. That is,
+ do include the optional header for Apollo target.
+ (COFF_MAGIC): Don't define as MC68MAGIC if it's already defined.
+
+ * config/tc-m68k.h [TE_DELTA] (LEX_PCT): Define as 1, so that `%'
+ can be used within a label name.
+
+ * config/tc-m68k.h (m68k_init_after_args): Declare.
+ (tc_init_after_args): Define as m68k_init_after_args.
+ * config/tc-m68k.c (m68k_init_after_args): New function,
+ containing one-shot code from md_assemble. Added warning for
+ combination of 68040 and 68851.
+ (md_assemble): Startup-time code deleted.
+
+
+ * doc/Makefile.in (clean, distclean): Remove asconfig.texi.
+
* config/tc-hppa.c (pa_ip): 13 bit immediate constant (for break
+ * config/tc-sparc.c (sparc_ip): Truncate args to %hi/%lo to 32 bits.
+
* expr.c (integer_constant): Fix computation of too_many_digits.
Variable digit_2 renamed to start. Fix check for whether number
will fit in 32 bits.
+ * config/tc-sparc.c (sparc_ip): Add default case to reloc switch.
+
* read.c (s_lcomm): Put small objects in .sbss for MIPS ELF as
well as MIPS ECOFF.
(get_stab_string_offset): Remove unused variable aligned.
* config/tc-hppa.c (tc_gen_reloc): ELF32_HPPA_R_ADDEND ->
HPPA_R_ADDEND.
+
+ * config/tc-sparc.c (sparc_ip): Fix up membar argument handling.
+
* config/tc-sh.c (build_Mbytes): Write the relocs to the correct
address. (md_assemble): Make error handling a bit more graceful.
(md_apply_fix): Don't warn on non aligned displacement.
-
+
* config/tc-z8k.c (get_specific, built_bytes): Understand all the
aspects of lda.
* config/tc-sparc.c (md_parse_option): Accept and ignore -sparc
option, which is used by the SunOS make default .s.o rule.
+
+ * config/tc-sparc.c (sparc_ip, label "immediate"): Convert relocs
+ operating on upper 32 bits of immediate constants to lower-half
+ relocs with adjusted constants.
+
* config/tc-m68k.h (AOUT_MACHTYPE): Define to be external variable
+ * config/obj-elf.h (elf_symbol): No longer special-cased on
+ sparcv9, since elf_symbol_type is now independent of size.
+
* config/obj-elf.h (elf_symbol): Fixed name of elf_symbol_type.
* config/tc-hppa.h (struct default_subspace_dict, struct
+ * config/tc-sparc.c (md_apply_fix, case BFD_RELOC_64): Fill in
+ bytes with real values, not zeros.
+ (md_pseudo_table): Call cons for .uaxword.
+
* config/obj-elf.c (obj_pseudo_table): Handle ".8byte".
* read.c (emit_expr): Use BFD_RELOC_64 for 8-byte expressions.
* app.c (do_scrub_next_char): Reset state to 0 after .appline if
file name is not seen.
+
+ * config/tc-sparc.c (md_apply_fix, case BFD_RELOC_64): New case,
+ parallel to BFD_RELOC_32.
+ (tc_gen_reloc): Accept BFD_RELOC_64.
+
* write.c: Don't use short int in a prototype.
+ * configure.in: Make sparc64-*-aout* use bfd gas.
+
* configure.in: Recognize h8300h.
+ * config/tc-sparc.h (ENV64): Define for v9.
+
+ * config/tc-sparc.c (s_xword): For numbers, call big_cons.
+ (sparc_ip, md_apply_fix, tc_gen_reloc): Handle some sparc64
+ relocation types.
+ (md_number_to_chars): Handle 8-byte value.
+
+ * config/obj-elf.h (elf_symbol): For v9, use 64-bit symbol type.
+
* as.h (valueT): Typedef moved here.
* struc-symbol.h (valueT): ...from here.
* write.c (write_object_file): Locals from_addr, to_addr,
+ * config/tc-sparc.c (md_begin): If sparcv9 is defined, reset
+ current_architecture to v9 automatically.
+
+ * config/tc-sparc.h (TARGET_FORMAT): Use elf64-sparc for v9.
+
+ * config/tc-sparc64.h, config/tc-sparc64.c: New files.
+ * configure.in: Use sparc64 cpu files for v9, and default to elf
+ format.
+
* write.c (relax_and_size_seg): Always fully process a section.
Section size is last frag's (vm)address plus its size. If no
relocations are present, force SEC_RELOC flag clear.
echo recursion lines.
(install-info, clean-info): collapse into the dvi and info rule.
+
+ * config/tc-sparc.c (sparc_ip): Remove support for %d/%q fp regs.
+ All fp regs are now specified as %f.
+ (priv_reg_table): fpq -> fq.
+
* as.h (strstr): Disable declaration for now.
* config/tc-vax.h (NO_RELOC): Define.
+
+ * config/tc-sparc.c (sparc_ip): Fix parsing of asi number.
+ Fix error message.
+
* gas/config/m88k-opcode.h (m88k_opcodes): Add 88110 instructions.
* Makefile.in: add installcheck target
+
+ * config/tc-sparc.c (sparc_ip): Add support for membar mask names.
+ Add missing colons in prefetch error messages. Add support for
+ ASI names.
+
* config/tc-mips.c (macro): Use $AT for any floating point load.
* CONTRIBUTORS: Update Gilmore entry.
+
+ * config/tc-sparc.c (sparc_ip): Correctly assemble prefetch
+ instructions. Accept integer prefetch function numbers.
+
+ * config/tc-sparc.c (s_xword): Now call big_cons, so large
+ constants are accepted, but symbolic values are not. GCC will not
+ generate the latter currently.
+
* frags.c (frag_init): New function.
(zero_address_frag, bss_address_frag): Now initialized at run
time.
+ * config/tc-sparc.c (sparc_ip): For %uhi and %ulo, if not
+ ENV64, emit no reloc.
+ (md_pseudo_table): For ".xword", call s_xword.
+ (s_xword): New function.
+
* config/tc-sparc.c (architecture_requested, warn_on_bump,
md_relax_table): Use default zero initialization.
(s_reserve): Since SEG_E2 is equivalent to SEG_BSS, just use the
\f
Local Variables:
+mode: indented-text
+left-margin: 8
+fill-column: 74
version-control: never
End:
projects / binutils.git / blobdiff