(md_longopts): Allow OPTION_MABI for ELF compilation only. RE-allow
OPTION_GP32, OPTION_GP64, OPTION_FP32 for non-ELF compilation.
Sort options a bit more logical.
(md_parse_option): Allow OPTION_32, OPTION_N32, OPTION_N64,
OPTION_MABI only for elf targets.
* gas/mips/mips.exp: Change naming of some conditionals to reflect
the object format they actually mean. Don't try mips-abi32 and
mips-abi32-pic tests for ecoff.
mips_set_options.
(mips_set_options): Add members gp32, fp32, abi.
(file_mips_gp32): New flag.
(file_mips_fp32): New flag.
(mips_opts): Initialize the new members.
(mips_gp32): Remove.
(mips_fp32): Remove.
(HAVE_32BIT_GPRS): Use the new values from mips_opts.
(HAVE_32BIT_FPRS): Likewise.
(HAVE_NEWABI): Likewise.
(HAVE_64BIT_OBJECTS): Likewise.
(md_begin): Likewise. Save default (file) values.
(md_parse_option): Use the new values from mips_opts.
(s_mipsset): Likewise. Fix logic to keep the ABI selection if
possible. Let .set mipsN work together with .set push/pop.
Enhance error messages.
(mips_elf_final_processing): Use file_mips_* for header processing.
2001-10-31 Chris Demetriou <cgd@demetriou.com>
* elf32-mips.c (_bfd_mips_elf_hi16_reloc): Handle PC-relative
relocations properly.
[ gas/ChangeLog ]
2001-10-31 Chris Demetriou <cgd@broadcom.com>
* config/tc-mips.c (HAVE_32BIT_ADDRESSES): If compiling embedded
PIC code, assume pointers the same size as GPRs.
(macro): In M_LA_AB handling for embedded PIC code, support
"la $treg,foo-bar($breg)". In load/store handling
(label ld_st) support "<op> $treg,<sym>-<local_sym>($breg)"
which is used by the compiler for switch statements.
In load/store double multi-instruction macro handling
(label ldd_std) add a comment that no special handling
is currently done for embedded PIC.
(mips_ip): In 'o' (16-bit offset) case, only accept 16
bit offsets.
[ gas/testsuite/ChangeLog ]
2001-10-31 Chris Demetriou <cgd@broadcom.com>
* gas/mips/empic.s: Undo damage inflicted on 2000-12-02.
* gas/mips/empic.d: Likewise.
* gas/mips/elempic.d: Likewise (it was copied into other files).
* gas/mips/telempic.d: Likewise.
* gas/mips/tempic.d: Likewise.
* gas/mips/empic2.s: New test to check new 'la' and 'lw' (and
related ops) syntax, test loads with large offsets.
* gas/mips/emcic2.d: Likewise.
* gas/mips/mips.exp: Run the new test on ELF platforms.
(mips_cpreturn_register): Likewise.
(mips_gp_register): Likewise.
(s_cpsetup): New function prototype.
(s_cplocal): Likewise.
(s_cpreturn): Likewise.
(s_gpvalue): Likewise.
(mips_pseudo_table): Add .cpsetup, .cplocal, .cpreturn, .gpvalue
pseudo-ops.
(macro): Don't warn about .cprestore for NewABI.
(md_pcrel_from): Code cleanup.
(mips_force_relocation): Force output of some NewABI relocations even
without a defined symbol.
(s_cpload): Ignore .cpload for NewABI.
(s_cpsetup): Handle .cpsetup.
(s_cplocal): Handle .cplocal.
(s_cprestore): Ignore .cprestore for NewABI.
(s_cpreturn): Handle .cpreturn.
(s_gpvalue): Handle .gpvalue.
(s_cpadd): Ignore .cpadd for NewABI.
(nopic_need_relax): Take g_switch_value into account as gp
optimization.
(tc_gen_reloc): Don't handle BFD_RELOC_MIPS_{CALL,GOT}* for NewABI.
(small_ex_type): Named this enum, more return values for
my_getSmallExpression.
(mips_ip): Allow SPC and HT between arguments. Handle some NewABI
triple relocations. Protect some parts with ifdef OBJ_ELF.
(percent_op_match): New struct, lookup table for %some_reloc().
(my_getSmallParser): New function, parses nested percent_ops also.
(my_getSmallExpression): Rewite to support nested percent_ops.
(load_address): Support both 32- and 64-bit addresses.
(macro): Call load_register correctly. Expand 64-bit loads ans stores.
(macro2): Call load_address correctly.
relocation triple.
(prev_insn_fixp): Likewise.
(append_insn): Changed prototype to accept a relocation pointer.
(imm_reloc): Make it an array.
(offset_reloc): Likewise.
(md_assemble): Handle triple relocations.
(append_insn): Likewise. Add handling for some NewABI relocations.
(mips_no_prev_insn): Handle triple relocations.
(macro_build): Likewise. Add handling for some NewABI relocations.
Move handling for the 'u' case to append_insn().
(mips16_macro_build): Handle triple relocations.
(macro_build_lui): Likewise. Don't handle _gp_disp as special symbol
for NewABI.
(mips_ip): Handle triple relocations.
(mips16_ip): Likewise.
(mips_force_relocation): Force handling of triple relocations
without symbols for NewABI.
(md_apply_fix): Add handling for some NewABI relocations.
(mips_target_format): Move downwards in file, use HAVE_64BIT_OBJECTS
in it.
(mips_abi_level, mips_abi): New enum.
(mips_32bit_abi): Remove.
(HAVE*PRS): Use mips_abi instead of mips_32bit_abi.
(HAVE_NEWABI): New define.
(HAVE_64BIT_OBJECTS): New define.
(HAVE_32BIT_ADDRESSES): Don't return true for 64bit objects.
(HAVE_64BIT_ADDRESSES): New define, inverse of HAVE_32BIT_ADDRESSES.
(support_64bit_objects): New prototype.
(md_begin): Use mips_abi instead of mips_32bit_abi. Don't write
.reginfo section for n32, use .MIPS.options instead.
(support_64bit_objects): New function, code from md_parse_option.
(md_longopts): Add -n32 option.
(md_parse_option): Use mips_abi instead of mips_32bit_abi/mips64.
Add -n32 option. Protect with OBJ_ELF.
(s_mipsset): Use mips_abi instead of mips_32bit_abi.
(mips_elf_final_processing): Likewise. Don't write .reginfo section
for n32, use .MIPS.options instead.
and rearrange for efficiency. For "PIC code" subtraction, use
"rightseg" rather than recalculating. For "symbol OP symbol"
subtract, set "retval" to absolute_section if symbols in same
section.
* symbols.c (resolve_symbol_value): Resolve "sym +/- expr" to an
O_symbol. Simplify a +/- b code. Allow equality and non-equality
comparisons on symbols from any section. Allow other comparison
operators as for subtraction.
(symbol_equated_reloc_p): New predicate function.
* symbols.h (symbol_equated_reloc_p): Declare.
* write.c (adjust_reloc_syms): Use symbol_equated_reloc_p.
(write_relocs): Likewise.
(write_object_file): Likewise.
(relax_segment <rs_machine_dependent>): Ensure segment for
expression syms is set correctly.
* config/tc-mips.c (md_estimate_size_before_relax): Likewise.
* config/tc-i386.c (md_assemble <Output jumps>): Don't lose part
of a complex expression when setting up frag_var.
value, not the word beyond maximum.
* tc_mips.c (macro_build_lui): Code cleanup.
(macro): Reflect change to MAX_GPREL_OFFSET.
(mips_ip): Check explicitly against S_EX_NONE.
(my_get_SmallExpression): parse for %gp_rel, not %gprel.
(md_apply_fix): Code cleanup.
return values.
(mips_ip): Use the new return values instead of characters. Add
support for %higher and %highest.
(LP): Remove.
(RP): Remove.
(my_getSmallExpression): Make parsing case insensitive and more
reliable. Add support for %higher and %highest. Further support to
parse %gprel and %neg is implemented but currently deactivated.
-mipsX is specified. Make both -mcpu/-march and -mcpu/-mtune pairs
mutually exclusive (if they are different).
(md_parse_option): Warn if an -march/-mtune/-mcpu/-m<cpu> option is
set more than once.
* config/tc-mips.c (mips_fp32, mips_32bit_abi): New static variables.
(md_long_opts): Add -mfp32 option.
(md_parse_option): Handle it. Set mips_32bit_abi given -mabi=32.
(md_show_usage): Show usage for -mfp32 and -mgp32.
(HAVE_32BIT_GPRS, HAVE_32BIT_FPRS): New macros.
(HAVE_64BIT_GPRS, HAVE_64BIT_FPRS): New macros, inverse of the above.
(HAVE_32BIT_ADDRESSES): New macro.
(load_register): Use HAVE_32BIT_GPRS to determine the register width.
(load_address): Use HAVE_32BIT_ADDRESSES to determine the address size.
(s_cprestore, s_cpadd): Likewise.
(macro): Use HAVE_32BIT_GPRS to determine the width of registers
used in branch and M_LI_D macros. Use HAVE_64BIT_FPRS to determine
the width registers used in M_LI_DD macros. Use HAVE_32BIT_ADDRESSES
to determine the width of addresses in load, store and jump macros.
(macro2): Use HAVE_32BIT_GPRS to determine the width of registers
used in set instructions; do not check the address size for them.
Use HAVE_32BIT_ADDRESSES to determine the width of addresses in
unaligned load and store macros.
(mips_ip): Use the new macros to check the width of a register when
processing float constants. Force a constant into memory if it is
destined for an FPR and the FPRs are wider than the GPRs. Warn about
odd FPR numbers if HAVE_32BIT_FPRS. Use HAVE_32BIT_GPRS rather
than mips_gp32 to select synthetic instructions.
(macro_build): Use HAVE_32BIT_GPRS rather than mips_gp32 to select
synthetic instructions.
* config/tc-mips.c (md_estimate_size_before_relax): Make sure
we treat weak like extern only for ELF.
(mips_fix_adjustable): Make sure we don't adjust extern/weak
symbols only for ELF.
* config/tc-mips.c (md_apply_fix): Prevent addend from becoming zero
if it's expected to be non-zero.
[gas/testsuite]
* gas/mips/elf-rel3.s: Add zero word to end of file.
* NEWS: Updated for the new -n option for the MIPS assembler.
* config/tc-mips.c (md_show_usage): Add -n.
* doc/as.texinfo: Document the new -n option.
* doc/c-mips.texi: Likewise.
* doc/as.1: Regenerated.
* config/tc-mips.c (warn_nops): New variable. Set to 0 to
disable warning about all NOPS that the assembler generates.
(macro): Warn NOPS generated only if warn_nops is not 0.
(md_shortopts): Add `n'.
(md_parse_option): Set warn_nops to 1 for `n'.
* config/tc-mips.c (md_apply_fix): Don't adjust common
extern/weak symbols for ELF.
(md_estimate_size_before_relax): Treat weak like extern for
ELF.
(mips_fix_adjustable): Don't adjust extern/weak symbols for
ELF.
