Certain insns have restrictions on fields. For example, the insn
mentioned in the PR, lqarx, must specify an even general purpose
register as its destination and that register cannot appear in
either of the base or index reg fields. This holds even when the RA0
field is 0 (meaning a zero rather than r0).
PR 21124
* ppc-opc.c (extract_esync, extract_ls, extract_ral, extract_ram)
(extract_raq, extract_ras, extract_rbx): New functions.
(powerpc_operands): Use opposite corresponding insert function.
(Q_MASK): Define.
(powerpc_opcodes): Apply Q_MASK to all quad insns with even
register restriction.
This commit adds support to GDB so that it can modify the disassembler-options
value that is passed to the disassembler, similar to objdump's -M option.
Currently, the only supported targets are ARM, PowerPC and S/390, but
adding support for a new target(s) is not difficult.
include/
* dis-asm.h (disasm_options_t): New typedef.
(parse_arm_disassembler_option): Remove prototype.
(set_arm_regname_option): Likewise.
(get_arm_regnames): Likewise.
(get_arm_regname_num_options): Likewise.
(disassemble_init_s390): New prototype.
(disassembler_options_powerpc): Likewise.
(disassembler_options_arm): Likewise.
(disassembler_options_s390): Likewise.
(remove_whitespace_and_extra_commas): Likewise.
(disassembler_options_cmp): Likewise.
(next_disassembler_option): New inline function.
(FOR_EACH_DISASSEMBLER_OPTION): New macro.
opcodes/
* disassemble.c Include "safe-ctype.h".
(disassemble_init_for_target): Handle s390 init.
(remove_whitespace_and_extra_commas): New function.
(disassembler_options_cmp): Likewise.
* arm-dis.c: Include "libiberty.h".
(NUM_ELEM): Delete.
(regnames): Use long disassembler style names.
Add force-thumb and no-force-thumb options.
(NUM_ARM_REGNAMES): Rename from this...
(NUM_ARM_OPTIONS): ...to this. Use ARRAY_SIZE.
(get_arm_regname_num_options): Delete.
(set_arm_regname_option): Likewise.
(get_arm_regnames): Likewise.
(parse_disassembler_options): Likewise.
(parse_arm_disassembler_option): Rename from this...
(parse_arm_disassembler_options): ...to this. Make static.
Use new FOR_EACH_DISASSEMBLER_OPTION macro to scan over options.
(print_insn): Use parse_arm_disassembler_options.
(disassembler_options_arm): New function.
(print_arm_disassembler_options): Handle updated regnames.
* ppc-dis.c: Include "libiberty.h".
(ppc_opts): Add "32" and "64" entries.
(ppc_parse_cpu): Use ARRAY_SIZE and disassembler_options_cmp.
(powerpc_init_dialect): Add break to switch statement.
Use new FOR_EACH_DISASSEMBLER_OPTION macro.
(disassembler_options_powerpc): New function.
(print_ppc_disassembler_options): Use ARRAY_SIZE.
Remove printing of "32" and "64".
* s390-dis.c: Include "libiberty.h".
(init_flag): Remove unneeded variable.
(struct s390_options_t): New structure type.
(options): New structure.
(init_disasm): Rename from this...
(disassemble_init_s390): ...to this. Add initializations for
current_arch_mask and option_use_insn_len_bits_p. Remove init_flag.
(print_insn_s390): Delete call to init_disasm.
(disassembler_options_s390): New function.
(print_s390_disassembler_options): Print using information from
struct 'options'.
* po/opcodes.pot: Regenerate.
binutils/
* objdump.c (main): Use remove_whitespace_and_extra_commas.
gdb/
* NEWS: Mention new set/show disassembler-options commands.
* doc/gdb.texinfo: Document new set/show disassembler-options commands.
* disasm.c: Include "arch-utils.h", "gdbcmd.h" and "safe-ctype.h".
(prospective_options): New static variable.
(gdb_disassembler::gdb_disassembler): Initialize
m_di.disassembler_options.
(gdb_buffered_insn_length_init_dis): Initilize di->disassembler_options.
(get_disassembler_options): New function.
(set_disassembler_options): Likewise.
(set_disassembler_options_sfunc): Likewise.
(show_disassembler_options_sfunc): Likewise.
(disassembler_options_completer): Likewise.
(_initialize_disasm): Likewise.
* disasm.h (get_disassembler_options): New prototype.
(set_disassembler_options): Likewise.
* gdbarch.sh (gdbarch_disassembler_options): New variable.
(gdbarch_verify_disassembler_options): Likewise.
* gdbarch.c: Regenerate.
* gdbarch.h: Likewise.
* arm-tdep.c (num_disassembly_options): Delete.
(set_disassembly_style): Likewise.
(arm_disassembler_options): New static variable.
(set_disassembly_style_sfunc): Convert short style name into long
option name. Call set_disassembler_options.
(show_disassembly_style_sfunc): New function.
(arm_gdbarch_init): Call set_gdbarch_disassembler_options and
set_gdbarch_verify_disassembler_options.
(_initialize_arm_tdep): Delete regnames variable and update callers.
(arm_disassembler_options): Initialize.
(disasm_options): New variable.
(num_disassembly_options): Rename from this...
(num_disassembly_styles): ...to this. Compute by scanning through
disasm_options.
(valid_disassembly_styles): Initialize using disasm_options.
Remove calls to parse_arm_disassembler_option, get_arm_regnames and
set_arm_regname_option.
Pass show_disassembly_style_sfunc to the "disassembler" setshow command.
* rs6000-tdep.c (powerpc_disassembler_options): New static variable.
(rs6000_gdbarch_init): Call set_gdbarch_disassembler_options and
set_gdbarch_verify_disassembler_options.
* s390-tdep.c (s390_disassembler_options): New static variable.
(s390_gdbarch_init):all set_gdbarch_disassembler_options and
set_gdbarch_verify_disassembler_options.
gdb/testsuite/
* gdb.arch/powerpc-power.exp: Delete test.
* gdb.arch/powerpc-power.s: Likewise.
* gdb.disasm/disassembler-options.exp: New test.
* gdb.arch/powerpc-altivec.exp: Likewise.
* gdb.arch/powerpc-altivec.s: Likewise.
* gdb.arch/powerpc-altivec2.exp: Likewise.
* gdb.arch/powerpc-altivec2.s: Likewise.
* gdb.arch/powerpc-altivec3.exp: Likewise.
* gdb.arch/powerpc-altivec3.s: Likewise.
* gdb.arch/powerpc-power7.exp: Likewise.
* gdb.arch/powerpc-power7.s: Likewise.
* gdb.arch/powerpc-power8.exp: Likewise.
* gdb.arch/powerpc-power8.s: Likewise.
* gdb.arch/powerpc-power9.exp: Likewise.
* gdb.arch/powerpc-power9.s: Likewise.
* gdb.arch/powerpc-vsx.exp: Likewise.
* gdb.arch/powerpc-vsx.s: Likewise.
* gdb.arch/powerpc-vsx2.exp: Likewise.
* gdb.arch/powerpc-vsx2.s: Likewise.
* gdb.arch/powerpc-vsx3.exp: Likewise.
* gdb.arch/powerpc-vsx3.s: Likewise.
* gdb.arch/arm-disassembler-options.exp: Likewise.
* gdb.arch/powerpc-disassembler-options.exp: Likewise.
* gdb.arch/s390-disassembler-options.exp: Likewise.
Just like REX.W affects operand size of the implicit rAX/rDX inputs to
PCMPESTR{I,M}, VEX.W does for VPCMPESTR{I,M}. Allow Q or L suffixes on
the instructions.
Similarly the disassembler needs to be adjusted to no longer require
VEX.W to be zero for the instructions to be valid, and to emit proper
suffixes.
Note, however, that this doesn't address the problem of there being no
way to control (at least) {,E}VEX.W for 32- or 16-bit code. Nor does it
address the problem of the many WIG instructions not getting properly
disassembled when VEX.W=1.
This patch supports some additions to the SVE architecture prior to
its public release.
include/
* opcode/aarch64.h (AARCH64_OPND_SVE_ADDR_RI_S4x16)
(AARCH64_OPND_SVE_IMM_ROT1, AARCH64_OPND_SVE_IMM_ROT2)
(AARCH64_OPND_SVE_Zm3_INDEX, AARCH64_OPND_SVE_Zm3_22_INDEX)
(AARCH64_OPND_SVE_Zm4_INDEX): New aarch64_opnds.
opcodes/
* aarch64-tbl.h (OP_SVE_HMH, OP_SVE_VMU_HSD, OP_SVE_VMVU_HSD)
(OP_SVE_VMVV_HSD, OP_SVE_VMVVU_HSD, OP_SVE_VM_HSD, OP_SVE_VUVV_HSD)
(OP_SVE_VUV_HSD, OP_SVE_VU_HSD, OP_SVE_VVVU_H, OP_SVE_VVVU_S)
(OP_SVE_VVVU_HSD, OP_SVE_VVV_D, OP_SVE_VVV_D_H, OP_SVE_VVV_H)
(OP_SVE_VVV_HSD, OP_SVE_VVV_S, OP_SVE_VVV_S_B, OP_SVE_VVV_SD_BH)
(OP_SVE_VV_BHSDQ, OP_SVE_VV_HSD, OP_SVE_VZVV_HSD, OP_SVE_VZV_HSD)
(OP_SVE_V_HSD): New macros.
(OP_SVE_VMU_SD, OP_SVE_VMVU_SD, OP_SVE_VM_SD, OP_SVE_VUVV_SD)
(OP_SVE_VU_SD, OP_SVE_VVVU_SD, OP_SVE_VVV_SD, OP_SVE_VZVV_SD)
(OP_SVE_VZV_SD, OP_SVE_V_SD): Delete.
(aarch64_opcode_table): Add new SVE instructions.
(aarch64_opcode_table): Use imm_rotate{1,2} instead of imm_rotate
for rotation operands. Add new SVE operands.
* aarch64-asm.h (ins_sve_addr_ri_s4): New inserter.
(ins_sve_quad_index): Likewise.
(ins_imm_rotate): Split into...
(ins_imm_rotate1, ins_imm_rotate2): ...these two inserters.
* aarch64-asm.c (aarch64_ins_imm_rotate): Split into...
(aarch64_ins_imm_rotate1, aarch64_ins_imm_rotate2): ...these two
functions.
(aarch64_ins_sve_addr_ri_s4): New function.
(aarch64_ins_sve_quad_index): Likewise.
(do_misc_encoding): Handle "MOV Zn.Q, Qm".
* aarch64-asm-2.c: Regenerate.
* aarch64-dis.h (ext_sve_addr_ri_s4): New extractor.
(ext_sve_quad_index): Likewise.
(ext_imm_rotate): Split into...
(ext_imm_rotate1, ext_imm_rotate2): ...these two extractors.
* aarch64-dis.c (aarch64_ext_imm_rotate): Split into...
(aarch64_ext_imm_rotate1, aarch64_ext_imm_rotate2): ...these two
functions.
(aarch64_ext_sve_addr_ri_s4): New function.
(aarch64_ext_sve_quad_index): Likewise.
(aarch64_ext_sve_index): Allow quad indices.
(do_misc_decoding): Likewise.
* aarch64-dis-2.c: Regenerate.
* aarch64-opc.h (FLD_SVE_i3h, FLD_SVE_rot1, FLD_SVE_rot2): New
aarch64_field_kinds.
(OPD_F_OD_MASK): Widen by one bit.
(OPD_F_NO_ZR): Bump accordingly.
(get_operand_field_width): New function.
* aarch64-opc.c (fields): Add new SVE fields.
(operand_general_constraint_met_p): Handle new SVE operands.
(aarch64_print_operand): Likewise.
* aarch64-opc-2.c: Regenerate.
gas/
* doc/c-aarch64.texi: Document that sve implies fp16, simd and compnum.
* config/tc-aarch64.c (parse_vector_type_for_operand): Allow .q
to be used with SVE registers.
(parse_operands): Handle new SVE operands.
(aarch64_features): Make "sve" require F16 rather than FP. Also
require COMPNUM.
* testsuite/gas/aarch64/sve.s: Add tests for new instructions.
Include compnum tests.
* testsuite/gas/aarch64/sve.d: Update accordingly.
* testsuite/gas/aarch64/sve-invalid.s: Add tests for new instructions.
* testsuite/gas/aarch64/sve-invalid.l: Update accordingly. Also
update expected output for new FMOV and MOV alternatives.
This patch adds a named "compnum" feature for the ARMv8.3-A FCADD
and FCMLA extensions.
include/
* opcode/aarch64.h (AARCH64_FEATURE_COMPNUM): New macro.
(AARCH64_ARCH_V8_3): Include AARCH64_FEATURE_COMPNUM.
opcodes/
* aarch64-tbl.h (aarch64_feature_simd_v8_3): Replace with...
(aarch64_feature_compnum): ...this.
(SIMD_V8_3): Replace with...
(COMPNUM): ...this.
(CNUM_INSN): New macro.
(aarch64_opcode_table): Use it for the complex number instructions.
gas/
* doc/c-aarch64.texi: Add a "compnum" entry.
* config/tc-aarch64.c (aarch64_features): Likewise,
* testsuite/gas/aarch64/advsimd-compnum.s: New test.
* testsuite/gas/aarch64/advsimd-compnum.d: Likewise.
Opcodes F6/1 and F7/1 are aliases of F6/0 and F7/0 in all modes. This
complements commit 8b89fe14b5 ("X86: Decode opcode 0x82 as opcode 0x80
in 32-bit mode"), just that here 64-bit mode is also covered.
With this change an architecture level bump due to assembly ASIs will show
up as a warning/error depending on options passed to gas.
Tested with sparc64-linux-gnu, and it does not introduce any regressions.
gas/ChangeLog:
Add support for associating SPARC ASIs with an architecture level.
* config/tc-sparc.c (parse_sparc_asi): New encode SPARC ASIs.
opcodes/ChangeLog:
Add support for associating SPARC ASIs with an architecture level.
* include/opcode/sparc.h (sparc_asi): New sparc_asi struct.
* opcodes/sparc-opc.c (asi_table): Updated asi_table and encoding/
decoding of SPARC ASIs.
No such ModR/M-less opcode has ever existed in public documentation
afaict, so I can't see why it had been added in the first place.
Furthermore opcode 77 is special only with implied prefix 0F.
Commit 93562a343c ("[AArch64] PR target/20666, fix wrong encoding of
new introduced BFC pseudo") changed the destination operand to 0,
making the whole function invocation a no-op. We really want to copy
operand 0 (a register) to operand 1 (an immediate before coming here),
even if right now this likely is only a latent bug.
This patch adds the SVE-specific system registers.
opcodes/
* aarch64-opc.c (aarch64_sys_regs): Add SVE registers.
(aarch64_sys_reg_supported_p): Handle them.
gas/
* testsuite/gas/aarch64/sve-sysreg.s,
testsuite/gas/aarch64/sve-sysreg.d,
testsuite/gas/aarch64/sve-sysreg-invalid.d,
testsuite/gas/aarch64/sve-sysreg-invalid.l: New tests.
opcodes * arc-regs.h: Distinguish some of the registers different on
ARC700 and HS38 cpus.
gas * testsuite/gas/arc/st.d: Update for 0xe having a name now
This stops powerpc gas blithely accepting such nonsense as
"addi %f4,%cr3,%r31".
PR 21118
gas/
* NEWS: Mention powerpc register checks.
* config/tc-ppc.c (struct pd_reg): Make value a short. Add flags.
(pre_defined_registers): Delete fpscr and pmr entries. Set
register type in flags.
(cr_names): Set type in flags.
(reg_name_search): Return pointer to struct pd_reg rather than value.
(register_name): Adjust to suit. Set X_md from flags.
(ppc_parse_name): Likewise.
(ppc_optimize_expr): New function.
(md_assemble): Verify expresion reg flags match operand.
* config/tc-ppc.h (md_optimize_expr): Define.
(ppc_optimize_expr): Declare.
include/
* opcode/ppc.h (PPC_OPERAND_*): Reassign values, regs first.
(PPC_OPERAND_SPR, PPC_OPERAND_GQR): Define.
opcodes/
* ppc-opc.c (powerpc_operands): Flag SPR, SPRG and TBR entries
with PPC_OPERAND_SPR. Flag PSQ and PSQM with PPC_OPERAND_GQR.
* cgen-opc.c (cgen_lookup_insn): Delete buf and base_insn temps.
Use insn_bytes_value and insn_int_value directly instead. Don't
free allocated memory until function exit.
PR 21096
bfd * coffcode.h (coff_write_object_contents): Enlarge size of
s_name_buf in order to avoid compile time warning about possible
integer truncation.
* elf32-nds32.c (nds32_elf_ex9_import_table): Mask off lower
32-bits of insn value before printing into buffer.
opcodes * aarch64-opc.c (print_register_list): Ensure that the register
list index will fir into the tb buffer.
(print_register_offset_address): Likewise.
* tic6x-dis.c (print_insn_tic6x): Increase size of func_unit_buf.
PR 21056
opcodes * tic6x-dis.c (print_insn_tic6x): Correct displaying of parallel
instructions when the previous fetch packet ends with a 32-bit
instruction.
gas * testsuite/gas/tic6x/insns16-parallel.s: New test case.
* testsuite/gas/tic6x/insns16-parallel.d: New test driver.
Bit 24 of the indexed element vcmla decode mask was incorrectly
left unset. This could cause incorrect disassembly of some
currently undefined instructions as vcmla.
Rotatation immediates were not printed correctly in the disassembly
(could print 170 and 280 instead of 180 and 270).
opcodes/
* arm-dis.c (coprocessor_opcodes): Fix vcmla mask and disassembly.
gas/
* testsuite/gas/arm/armv8_3-a-simd.s: Add vcmla tests.
* testsuite/gas/arm/armv8_3-a-simd.d: Update.
When I inspect the return values of disassmblers, I happen to see
various -1/-2/-3 magic numbers are used in m68k-dis.c. This patch
is to replace them with enum.
-1 and -2 is "clearly documented" in print_ins_arg's comments, but
-3 isn't. In fact, -3 is returned when FETCH_DATA returns false,
which means memory error (because fetch_data return 0 on memory
error). So I name enum PRINT_INSN_ARG_MEMORY_ERROR for -3.
This patch is a refactor patch, doesn't affect any functionality.
opcodes:
2017-01-13 Yao Qi <yao.qi@linaro.org>
* m68k-dis.c (enum print_insn_arg_error): New.
(NEXTBYTE): Replace -3 with
PRINT_INSN_ARG_MEMORY_ERROR.
(NEXTULONG): Likewise.
(NEXTSINGLE): Likewise.
(NEXTDOUBLE): Likewise.
(NEXTDOUBLE): Likewise.
(NEXTPACKED): Likewise.
(FETCH_ARG): Likewise.
(FETCH_DATA): Update comments.
(print_insn_arg): Update comments. Replace magic numbers with
enum.
(match_insn_m68k): Likewise.
Disassemblers in opcodes return -1 on memory error, but msp430 doesn't
follow this convention. If I change GDB not to throw exception in
disassemble_info.memory_error_func and rely on the return value of
disassembler, I'll get the following output.
(gdb) disassemble 0x0,+8
Dump of assembler code from 0x0 to 0x8:
0x00000000: .word 0xffff; ????
0x00000002: .word 0xffff; ????
0x00000004: .word 0xffff; ????
0x00000006: .word 0xffff; ????
End of assembler dump.
This patch teaches print_insn_msp430 and its callees to return -1
on memory error.
opcodes:
2017-01-12 Yao Qi <yao.qi@linaro.org>
* msp430-dis.c (msp430_singleoperand): Return -1 if
msp430dis_opcode_signed returns false.
(msp430_doubleoperand): Likewise.
(msp430_branchinstr): Return -1 if
msp430dis_opcode_unsigned returns false.
(msp430x_calla_instr): Likewise.
(print_insn_msp430): Likewise.
PR 20946
* frv-desc.c (lookup_mach_via_bfd_name): Return NULL if the name
could not be matched.
(frv_cgen_cpu_open): Allow for lookup_mach_via_bfd_name returning
NULL.
The weaker release consistency support of ARMv8.3-A is allowed as an optional
extension for ARMv8.2-A, so separate command line option and feature flag is
added: -march=armv8.2-a+rcpc turns LDAPR, LDAPRB, LDAPRH instructions on.
opcodes/
* aarch64-tbl.h (RCPC, RCPC_INSN): Define.
(aarch64_opcode_table): Use RCPC_INSN.
include/
* opcode/aarch64.h (AARCH64_FEATURE_RCPC): Define.
(AARCH64_ARCH_V8_3): Update.
gas/
* config/tc-aarch64.c (aarch64_features): Add rcpc.
* doc/c-aarch64.texi (AArch64 Extensions): Document rcpc.
* testsuite/gas/aarch64/ldst-exclusive-armv8_3.d: Rename to ...
* testsuite/gas/aarch64/ldst-rcpc.d: This.
* testsuite/gas/aarch64/ldst-exclusive-armv8_3.s: Rename to ...
* testsuite/gas/aarch64/ldst-rcpc.s: This.
* testsuite/gas/aarch64/ldst-rcpc-armv8_2.d: New test.
gas * config/tc-riscv.c (riscv_set_arch): Whitelist the "q" ISA
extension.
(riscv_after_parse_args): Set FLOAT_ABI_QUAD when the Q ISA is
enabled and no other ABI is specified.
include * opcode/riscv-opc.h: Add support for the "q" ISA extension.
opcodes * riscv-opc.c (riscv-opcodes): Add support for the "q" ISA
extension.
* riscv-opcodes/all-opcodes: Likewise.
avrdis_opcode return type is unsigned short, but -1 at the end of
this function is returned. Additionally, print_insn_avr doesn't
handle when -1 (in case of memory error) is returned from
avrdis_opcode.
This patch changes avrdis_opcode returning int indicating the error,
and adds a new argument for instruction we got on success. The
opcode is 16-bit, so I change local variables type to uint16_t,
and include "bfd_stdint.h" as a result. On memory error,
print_insn_avr returns -1, which is a common convention among most
of print_insn_$ARCH functions.
opcodes:
2016-12-29 Yao Qi <yao.qi@linaro.org>
* avr-dis.c: Include "bfd_stdint.h"
(avrdis_opcode): Change return type to int, add argument
insn. Set *INSN on success.
(print_insn_avr): Check return value of avrdis_opcode, and
return -1 on error.
Add ASMACRO instruction support as per the MIPS16e ASE architecture
specifications [1][2], completing MIPS16e instruction set support.
[1] "MIPS32 Architecture for Programmers, Volume IV-a: The MIPS16e
Application-Specific Extension to the MIPS32 Architecture", MIPS
Technologies, Inc., Document Number: MD00076, Revision 2.63, July
16, 2013, Section 4.1 "MIPS16e Instruction Descriptions", p. 65
[2] "MIPS64 Architecture for Programmers, Volume IV-a: The MIPS16e
Application-Specific Extension to the MIPS64 Architecture", MIPS
Technologies, Inc., Document Number: MD00077, Revision 2.60, June
25, 2008, Section 1.1 "MIPS16e Instruction Descriptions", p. 66
include/
* opcode/mips.h: Document `0', `1', `2', `3', `4' and `s'
operand codes.
opcodes/
* mips16-opc.c (decode_mips16_operand): Add `0', `1', `2', `3',
`4' and `s' operand codes.
(mips16_opcodes): Add "asmacro" entry.
binutils/
* testsuite/binutils-all/mips/mips16-extend-insn.d: Update for
ASMACRO support.
gas/
* testsuite/gas/mips/mips16-asmacro.d: New test.
* testsuite/gas/mips/mips16-32@mips16-asmacro.d: New test.
* testsuite/gas/mips/mips16-64@mips16-asmacro.d: New test.
* testsuite/gas/mips/mips16-asmacro.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new tests.
Simplify extended operand handling and only specially process immediates
which require bit shuffling, using the generic operand insertion and
extraction handlers for the '<' (5-bit shift amount) operand code in
particular. Require the least significant bit of all extended operand
forms to be (artificially) set to 0 for their special processing to
trigger.
gas/
* config/tc-mips.c (mips16_immed): Limit `mips16_immed_extend'
use to operands whose LSB position is zero.
opcodes/
* mips-dis.c (print_mips16_insn_arg): Simplify processing of
extended operands.
* mips16-opc.c (decode_mips16_operand): Switch the extended
form of the `<' operand type to LSB position 22.
Replace `0' and `4' operand codes with `.' and `F' respectively to free
up the `0'-`4' consecutive range. No functional change.
gas/
* config/tc-mips.c (mips16_macro_build): Replace `0' and `4'
operand codes with `.' and `F' respectively.
(mips16_macro): Likewise.
include/
* opcode/mips.h: Replace `0' and `4' operand codes with `.' and
`F' respectively.
opcodes/
* mips16-opc.c (decode_mips16_operand): Replace `0' and `4'
operand codes with `.' and `F' respectively.
(mips16_opcodes): Likewise.
Identify non-extensible instructions in the MIPS16 opcode table and
disallow their use with the `.e' instruction size suffix in assembly and
do not interpret any EXTEND prefix present as a part of the instruction
in disassembly.
According to all versions of the MIPS16 ASE specifications the following
instructions encodings are not extensible [1][2][3][4][5][6]: I8/MOV32R,
I8/MOVR32, all RRR minor opcodes, all RR minor opcodes except from DSRA
and DSRL, and EXTEND itself, and as from revision 2.50 of the MIPS16e
ASE specifications it has been further clarified what was previously
implied, that non-extesiable instructions when preceded with an EXTEND
prefix must cause a Reserved Instruction exception [3][5].
Therefore in the presence of an EXTEND prefix none of these instructions
are supposed to be handled as extended instructions and supporting these
forms in disassembly causes confusion, and in the case of the RRR major
opcode it also clashes with the ASMACRO encoding.
References:
[1] "Product Description, MIPS16 Application-Specific Extension",
Version 1.3, MIPS Technologies, Inc., 970130, Table 3. "MIPS16
Instruction Set Summary", p. 5
[2] same, Table 5 "RR Minor Opcodes (RR-type instructions)", p.10
[3] "MIPS32 Architecture for Programmers, Volume IV-a: The MIPS16e
Application-Specific Extension to the MIPS32 Architecture", MIPS
Technologies, Inc., Document Number: MD00076, Revision 2.63, July
16, 2013, Section 3.9 "MIPS16e Instruction Summaries", pp. 37-39
[4] same, Section 3.15 "Instruction Bit Encoding", pp. 46-49
[5] "MIPS64 Architecture for Programmers, Volume IV-a: The MIPS16e
Application-Specific Extension to the MIPS64 Architecture", MIPS
Technologies, Inc., Document Number: MD00077, Revision 2.60, June
25, 2008, Section 1.9 "MIPS16e Instruction Summaries", pp. 38-41
[6] same, Section 1.15 "Instruction Bit Encoding", pp. 48-51
include/
* opcode/mips.h (INSN2_SHORT_ONLY): New macro.
gas/
* config/tc-mips.c (is_size_valid_16): Disallow a `.e' suffix
instruction size override for INSN2_SHORT_ONLY opcode table
entries.
* testsuite/gas/mips/mips16-extend-swap.d: Adjust output.
* testsuite/gas/mips/mips16-macro-e.l: Adjust error messages.
* testsuite/gas/mips/mips16-32@mips16-macro-e.l: Adjust error
messages.
* testsuite/gas/mips/mips16e-32@mips16-macro-e.l: Adjust error
messages.
* testsuite/gas/mips/mips16-insn-e.d: New test.
* testsuite/gas/mips/mips16-insn-t.d: New test.
* testsuite/gas/mips/mips16-32@mips16-insn-e.d: New test.
* testsuite/gas/mips/mips16-64@mips16-insn-e.d: New test.
* testsuite/gas/mips/mips16e-32@mips16-insn-e.d: New test.
* testsuite/gas/mips/mips16-32@mips16-insn-t.d: New test.
* testsuite/gas/mips/mips16-64@mips16-insn-t.d: New test.
* testsuite/gas/mips/mips16e-32@mips16-insn-t.d: New test.
* testsuite/gas/mips/mips16-insn-e.l: New stderr output.
* testsuite/gas/mips/mips16-insn-t.l: New stderr output.
* testsuite/gas/mips/mips16-32@mips16-insn-e.l: New stderr
output.
* testsuite/gas/mips/mips16-64@mips16-insn-e.l: New stderr
output.
* testsuite/gas/mips/mips16e-32@mips16-insn-e.l: New stderr
output.
* testsuite/gas/mips/mips16-32@mips16-insn-t.l: New stderr
output.
* testsuite/gas/mips/mips16-64@mips16-insn-t.l: New stderr
output.
* testsuite/gas/mips/mips16e-32@mips16-insn-t.l: New stderr
output.
* testsuite/gas/mips/mips16-insn-e.s: New test source.
* testsuite/gas/mips/mips16-insn-t.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new tests.
opcodes/
* mips-dis.c (print_insn_mips16): Disallow EXTEND prefix
matching for INSN2_SHORT_ONLY opcode table entries.
* mips16-opc.c (SH): New macro.
(mips16_opcodes): Set SH in `pinfo2' for non-extensible
instruction entries: "nop", "addu", "and", "break", "cmp",
"daddu", "ddiv", "ddivu", "div", "divu", "dmult", "dmultu",
"drem", "dremu", "dsllv", "dsll", "dsrav", "dsra", "dsrlv",
"dsrl", "dsubu", "exit", "entry", "jalr", "jal", "jr", "j",
"jalrc", "jrc", "mfhi", "mflo", "move", "mult", "multu", "neg",
"not", "or", "rem", "remu", "sllv", "sll", "slt", "sltu",
"srav", "sra", "srlv", "srl", "subu", "xor", "sdbbp", "seb",
"seh", "sew", "zeb", "zeh", "zew" and "extend".
binutils/
* testsuite/binutils-all/mips/mips16-extend-insn.d: New test.
* testsuite/binutils-all/mips/mips16-extend-insn.s: New test
source.
* testsuite/binutils-all/mips/mips.exp: Run the new tests.
Remove special casing for the `6' operand code used for the embedded
trap code of the BREAK and the SDBBP instructions to support supposedly
extended forms of these instructions.
According to all versions of the MIPS16 ASE specifications these
instructions are not extensible [1][2][3][4][5][7][8][10][11], and as
from revision 2.50 of the MIPS16e ASE specifications it has been further
clarified what was previously implied, that non-extesiable instructions
when preceded with an EXTEND prefix must cause a Reserved Instruction
exception [5][6][9][10].
Therefore supposedly extended BREAK and SDBBP instructions do not serve
their purpose anymore as they do not cause a Bp and a Debug exception
respectively and supporting these forms in disassembly only causes
confusion.
References:
[1] "Product Description, MIPS16 Application-Specific Extension",
Version 1.3, MIPS Technologies, Inc., 970130, Table 3. "MIPS16
Instruction Set Summary", p. 5
[2] same, Table 5 "RR Minor Opcodes (RR-type instructions)", p.10
[3] same, Table 18. "Extendable MIPS16 Instructions", p. 24
[4] "MIPS32 Architecture for Programmers, Volume IV-a: The MIPS16e
Application-Specific Extension to the MIPS32 Architecture", MIPS
Technologies, Inc., Document Number: MD00076, Revision 2.63, July
16, 2013, Table 3.8 "MIPS16e Special Instructions", p. 38
[5] same, Section 3.11 "MIPS16e Extensible Instructions, p. 41
[6] same, Table 3.15 "MIPS16e Extensible Instructions", p. 41
[7] same, Table 3.24 "MIPS16e RR Encoding of the Funct Field", p. 49
[8] "MIPS64 Architecture for Programmers, Volume IV-a: The MIPS16e
Application-Specific Extension to the MIPS64 Architecture", MIPS
Technologies, Inc., Document Number: MD00077, Revision 2.60, June
25, 2008, Table 1.8 "MIPS16e Special Instructions", p. 39
[9] same, Section 1.11 "MIPS16e Extensible Instructions", p. 42
[10] same, Table 1.15 "MIPS16e Extensible Instructions", pp. 42-43
[11] same, Table 1.24 "MIPS16e RR Encoding of the Funct Field", p. 50
gas/
* config/tc-mips.c (match_mips16_insn): Remove the `6' operand
code special case and its associated comment.
opcodes/
* mips16-opc.c (decode_mips16_operand) <'6'>: Remove extended
encoding support.
Do not allow any explicitly coded EXTEND instruction to be automatically
scheduled into a jump delay slot, as an EXTEND prefix is coupled with
the next regular MIPS16 instruction and therefore swapping it with a
jump would change program's semantics; EXTEND is not architecturally
allowed to be present in a jump delay slot anyway.
opcodes/
* mips16-opc.c (mips16_opcodes): Set NODS in `pinfo' for
"extend".
gas/
* testsuite/gas/mips/mips16-extend-swap.d: New test.
* testsuite/gas/mips/mips16-extend-swap.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new test.
Fix a regression introduced with commit 5e7fc731f8 ("MIPS/opcodes:
Also set disassembler's ASE flags from ELF structures"), further updated
with commit 4df995c771 ("MIPS/opcodes: Also set disassembler's ASE
flags from ELF structures"), and use autoconf to check for the presence
of `bfd_mips_elf_get_abiflags' in BFD.
opcodes/
* mips-dis.c (set_default_mips_dis_options): Use
HAVE_BFD_MIPS_ELF_GET_ABIFLAGS rather than BFD64 to guard the
call to `bfd_mips_elf_get_abiflags'.
* configure.ac: Check for `bfd_mips_elf_get_abiflags' in BFD.
* Makefile.am (CONFIG_STATUS_DEPENDENCIES): Add `libbfd.la'.
* aclocal.m4: Regenerate.
* configure: Regenerate.
* config.in: Regenerate.
* Makefile.in: Regenerate.
Instructions like "jal t0, foo" were erroneously creating symbol table
entries for t0 as well as foo, which causes linking problems. Fix by
reordering instruction alternatives so that t0 is first attempted to
be parsed as a register, rather than as a symbol.
* riscv-opc.c (riscv_opcodes): Reorder jal and call entries.
Limit MIPS16 instruction disassembly according to the ISA level and ASE
set selected, as with the regular MIPS and microMIPS instruction sets.
Retain the property of `objdump -m mips:16' disassembling all MIPS16
instructions however, regardless of any ISA level recorded in the binary
examined.
To validate the disassembler use the GAS test suite for its convenience
of running tests across multiple ISAs, even though placing the tests in
the binutils test suite would be more appropriate. Adjust the single
binutils test which depends on 64-bit instruction disassembly to have
the ISA level required actually recorded in the binary examined.
opcodes/
* mips-dis.c (mips_arch_choices): Use ISA_MIPS64 rather than
ISA_MIPS3 as the `isa' selection in the `bfd_mach_mips16' entry.
(print_insn_mips16): Check opcode entries for validity against
the ISA level and ASE set selected.
binutils/
* testsuite/binutils-all/mips/mips16-undecoded.s: Use `.module'
rather than `.set' to set the ISA level.
gas/
* testsuite/gas/mips/mips16-sub.d: New test.
* testsuite/gas/mips/mips16-32@mips16-sub.d: New test.
* testsuite/gas/mips/mips16e-32@mips16-sub.d: New test.
* testsuite/gas/mips/mips16e-sub.d: New test.
* testsuite/gas/mips/mips16-32@mips16e-sub.d: New test.
* testsuite/gas/mips/mips16-64@mips16e-sub.d: New test.
* testsuite/gas/mips/mips16e-64-sub.d: New test.
* testsuite/gas/mips/mips16-32@mips16e-64-sub.d: New test.
* testsuite/gas/mips/mips16-64@mips16e-64-sub.d: New test.
* testsuite/gas/mips/mips16e-32@mips16e-64-sub.d: New test.
* testsuite/gas/mips/mips16-sub.s: New test source.
* testsuite/gas/mips/mips16e-sub.s: New test source.
* testsuite/gas/mips/mips16e-64-sub.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new tests.
