1999-05-03 15:29:11 +08:00
|
|
|
/* Instruction printing code for the ARC.
|
2020-01-01 15:57:01 +08:00
|
|
|
Copyright (C) 1994-2020 Free Software Foundation, Inc.
|
2015-10-07 21:20:19 +08:00
|
|
|
|
|
|
|
Contributed by Claudiu Zissulescu (claziss@synopsys.com)
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2007-07-05 17:49:03 +08:00
|
|
|
This file is part of libopcodes.
|
|
|
|
|
|
|
|
This library is free software; you can redistribute it and/or modify
|
2001-01-12 05:20:20 +08:00
|
|
|
it under the terms of the GNU General Public License as published by
|
2007-07-05 17:49:03 +08:00
|
|
|
the Free Software Foundation; either version 3, or (at your option)
|
|
|
|
any later version.
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2007-07-05 17:49:03 +08:00
|
|
|
It is distributed in the hope that it will be useful, but WITHOUT
|
|
|
|
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
|
|
|
|
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public
|
|
|
|
License for more details.
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2001-01-12 05:20:20 +08:00
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
along with this program; if not, write to the Free Software
|
2005-07-01 19:16:33 +08:00
|
|
|
Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
|
|
|
|
MA 02110-1301, USA. */
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2012-05-18 09:59:38 +08:00
|
|
|
#include "sysdep.h"
|
2015-10-07 21:20:19 +08:00
|
|
|
#include <stdio.h>
|
|
|
|
#include <assert.h>
|
1999-05-03 15:29:11 +08:00
|
|
|
#include "dis-asm.h"
|
|
|
|
#include "opcode/arc.h"
|
2016-07-21 00:08:07 +08:00
|
|
|
#include "elf/arc.h"
|
2001-01-12 05:20:20 +08:00
|
|
|
#include "arc-dis.h"
|
|
|
|
#include "arc-ext.h"
|
2016-07-21 00:08:07 +08:00
|
|
|
#include "elf-bfd.h"
|
|
|
|
#include "libiberty.h"
|
|
|
|
#include "opintl.h"
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
/* Structure used to iterate over, and extract the values for, operands of
|
|
|
|
an opcode. */
|
|
|
|
|
|
|
|
struct arc_operand_iterator
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
/* The complete instruction value to extract operands from. */
|
|
|
|
unsigned long long insn;
|
2016-06-02 21:03:23 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
/* The LIMM if this is being tracked separately. This field is only
|
|
|
|
valid if we find the LIMM operand in the operand list. */
|
|
|
|
unsigned limm;
|
2016-06-02 21:03:23 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
/* The opcode this iterator is operating on. */
|
|
|
|
const struct arc_opcode *opcode;
|
2016-06-02 21:03:23 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
/* The index into the opcodes operand index list. */
|
|
|
|
const unsigned char *opidx;
|
2016-06-02 21:03:23 +08:00
|
|
|
};
|
1999-05-03 15:29:11 +08:00
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
/* A private data used by ARC decoder. */
|
|
|
|
struct arc_disassemble_info
|
|
|
|
{
|
|
|
|
/* The current disassembled arc opcode. */
|
|
|
|
const struct arc_opcode *opcode;
|
|
|
|
|
|
|
|
/* Instruction length w/o limm field. */
|
|
|
|
unsigned insn_len;
|
|
|
|
|
|
|
|
/* TRUE if we have limm. */
|
|
|
|
bfd_boolean limm_p;
|
|
|
|
|
|
|
|
/* LIMM value, if exists. */
|
|
|
|
unsigned limm;
|
|
|
|
|
|
|
|
/* Condition code, if exists. */
|
|
|
|
unsigned condition_code;
|
|
|
|
|
|
|
|
/* Writeback mode. */
|
|
|
|
unsigned writeback_mode;
|
|
|
|
|
|
|
|
/* Number of operands. */
|
|
|
|
unsigned operands_count;
|
|
|
|
|
|
|
|
struct arc_insn_operand operands[MAX_INSN_ARGS];
|
|
|
|
};
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Globals variables. */
|
2005-03-03 23:42:05 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
static const char * const regnames[64] =
|
2005-07-01 19:16:33 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
"r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7",
|
|
|
|
"r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15",
|
|
|
|
"r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
|
|
|
|
"r24", "r25", "gp", "fp", "sp", "ilink", "r30", "blink",
|
|
|
|
|
|
|
|
"r32", "r33", "r34", "r35", "r36", "r37", "r38", "r39",
|
|
|
|
"r40", "r41", "r42", "r43", "r44", "r45", "r46", "r47",
|
|
|
|
"r48", "r49", "r50", "r51", "r52", "r53", "r54", "r55",
|
2020-01-13 16:21:30 +08:00
|
|
|
"r56", "r57", "r58", "r59", "lp_count", "reserved", "LIMM", "pcl"
|
2015-10-07 21:20:19 +08:00
|
|
|
};
|
|
|
|
|
2016-07-27 22:57:18 +08:00
|
|
|
static const char * const addrtypenames[ARC_NUM_ADDRTYPES] =
|
|
|
|
{
|
|
|
|
"bd", "jid", "lbd", "mbd", "sd", "sm", "xa", "xd",
|
|
|
|
"cd", "cbd", "cjid", "clbd", "cm", "csd", "cxa", "cxd"
|
|
|
|
};
|
|
|
|
|
|
|
|
static int addrtypenames_max = ARC_NUM_ADDRTYPES - 1;
|
|
|
|
|
|
|
|
static const char * const addrtypeunknown = "unknown";
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* This structure keeps track which instruction class(es)
|
|
|
|
should be ignored durring disassembling. */
|
|
|
|
|
|
|
|
typedef struct skipclass
|
|
|
|
{
|
|
|
|
insn_class_t insn_class;
|
|
|
|
insn_subclass_t subclass;
|
|
|
|
struct skipclass *nxt;
|
|
|
|
} skipclass_t, *linkclass;
|
|
|
|
|
|
|
|
/* Intial classes of instructions to be consider first when
|
|
|
|
disassembling. */
|
|
|
|
static linkclass decodelist = NULL;
|
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
/* ISA mask value enforced via disassembler info options. ARC_OPCODE_NONE
|
|
|
|
value means that no CPU is enforced. */
|
|
|
|
|
|
|
|
static unsigned enforced_isa_mask = ARC_OPCODE_NONE;
|
|
|
|
|
2017-11-03 22:36:42 +08:00
|
|
|
/* True if we want to print using only hex numbers. */
|
|
|
|
static bfd_boolean print_hex = FALSE;
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Macros section. */
|
|
|
|
|
|
|
|
#ifdef DEBUG
|
|
|
|
# define pr_debug(fmt, args...) fprintf (stderr, fmt, ##args)
|
|
|
|
#else
|
|
|
|
# define pr_debug(fmt, args...)
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define ARRANGE_ENDIAN(info, buf) \
|
|
|
|
(info->endian == BFD_ENDIAN_LITTLE ? bfd_getm32 (bfd_getl32 (buf)) \
|
|
|
|
: bfd_getb32 (buf))
|
|
|
|
|
2019-12-11 14:15:14 +08:00
|
|
|
#define BITS(word,s,e) (((word) >> (s)) & ((1ull << ((e) - (s)) << 1) - 1))
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
#define OPCODE_32BIT_INSN(word) (BITS ((word), 27, 31))
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Functions implementation. */
|
1999-05-03 15:29:11 +08:00
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
/* Initialize private data. */
|
|
|
|
static bfd_boolean
|
|
|
|
init_arc_disasm_info (struct disassemble_info *info)
|
|
|
|
{
|
|
|
|
struct arc_disassemble_info *arc_infop
|
|
|
|
= calloc (sizeof (*arc_infop), 1);
|
|
|
|
|
|
|
|
if (arc_infop == NULL)
|
|
|
|
return FALSE;
|
|
|
|
|
|
|
|
info->private_data = arc_infop;
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* Add a new element to the decode list. */
|
|
|
|
|
|
|
|
static void
|
|
|
|
add_to_decodelist (insn_class_t insn_class,
|
|
|
|
insn_subclass_t subclass)
|
|
|
|
{
|
|
|
|
linkclass t = (linkclass) xmalloc (sizeof (skipclass_t));
|
|
|
|
|
|
|
|
t->insn_class = insn_class;
|
|
|
|
t->subclass = subclass;
|
|
|
|
t->nxt = decodelist;
|
|
|
|
decodelist = t;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Return TRUE if we need to skip the opcode from being
|
|
|
|
disassembled. */
|
|
|
|
|
|
|
|
static bfd_boolean
|
2016-10-06 18:05:27 +08:00
|
|
|
skip_this_opcode (const struct arc_opcode *opcode)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
|
|
|
linkclass t = decodelist;
|
|
|
|
|
|
|
|
/* Check opcode for major 0x06, return if it is not in. */
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (arc_opcode_len (opcode) == 4
|
2019-07-24 21:52:23 +08:00
|
|
|
&& (OPCODE_32BIT_INSN (opcode->opcode) != 0x06
|
|
|
|
/* Can be an APEX extensions. */
|
|
|
|
&& OPCODE_32BIT_INSN (opcode->opcode) != 0x07))
|
2016-07-21 00:08:07 +08:00
|
|
|
return FALSE;
|
|
|
|
|
2016-10-06 18:05:27 +08:00
|
|
|
/* or not a known truble class. */
|
|
|
|
switch (opcode->insn_class)
|
|
|
|
{
|
|
|
|
case FLOAT:
|
|
|
|
case DSP:
|
[ARC] Sync opcode data base.
New EM and HS variants are developed, sync the data base to match them.
opcodes/
2017-11-03 Claudiu Zissulescu <claziss@synopsys.com>
* arc-tbl.h (abss, abssh, adc, adcs, adds, aslacc, asls, aslsacc)
(asrs, asrsr, cbflyhf0r, cbflyhf1r, cmacchfr, cmacchnfr, cmachfr)
(cmachnfr, cmpychfr, cmpychnfr, cmpyhfmr, cmpyhfr, cmpyhnfr, divf)
(dmachbl, dmachbm, dmachf, dmachfr, dmacwhf, dmpyhbl, dmpyhbm)
(dmpyhf, dmpyhfr, dmpyhwf, dmpywhf, dsync, flagacc, getacc, macdf)
(macf, macfr, macwhfl, macwhflr, macwhfm, macwhfmr, macwhkl)
(macwhkul, macwhl, macwhul, mpydf, mpyf, mpyfr, mpywhfl, mpywhflr)
(mpywhfm, mpywhfmr, mpywhkl, mpywhkul, mpywhl, mpywhul, msubdf)
(msubf, msubfr, msubwhfl, msubwhflr, msubwhfm, msubwhfmr, mul64)
(negs, negsh, normacc, qmachf, qmpyh, qmpyhf, rndh, satf, sath)
(sbcs, setacc, sflag, sqrt, sqrtf, subs, swi_s, vabs2h, vabss2h)
(vadd4b, vadds2, vadds2h, vadds4h, vaddsubs, vaddsubs2h)
(vaddsubs4h, valgn2h, vasl2h, vasls2h, vasr2h, vasrs2h, vasrsr2h)
(vext2bhl, vext2bhlf, vext2bhm, vext2bhmf, vlsr2h, vmac2hf)
(vmac2hfr, vmac2hnfr, vmax2h, vmin2h, vmpy2h, vmpy2hf, vmpy2hfr)
(vmpy2hwf, vmsub2hf, vmsub2hfr, vmsub2hnfr, vneg2h, vnegs2h)
(vnorm2h, vpack2hbl, vpack2hblf, vpack2hbm, vpack2hbmf, vpack2hl)
(vpack2hm, vperm, vrep2hl, vrep2hm, vsext2bhl, vsext2bhm, vsub4b)
(vsubadds, vsubadds2h, vsubadds4h, vsubs2, vsubs2h, vsubs4h):
Changed opcodes.
(prealloc, prefetch*): Place them before ld instruction.
* arc-opc.c (skip_this_opcode): Add ARITH class.
2017-11-03 21:38:05 +08:00
|
|
|
case ARITH:
|
2019-07-24 21:52:23 +08:00
|
|
|
case MPY:
|
2016-10-06 18:05:27 +08:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
while (t != NULL)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
|
|
|
if ((t->insn_class == opcode->insn_class)
|
|
|
|
&& (t->subclass == opcode->subclass))
|
2016-10-06 18:05:27 +08:00
|
|
|
return FALSE;
|
2016-07-21 00:08:07 +08:00
|
|
|
t = t->nxt;
|
|
|
|
}
|
|
|
|
|
2016-10-06 18:05:27 +08:00
|
|
|
return TRUE;
|
2016-07-21 00:08:07 +08:00
|
|
|
}
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
static bfd_vma
|
|
|
|
bfd_getm32 (unsigned int data)
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
bfd_vma value = 0;
|
2001-01-12 05:20:20 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
value = ((data & 0xff00) | (data & 0xff)) << 16;
|
|
|
|
value |= ((data & 0xff0000) | (data & 0xff000000)) >> 16;
|
|
|
|
return value;
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
static bfd_boolean
|
2015-10-07 21:20:19 +08:00
|
|
|
special_flag_p (const char *opname,
|
|
|
|
const char *flgname)
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
const struct arc_flag_special *flg_spec;
|
|
|
|
unsigned i, j, flgidx;
|
2001-01-12 05:20:20 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
for (i = 0; i < arc_num_flag_special; i++)
|
1999-05-03 15:29:11 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
flg_spec = &arc_flag_special_cases[i];
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-12-04 18:49:57 +08:00
|
|
|
if (strcmp (opname, flg_spec->name))
|
2015-10-07 21:20:19 +08:00
|
|
|
continue;
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Found potential special case instruction. */
|
|
|
|
for (j=0;; ++j)
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
flgidx = flg_spec->flags[j];
|
|
|
|
if (flgidx == 0)
|
|
|
|
break; /* End of the array. */
|
2001-01-12 05:20:20 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
if (strcmp (flgname, arc_flag_operands[flgidx].name) == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
return TRUE;
|
1999-05-03 15:29:11 +08:00
|
|
|
}
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
2016-07-21 00:08:07 +08:00
|
|
|
return FALSE;
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
/* Find opcode from ARC_TABLE given the instruction described by INSN and
|
|
|
|
INSNLEN. The ISA_MASK restricts the possible matches in ARC_TABLE. */
|
|
|
|
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
static const struct arc_opcode *
|
2016-07-21 00:08:07 +08:00
|
|
|
find_format_from_table (struct disassemble_info *info,
|
|
|
|
const struct arc_opcode *arc_table,
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
unsigned long long insn,
|
2016-07-21 00:08:07 +08:00
|
|
|
unsigned int insn_len,
|
|
|
|
unsigned isa_mask,
|
|
|
|
bfd_boolean *has_limm,
|
|
|
|
bfd_boolean overlaps)
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
{
|
|
|
|
unsigned int i = 0;
|
|
|
|
const struct arc_opcode *opcode = NULL;
|
2016-10-06 18:05:27 +08:00
|
|
|
const struct arc_opcode *t_op = NULL;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
const unsigned char *opidx;
|
|
|
|
const unsigned char *flgidx;
|
2016-10-06 18:05:27 +08:00
|
|
|
bfd_boolean warn_p = FALSE;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
do
|
|
|
|
{
|
|
|
|
bfd_boolean invalid = FALSE;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
opcode = &arc_table[i++];
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (!(opcode->cpu & isa_mask))
|
2016-07-21 00:08:07 +08:00
|
|
|
continue;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (arc_opcode_len (opcode) != (int) insn_len)
|
2016-07-21 00:08:07 +08:00
|
|
|
continue;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if ((insn & opcode->mask) != opcode->opcode)
|
2016-07-21 00:08:07 +08:00
|
|
|
continue;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
*has_limm = FALSE;
|
2016-06-02 21:03:23 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* Possible candidate, check the operands. */
|
|
|
|
for (opidx = opcode->operands; *opidx; opidx++)
|
|
|
|
{
|
2016-10-13 21:01:19 +08:00
|
|
|
int value, limmind;
|
2016-07-21 00:08:07 +08:00
|
|
|
const struct arc_operand *operand = &arc_operands[*opidx];
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
if (operand->flags & ARC_OPERAND_FAKE)
|
|
|
|
continue;
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
if (operand->extract)
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = (*operand->extract) (insn, &invalid);
|
2016-07-21 00:08:07 +08:00
|
|
|
else
|
2019-12-10 16:39:47 +08:00
|
|
|
value = (insn >> operand->shift) & ((1ull << operand->bits) - 1);
|
2016-07-21 00:08:07 +08:00
|
|
|
|
|
|
|
/* Check for LIMM indicator. If it is there, then make sure
|
|
|
|
we pick the right format. */
|
2016-10-13 21:01:19 +08:00
|
|
|
limmind = (isa_mask & ARC_OPCODE_ARCV2) ? 0x1E : 0x3E;
|
2016-07-21 00:08:07 +08:00
|
|
|
if (operand->flags & ARC_OPERAND_IR
|
|
|
|
&& !(operand->flags & ARC_OPERAND_LIMM))
|
|
|
|
{
|
|
|
|
if ((value == 0x3E && insn_len == 4)
|
2016-10-13 21:01:19 +08:00
|
|
|
|| (value == limmind && insn_len == 2))
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
|
|
|
invalid = TRUE;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (operand->flags & ARC_OPERAND_LIMM
|
|
|
|
&& !(operand->flags & ARC_OPERAND_DUPLICATE))
|
|
|
|
*has_limm = TRUE;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Check the flags. */
|
|
|
|
for (flgidx = opcode->flags; *flgidx; flgidx++)
|
|
|
|
{
|
|
|
|
/* Get a valid flag class. */
|
|
|
|
const struct arc_flag_class *cl_flags = &arc_flag_classes[*flgidx];
|
|
|
|
const unsigned *flgopridx;
|
|
|
|
int foundA = 0, foundB = 0;
|
|
|
|
unsigned int value;
|
|
|
|
|
|
|
|
/* Check first the extensions. */
|
|
|
|
if (cl_flags->flag_class & F_CLASS_EXTEND)
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = (insn & 0x1F);
|
2016-07-21 00:08:07 +08:00
|
|
|
if (arcExtMap_condCodeName (value))
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
/* Check for the implicit flags. */
|
|
|
|
if (cl_flags->flag_class & F_CLASS_IMPLICIT)
|
|
|
|
continue;
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
for (flgopridx = cl_flags->flags; *flgopridx; ++flgopridx)
|
|
|
|
{
|
|
|
|
const struct arc_flag_operand *flg_operand =
|
|
|
|
&arc_flag_operands[*flgopridx];
|
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = (insn >> flg_operand->shift)
|
2016-07-21 00:08:07 +08:00
|
|
|
& ((1 << flg_operand->bits) - 1);
|
|
|
|
if (value == flg_operand->code)
|
|
|
|
foundA = 1;
|
|
|
|
if (value)
|
|
|
|
foundB = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!foundA && foundB)
|
|
|
|
{
|
|
|
|
invalid = TRUE;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (invalid)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
if (insn_len == 4
|
2016-10-06 18:05:27 +08:00
|
|
|
&& overlaps)
|
|
|
|
{
|
|
|
|
warn_p = TRUE;
|
|
|
|
t_op = opcode;
|
|
|
|
if (skip_this_opcode (opcode))
|
|
|
|
continue;
|
|
|
|
}
|
2016-07-21 00:08:07 +08:00
|
|
|
|
|
|
|
/* The instruction is valid. */
|
|
|
|
return opcode;
|
|
|
|
}
|
|
|
|
while (opcode->mask);
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
|
2016-10-06 18:05:27 +08:00
|
|
|
if (warn_p)
|
|
|
|
{
|
|
|
|
info->fprintf_func (info->stream,
|
|
|
|
_("\nWarning: disassembly may be wrong due to "
|
|
|
|
"guessed opcode class choice.\n"
|
|
|
|
"Use -M<class[,class]> to select the correct "
|
|
|
|
"opcode class(es).\n\t\t\t\t"));
|
|
|
|
return t_op;
|
|
|
|
}
|
|
|
|
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
/* Find opcode for INSN, trying various different sources. The instruction
|
|
|
|
length in INSN_LEN will be updated if the instruction requires a LIMM
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
extension.
|
2016-06-02 21:03:23 +08:00
|
|
|
|
|
|
|
A pointer to the opcode is placed into OPCODE_RESULT, and ITER is
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
initialised, ready to iterate over the operands of the found opcode. If
|
|
|
|
the found opcode requires a LIMM then the LIMM value will be loaded into a
|
|
|
|
field of ITER.
|
2016-06-02 21:03:23 +08:00
|
|
|
|
|
|
|
This function returns TRUE in almost all cases, FALSE is reserved to
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
indicate an error (failing to find an opcode is not an error) a returned
|
|
|
|
result of FALSE would indicate that the disassembler can't continue.
|
2016-06-02 21:03:23 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
If no matching opcode is found then the returned result will be TRUE, the
|
|
|
|
value placed into OPCODE_RESULT will be NULL, ITER will be undefined, and
|
|
|
|
INSN_LEN will be unchanged.
|
2016-06-02 21:03:23 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
If a matching opcode is found, then the returned result will be TRUE, the
|
|
|
|
opcode pointer is placed into OPCODE_RESULT, INSN_LEN will be increased by
|
|
|
|
4 if the instruction requires a LIMM, and the LIMM value will have been
|
|
|
|
loaded into a field of ITER. Finally, ITER will have been initialised so
|
|
|
|
that calls to OPERAND_ITERATOR_NEXT will iterate over the opcode's
|
|
|
|
operands. */
|
2016-06-02 21:03:23 +08:00
|
|
|
|
|
|
|
static bfd_boolean
|
2016-07-21 00:08:07 +08:00
|
|
|
find_format (bfd_vma memaddr,
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
unsigned long long insn,
|
2016-07-21 00:08:07 +08:00
|
|
|
unsigned int * insn_len,
|
|
|
|
unsigned isa_mask,
|
|
|
|
struct disassemble_info * info,
|
|
|
|
const struct arc_opcode ** opcode_result,
|
|
|
|
struct arc_operand_iterator * iter)
|
2016-06-02 21:03:23 +08:00
|
|
|
{
|
2016-07-21 00:08:07 +08:00
|
|
|
const struct arc_opcode *opcode = NULL;
|
2016-06-02 21:03:23 +08:00
|
|
|
bfd_boolean needs_limm;
|
2016-09-15 18:24:24 +08:00
|
|
|
const extInstruction_t *einsn, *i;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
unsigned limm = 0;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
struct arc_disassemble_info *arc_infop = info->private_data;
|
2016-06-02 21:03:23 +08:00
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* First, try the extension instructions. */
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (*insn_len == 4)
|
2016-06-02 21:03:23 +08:00
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
einsn = arcExtMap_insn (OPCODE_32BIT_INSN (insn), insn);
|
|
|
|
for (i = einsn; (i != NULL) && (opcode == NULL); i = i->next)
|
2016-06-02 21:03:23 +08:00
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
const char *errmsg = NULL;
|
|
|
|
|
|
|
|
opcode = arcExtMap_genOpcode (i, isa_mask, &errmsg);
|
|
|
|
if (opcode == NULL)
|
|
|
|
{
|
|
|
|
(*info->fprintf_func) (info->stream, "\
|
2016-07-21 00:08:07 +08:00
|
|
|
An error occured while generating the extension instruction operations");
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
*opcode_result = NULL;
|
|
|
|
return FALSE;
|
|
|
|
}
|
2016-07-21 00:08:07 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
opcode = find_format_from_table (info, opcode, insn, *insn_len,
|
|
|
|
isa_mask, &needs_limm, FALSE);
|
|
|
|
}
|
2016-06-02 21:03:23 +08:00
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* Then, try finding the first match in the opcode table. */
|
|
|
|
if (opcode == NULL)
|
|
|
|
opcode = find_format_from_table (info, arc_opcodes, insn, *insn_len,
|
|
|
|
isa_mask, &needs_limm, TRUE);
|
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
if (needs_limm && opcode != NULL)
|
|
|
|
{
|
|
|
|
bfd_byte buffer[4];
|
|
|
|
int status;
|
|
|
|
|
|
|
|
status = (*info->read_memory_func) (memaddr + *insn_len, buffer,
|
|
|
|
4, info);
|
|
|
|
if (status != 0)
|
|
|
|
{
|
|
|
|
opcode = NULL;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
limm = ARRANGE_ENDIAN (info, buffer);
|
2016-06-02 21:03:23 +08:00
|
|
|
*insn_len += 4;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (opcode != NULL)
|
2016-06-02 21:03:23 +08:00
|
|
|
{
|
|
|
|
iter->insn = insn;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
iter->limm = limm;
|
|
|
|
iter->opcode = opcode;
|
|
|
|
iter->opidx = opcode->operands;
|
2016-06-02 21:03:23 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
*opcode_result = opcode;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
|
|
|
|
/* Update private data. */
|
|
|
|
arc_infop->opcode = opcode;
|
|
|
|
arc_infop->limm = (needs_limm) ? limm : 0;
|
|
|
|
arc_infop->limm_p = needs_limm;
|
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2016-04-06 22:08:04 +08:00
|
|
|
static void
|
|
|
|
print_flags (const struct arc_opcode *opcode,
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
unsigned long long *insn,
|
2016-04-06 22:08:04 +08:00
|
|
|
struct disassemble_info *info)
|
|
|
|
{
|
|
|
|
const unsigned char *flgidx;
|
|
|
|
unsigned int value;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
struct arc_disassemble_info *arc_infop = info->private_data;
|
2016-04-06 22:08:04 +08:00
|
|
|
|
|
|
|
/* Now extract and print the flags. */
|
|
|
|
for (flgidx = opcode->flags; *flgidx; flgidx++)
|
|
|
|
{
|
|
|
|
/* Get a valid flag class. */
|
|
|
|
const struct arc_flag_class *cl_flags = &arc_flag_classes[*flgidx];
|
|
|
|
const unsigned *flgopridx;
|
|
|
|
|
|
|
|
/* Check first the extensions. */
|
2016-05-23 23:25:46 +08:00
|
|
|
if (cl_flags->flag_class & F_CLASS_EXTEND)
|
2016-04-06 22:08:04 +08:00
|
|
|
{
|
|
|
|
const char *name;
|
|
|
|
value = (insn[0] & 0x1F);
|
|
|
|
|
|
|
|
name = arcExtMap_condCodeName (value);
|
|
|
|
if (name)
|
|
|
|
{
|
|
|
|
(*info->fprintf_func) (info->stream, ".%s", name);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (flgopridx = cl_flags->flags; *flgopridx; ++flgopridx)
|
|
|
|
{
|
|
|
|
const struct arc_flag_operand *flg_operand =
|
|
|
|
&arc_flag_operands[*flgopridx];
|
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
/* Implicit flags are only used for the insn decoder. */
|
|
|
|
if (cl_flags->flag_class & F_CLASS_IMPLICIT)
|
|
|
|
{
|
|
|
|
if (cl_flags->flag_class & F_CLASS_COND)
|
|
|
|
arc_infop->condition_code = flg_operand->code;
|
|
|
|
else if (cl_flags->flag_class & F_CLASS_WB)
|
|
|
|
arc_infop->writeback_mode = flg_operand->code;
|
|
|
|
else if (cl_flags->flag_class & F_CLASS_ZZ)
|
|
|
|
info->data_size = flg_operand->code;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2016-04-06 22:08:04 +08:00
|
|
|
if (!flg_operand->favail)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
value = (insn[0] >> flg_operand->shift)
|
|
|
|
& ((1 << flg_operand->bits) - 1);
|
|
|
|
if (value == flg_operand->code)
|
|
|
|
{
|
|
|
|
/* FIXME!: print correctly nt/t flag. */
|
|
|
|
if (!special_flag_p (opcode->name, flg_operand->name))
|
|
|
|
(*info->fprintf_func) (info->stream, ".");
|
|
|
|
else if (info->insn_type == dis_dref)
|
|
|
|
{
|
|
|
|
switch (flg_operand->name[0])
|
|
|
|
{
|
|
|
|
case 'b':
|
|
|
|
info->data_size = 1;
|
|
|
|
break;
|
|
|
|
case 'h':
|
|
|
|
case 'w':
|
|
|
|
info->data_size = 2;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
info->data_size = 4;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
[ARC] Update instruction type and delay slot info.
This patch corrects the instructioninformation passed into the
disassebler_info structure.
include/
2016-05-23 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (insn_subclass_t): Add COND.
(flag_class_t): Add F_CLASS_EXTEND.
opcodes/
2016-05-23 Claudiu Zissulescu <claziss@synopsys.com>
* arc-dis.c (print_flags): Set branch_delay_insns, and insn_type
information.
(print_insn_arc): Set insn_type information.
* arc-opc.c (C_CC): Add F_CLASS_COND.
* arc-tbl.h (bbit0, bbit1): Update subclass to COND.
(beq_s, bge_s, bgt_s, bhi_s, bhs_s): Likewise.
(ble_s, blo_s, bls_s, blt_s, bne_s): Likewise.
(breq, breq_s, brge, brhs, brlo, brlt): Likewise.
(brne, brne_s, jeq_s, jne_s): Likewise.
2016-05-19 18:19:32 +08:00
|
|
|
if (flg_operand->name[0] == 'd'
|
|
|
|
&& flg_operand->name[1] == 0)
|
|
|
|
info->branch_delay_insns = 1;
|
|
|
|
|
|
|
|
/* Check if it is a conditional flag. */
|
|
|
|
if (cl_flags->flag_class & F_CLASS_COND)
|
|
|
|
{
|
|
|
|
if (info->insn_type == dis_jsr)
|
|
|
|
info->insn_type = dis_condjsr;
|
|
|
|
else if (info->insn_type == dis_branch)
|
|
|
|
info->insn_type = dis_condbranch;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
arc_infop->condition_code = flg_operand->code;
|
[ARC] Update instruction type and delay slot info.
This patch corrects the instructioninformation passed into the
disassebler_info structure.
include/
2016-05-23 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (insn_subclass_t): Add COND.
(flag_class_t): Add F_CLASS_EXTEND.
opcodes/
2016-05-23 Claudiu Zissulescu <claziss@synopsys.com>
* arc-dis.c (print_flags): Set branch_delay_insns, and insn_type
information.
(print_insn_arc): Set insn_type information.
* arc-opc.c (C_CC): Add F_CLASS_COND.
* arc-tbl.h (bbit0, bbit1): Update subclass to COND.
(beq_s, bge_s, bgt_s, bhi_s, bhs_s): Likewise.
(ble_s, blo_s, bls_s, blt_s, bne_s): Likewise.
(breq, breq_s, brge, brhs, brlo, brlt): Likewise.
(brne, brne_s, jeq_s, jne_s): Likewise.
2016-05-19 18:19:32 +08:00
|
|
|
}
|
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
/* Check for the write back modes. */
|
|
|
|
if (cl_flags->flag_class & F_CLASS_WB)
|
|
|
|
arc_infop->writeback_mode = flg_operand->code;
|
|
|
|
|
2016-04-06 22:08:04 +08:00
|
|
|
(*info->fprintf_func) (info->stream, "%s", flg_operand->name);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static const char *
|
|
|
|
get_auxreg (const struct arc_opcode *opcode,
|
|
|
|
int value,
|
|
|
|
unsigned isa_mask)
|
|
|
|
{
|
|
|
|
const char *name;
|
|
|
|
unsigned int i;
|
|
|
|
const struct arc_aux_reg *auxr = &arc_aux_regs[0];
|
|
|
|
|
2016-05-23 23:25:46 +08:00
|
|
|
if (opcode->insn_class != AUXREG)
|
2016-04-06 22:08:04 +08:00
|
|
|
return NULL;
|
|
|
|
|
|
|
|
name = arcExtMap_auxRegName (value);
|
|
|
|
if (name)
|
|
|
|
return name;
|
|
|
|
|
|
|
|
for (i = 0; i < arc_num_aux_regs; i++, auxr++)
|
|
|
|
{
|
|
|
|
if (!(auxr->cpu & isa_mask))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
if (auxr->subclass != NONE)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
if (auxr->address == value)
|
|
|
|
return auxr->name;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
2016-03-25 00:54:37 +08:00
|
|
|
|
2016-07-27 22:57:18 +08:00
|
|
|
/* Convert a value representing an address type to a string used to refer to
|
|
|
|
the address type in assembly code. */
|
|
|
|
|
|
|
|
static const char *
|
|
|
|
get_addrtype (int value)
|
|
|
|
{
|
|
|
|
if (value < 0 || value > addrtypenames_max)
|
|
|
|
return addrtypeunknown;
|
|
|
|
|
|
|
|
return addrtypenames[value];
|
|
|
|
}
|
|
|
|
|
2016-03-25 00:54:37 +08:00
|
|
|
/* Calculate the instruction length for an instruction starting with MSB
|
|
|
|
and LSB, the most and least significant byte. The ISA_MASK is used to
|
|
|
|
filter the instructions considered to only those that are part of the
|
|
|
|
current architecture.
|
|
|
|
|
|
|
|
The instruction lengths are calculated from the ARC_OPCODE table, and
|
|
|
|
cached for later use. */
|
|
|
|
|
|
|
|
static unsigned int
|
2016-06-02 21:03:23 +08:00
|
|
|
arc_insn_length (bfd_byte msb, bfd_byte lsb, struct disassemble_info *info)
|
2016-03-25 00:54:37 +08:00
|
|
|
{
|
|
|
|
bfd_byte major_opcode = msb >> 3;
|
|
|
|
|
|
|
|
switch (info->mach)
|
|
|
|
{
|
2016-06-21 21:03:08 +08:00
|
|
|
case bfd_mach_arc_arc700:
|
|
|
|
/* The nps400 extension set requires this special casing of the
|
|
|
|
instruction length calculation. Right now this is not causing any
|
|
|
|
problems as none of the known extensions overlap in opcode space,
|
|
|
|
but, if they ever do then we might need to start carrying
|
|
|
|
information around in the elf about which extensions are in use. */
|
2016-06-02 21:03:23 +08:00
|
|
|
if (major_opcode == 0xb)
|
|
|
|
{
|
|
|
|
bfd_byte minor_opcode = lsb & 0x1f;
|
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (minor_opcode < 4)
|
|
|
|
return 6;
|
|
|
|
else if (minor_opcode == 0x10 || minor_opcode == 0x11)
|
|
|
|
return 8;
|
2016-06-02 21:03:23 +08:00
|
|
|
}
|
2016-10-19 03:10:25 +08:00
|
|
|
if (major_opcode == 0xa)
|
|
|
|
{
|
|
|
|
return 8;
|
|
|
|
}
|
2016-10-05 15:47:02 +08:00
|
|
|
/* Fall through. */
|
2016-03-25 00:54:37 +08:00
|
|
|
case bfd_mach_arc_arc600:
|
|
|
|
return (major_opcode > 0xb) ? 2 : 4;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case bfd_mach_arc_arcv2:
|
|
|
|
return (major_opcode > 0x7) ? 2 : 4;
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
2019-08-08 00:22:29 +08:00
|
|
|
return 0;
|
2016-03-25 00:54:37 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
/* Extract and return the value of OPERAND from the instruction whose value
|
|
|
|
is held in the array INSN. */
|
|
|
|
|
|
|
|
static int
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
extract_operand_value (const struct arc_operand *operand,
|
|
|
|
unsigned long long insn,
|
|
|
|
unsigned limm)
|
2016-06-02 21:03:23 +08:00
|
|
|
{
|
|
|
|
int value;
|
|
|
|
|
|
|
|
/* Read the limm operand, if required. */
|
|
|
|
if (operand->flags & ARC_OPERAND_LIMM)
|
|
|
|
/* The second part of the instruction value will have been loaded as
|
|
|
|
part of the find_format call made earlier. */
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = limm;
|
2016-06-02 21:03:23 +08:00
|
|
|
else
|
|
|
|
{
|
|
|
|
if (operand->extract)
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = (*operand->extract) (insn, (int *) NULL);
|
2016-06-02 21:03:23 +08:00
|
|
|
else
|
|
|
|
{
|
|
|
|
if (operand->flags & ARC_OPERAND_ALIGNED32)
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = (insn >> operand->shift)
|
2016-06-02 21:03:23 +08:00
|
|
|
& ((1 << (operand->bits - 2)) - 1);
|
|
|
|
value = value << 2;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
value = (insn >> operand->shift) & ((1 << operand->bits) - 1);
|
2016-06-02 21:03:23 +08:00
|
|
|
}
|
|
|
|
if (operand->flags & ARC_OPERAND_SIGNED)
|
|
|
|
{
|
|
|
|
int signbit = 1 << (operand->bits - 1);
|
|
|
|
value = (value ^ signbit) - signbit;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return value;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Find the next operand, and the operands value from ITER. Return TRUE if
|
|
|
|
there is another operand, otherwise return FALSE. If there is an
|
|
|
|
operand returned then the operand is placed into OPERAND, and the value
|
|
|
|
into VALUE. If there is no operand returned then OPERAND and VALUE are
|
|
|
|
unchanged. */
|
|
|
|
|
|
|
|
static bfd_boolean
|
|
|
|
operand_iterator_next (struct arc_operand_iterator *iter,
|
|
|
|
const struct arc_operand **operand,
|
|
|
|
int *value)
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (*iter->opidx == 0)
|
2016-06-02 21:03:23 +08:00
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
*operand = NULL;
|
|
|
|
return FALSE;
|
2016-06-02 21:03:23 +08:00
|
|
|
}
|
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
*operand = &arc_operands[*iter->opidx];
|
|
|
|
*value = extract_operand_value (*operand, iter->insn, iter->limm);
|
|
|
|
iter->opidx++;
|
2016-06-02 21:03:23 +08:00
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* Helper for parsing the options. */
|
|
|
|
|
|
|
|
static void
|
2017-04-06 02:21:33 +08:00
|
|
|
parse_option (const char *option)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
if (disassembler_options_cmp (option, "dsp") == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
add_to_decodelist (DSP, NONE);
|
|
|
|
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
else if (disassembler_options_cmp (option, "spfp") == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
add_to_decodelist (FLOAT, SPX);
|
|
|
|
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
else if (disassembler_options_cmp (option, "dpfp") == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
add_to_decodelist (FLOAT, DPX);
|
|
|
|
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
else if (disassembler_options_cmp (option, "quarkse_em") == 0)
|
2016-10-06 18:05:27 +08:00
|
|
|
{
|
|
|
|
add_to_decodelist (FLOAT, DPX);
|
|
|
|
add_to_decodelist (FLOAT, SPX);
|
[ARC] Object attributes.
gas/
2017-05-10 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/attr-arc600.d: New file.
* testsuite/gas/arc/attr-arc600_mul32x16.d: Likewise.
* testsuite/gas/arc/attr-arc600_norm.d: Likewise.
* testsuite/gas/arc/attr-arc601.d: Likewise.
* testsuite/gas/arc/attr-arc601_mul32x16.d: Likewise.
* testsuite/gas/arc/attr-arc601_mul64.d: Likewise.
* testsuite/gas/arc/attr-arc601_norm.d: Likewise.
* testsuite/gas/arc/attr-arc700.d: Likewise.
* testsuite/gas/arc/attr-arcem.d: Likewise.
* testsuite/gas/arc/attr-archs.d: Likewise.
* testsuite/gas/arc/attr-autodetect-1.d: Likewise.
* testsuite/gas/arc/attr-autodetect-1.s: Likewise.
* testsuite/gas/arc/attr-cpu-a601.d: Likewise.
* testsuite/gas/arc/attr-cpu-a601.s: Likewise.
* testsuite/gas/arc/attr-cpu-a700.d: Likewise.
* testsuite/gas/arc/attr-cpu-a700.s: Likewise.
* testsuite/gas/arc/attr-cpu-em.d: Likewise.
* testsuite/gas/arc/attr-cpu-em.s: Likewise.
* testsuite/gas/arc/attr-cpu-hs.d: Likewise.
* testsuite/gas/arc/attr-cpu-hs.s: Likewise.
* testsuite/gas/arc/attr-em.d: Likewise.
* testsuite/gas/arc/attr-em4.d: Likewise.
* testsuite/gas/arc/attr-em4_dmips.d: Likewise.
* testsuite/gas/arc/attr-em4_fpuda.d: Likewise.
* testsuite/gas/arc/attr-em4_fpus.d: Likewise.
* testsuite/gas/arc/attr-hs.d: Likewise.
* testsuite/gas/arc/attr-hs34.d: Likewise.
* testsuite/gas/arc/attr-hs38.d: Likewise.
* testsuite/gas/arc/attr-hs38_linux.d: Likewise.
* testsuite/gas/arc/attr-mul64.d: Likewise.
* testsuite/gas/arc/attr-name.d: Likewise.
* testsuite/gas/arc/attr-name.s: Likewise.
* testsuite/gas/arc/attr-nps400.d: Likewise.
* testsuite/gas/arc/attr-override-mcpu.d: Likewise.
* testsuite/gas/arc/attr-override-mcpu.s
* testsuite/gas/arc/attr-quarkse_em.d: Likewise.
* testsuite/gas/arc/blank.s: Likewise.
* testsuite/gas/elf/section2.e-arc: Likewise.
* testsuite/gas/arc/cpu-pseudop-1.d: Update test.
* testsuite/gas/arc/cpu-pseudop-2.d: Likewise.
* testsuite/gas/arc/nps400-0.d: Likewise.
* testsuite/gas/elf/elf.exp: Set target_machine for ARC.
* config/tc-arc.c (opcode/arc-attrs.h): Include.
(ARC_GET_FLAG, ARC_SET_FLAG, streq): Define.
(arc_attribute): Declare new function.
(md_pseudo_table): Add arc_attribute.
(cpu_types): Rename default cpu features.
(selected_cpu): Set the default OSABI flag.
(mpy_option): New variable.
(pic_option): Likewise.
(sda_option): Likewise.
(tls_option): Likewise.
(feature_type, feature_list): Remove.
(arc_initial_eflag): Likewise.
(attributes_set_explicitly): New variable.
(arc_check_feature): Check also for the conflicting features.
(arc_select_cpu): Refactor assignment of selected_cpu.eflags.
(arc_option): Remove setting of private flags and architecture.
(check_cpu_feature): Refactor feature names.
(autodetect_attributes): New function.
(assemble_tokens): Use above function.
(md_parse_option): Refactor feature names.
(arc_attribute): New function.
(arc_set_attribute_int): Likewise.
(arc_set_attribute_string): Likewise.
(arc_stralloc): Likewise.
(arc_set_public_attributes): Likewise.
(arc_md_end): Likewise.
(arc_copy_symbol_attributes): Likewise.
(rc_convert_symbolic_attribute): Likewise.
* config/tc-arc.h (md_end): Define.
(CONVERT_SYMBOLIC_ATTRIBUTE): Likewise.
(TC_COPY_SYMBOL_ATTRIBUTES): Likewise.
* doc/c-arc.texi: Document ARC object attributes.
binutils/
2017-05-10 Claudiu Zissulescu <claziss@synopsys.com>
* readelf.c (decode_ARC_machine_flags): Recognize OSABI v4.
(get_arc_section_type_name): New function.
(get_section_type_name): Use the above function.
(display_arc_attribute): New function.
(process_arc_specific): Likewise.
(process_arch_specific): Handle ARC specific information.
* testsuite/binutils-all/strip-3.d: Consider ARC.attributes
section.
include/
2017-05-10 Claudiu Zissulescu <claziss@synopsys.com>
* elf/arc.h (SHT_ARC_ATTRIBUTES): Define.
(Tag_ARC_*): Define.
(E_ARC_OSABI_V4): Define.
(E_ARC_OSABI_CURRENT): Reassign it.
(TAG_CPU_*): Define.
* opcode/arc-attrs.h: New file.
* opcode/arc.h (insn_subclass_t): Assign enum values.
(insn_subclass_t): Update enum with QUARKSE1, QUARKSE2, and LL64.
(ARC_EA, ARC_CD, ARC_LLOCK, ARC_ATOMIC, ARC_MPY, ARC_MULT)
(ARC_NPS400, ARC_DPFP, ARC_SPFP, ARC_FPU, ARC_FPUDA, ARC_SWAP)
(ARC_NORM, ARC_BSCAN, ARC_UIX, ARC_TSTAMP, ARC_VBFDW)
(ARC_BARREL, ARC_DSPA, ARC_SHIFT, ARC_INTR, ARC_DIV, ARC_XMAC)
(ARC_CRC): Delete.
bfd/
2017-05-10 Claudiu Zissulescu <claziss@synopsys.com>
* elf32-arc.c (FEATURE_LIST_NAME): Define.
(CONFLICT_LIST): Likewise.
(opcode/arc-attrs.h): Include.
(arc_elf_print_private_bfd_data): Print OSABI v4 flag.
(arc_extract_features): New file.
(arc_stralloc): Likewise.
(arc_elf_merge_attributes): Likewise.
(arc_elf_merge_private_bfd_data): Use object attributes.
(bfd_arc_get_mach_from_attributes): New function.
(arc_elf_object_p): Use object attributes.
(arc_elf_final_write_processing): Likewise.
(elf32_arc_obj_attrs_arg_type): New function.
(elf32_arc_obj_attrs_handle_unknown): Likewise.
(elf32_arc_section_from_shdr): Likewise.
(elf_backend_obj_attrs_vendor): Define.
(elf_backend_obj_attrs_section): Likewise.
(elf_backend_obj_attrs_arg_type): Likewise.
(elf_backend_obj_attrs_section_type): Likewise.
(elf_backend_obj_attrs_handle_unknown): Likewise.
(elf_backend_section_from_shdr): Likewise.
ld/
2017-05-10 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/ld-arc/attr-merge-0.d: New file.
* testsuite/ld-arc/attr-merge-0.s: Likewise.
* testsuite/ld-arc/attr-merge-0e.s: Likewise.
* testsuite/ld-arc/attr-merge-1.d: Likewise.
* testsuite/ld-arc/attr-merge-1.s: Likewise.
* testsuite/ld-arc/attr-merge-1e.s: Likewise.
* testsuite/ld-arc/attr-merge-2.d: Likewise.
* testsuite/ld-arc/attr-merge-2.s: Likewise.
* testsuite/ld-arc/attr-merge-3.d: Likewise.
* testsuite/ld-arc/attr-merge-3.s: Likewise.
* testsuite/ld-arc/attr-merge-3e.s: Likewise.
* testsuite/ld-arc/attr-merge-4.s: Likewise.
* testsuite/ld-arc/attr-merge-5.d: Likewise.
* testsuite/ld-arc/attr-merge-5a.s: Likewise.
* testsuite/ld-arc/attr-merge-5b.s: Likewise.
* testsuite/ld-arc/attr-merge-conflict-isa.d: Likewise.
* testsuite/ld-arc/attr-merge-err-isa.d: Likewise.
* testsuite/ld-arc/attr-merge-incompatible-cpu.d: Likewise.
* testsuite/ld-arc/got-01.d: Update test.
* testsuite/ld-arc/attr-merge-err-quarkse.d: New file.
* testsuite/ld-arc/attr-quarkse.s: Likewise.
* testsuite/ld-arc/attr-quarkse2.s: Likewise.
opcodes/
2017-05-10 Claudiu Zissulescu <claziss@synopsys.com>
* arc-dis.c (parse_option): Update quarkse_em option..
* arc-ext-tbl.h (dsp_fp_flt2i, dsp_fp_i2flt): Change subclass to
QUARKSE1.
(dsp_fp_div, dsp_fp_cmp): Change subclass to QUARKSE2.
2017-05-10 20:42:22 +08:00
|
|
|
add_to_decodelist (FLOAT, QUARKSE1);
|
|
|
|
add_to_decodelist (FLOAT, QUARKSE2);
|
2016-10-06 18:05:27 +08:00
|
|
|
}
|
2016-07-21 00:08:07 +08:00
|
|
|
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
else if (disassembler_options_cmp (option, "fpuda") == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
add_to_decodelist (FLOAT, DPA);
|
|
|
|
|
2019-07-24 21:52:23 +08:00
|
|
|
else if (disassembler_options_cmp (option, "nps400") == 0)
|
|
|
|
{
|
|
|
|
add_to_decodelist (ACL, NPS400);
|
|
|
|
add_to_decodelist (ARITH, NPS400);
|
|
|
|
add_to_decodelist (BITOP, NPS400);
|
|
|
|
add_to_decodelist (BMU, NPS400);
|
|
|
|
add_to_decodelist (CONTROL, NPS400);
|
|
|
|
add_to_decodelist (DMA, NPS400);
|
|
|
|
add_to_decodelist (DPI, NPS400);
|
|
|
|
add_to_decodelist (MEMORY, NPS400);
|
|
|
|
add_to_decodelist (MISC, NPS400);
|
|
|
|
add_to_decodelist (NET, NPS400);
|
|
|
|
add_to_decodelist (PMU, NPS400);
|
|
|
|
add_to_decodelist (PROTOCOL_DECODE, NPS400);
|
|
|
|
add_to_decodelist (ULTRAIP, NPS400);
|
|
|
|
}
|
|
|
|
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
else if (disassembler_options_cmp (option, "fpus") == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
|
|
|
add_to_decodelist (FLOAT, SP);
|
|
|
|
add_to_decodelist (FLOAT, CVT);
|
|
|
|
}
|
|
|
|
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
else if (disassembler_options_cmp (option, "fpud") == 0)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
|
|
|
add_to_decodelist (FLOAT, DP);
|
|
|
|
add_to_decodelist (FLOAT, CVT);
|
|
|
|
}
|
2017-11-03 22:36:42 +08:00
|
|
|
else if (CONST_STRNEQ (option, "hex"))
|
|
|
|
print_hex = TRUE;
|
2016-07-21 00:08:07 +08:00
|
|
|
else
|
opcodes error messages
Another patch aimed at making binutils comply with the GNU coding
standard. The generated files require
https://sourceware.org/ml/cgen/2018-q1/msg00004.html
cpu/
* frv.opc: Include opintl.h.
(add_next_to_vliw): Use opcodes_error_handler to print error.
Standardize error message.
(fr500_check_insn_major_constraints, frv_vliw_add_insn): Likewise.
opcodes/
* sysdep.h (opcodes_error_handler): Define.
(_bfd_error_handler): Declare.
* Makefile.am: Remove stray #.
* opc2c.c (main): Remove bogus -l arg handling. Print "DO NOT
EDIT" comment.
* aarch64-dis.c, * arc-dis.c, * arm-dis.c, * avr-dis.c,
* d30v-dis.c, * h8300-dis.c, * mmix-dis.c, * ppc-dis.c,
* riscv-dis.c, * s390-dis.c, * sparc-dis.c, * v850-dis.c: Use
opcodes_error_handler to print errors. Standardize error messages.
* msp430-decode.opc, * nios2-dis.c, * rl78-decode.opc: Likewise,
and include opintl.h.
* nds32-asm.c: Likewise, and include sysdep.h and opintl.h.
* i386-gen.c: Standardize error messages.
* msp430-decode.c, * rl78-decode.c, rx-decode.c: Regenerate.
* Makefile.in: Regenerate.
* epiphany-asm.c, * epiphany-desc.c, * epiphany-dis.c,
* epiphany-ibld.c, * fr30-asm.c, * fr30-desc.c, * fr30-dis.c,
* fr30-ibld.c, * frv-asm.c, * frv-desc.c, * frv-dis.c, * frv-ibld.c,
* frv-opc.c, * ip2k-asm.c, * ip2k-desc.c, * ip2k-dis.c, * ip2k-ibld.c,
* iq2000-asm.c, * iq2000-desc.c, * iq2000-dis.c, * iq2000-ibld.c,
* lm32-asm.c, * lm32-desc.c, * lm32-dis.c, * lm32-ibld.c,
* m32c-asm.c, * m32c-desc.c, * m32c-dis.c, * m32c-ibld.c,
* m32r-asm.c, * m32r-desc.c, * m32r-dis.c, * m32r-ibld.c,
* mep-asm.c, * mep-desc.c, * mep-dis.c, * mep-ibld.c, * mt-asm.c,
* mt-desc.c, * mt-dis.c, * mt-ibld.c, * or1k-asm.c, * or1k-desc.c,
* or1k-dis.c, * or1k-ibld.c, * xc16x-asm.c, * xc16x-desc.c,
* xc16x-dis.c, * xc16x-ibld.c, * xstormy16-asm.c, * xstormy16-desc.c,
* xstormy16-dis.c, * xstormy16-ibld.c: Regenerate.
2018-03-02 05:53:50 +08:00
|
|
|
/* xgettext:c-format */
|
|
|
|
opcodes_error_handler (_("unrecognised disassembler option: %s"), option);
|
2016-07-21 00:08:07 +08:00
|
|
|
}
|
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
#define ARC_CPU_TYPE_A6xx(NAME,EXTRA) \
|
|
|
|
{ #NAME, ARC_OPCODE_ARC600, "ARC600" }
|
|
|
|
#define ARC_CPU_TYPE_A7xx(NAME,EXTRA) \
|
|
|
|
{ #NAME, ARC_OPCODE_ARC700, "ARC700" }
|
|
|
|
#define ARC_CPU_TYPE_AV2EM(NAME,EXTRA) \
|
|
|
|
{ #NAME, ARC_OPCODE_ARCv2EM, "ARC EM" }
|
|
|
|
#define ARC_CPU_TYPE_AV2HS(NAME,EXTRA) \
|
|
|
|
{ #NAME, ARC_OPCODE_ARCv2HS, "ARC HS" }
|
|
|
|
#define ARC_CPU_TYPE_NONE \
|
|
|
|
{ 0, 0, 0 }
|
|
|
|
|
|
|
|
/* A table of CPU names and opcode sets. */
|
|
|
|
static const struct cpu_type
|
|
|
|
{
|
|
|
|
const char *name;
|
|
|
|
unsigned flags;
|
|
|
|
const char *isa;
|
|
|
|
}
|
|
|
|
cpu_types[] =
|
|
|
|
{
|
|
|
|
#include "elf/arc-cpu.def"
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Helper for parsing the CPU options. Accept any of the ARC architectures
|
|
|
|
values. OPTION should be a value passed to cpu=. */
|
|
|
|
|
|
|
|
static unsigned
|
|
|
|
parse_cpu_option (const char *option)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; cpu_types[i].name; ++i)
|
|
|
|
{
|
[ARC] Fix handling of cpu=... disassembler option value
There is a bug in handling of cpu=... disassembler option in case there are
other options after it, for example, `cpu=EM,dsp'. In this case `EM,dsp' is
treated as an option value, and strcasecmp reports is as non-equal to "EM".
This is fixed by using disassembler_options_cmp function, which compares string
treating `,' the same way as `\0'.
This function also solves a problem with option order in parse_option.
Previously, if several option had same prefix (e.g. fpud, fpuda), then the
longer one should have been compared first, otherwise when longer option is
passed it would be treated as a short one, because
CONST_STRNEQ ("fpud", "fpuda")
would be true. The order of options was correct for ARC, so there were no
bugs per se, but with disassembler_option_cmp there is no risk of such a bug
being introduced in the future.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc-dis.c (parse_option): Use disassembler_options_cmp to compare
disassembler option strings.
(parse_cpu_option): Likewise.
binutils/ChangeLog
yyyy-mm-dd Anton Kolesov <Anton.Kolesov@synopsys.com>
* testsuite/binutils-all/arc/double_store.s: New file.
* testsuite/binutils-all/arc/objdump.exp: Tests for disassembler
options.
(do_objfile): New function.
(check_assembly): Likewise.
2017-06-15 20:58:32 +08:00
|
|
|
if (!disassembler_options_cmp (cpu_types[i].name, option))
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
{
|
|
|
|
return cpu_types[i].flags;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
opcodes error messages
Another patch aimed at making binutils comply with the GNU coding
standard. The generated files require
https://sourceware.org/ml/cgen/2018-q1/msg00004.html
cpu/
* frv.opc: Include opintl.h.
(add_next_to_vliw): Use opcodes_error_handler to print error.
Standardize error message.
(fr500_check_insn_major_constraints, frv_vliw_add_insn): Likewise.
opcodes/
* sysdep.h (opcodes_error_handler): Define.
(_bfd_error_handler): Declare.
* Makefile.am: Remove stray #.
* opc2c.c (main): Remove bogus -l arg handling. Print "DO NOT
EDIT" comment.
* aarch64-dis.c, * arc-dis.c, * arm-dis.c, * avr-dis.c,
* d30v-dis.c, * h8300-dis.c, * mmix-dis.c, * ppc-dis.c,
* riscv-dis.c, * s390-dis.c, * sparc-dis.c, * v850-dis.c: Use
opcodes_error_handler to print errors. Standardize error messages.
* msp430-decode.opc, * nios2-dis.c, * rl78-decode.opc: Likewise,
and include opintl.h.
* nds32-asm.c: Likewise, and include sysdep.h and opintl.h.
* i386-gen.c: Standardize error messages.
* msp430-decode.c, * rl78-decode.c, rx-decode.c: Regenerate.
* Makefile.in: Regenerate.
* epiphany-asm.c, * epiphany-desc.c, * epiphany-dis.c,
* epiphany-ibld.c, * fr30-asm.c, * fr30-desc.c, * fr30-dis.c,
* fr30-ibld.c, * frv-asm.c, * frv-desc.c, * frv-dis.c, * frv-ibld.c,
* frv-opc.c, * ip2k-asm.c, * ip2k-desc.c, * ip2k-dis.c, * ip2k-ibld.c,
* iq2000-asm.c, * iq2000-desc.c, * iq2000-dis.c, * iq2000-ibld.c,
* lm32-asm.c, * lm32-desc.c, * lm32-dis.c, * lm32-ibld.c,
* m32c-asm.c, * m32c-desc.c, * m32c-dis.c, * m32c-ibld.c,
* m32r-asm.c, * m32r-desc.c, * m32r-dis.c, * m32r-ibld.c,
* mep-asm.c, * mep-desc.c, * mep-dis.c, * mep-ibld.c, * mt-asm.c,
* mt-desc.c, * mt-dis.c, * mt-ibld.c, * or1k-asm.c, * or1k-desc.c,
* or1k-dis.c, * or1k-ibld.c, * xc16x-asm.c, * xc16x-desc.c,
* xc16x-dis.c, * xc16x-ibld.c, * xstormy16-asm.c, * xstormy16-desc.c,
* xstormy16-dis.c, * xstormy16-ibld.c: Regenerate.
2018-03-02 05:53:50 +08:00
|
|
|
/* xgettext:c-format */
|
|
|
|
opcodes_error_handler (_("unrecognised disassembler CPU option: %s"), option);
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
return ARC_OPCODE_NONE;
|
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
/* Go over the options list and parse it. */
|
|
|
|
|
|
|
|
static void
|
2017-04-06 02:21:33 +08:00
|
|
|
parse_disassembler_options (const char *options)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
2017-06-19 18:27:52 +08:00
|
|
|
const char *option;
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
if (options == NULL)
|
|
|
|
return;
|
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
/* Disassembler might be reused for difference CPU's, and cpu option set for
|
|
|
|
the first one shouldn't be applied to second (which might not have
|
|
|
|
explicit cpu in its options. Therefore it is required to reset enforced
|
|
|
|
CPU when new options are being parsed. */
|
|
|
|
enforced_isa_mask = ARC_OPCODE_NONE;
|
|
|
|
|
2017-06-19 18:27:52 +08:00
|
|
|
FOR_EACH_DISASSEMBLER_OPTION (option, options)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
/* A CPU option? Cannot use STRING_COMMA_LEN because strncmp is also a
|
|
|
|
preprocessor macro. */
|
2017-06-19 18:27:52 +08:00
|
|
|
if (strncmp (option, "cpu=", 4) == 0)
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
/* Strip leading `cpu=`. */
|
2017-06-19 18:27:52 +08:00
|
|
|
enforced_isa_mask = parse_cpu_option (option + 4);
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
else
|
2017-06-19 18:27:52 +08:00
|
|
|
parse_option (option);
|
2016-07-21 00:08:07 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
/* Return the instruction type for an instruction described by OPCODE. */
|
|
|
|
|
|
|
|
static enum dis_insn_type
|
|
|
|
arc_opcode_to_insn_type (const struct arc_opcode *opcode)
|
|
|
|
{
|
|
|
|
enum dis_insn_type insn_type;
|
|
|
|
|
|
|
|
switch (opcode->insn_class)
|
|
|
|
{
|
|
|
|
case BRANCH:
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
case BBIT0:
|
|
|
|
case BBIT1:
|
|
|
|
case BI:
|
|
|
|
case BIH:
|
|
|
|
case BRCC:
|
|
|
|
case EI:
|
|
|
|
case JLI:
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
case JUMP:
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
case LOOP:
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
if (!strncmp (opcode->name, "bl", 2)
|
|
|
|
|| !strncmp (opcode->name, "jl", 2))
|
|
|
|
{
|
|
|
|
if (opcode->subclass == COND)
|
|
|
|
insn_type = dis_condjsr;
|
|
|
|
else
|
|
|
|
insn_type = dis_jsr;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (opcode->subclass == COND)
|
|
|
|
insn_type = dis_condbranch;
|
|
|
|
else
|
|
|
|
insn_type = dis_branch;
|
|
|
|
}
|
|
|
|
break;
|
[ARC] Add checking for LP_COUNT reg usage, improve error reporting.
gas/
2016-11-29 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (find_opcode_match): New function argument
errmsg.
(assemble_tokens): Collect and report the eventual error message
found during opcode matching process.
* testsuite/gas/arc/lpcount-err.s: New file.
* testsuite/gas/arc/add_s-err.s: Update error message.
opcode/
2016-11-29 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (insert_ra_chk): New function.
(insert_rb_chk): Likewise.
(insert_rad): Update text error message.
(insert_rcd): Likewise.
(insert_rhv2): Likewise.
(insert_r0): Likewise.
(insert_r1): Likewise.
(insert_r2): Likewise.
(insert_r3): Likewise.
(insert_sp): Likewise.
(insert_gp): Likewise.
(insert_pcl): Likewise.
(insert_blink): Likewise.
(insert_ilink1): Likewise.
(insert_ilink2): Likewise.
(insert_ras): Likewise.
(insert_rbs): Likewise.
(insert_rcs): Likewise.
(insert_simm3s): Likewise.
(insert_rrange): Likewise.
(insert_fpel): Likewise.
(insert_blinkel): Likewise.
(insert_pcel): Likewise.
(insert_nps_3bit_dst): Likewise.
(insert_nps_3bit_dst_short): Likewise.
(insert_nps_3bit_src2_short): Likewise.
(insert_nps_bitop_size_2b): Likewise.
(MAKE_SRC_POS_INSERT_EXTRACT_FUNCS): Likewise.
(RA_CHK): Define.
(RB): Adjust.
(RB_CHK): Define.
(RC): Adjust.
* arc-dis.c (print_insn_arc): Add LOAD and STORE class.
* arc-tbl.h (div, divu): All instructions are DIVREM class.
Change first insn argument to check for LP_COUNT usage.
(rem): Likewise.
(ld, ldd): All instructions are LOAD class. Change first insn
argument to check for LP_COUNT usage.
(st, std): All instructions are STORE class.
(mac, mpy, dmac, mul, dmpy): All instructions are MPY class.
Change first insn argument to check for LP_COUNT usage.
(mov): All instructions are MOVE class. Change first insn
argument to check for LP_COUNT usage.
include/
2016-11-29 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (insn_class_t): Add DIVREM, LOAD, MOVE, MPY, STORE
instruction classes.
2016-11-15 22:11:47 +08:00
|
|
|
case LOAD:
|
|
|
|
case STORE:
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
case MEMORY:
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
case ENTER:
|
|
|
|
case PUSH:
|
|
|
|
case POP:
|
[ARC] Add checking for LP_COUNT reg usage, improve error reporting.
gas/
2016-11-29 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (find_opcode_match): New function argument
errmsg.
(assemble_tokens): Collect and report the eventual error message
found during opcode matching process.
* testsuite/gas/arc/lpcount-err.s: New file.
* testsuite/gas/arc/add_s-err.s: Update error message.
opcode/
2016-11-29 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (insert_ra_chk): New function.
(insert_rb_chk): Likewise.
(insert_rad): Update text error message.
(insert_rcd): Likewise.
(insert_rhv2): Likewise.
(insert_r0): Likewise.
(insert_r1): Likewise.
(insert_r2): Likewise.
(insert_r3): Likewise.
(insert_sp): Likewise.
(insert_gp): Likewise.
(insert_pcl): Likewise.
(insert_blink): Likewise.
(insert_ilink1): Likewise.
(insert_ilink2): Likewise.
(insert_ras): Likewise.
(insert_rbs): Likewise.
(insert_rcs): Likewise.
(insert_simm3s): Likewise.
(insert_rrange): Likewise.
(insert_fpel): Likewise.
(insert_blinkel): Likewise.
(insert_pcel): Likewise.
(insert_nps_3bit_dst): Likewise.
(insert_nps_3bit_dst_short): Likewise.
(insert_nps_3bit_src2_short): Likewise.
(insert_nps_bitop_size_2b): Likewise.
(MAKE_SRC_POS_INSERT_EXTRACT_FUNCS): Likewise.
(RA_CHK): Define.
(RB): Adjust.
(RB_CHK): Define.
(RC): Adjust.
* arc-dis.c (print_insn_arc): Add LOAD and STORE class.
* arc-tbl.h (div, divu): All instructions are DIVREM class.
Change first insn argument to check for LP_COUNT usage.
(rem): Likewise.
(ld, ldd): All instructions are LOAD class. Change first insn
argument to check for LP_COUNT usage.
(st, std): All instructions are STORE class.
(mac, mpy, dmac, mul, dmpy): All instructions are MPY class.
Change first insn argument to check for LP_COUNT usage.
(mov): All instructions are MOVE class. Change first insn
argument to check for LP_COUNT usage.
include/
2016-11-29 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (insn_class_t): Add DIVREM, LOAD, MOVE, MPY, STORE
instruction classes.
2016-11-15 22:11:47 +08:00
|
|
|
insn_type = dis_dref;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
break;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
case LEAVE:
|
|
|
|
insn_type = dis_branch;
|
|
|
|
break;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
default:
|
|
|
|
insn_type = dis_nonbranch;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return insn_type;
|
|
|
|
}
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Disassemble ARC instructions. */
|
2001-01-12 05:20:20 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
static int
|
|
|
|
print_insn_arc (bfd_vma memaddr,
|
|
|
|
struct disassemble_info *info)
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
bfd_byte buffer[8];
|
|
|
|
unsigned int highbyte, lowbyte;
|
2015-10-07 21:20:19 +08:00
|
|
|
int status;
|
2016-06-02 21:03:23 +08:00
|
|
|
unsigned int insn_len;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
unsigned long long insn = 0;
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
unsigned isa_mask = ARC_OPCODE_NONE;
|
2015-10-07 21:20:19 +08:00
|
|
|
const struct arc_opcode *opcode;
|
|
|
|
bfd_boolean need_comma;
|
|
|
|
bfd_boolean open_braket;
|
2015-12-04 18:49:57 +08:00
|
|
|
int size;
|
2016-06-02 21:03:23 +08:00
|
|
|
const struct arc_operand *operand;
|
2017-11-21 21:03:03 +08:00
|
|
|
int value, vpcl;
|
2016-06-02 21:03:23 +08:00
|
|
|
struct arc_operand_iterator iter;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
struct arc_disassemble_info *arc_infop;
|
2017-11-21 21:03:03 +08:00
|
|
|
bfd_boolean rpcl = FALSE, rset = FALSE;
|
2016-07-21 00:08:07 +08:00
|
|
|
|
|
|
|
if (info->disassembler_options)
|
|
|
|
{
|
|
|
|
parse_disassembler_options (info->disassembler_options);
|
|
|
|
|
|
|
|
/* Avoid repeated parsing of the options. */
|
|
|
|
info->disassembler_options = NULL;
|
|
|
|
}
|
2001-01-12 05:20:20 +08:00
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
if (info->private_data == NULL && !init_arc_disasm_info (info))
|
|
|
|
return -1;
|
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
memset (&iter, 0, sizeof (iter));
|
2016-10-13 16:34:16 +08:00
|
|
|
highbyte = ((info->endian == BFD_ENDIAN_LITTLE) ? 1 : 0);
|
|
|
|
lowbyte = ((info->endian == BFD_ENDIAN_LITTLE) ? 0 : 1);
|
2001-01-12 05:20:20 +08:00
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
/* Figure out CPU type, unless it was enforced via disassembler options. */
|
|
|
|
if (enforced_isa_mask == ARC_OPCODE_NONE)
|
2015-10-07 21:20:19 +08:00
|
|
|
{
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
Elf_Internal_Ehdr *header = NULL;
|
2001-01-12 05:20:20 +08:00
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
if (info->section && info->section->owner)
|
|
|
|
header = elf_elfheader (info->section->owner);
|
2001-01-12 05:20:20 +08:00
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
switch (info->mach)
|
2016-07-21 00:08:07 +08:00
|
|
|
{
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
case bfd_mach_arc_arc700:
|
|
|
|
isa_mask = ARC_OPCODE_ARC700;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case bfd_mach_arc_arc600:
|
|
|
|
isa_mask = ARC_OPCODE_ARC600;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case bfd_mach_arc_arcv2:
|
|
|
|
default:
|
|
|
|
isa_mask = ARC_OPCODE_ARCv2EM;
|
|
|
|
/* TODO: Perhaps remove definition of header since it is only used at
|
|
|
|
this location. */
|
|
|
|
if (header != NULL
|
|
|
|
&& (header->e_flags & EF_ARC_MACH_MSK) == EF_ARC_CPU_ARCV2HS)
|
|
|
|
isa_mask = ARC_OPCODE_ARCv2HS;
|
|
|
|
break;
|
2016-07-21 00:08:07 +08:00
|
|
|
}
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
isa_mask = enforced_isa_mask;
|
|
|
|
|
|
|
|
if (isa_mask == ARC_OPCODE_ARCv2HS)
|
|
|
|
{
|
|
|
|
/* FPU instructions are not extensions for HS. */
|
|
|
|
add_to_decodelist (FLOAT, SP);
|
|
|
|
add_to_decodelist (FLOAT, DP);
|
|
|
|
add_to_decodelist (FLOAT, CVT);
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
|
|
|
|
2015-12-04 18:49:57 +08:00
|
|
|
/* This variable may be set by the instruction decoder. It suggests
|
|
|
|
the number of bytes objdump should display on a single line. If
|
|
|
|
the instruction decoder sets this, it should always set it to
|
|
|
|
the same value in order to get reasonable looking output. */
|
|
|
|
info->bytes_per_line = 8;
|
|
|
|
|
|
|
|
/* In the next lines, we set two info variables control the way
|
|
|
|
objdump displays the raw data. For example, if bytes_per_line is
|
|
|
|
8 and bytes_per_chunk is 4, the output will look like this:
|
|
|
|
00: 00000000 00000000
|
|
|
|
with the chunks displayed according to "display_endian". */
|
|
|
|
if (info->section
|
|
|
|
&& !(info->section->flags & SEC_CODE))
|
|
|
|
{
|
|
|
|
/* This is not a CODE section. */
|
|
|
|
switch (info->section->size)
|
|
|
|
{
|
|
|
|
case 1:
|
|
|
|
case 2:
|
|
|
|
case 4:
|
|
|
|
size = info->section->size;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
size = (info->section->size & 0x01) ? 1 : 4;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
info->bytes_per_chunk = 1;
|
|
|
|
info->display_endian = info->endian;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
size = 2;
|
|
|
|
info->bytes_per_chunk = 2;
|
|
|
|
info->display_endian = info->endian;
|
|
|
|
}
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Read the insn into a host word. */
|
2015-12-04 18:49:57 +08:00
|
|
|
status = (*info->read_memory_func) (memaddr, buffer, size, info);
|
2017-11-21 21:03:03 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
if (status != 0)
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
(*info->memory_error_func) (status, memaddr, info);
|
|
|
|
return -1;
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
if (info->section
|
|
|
|
&& !(info->section->flags & SEC_CODE))
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-12-04 18:49:57 +08:00
|
|
|
/* Data section. */
|
|
|
|
unsigned long data;
|
|
|
|
|
|
|
|
data = bfd_get_bits (buffer, size * 8,
|
|
|
|
info->display_endian == BFD_ENDIAN_BIG);
|
|
|
|
switch (size)
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-12-04 18:49:57 +08:00
|
|
|
case 1:
|
|
|
|
(*info->fprintf_func) (info->stream, ".byte\t0x%02lx", data);
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
(*info->fprintf_func) (info->stream, ".short\t0x%04lx", data);
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
(*info->fprintf_func) (info->stream, ".word\t0x%08lx", data);
|
|
|
|
break;
|
|
|
|
default:
|
2019-08-08 00:22:29 +08:00
|
|
|
return -1;
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
2015-12-04 18:49:57 +08:00
|
|
|
return size;
|
2015-10-07 21:20:19 +08:00
|
|
|
}
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2016-10-13 16:34:16 +08:00
|
|
|
insn_len = arc_insn_length (buffer[highbyte], buffer[lowbyte], info);
|
2016-06-02 21:03:23 +08:00
|
|
|
pr_debug ("instruction length = %d bytes\n", insn_len);
|
2019-08-08 00:22:29 +08:00
|
|
|
if (insn_len == 0)
|
|
|
|
return -1;
|
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
arc_infop = info->private_data;
|
|
|
|
arc_infop->insn_len = insn_len;
|
2016-07-21 00:08:07 +08:00
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
switch (insn_len)
|
2015-10-07 21:20:19 +08:00
|
|
|
{
|
2016-03-25 00:54:37 +08:00
|
|
|
case 2:
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
insn = (buffer[highbyte] << 8) | buffer[lowbyte];
|
2016-03-25 00:54:37 +08:00
|
|
|
break;
|
2015-10-07 21:20:19 +08:00
|
|
|
|
2016-03-25 00:54:37 +08:00
|
|
|
case 4:
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
{
|
|
|
|
/* This is a long instruction: Read the remaning 2 bytes. */
|
|
|
|
status = (*info->read_memory_func) (memaddr + 2, &buffer[2], 2, info);
|
|
|
|
if (status != 0)
|
|
|
|
{
|
|
|
|
(*info->memory_error_func) (status, memaddr + 2, info);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
insn = (unsigned long long) ARRANGE_ENDIAN (info, buffer);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 6:
|
|
|
|
{
|
|
|
|
status = (*info->read_memory_func) (memaddr + 2, &buffer[2], 4, info);
|
|
|
|
if (status != 0)
|
|
|
|
{
|
|
|
|
(*info->memory_error_func) (status, memaddr + 2, info);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
insn = (unsigned long long) ARRANGE_ENDIAN (info, &buffer[2]);
|
|
|
|
insn |= ((unsigned long long) buffer[highbyte] << 40)
|
|
|
|
| ((unsigned long long) buffer[lowbyte] << 32);
|
|
|
|
}
|
2016-03-25 00:54:37 +08:00
|
|
|
break;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
|
|
|
|
case 8:
|
|
|
|
{
|
|
|
|
status = (*info->read_memory_func) (memaddr + 2, &buffer[2], 6, info);
|
|
|
|
if (status != 0)
|
|
|
|
{
|
|
|
|
(*info->memory_error_func) (status, memaddr + 2, info);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
insn =
|
|
|
|
((((unsigned long long) ARRANGE_ENDIAN (info, buffer)) << 32)
|
|
|
|
| ((unsigned long long) ARRANGE_ENDIAN (info, &buffer[4])));
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
|
|
|
/* There is no instruction whose length is not 2, 4, 6, or 8. */
|
2019-08-08 00:22:29 +08:00
|
|
|
return -1;
|
2015-10-07 21:20:19 +08:00
|
|
|
}
|
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
pr_debug ("instruction value = %llx\n", insn);
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Set some defaults for the insn info. */
|
|
|
|
info->insn_info_valid = 1;
|
|
|
|
info->branch_delay_insns = 0;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
info->data_size = 4;
|
2015-10-07 21:20:19 +08:00
|
|
|
info->insn_type = dis_nonbranch;
|
|
|
|
info->target = 0;
|
|
|
|
info->target2 = 0;
|
|
|
|
|
|
|
|
/* FIXME to be moved in dissasemble_init_for_target. */
|
|
|
|
info->disassembler_needs_relocs = TRUE;
|
|
|
|
|
|
|
|
/* Find the first match in the opcode table. */
|
2016-06-02 21:03:23 +08:00
|
|
|
if (!find_format (memaddr, insn, &insn_len, isa_mask, info, &opcode, &iter))
|
|
|
|
return -1;
|
2015-10-07 21:20:19 +08:00
|
|
|
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
if (!opcode)
|
|
|
|
{
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
switch (insn_len)
|
|
|
|
{
|
|
|
|
case 2:
|
2017-11-21 21:03:03 +08:00
|
|
|
(*info->fprintf_func) (info->stream, ".shor\t%#04llx",
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
insn & 0xffff);
|
|
|
|
break;
|
2019-08-08 00:22:29 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
case 4:
|
2017-11-21 21:03:03 +08:00
|
|
|
(*info->fprintf_func) (info->stream, ".word\t%#08llx",
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
insn & 0xffffffff);
|
|
|
|
break;
|
2019-08-08 00:22:29 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
case 6:
|
2017-11-21 21:03:03 +08:00
|
|
|
(*info->fprintf_func) (info->stream, ".long\t%#08llx",
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
insn & 0xffffffff);
|
2017-11-21 21:03:03 +08:00
|
|
|
(*info->fprintf_func) (info->stream, ".long\t%#04llx",
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
(insn >> 32) & 0xffff);
|
|
|
|
break;
|
2019-08-08 00:22:29 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
case 8:
|
2017-11-21 21:03:03 +08:00
|
|
|
(*info->fprintf_func) (info->stream, ".long\t%#08llx",
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
insn & 0xffffffff);
|
2017-11-21 21:03:03 +08:00
|
|
|
(*info->fprintf_func) (info->stream, ".long\t%#08llx",
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
insn >> 32);
|
|
|
|
break;
|
2019-08-08 00:22:29 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
default:
|
2019-08-08 00:22:29 +08:00
|
|
|
return -1;
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
}
|
2015-10-07 21:20:19 +08:00
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
info->insn_type = dis_noninsn;
|
|
|
|
return insn_len;
|
2015-10-07 21:20:19 +08:00
|
|
|
}
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Print the mnemonic. */
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", opcode->name);
|
|
|
|
|
|
|
|
/* Preselect the insn class. */
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
info->insn_type = arc_opcode_to_insn_type (opcode);
|
2001-08-21 16:51:12 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
pr_debug ("%s: 0x%08llx\n", opcode->name, opcode->opcode);
|
2001-08-21 16:51:12 +08:00
|
|
|
|
arc: Change max instruction length to 64-bits
The current handling for arc instructions longer than 32-bits is all
handled as a special case in both the assembler and disassembler.
The problem with this approach is that it leads to code duplication,
selecting a long instruction is exactly the same process as selecting a
short instruction, except over more bits, in both cases we select based
on bit comparison, and initial operand insertion and extraction.
This commit unifies both the long and short instruction worlds,
converting the core opcodes library from being largely 32-bit focused,
to being largely 64-bit focused.
The changes are, on the whole, not too much. There's obviously a lot of
type changes but otherwise the bulk of the code just works. Most of the
actual functional changes are to code that previously handled the longer
48 or 64 bit instructions. The insert/extract handlers for these have
now been brought into line with the short instruction insert/extract
handlers.
All of the special case handling code that was previously added has now
been removed again. Overall, this commit reduces the amount of code in
the arc assembler and disassembler.
gas/ChangeLog:
* config/tc-arc.c (struct arc_insn): Change type of insn field.
(md_number_to_chars_midend): Support 6- and 8-byte values.
(emit_insn0): Update debug output.
(find_opcode_match): Likewise.
(build_fake_opcode_hash_entry): Delete.
(find_special_case_long_opcode): Delete.
(find_special_case): Remove long format special case handling.
(insert_operand): Change instruction type and update debug print
format.
(assemble_insn): Change instruction type, update debug print
formats, and remove unneeded assert.
include/ChangeLog:
* opcode/arc.h (struct arc_opcode): Change type of opcode and mask
fields.
(struct arc_long_opcode): Delete.
(struct arc_operand): Change types for insert and extract
handlers.
opcodes/ChangeLog:
* arc-dis.c (struct arc_operand_iterator): Remove all fields
relating to long instruction processing, add new limm field.
(OPCODE): Rename to...
(OPCODE_32BIT_INSN): ...this.
(OPCODE_AC): Delete.
(skip_this_opcode): Handle different instruction lengths, update
macro name.
(special_flag_p): Update parameter type.
(find_format_from_table): Update for more instruction lengths.
(find_format_long_instructions): Delete.
(find_format): Update for more instruction lengths.
(arc_insn_length): Likewise.
(extract_operand_value): Update for more instruction lengths.
(operand_iterator_next): Remove code relating to long
instructions.
(arc_opcode_to_insn_type): New function.
(print_insn_arc):Update for more instructions lengths.
* arc-ext.c (extInstruction_t): Change argument type.
* arc-ext.h (extInstruction_t): Change argument type.
* arc-fxi.h: Change type unsigned to unsigned long long
extensively throughout.
* arc-nps400-tbl.h: Add long instructions taken from
arc_long_opcodes table in arc-opc.c.
* arc-opc.c: Update parameter types on insert/extract handlers.
(arc_long_opcodes): Delete.
(arc_num_long_opcodes): Delete.
(arc_opcode_len): Update for more instruction lengths.
2016-07-07 02:39:55 +08:00
|
|
|
print_flags (opcode, &insn, info);
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
if (opcode->operands[0] != 0)
|
|
|
|
(*info->fprintf_func) (info->stream, "\t");
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
need_comma = FALSE;
|
|
|
|
open_braket = FALSE;
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
arc_infop->operands_count = 0;
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Now extract and print the operands. */
|
2016-06-02 21:03:23 +08:00
|
|
|
operand = NULL;
|
2017-11-21 21:03:03 +08:00
|
|
|
vpcl = 0;
|
2016-06-02 21:03:23 +08:00
|
|
|
while (operand_iterator_next (&iter, &operand, &value))
|
2015-10-07 21:20:19 +08:00
|
|
|
{
|
|
|
|
if (open_braket && (operand->flags & ARC_OPERAND_BRAKET))
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
(*info->fprintf_func) (info->stream, "]");
|
|
|
|
open_braket = FALSE;
|
|
|
|
continue;
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Only take input from real operands. */
|
2016-07-27 22:57:18 +08:00
|
|
|
if (ARC_OPERAND_IS_FAKE (operand))
|
2015-10-07 21:20:19 +08:00
|
|
|
continue;
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
if ((operand->flags & ARC_OPERAND_IGNORE)
|
|
|
|
&& (operand->flags & ARC_OPERAND_IR)
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
&& value == -1)
|
2015-10-07 21:20:19 +08:00
|
|
|
continue;
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2016-07-27 22:57:18 +08:00
|
|
|
if (operand->flags & ARC_OPERAND_COLON)
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
{
|
|
|
|
(*info->fprintf_func) (info->stream, ":");
|
|
|
|
continue;
|
|
|
|
}
|
2016-07-27 22:57:18 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
if (need_comma)
|
|
|
|
(*info->fprintf_func) (info->stream, ",");
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
if (!open_braket && (operand->flags & ARC_OPERAND_BRAKET))
|
2001-01-12 05:20:20 +08:00
|
|
|
{
|
2015-10-07 21:20:19 +08:00
|
|
|
(*info->fprintf_func) (info->stream, "[");
|
|
|
|
open_braket = TRUE;
|
|
|
|
need_comma = FALSE;
|
|
|
|
continue;
|
2001-01-12 05:20:20 +08:00
|
|
|
}
|
2015-10-07 21:20:19 +08:00
|
|
|
|
2016-07-27 22:57:18 +08:00
|
|
|
need_comma = TRUE;
|
|
|
|
|
2017-11-21 21:03:03 +08:00
|
|
|
if (operand->flags & ARC_OPERAND_PCREL)
|
|
|
|
{
|
|
|
|
rpcl = TRUE;
|
|
|
|
vpcl = value;
|
|
|
|
rset = TRUE;
|
|
|
|
|
|
|
|
info->target = (bfd_vma) (memaddr & ~3) + value;
|
|
|
|
}
|
|
|
|
else if (!(operand->flags & ARC_OPERAND_IR))
|
|
|
|
{
|
|
|
|
vpcl = value;
|
|
|
|
rset = TRUE;
|
|
|
|
}
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Print the operand as directed by the flags. */
|
|
|
|
if (operand->flags & ARC_OPERAND_IR)
|
|
|
|
{
|
2016-04-06 22:08:04 +08:00
|
|
|
const char *rname;
|
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
assert (value >=0 && value < 64);
|
2016-04-06 22:08:04 +08:00
|
|
|
rname = arcExtMap_coreRegName (value);
|
|
|
|
if (!rname)
|
|
|
|
rname = regnames[value];
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", rname);
|
2015-10-07 21:20:19 +08:00
|
|
|
if (operand->flags & ARC_OPERAND_TRUNCATE)
|
2016-04-06 22:08:04 +08:00
|
|
|
{
|
|
|
|
rname = arcExtMap_coreRegName (value + 1);
|
|
|
|
if (!rname)
|
|
|
|
rname = regnames[value + 1];
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", rname);
|
|
|
|
}
|
2017-11-21 21:03:03 +08:00
|
|
|
if (value == 63)
|
|
|
|
rpcl = TRUE;
|
|
|
|
else
|
|
|
|
rpcl = FALSE;
|
2015-10-07 21:20:19 +08:00
|
|
|
}
|
|
|
|
else if (operand->flags & ARC_OPERAND_LIMM)
|
|
|
|
{
|
2016-06-02 21:03:23 +08:00
|
|
|
const char *rname = get_auxreg (opcode, value, isa_mask);
|
2016-07-27 22:57:18 +08:00
|
|
|
|
2016-04-06 22:08:04 +08:00
|
|
|
if (rname && open_braket)
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", rname);
|
|
|
|
else
|
|
|
|
{
|
2016-06-02 21:03:23 +08:00
|
|
|
(*info->fprintf_func) (info->stream, "%#x", value);
|
2016-04-06 22:08:04 +08:00
|
|
|
if (info->insn_type == dis_branch
|
|
|
|
|| info->insn_type == dis_jsr)
|
2016-06-02 21:03:23 +08:00
|
|
|
info->target = (bfd_vma) value;
|
2016-04-06 22:08:04 +08:00
|
|
|
}
|
2015-10-07 21:20:19 +08:00
|
|
|
}
|
|
|
|
else if (operand->flags & ARC_OPERAND_SIGNED)
|
2016-04-06 22:08:04 +08:00
|
|
|
{
|
|
|
|
const char *rname = get_auxreg (opcode, value, isa_mask);
|
|
|
|
if (rname && open_braket)
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", rname);
|
|
|
|
else
|
2017-11-03 22:36:42 +08:00
|
|
|
{
|
|
|
|
if (print_hex)
|
|
|
|
(*info->fprintf_func) (info->stream, "%#x", value);
|
|
|
|
else
|
|
|
|
(*info->fprintf_func) (info->stream, "%d", value);
|
|
|
|
}
|
2016-04-06 22:08:04 +08:00
|
|
|
}
|
2016-07-27 22:57:18 +08:00
|
|
|
else if (operand->flags & ARC_OPERAND_ADDRTYPE)
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
{
|
|
|
|
const char *addrtype = get_addrtype (value);
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", addrtype);
|
|
|
|
/* A colon follow an address type. */
|
|
|
|
need_comma = FALSE;
|
|
|
|
}
|
2015-10-07 21:20:19 +08:00
|
|
|
else
|
2016-04-06 22:08:04 +08:00
|
|
|
{
|
|
|
|
if (operand->flags & ARC_OPERAND_TRUNCATE
|
|
|
|
&& !(operand->flags & ARC_OPERAND_ALIGNED32)
|
|
|
|
&& !(operand->flags & ARC_OPERAND_ALIGNED16)
|
2017-04-25 23:07:00 +08:00
|
|
|
&& value >= 0 && value <= 14)
|
|
|
|
{
|
2017-11-21 21:03:03 +08:00
|
|
|
/* Leave/Enter mnemonics. */
|
2017-04-25 23:07:00 +08:00
|
|
|
switch (value)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
need_comma = FALSE;
|
|
|
|
break;
|
|
|
|
case 1:
|
|
|
|
(*info->fprintf_func) (info->stream, "r13");
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
(*info->fprintf_func) (info->stream, "r13-%s",
|
|
|
|
regnames[13 + value - 1]);
|
|
|
|
break;
|
|
|
|
}
|
2017-11-21 21:03:03 +08:00
|
|
|
rpcl = FALSE;
|
|
|
|
rset = FALSE;
|
2017-04-25 23:07:00 +08:00
|
|
|
}
|
2016-04-06 22:08:04 +08:00
|
|
|
else
|
|
|
|
{
|
|
|
|
const char *rname = get_auxreg (opcode, value, isa_mask);
|
|
|
|
if (rname && open_braket)
|
|
|
|
(*info->fprintf_func) (info->stream, "%s", rname);
|
|
|
|
else
|
|
|
|
(*info->fprintf_func) (info->stream, "%#x", value);
|
|
|
|
}
|
|
|
|
}
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
|
|
|
|
if (operand->flags & ARC_OPERAND_LIMM)
|
|
|
|
{
|
|
|
|
arc_infop->operands[arc_infop->operands_count].kind
|
|
|
|
= ARC_OPERAND_KIND_LIMM;
|
|
|
|
/* It is not important to have exactly the LIMM indicator
|
|
|
|
here. */
|
|
|
|
arc_infop->operands[arc_infop->operands_count].value = 63;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
arc_infop->operands[arc_infop->operands_count].value = value;
|
|
|
|
arc_infop->operands[arc_infop->operands_count].kind
|
|
|
|
= (operand->flags & ARC_OPERAND_IR
|
|
|
|
? ARC_OPERAND_KIND_REG
|
|
|
|
: ARC_OPERAND_KIND_SHIMM);
|
|
|
|
}
|
|
|
|
arc_infop->operands_count ++;
|
1999-05-03 15:29:11 +08:00
|
|
|
}
|
2001-08-21 16:51:12 +08:00
|
|
|
|
2017-11-21 21:03:03 +08:00
|
|
|
/* Pretty print extra info for pc-relative operands. */
|
|
|
|
if (rpcl && rset)
|
|
|
|
{
|
|
|
|
if (info->flags & INSN_HAS_RELOC)
|
|
|
|
/* If the instruction has a reloc associated with it, then the
|
|
|
|
offset field in the instruction will actually be the addend
|
|
|
|
for the reloc. (We are using REL type relocs). In such
|
|
|
|
cases, we can ignore the pc when computing addresses, since
|
|
|
|
the addend is not currently pc-relative. */
|
|
|
|
memaddr = 0;
|
|
|
|
|
|
|
|
(*info->fprintf_func) (info->stream, "\t;");
|
|
|
|
(*info->print_address_func) ((memaddr & ~3) + vpcl, info);
|
|
|
|
}
|
|
|
|
|
2016-06-02 21:03:23 +08:00
|
|
|
return insn_len;
|
1999-05-03 15:29:11 +08:00
|
|
|
}
|
|
|
|
|
2005-07-01 19:16:33 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
disassembler_ftype
|
|
|
|
arc_get_disassembler (bfd *abfd)
|
1999-05-03 15:29:11 +08:00
|
|
|
{
|
2016-09-14 19:20:13 +08:00
|
|
|
/* BFD my be absent, if opcodes is invoked from the debugger that
|
|
|
|
has connected to remote target and doesn't have an ELF file. */
|
|
|
|
if (abfd != NULL)
|
|
|
|
{
|
|
|
|
/* Read the extension insns and registers, if any. */
|
|
|
|
build_ARC_extmap (abfd);
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
#ifdef DEBUG
|
2016-09-14 19:20:13 +08:00
|
|
|
dump_ARC_extmap ();
|
Add support for .extInstruction pseudo-op.
gas/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* testsuite/gas/arc/textinsn-errors.d: New File.
* testsuite/gas/arc/textinsn-errors.err: Likewise.
* testsuite/gas/arc/textinsn-errors.s: Likewise.
* testsuite/gas/arc/textinsn2op.d: Likewise.
* testsuite/gas/arc/textinsn2op.s: Likewise.
* testsuite/gas/arc/textinsn2op01.d: Likewise.
* testsuite/gas/arc/textinsn2op01.s: Likewise.
* testsuite/gas/arc/textinsn3op.d: Likewise.
* testsuite/gas/arc/textinsn3op.s: Likewise.
* doc/c-arc.texi (ARC Directives): Add .extInstruction
documentation.
* config/tc-arc.c (arcext_section): New variable.
(arc_extinsn): New function.
(md_pseudo_table): Add .extInstruction pseudo op.
(attributes_t): New type.
(suffixclass, syntaxclass, syntaxclassmod): New constant
structures.
(find_opcode_match): Remove arc_num_opcodes.
(md_begin): Likewise.
(tokenize_extinsn): New function.
(arc_set_ext_seg): Likewise.
(create_extinst_section): Likewise.
include/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* opcode/arc.h (arc_num_opcodes): Remove.
(ARC_SYNTAX_3OP, ARC_SYNTAX_2OP, ARC_OP1_MUST_BE_IMM)
(ARC_OP1_IMM_IMPLIED, ARC_SUFFIX_NONE, ARC_SUFFIX_COND)
(ARC_SUFFIX_FLAG): Define.
(flags_none, flags_f, flags_cc, flags_ccf): Declare.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
opcodes/
2016-04-04 Claudiu Zissulescu <claziss@synopsys.com>
* arc-opc.c (flags_none, flags_f, flags_cc, flags_ccf):
Initialize.
(arg_none, arg_32bit_rarbrc, arg_32bit_zarbrc, arg_32bit_rbrbrc)
(arg_32bit_rarbu6, arg_32bit_zarbu6, arg_32bit_rbrbu6)
(arg_32bit_rbrbs12, arg_32bit_ralimmrc, arg_32bit_rarblimm)
(arg_32bit_zalimmrc, arg_32bit_zarblimm, arg_32bit_rbrblimm)
(arg_32bit_ralimmu6, arg_32bit_zalimmu6, arg_32bit_zalimms12)
(arg_32bit_ralimmlimm, arg_32bit_zalimmlimm, arg_32bit_rbrc)
(arg_32bit_zarc, arg_32bit_rbu6, arg_32bit_zau6, arg_32bit_rblimm)
(arg_32bit_zalimm, arg_32bit_limmrc, arg_32bit_limmu6)
(arg_32bit_limms12, arg_32bit_limmlimm): Likewise.
(arc_opcode arc_opcodes): Null terminate the array.
(arc_num_opcodes): Remove.
* arc-ext.h (INSERT_XOP): Define.
(extInstruction_t): Likewise.
(arcExtMap_instName): Delete.
(arcExtMap_insn): New function.
(arcExtMap_genOpcode): Likewise.
* arc-ext.c (ExtInstruction): Remove.
(create_map): Zero initialize instruction fields.
(arcExtMap_instName): Remove.
(arcExtMap_insn): New function.
(dump_ARC_extmap): More info while debuging.
(arcExtMap_genOpcode): New function.
* arc-dis.c (find_format): New function.
(print_insn_arc): Use find_format.
(arc_get_disassembler): Enable dump_ARC_extmap only when
debugging.
Signed-off-by: Claudiu Zissulescu <claziss@synopsys.com>
2016-04-04 22:03:53 +08:00
|
|
|
#endif
|
2016-09-14 19:20:13 +08:00
|
|
|
}
|
1999-05-03 15:29:11 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
return print_insn_arc;
|
1999-05-03 15:29:11 +08:00
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
void
|
|
|
|
print_arc_disassembler_options (FILE *stream)
|
|
|
|
{
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
int i;
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
fprintf (stream, _("\n\
|
|
|
|
The following ARC specific disassembler options are supported for use \n\
|
|
|
|
with -M switch (multiple options should be separated by commas):\n"));
|
|
|
|
|
[ARC] Allow CPU to be enforced via disassemble_info options
Currently print_insn_arc relies on BFD mach and ELF private headers to
distinguish between various ARC architectures. Sometimes those values are not
correct or available, mainly in the case of debugging targets without and ELF
file available. Changing a BFD mach is not a problem for the debugger, because
this is a generic BFD field, and GDB, for example, already sets it according to
information provided in XML target description or specified via GDB 'set arch'
command. However, things are more complicated for ELF private headers, since
it requires existing of an actual ELF file. To workaround this problem this
patch allows CPU model to be specified via disassemble info options. If CPU is
specified in options, then it will take a higher precedence than whatever might
be specified in ELF file.
This is mostly needed for ARC EM and ARC HS, because they have the same
"architecture" (mach) ARCv2 and differ in their private ELF headers. Other ARC
architectures can be distinguished between each other purely via "mach" field.
Proposed disassemble option format is "cpu=<CPU>", where CPU can be any valid
ARC CPU name as supported by GAS. Note that this creates a seeming redundancy
with objdump -m/--architecture option, however -mEM and -mHS still result in
"ARCv2" architecture internally, while -Mcpu={HS,EM} would have an actual
effect on disassembler.
opcodes/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (enforced_isa_mask): Declare.
(cpu_types): Likewise.
(parse_cpu_option): New function.
(parse_disassembler_options): Use it.
(print_insn_arc): Use enforced_isa_mask.
(print_arc_disassembler_options): Document new options.
binutils/ChangeLog:
yyyy-mm-dd Anton Kolesov <anton.kolesov@synopsys.com>
* doc/binutils.texi: Document new cpu=... disassembler options for ARC.
2017-03-16 20:21:31 +08:00
|
|
|
/* cpu=... options. */
|
|
|
|
for (i = 0; cpu_types[i].name; ++i)
|
|
|
|
{
|
|
|
|
/* As of now all value CPU values are less than 16 characters. */
|
|
|
|
fprintf (stream, " cpu=%-16s\tEnforce %s ISA.\n",
|
|
|
|
cpu_types[i].name, cpu_types[i].isa);
|
|
|
|
}
|
|
|
|
|
2016-07-21 00:08:07 +08:00
|
|
|
fprintf (stream, _("\
|
|
|
|
dsp Recognize DSP instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
|
|
|
spfp Recognize FPX SP instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
|
|
|
dpfp Recognize FPX DP instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
|
|
|
quarkse_em Recognize FPU QuarkSE-EM instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
|
|
|
fpuda Recognize double assist FPU instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
|
|
|
fpus Recognize single precision FPU instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
|
|
|
fpud Recognize double precision FPU instructions.\n"));
|
2017-11-03 22:36:42 +08:00
|
|
|
fprintf (stream, _("\
|
2019-07-24 21:52:23 +08:00
|
|
|
nps400 Recognize NPS400 instructions.\n"));
|
|
|
|
fprintf (stream, _("\
|
2017-11-03 22:36:42 +08:00
|
|
|
hex Use only hexadecimal number to print immediates.\n"));
|
2016-07-21 00:08:07 +08:00
|
|
|
}
|
|
|
|
|
[ARC] Provide an interface to decode ARC instructions.
gas/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
* config/tc-arc.c (parse_opcode_flags): Ignore implicit flags.
include/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* opcode/arc.h (insn_class_t): Add ENTER, LEAVE, POP, PUSH, BBIT0,
BBIT1, BI, BIH, BRCC, EI, JLI, and SUB instruction classes.
(flag_class_t): Add F_CLASS_WB, F_CLASS_ZZ, and F_CLASS_IMPLICIT
flag classes.
opcode/
2017-02-06 Claudiu Zissulescu <claziss@synopsys.com>
Anton Kolesov <anton.kolesov@synopsys.com>
* arc-dis.c (arc_disassemble_info): New structure.
(init_arc_disasm_info): New function.
(find_format_from_table): Ignore implicit flags.
(find_format): Update dissassembler private data.
(print_flags): Likewise.
(print_insn_arc): Likewise.
(arc_opcode_to_insn_type): Consider the new added instruction
classes.
(arcAnalyzeInstr): Remove.
(arc_insn_decode): New function.
* arc-dis.h (arc_ldst_writeback_mode): New enum.
(arc_ldst_data_size): Likewise.
(arc_condition_code): Likewise.
(arc_operand_kind): Likewise.
(arc_insn_kind): New struct.
(arc_instruction): Likewise.
(arc_insn_decode): Declare function.
(ARC_Debugger_OperandType): Deleted.
(Flow): Likewise.
(NullifyMode): Likewise.
(allOperandsSize): Likewise.
(arcDisState): Likewise.
(arcAnalyzeInstr): Likewise.
* arc-dis.c (arc_opcode_to_insn_type): Handle newly introduced
insn_class_t enums.
* arc-opc.c (F_SIZED): New define.
(C_CC_EQ, C_CC_GE, C_CC_GT, C_CC_HI, C_CC_HS): Likewise.
(C_CC_LE, C_CC_LO, C_CC_LS, C_CC_LT, C_CC_NE): Likewise.
(C_CC_NE, C_AA_AB, C_AA_AW, C_ZZ_D, C_ZZ_H, C_ZZ_B): Likewise.
(arc_flag_classes): Add F_CLASS_COND/F_CLASS_IMPLICIT flags.
* opcodes/arc-tbl.h: Update instructions to include new
F_CLASS_IMPLICIT flags.
(bbit0, lp): Change class.
(bbit1, bi, bih, br*, ei_s, jli_s): Likewsie
2017-02-06 18:26:13 +08:00
|
|
|
void arc_insn_decode (bfd_vma addr,
|
|
|
|
struct disassemble_info *info,
|
|
|
|
disassembler_ftype disasm_func,
|
|
|
|
struct arc_instruction *insn)
|
|
|
|
{
|
|
|
|
const struct arc_opcode *opcode;
|
|
|
|
struct arc_disassemble_info *arc_infop;
|
|
|
|
|
|
|
|
/* Ensure that insn would be in the reset state. */
|
|
|
|
memset (insn, 0, sizeof (struct arc_instruction));
|
|
|
|
|
|
|
|
/* There was an error when disassembling, for example memory read error. */
|
|
|
|
if (disasm_func (addr, info) < 0)
|
|
|
|
{
|
|
|
|
insn->valid = FALSE;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
assert (info->private_data != NULL);
|
|
|
|
arc_infop = info->private_data;
|
|
|
|
|
|
|
|
insn->length = arc_infop->insn_len;;
|
|
|
|
insn->address = addr;
|
|
|
|
|
|
|
|
/* Quick exit if memory at this address is not an instruction. */
|
|
|
|
if (info->insn_type == dis_noninsn)
|
|
|
|
{
|
|
|
|
insn->valid = FALSE;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
insn->valid = TRUE;
|
|
|
|
|
|
|
|
opcode = (const struct arc_opcode *) arc_infop->opcode;
|
|
|
|
insn->insn_class = opcode->insn_class;
|
|
|
|
insn->limm_value = arc_infop->limm;
|
|
|
|
insn->limm_p = arc_infop->limm_p;
|
|
|
|
|
|
|
|
insn->is_control_flow = (info->insn_type == dis_branch
|
|
|
|
|| info->insn_type == dis_condbranch
|
|
|
|
|| info->insn_type == dis_jsr
|
|
|
|
|| info->insn_type == dis_condjsr);
|
|
|
|
|
|
|
|
insn->has_delay_slot = info->branch_delay_insns;
|
|
|
|
insn->writeback_mode
|
|
|
|
= (enum arc_ldst_writeback_mode) arc_infop->writeback_mode;
|
|
|
|
insn->data_size_mode = info->data_size;
|
|
|
|
insn->condition_code = arc_infop->condition_code;
|
|
|
|
memcpy (insn->operands, arc_infop->operands,
|
|
|
|
sizeof (struct arc_insn_operand) * MAX_INSN_ARGS);
|
|
|
|
insn->operands_count = arc_infop->operands_count;
|
|
|
|
}
|
2016-07-21 00:08:07 +08:00
|
|
|
|
2015-10-07 21:20:19 +08:00
|
|
|
/* Local variables:
|
|
|
|
eval: (c-set-style "gnu")
|
|
|
|
indent-tabs-mode: t
|
|
|
|
End: */
|