xref: /qemu/include/disas/dis-asm.h (revision ab9056ff)
1 /* Interface between the opcode library and its callers.
2    Written by Cygnus Support, 1993.
3 
4    The opcode library (libopcodes.a) provides instruction decoders for
5    a large variety of instruction sets, callable with an identical
6    interface, for making instruction-processing programs more independent
7    of the instruction set being processed.  */
8 
9 #ifndef DISAS_DIS_ASM_H
10 #define DISAS_DIS_ASM_H
11 
12 typedef void *PTR;
13 typedef uint64_t bfd_vma;
14 typedef int64_t bfd_signed_vma;
15 typedef uint8_t bfd_byte;
16 #define sprintf_vma(s,x) sprintf (s, "%0" PRIx64, x)
17 #define snprintf_vma(s,ss,x) snprintf (s, ss, "%0" PRIx64, x)
18 
19 #define BFD64
20 
21 enum bfd_flavour {
22   bfd_target_unknown_flavour,
23   bfd_target_aout_flavour,
24   bfd_target_coff_flavour,
25   bfd_target_ecoff_flavour,
26   bfd_target_elf_flavour,
27   bfd_target_ieee_flavour,
28   bfd_target_nlm_flavour,
29   bfd_target_oasys_flavour,
30   bfd_target_tekhex_flavour,
31   bfd_target_srec_flavour,
32   bfd_target_ihex_flavour,
33   bfd_target_som_flavour,
34   bfd_target_os9k_flavour,
35   bfd_target_versados_flavour,
36   bfd_target_msdos_flavour,
37   bfd_target_evax_flavour
38 };
39 
40 enum bfd_endian { BFD_ENDIAN_BIG, BFD_ENDIAN_LITTLE, BFD_ENDIAN_UNKNOWN };
41 
42 enum bfd_architecture
43 {
44   bfd_arch_unknown,    /* File arch not known */
45   bfd_arch_obscure,    /* Arch known, not one of these */
46   bfd_arch_m68k,       /* Motorola 68xxx */
47 #define bfd_mach_m68000 1
48 #define bfd_mach_m68008 2
49 #define bfd_mach_m68010 3
50 #define bfd_mach_m68020 4
51 #define bfd_mach_m68030 5
52 #define bfd_mach_m68040 6
53 #define bfd_mach_m68060 7
54 #define bfd_mach_cpu32  8
55 #define bfd_mach_mcf5200  9
56 #define bfd_mach_mcf5206e 10
57 #define bfd_mach_mcf5307  11
58 #define bfd_mach_mcf5407  12
59 #define bfd_mach_mcf528x  13
60 #define bfd_mach_mcfv4e   14
61 #define bfd_mach_mcf521x   15
62 #define bfd_mach_mcf5249   16
63 #define bfd_mach_mcf547x   17
64 #define bfd_mach_mcf548x   18
65   bfd_arch_vax,        /* DEC Vax */
66   bfd_arch_i960,       /* Intel 960 */
67      /* The order of the following is important.
68        lower number indicates a machine type that
69        only accepts a subset of the instructions
70        available to machines with higher numbers.
71        The exception is the "ca", which is
72        incompatible with all other machines except
73        "core". */
74 
75 #define bfd_mach_i960_core      1
76 #define bfd_mach_i960_ka_sa     2
77 #define bfd_mach_i960_kb_sb     3
78 #define bfd_mach_i960_mc        4
79 #define bfd_mach_i960_xa        5
80 #define bfd_mach_i960_ca        6
81 #define bfd_mach_i960_jx        7
82 #define bfd_mach_i960_hx        8
83 
84   bfd_arch_a29k,       /* AMD 29000 */
85   bfd_arch_sparc,      /* SPARC */
86 #define bfd_mach_sparc                 1
87 /* The difference between v8plus and v9 is that v9 is a true 64 bit env.  */
88 #define bfd_mach_sparc_sparclet        2
89 #define bfd_mach_sparc_sparclite       3
90 #define bfd_mach_sparc_v8plus          4
91 #define bfd_mach_sparc_v8plusa         5 /* with ultrasparc add'ns.  */
92 #define bfd_mach_sparc_sparclite_le    6
93 #define bfd_mach_sparc_v9              7
94 #define bfd_mach_sparc_v9a             8 /* with ultrasparc add'ns.  */
95 #define bfd_mach_sparc_v8plusb         9 /* with cheetah add'ns.  */
96 #define bfd_mach_sparc_v9b             10 /* with cheetah add'ns.  */
97 /* Nonzero if MACH has the v9 instruction set.  */
98 #define bfd_mach_sparc_v9_p(mach) \
99   ((mach) >= bfd_mach_sparc_v8plus && (mach) <= bfd_mach_sparc_v9b \
100    && (mach) != bfd_mach_sparc_sparclite_le)
101   bfd_arch_mips,       /* MIPS Rxxxx */
102 #define bfd_mach_mips3000              3000
103 #define bfd_mach_mips3900              3900
104 #define bfd_mach_mips4000              4000
105 #define bfd_mach_mips4010              4010
106 #define bfd_mach_mips4100              4100
107 #define bfd_mach_mips4300              4300
108 #define bfd_mach_mips4400              4400
109 #define bfd_mach_mips4600              4600
110 #define bfd_mach_mips4650              4650
111 #define bfd_mach_mips5000              5000
112 #define bfd_mach_mips6000              6000
113 #define bfd_mach_mips8000              8000
114 #define bfd_mach_mips10000             10000
115 #define bfd_mach_mips16                16
116   bfd_arch_i386,       /* Intel 386 */
117 #define bfd_mach_i386_i386 0
118 #define bfd_mach_i386_i8086 1
119 #define bfd_mach_i386_i386_intel_syntax 2
120 #define bfd_mach_x86_64 3
121 #define bfd_mach_x86_64_intel_syntax 4
122   bfd_arch_we32k,      /* AT&T WE32xxx */
123   bfd_arch_tahoe,      /* CCI/Harris Tahoe */
124   bfd_arch_i860,       /* Intel 860 */
125   bfd_arch_romp,       /* IBM ROMP PC/RT */
126   bfd_arch_alliant,    /* Alliant */
127   bfd_arch_convex,     /* Convex */
128   bfd_arch_m88k,       /* Motorola 88xxx */
129   bfd_arch_pyramid,    /* Pyramid Technology */
130   bfd_arch_h8300,      /* Hitachi H8/300 */
131 #define bfd_mach_h8300   1
132 #define bfd_mach_h8300h  2
133 #define bfd_mach_h8300s  3
134   bfd_arch_powerpc,    /* PowerPC */
135 #define bfd_mach_ppc           0
136 #define bfd_mach_ppc64         1
137 #define bfd_mach_ppc_403       403
138 #define bfd_mach_ppc_403gc     4030
139 #define bfd_mach_ppc_e500      500
140 #define bfd_mach_ppc_505       505
141 #define bfd_mach_ppc_601       601
142 #define bfd_mach_ppc_602       602
143 #define bfd_mach_ppc_603       603
144 #define bfd_mach_ppc_ec603e    6031
145 #define bfd_mach_ppc_604       604
146 #define bfd_mach_ppc_620       620
147 #define bfd_mach_ppc_630       630
148 #define bfd_mach_ppc_750       750
149 #define bfd_mach_ppc_860       860
150 #define bfd_mach_ppc_a35       35
151 #define bfd_mach_ppc_rs64ii    642
152 #define bfd_mach_ppc_rs64iii   643
153 #define bfd_mach_ppc_7400      7400
154   bfd_arch_rs6000,     /* IBM RS/6000 */
155   bfd_arch_hppa,       /* HP PA RISC */
156 #define bfd_mach_hppa10        10
157 #define bfd_mach_hppa11        11
158 #define bfd_mach_hppa20        20
159 #define bfd_mach_hppa20w       25
160   bfd_arch_d10v,       /* Mitsubishi D10V */
161   bfd_arch_z8k,        /* Zilog Z8000 */
162 #define bfd_mach_z8001         1
163 #define bfd_mach_z8002         2
164   bfd_arch_h8500,      /* Hitachi H8/500 */
165   bfd_arch_sh,         /* Hitachi SH */
166 #define bfd_mach_sh            1
167 #define bfd_mach_sh2        0x20
168 #define bfd_mach_sh_dsp     0x2d
169 #define bfd_mach_sh2a       0x2a
170 #define bfd_mach_sh2a_nofpu 0x2b
171 #define bfd_mach_sh2e       0x2e
172 #define bfd_mach_sh3        0x30
173 #define bfd_mach_sh3_nommu  0x31
174 #define bfd_mach_sh3_dsp    0x3d
175 #define bfd_mach_sh3e       0x3e
176 #define bfd_mach_sh4        0x40
177 #define bfd_mach_sh4_nofpu  0x41
178 #define bfd_mach_sh4_nommu_nofpu  0x42
179 #define bfd_mach_sh4a       0x4a
180 #define bfd_mach_sh4a_nofpu 0x4b
181 #define bfd_mach_sh4al_dsp  0x4d
182 #define bfd_mach_sh5        0x50
183   bfd_arch_alpha,      /* Dec Alpha */
184 #define bfd_mach_alpha 1
185 #define bfd_mach_alpha_ev4  0x10
186 #define bfd_mach_alpha_ev5  0x20
187 #define bfd_mach_alpha_ev6  0x30
188   bfd_arch_arm,        /* Advanced Risc Machines ARM */
189 #define bfd_mach_arm_unknown	0
190 #define bfd_mach_arm_2		1
191 #define bfd_mach_arm_2a		2
192 #define bfd_mach_arm_3		3
193 #define bfd_mach_arm_3M 	4
194 #define bfd_mach_arm_4 		5
195 #define bfd_mach_arm_4T 	6
196 #define bfd_mach_arm_5 		7
197 #define bfd_mach_arm_5T		8
198 #define bfd_mach_arm_5TE	9
199 #define bfd_mach_arm_XScale	10
200 #define bfd_mach_arm_ep9312	11
201 #define bfd_mach_arm_iWMMXt	12
202 #define bfd_mach_arm_iWMMXt2	13
203   bfd_arch_ns32k,      /* National Semiconductors ns32000 */
204   bfd_arch_w65,        /* WDC 65816 */
205   bfd_arch_tic30,      /* Texas Instruments TMS320C30 */
206   bfd_arch_v850,       /* NEC V850 */
207 #define bfd_mach_v850          0
208   bfd_arch_arc,        /* Argonaut RISC Core */
209 #define bfd_mach_arc_base 0
210   bfd_arch_m32r,       /* Mitsubishi M32R/D */
211 #define bfd_mach_m32r          0  /* backwards compatibility */
212   bfd_arch_mn10200,    /* Matsushita MN10200 */
213   bfd_arch_mn10300,    /* Matsushita MN10300 */
214   bfd_arch_cris,       /* Axis CRIS */
215 #define bfd_mach_cris_v0_v10   255
216 #define bfd_mach_cris_v32      32
217 #define bfd_mach_cris_v10_v32  1032
218   bfd_arch_microblaze, /* Xilinx MicroBlaze.  */
219   bfd_arch_moxie,      /* The Moxie core.  */
220   bfd_arch_ia64,      /* HP/Intel ia64 */
221 #define bfd_mach_ia64_elf64    64
222 #define bfd_mach_ia64_elf32    32
223   bfd_arch_nios2,	/* Nios II */
224 #define bfd_mach_nios2          0
225 #define bfd_mach_nios2r1        1
226 #define bfd_mach_nios2r2        2
227   bfd_arch_lm32,       /* Lattice Mico32 */
228 #define bfd_mach_lm32 1
229   bfd_arch_last
230   };
231 #define bfd_mach_s390_31 31
232 #define bfd_mach_s390_64 64
233 
234 typedef struct symbol_cache_entry
235 {
236     const char *name;
237     union
238     {
239         PTR p;
240         bfd_vma i;
241     } udata;
242 } asymbol;
243 
244 typedef int (*fprintf_function)(FILE *f, const char *fmt, ...)
245     GCC_FMT_ATTR(2, 3);
246 
247 enum dis_insn_type {
248   dis_noninsn,			/* Not a valid instruction */
249   dis_nonbranch,		/* Not a branch instruction */
250   dis_branch,			/* Unconditional branch */
251   dis_condbranch,		/* Conditional branch */
252   dis_jsr,			/* Jump to subroutine */
253   dis_condjsr,			/* Conditional jump to subroutine */
254   dis_dref,			/* Data reference instruction */
255   dis_dref2			/* Two data references in instruction */
256 };
257 
258 /* This struct is passed into the instruction decoding routine,
259    and is passed back out into each callback.  The various fields are used
260    for conveying information from your main routine into your callbacks,
261    for passing information into the instruction decoders (such as the
262    addresses of the callback functions), or for passing information
263    back from the instruction decoders to their callers.
264 
265    It must be initialized before it is first passed; this can be done
266    by hand, or using one of the initialization macros below.  */
267 
268 typedef struct disassemble_info {
269   fprintf_function fprintf_func;
270   FILE *stream;
271   PTR application_data;
272 
273   /* Target description.  We could replace this with a pointer to the bfd,
274      but that would require one.  There currently isn't any such requirement
275      so to avoid introducing one we record these explicitly.  */
276   /* The bfd_flavour.  This can be bfd_target_unknown_flavour.  */
277   enum bfd_flavour flavour;
278   /* The bfd_arch value.  */
279   enum bfd_architecture arch;
280   /* The bfd_mach value.  */
281   unsigned long mach;
282   /* Endianness (for bi-endian cpus).  Mono-endian cpus can ignore this.  */
283   enum bfd_endian endian;
284 
285   /* An array of pointers to symbols either at the location being disassembled
286      or at the start of the function being disassembled.  The array is sorted
287      so that the first symbol is intended to be the one used.  The others are
288      present for any misc. purposes.  This is not set reliably, but if it is
289      not NULL, it is correct.  */
290   asymbol **symbols;
291   /* Number of symbols in array.  */
292   int num_symbols;
293 
294   /* For use by the disassembler.
295      The top 16 bits are reserved for public use (and are documented here).
296      The bottom 16 bits are for the internal use of the disassembler.  */
297   unsigned long flags;
298 #define INSN_HAS_RELOC	0x80000000
299 #define INSN_ARM_BE32	0x00010000
300   PTR private_data;
301 
302   /* Function used to get bytes to disassemble.  MEMADDR is the
303      address of the stuff to be disassembled, MYADDR is the address to
304      put the bytes in, and LENGTH is the number of bytes to read.
305      INFO is a pointer to this struct.
306      Returns an errno value or 0 for success.  */
307   int (*read_memory_func)
308     (bfd_vma memaddr, bfd_byte *myaddr, int length,
309 	     struct disassemble_info *info);
310 
311   /* Function which should be called if we get an error that we can't
312      recover from.  STATUS is the errno value from read_memory_func and
313      MEMADDR is the address that we were trying to read.  INFO is a
314      pointer to this struct.  */
315   void (*memory_error_func)
316     (int status, bfd_vma memaddr, struct disassemble_info *info);
317 
318   /* Function called to print ADDR.  */
319   void (*print_address_func)
320     (bfd_vma addr, struct disassemble_info *info);
321 
322     /* Function called to print an instruction. The function is architecture
323      * specific.
324      */
325     int (*print_insn)(bfd_vma addr, struct disassemble_info *info);
326 
327   /* Function called to determine if there is a symbol at the given ADDR.
328      If there is, the function returns 1, otherwise it returns 0.
329      This is used by ports which support an overlay manager where
330      the overlay number is held in the top part of an address.  In
331      some circumstances we want to include the overlay number in the
332      address, (normally because there is a symbol associated with
333      that address), but sometimes we want to mask out the overlay bits.  */
334   int (* symbol_at_address_func)
335     (bfd_vma addr, struct disassemble_info * info);
336 
337   /* These are for buffer_read_memory.  */
338   bfd_byte *buffer;
339   bfd_vma buffer_vma;
340   int buffer_length;
341 
342   /* This variable may be set by the instruction decoder.  It suggests
343       the number of bytes objdump should display on a single line.  If
344       the instruction decoder sets this, it should always set it to
345       the same value in order to get reasonable looking output.  */
346   int bytes_per_line;
347 
348   /* the next two variables control the way objdump displays the raw data */
349   /* For example, if bytes_per_line is 8 and bytes_per_chunk is 4, the */
350   /* output will look like this:
351      00:   00000000 00000000
352      with the chunks displayed according to "display_endian". */
353   int bytes_per_chunk;
354   enum bfd_endian display_endian;
355 
356   /* Results from instruction decoders.  Not all decoders yet support
357      this information.  This info is set each time an instruction is
358      decoded, and is only valid for the last such instruction.
359 
360      To determine whether this decoder supports this information, set
361      insn_info_valid to 0, decode an instruction, then check it.  */
362 
363   char insn_info_valid;		/* Branch info has been set. */
364   char branch_delay_insns;	/* How many sequential insn's will run before
365 				   a branch takes effect.  (0 = normal) */
366   char data_size;		/* Size of data reference in insn, in bytes */
367   enum dis_insn_type insn_type;	/* Type of instruction */
368   bfd_vma target;		/* Target address of branch or dref, if known;
369 				   zero if unknown.  */
370   bfd_vma target2;		/* Second target address for dref2 */
371 
372   /* Command line options specific to the target disassembler.  */
373   char * disassembler_options;
374 
375   /* Options for Capstone disassembly.  */
376   int cap_arch;
377   int cap_mode;
378   int cap_insn_unit;
379   int cap_insn_split;
380 
381 } disassemble_info;
382 
383 
384 /* Standard disassemblers.  Disassemble one instruction at the given
385    target address.  Return number of bytes processed.  */
386 typedef int (*disassembler_ftype) (bfd_vma, disassemble_info *);
387 
388 int print_insn_tci(bfd_vma, disassemble_info*);
389 int print_insn_big_mips         (bfd_vma, disassemble_info*);
390 int print_insn_little_mips      (bfd_vma, disassemble_info*);
391 int print_insn_nanomips         (bfd_vma, disassemble_info*);
392 int print_insn_i386             (bfd_vma, disassemble_info*);
393 int print_insn_m68k             (bfd_vma, disassemble_info*);
394 int print_insn_z8001            (bfd_vma, disassemble_info*);
395 int print_insn_z8002            (bfd_vma, disassemble_info*);
396 int print_insn_h8300            (bfd_vma, disassemble_info*);
397 int print_insn_h8300h           (bfd_vma, disassemble_info*);
398 int print_insn_h8300s           (bfd_vma, disassemble_info*);
399 int print_insn_h8500            (bfd_vma, disassemble_info*);
400 int print_insn_arm_a64          (bfd_vma, disassemble_info*);
401 int print_insn_alpha            (bfd_vma, disassemble_info*);
402 disassembler_ftype arc_get_disassembler (int, int);
403 int print_insn_arm              (bfd_vma, disassemble_info*);
404 int print_insn_sparc            (bfd_vma, disassemble_info*);
405 int print_insn_big_a29k         (bfd_vma, disassemble_info*);
406 int print_insn_little_a29k      (bfd_vma, disassemble_info*);
407 int print_insn_i960             (bfd_vma, disassemble_info*);
408 int print_insn_sh               (bfd_vma, disassemble_info*);
409 int print_insn_shl              (bfd_vma, disassemble_info*);
410 int print_insn_hppa             (bfd_vma, disassemble_info*);
411 int print_insn_m32r             (bfd_vma, disassemble_info*);
412 int print_insn_m88k             (bfd_vma, disassemble_info*);
413 int print_insn_mn10200          (bfd_vma, disassemble_info*);
414 int print_insn_mn10300          (bfd_vma, disassemble_info*);
415 int print_insn_moxie            (bfd_vma, disassemble_info*);
416 int print_insn_ns32k            (bfd_vma, disassemble_info*);
417 int print_insn_big_powerpc      (bfd_vma, disassemble_info*);
418 int print_insn_little_powerpc   (bfd_vma, disassemble_info*);
419 int print_insn_rs6000           (bfd_vma, disassemble_info*);
420 int print_insn_w65              (bfd_vma, disassemble_info*);
421 int print_insn_d10v             (bfd_vma, disassemble_info*);
422 int print_insn_v850             (bfd_vma, disassemble_info*);
423 int print_insn_tic30            (bfd_vma, disassemble_info*);
424 int print_insn_ppc              (bfd_vma, disassemble_info*);
425 int print_insn_s390             (bfd_vma, disassemble_info*);
426 int print_insn_crisv32          (bfd_vma, disassemble_info*);
427 int print_insn_crisv10          (bfd_vma, disassemble_info*);
428 int print_insn_microblaze       (bfd_vma, disassemble_info*);
429 int print_insn_ia64             (bfd_vma, disassemble_info*);
430 int print_insn_lm32             (bfd_vma, disassemble_info*);
431 int print_insn_big_nios2        (bfd_vma, disassemble_info*);
432 int print_insn_little_nios2     (bfd_vma, disassemble_info*);
433 int print_insn_xtensa           (bfd_vma, disassemble_info*);
434 int print_insn_riscv32          (bfd_vma, disassemble_info*);
435 int print_insn_riscv64          (bfd_vma, disassemble_info*);
436 
437 #if 0
438 /* Fetch the disassembler for a given BFD, if that support is available.  */
439 disassembler_ftype disassembler(bfd *);
440 #endif
441 
442 
443 /* This block of definitions is for particular callers who read instructions
444    into a buffer before calling the instruction decoder.  */
445 
446 /* Here is a function which callers may wish to use for read_memory_func.
447    It gets bytes from a buffer.  */
448 int buffer_read_memory(bfd_vma, bfd_byte *, int, struct disassemble_info *);
449 
450 /* This function goes with buffer_read_memory.
451    It prints a message using info->fprintf_func and info->stream.  */
452 void perror_memory(int, bfd_vma, struct disassemble_info *);
453 
454 
455 /* Just print the address in hex.  This is included for completeness even
456    though both GDB and objdump provide their own (to print symbolic
457    addresses).  */
458 void generic_print_address(bfd_vma, struct disassemble_info *);
459 
460 /* Always true.  */
461 int generic_symbol_at_address(bfd_vma, struct disassemble_info *);
462 
463 /* Macro to initialize a disassemble_info struct.  This should be called
464    by all applications creating such a struct.  */
465 #define INIT_DISASSEMBLE_INFO(INFO, STREAM, FPRINTF_FUNC) \
466   (INFO).flavour = bfd_target_unknown_flavour, \
467   (INFO).arch = bfd_arch_unknown, \
468   (INFO).mach = 0, \
469   (INFO).endian = BFD_ENDIAN_UNKNOWN, \
470   INIT_DISASSEMBLE_INFO_NO_ARCH(INFO, STREAM, FPRINTF_FUNC)
471 
472 /* Call this macro to initialize only the internal variables for the
473    disassembler.  Architecture dependent things such as byte order, or machine
474    variant are not touched by this macro.  This makes things much easier for
475    GDB which must initialize these things separately.  */
476 
477 #define INIT_DISASSEMBLE_INFO_NO_ARCH(INFO, STREAM, FPRINTF_FUNC) \
478   (INFO).fprintf_func = (FPRINTF_FUNC), \
479   (INFO).stream = (STREAM), \
480   (INFO).symbols = NULL, \
481   (INFO).num_symbols = 0, \
482   (INFO).private_data = NULL, \
483   (INFO).buffer = NULL, \
484   (INFO).buffer_vma = 0, \
485   (INFO).buffer_length = 0, \
486   (INFO).read_memory_func = buffer_read_memory, \
487   (INFO).memory_error_func = perror_memory, \
488   (INFO).print_address_func = generic_print_address, \
489   (INFO).print_insn = NULL, \
490   (INFO).symbol_at_address_func = generic_symbol_at_address, \
491   (INFO).flags = 0, \
492   (INFO).bytes_per_line = 0, \
493   (INFO).bytes_per_chunk = 0, \
494   (INFO).display_endian = BFD_ENDIAN_UNKNOWN, \
495   (INFO).disassembler_options = NULL, \
496   (INFO).insn_info_valid = 0
497 
498 #ifndef ATTRIBUTE_UNUSED
499 #define ATTRIBUTE_UNUSED __attribute__((unused))
500 #endif
501 
502 /* from libbfd */
503 
504 bfd_vma bfd_getl64 (const bfd_byte *addr);
505 bfd_vma bfd_getl32 (const bfd_byte *addr);
506 bfd_vma bfd_getb32 (const bfd_byte *addr);
507 bfd_vma bfd_getl16 (const bfd_byte *addr);
508 bfd_vma bfd_getb16 (const bfd_byte *addr);
509 typedef bool bfd_boolean;
510 
511 #endif /* DISAS_DIS_ASM_H */
512