print_insn and streams

print_insn and streams

[Keith Seitz]

Hi,

printcmd.c currently defines print_insn as:

static int
print_insn (CORE_ADDR memaddr, struct ui_file *stream)
{
  if (TARGET_BYTE_ORDER == BIG_ENDIAN)
    TARGET_PRINT_INSN_INFO->endian = BFD_ENDIAN_BIG;
  else
    TARGET_PRINT_INSN_INFO->endian = BFD_ENDIAN_LITTLE;

  if (TARGET_ARCHITECTURE != NULL)
    TARGET_PRINT_INSN_INFO->mach = TARGET_ARCHITECTURE->mach;
  /* else: should set .mach=0 but some disassemblers don't grok this */

  return TARGET_PRINT_INSN (memaddr, TARGET_PRINT_INSN_INFO);
}

Is there some reason that it is ignoring the parameter stream? Can it not
just be set in the disasm info? (TARGET_PRINT_INSN_INFO->stream = stream;)

Curious.
Keith

Re: print_insn and streams

[John R. Moore]

> printcmd.c currently defines print_insn as:
>
> static int
> print_insn (CORE_ADDR memaddr, struct ui_file *stream)
> {
>   if (TARGET_BYTE_ORDER == BIG_ENDIAN)
>     TARGET_PRINT_INSN_INFO->endian = BFD_ENDIAN_BIG;
>   else
>     TARGET_PRINT_INSN_INFO->endian = BFD_ENDIAN_LITTLE;
>
>   if (TARGET_ARCHITECTURE != NULL)
>     TARGET_PRINT_INSN_INFO->mach = TARGET_ARCHITECTURE->mach;
>   /* else: should set .mach=0 but some disassemblers don't grok this */
>
>   return TARGET_PRINT_INSN (memaddr, TARGET_PRINT_INSN_INFO);
> }
>
> Is there some reason that it is ignoring the parameter stream? Can it not
> just be set in the disasm info? (TARGET_PRINT_INSN_INFO->stream = stream;)
>
> Curious.
> Keith

When print_insn was first put in printcmd.c (-r1.108):

--------------------------------------------------------------

/* Print the instruction at address MEMADDR in debugged memory,
   on STREAM.  Returns length of the instruction, in bytes.  */

int
print_insn (memaddr, stream)
     CORE_ADDR memaddr;
     GDB_FILE *stream;
{
  disassemble_info info;

#define GDB_INIT_DISASSEMBLE_INFO(INFO, STREAM) \
  (INFO).fprintf_func = (fprintf_ftype)fprintf_filtered, \
  (INFO).stream = (STREAM), \
  (INFO).read_memory_func = dis_asm_read_memory, \
  (INFO).memory_error_func = dis_asm_memory_error, \
  (INFO).print_address_func = dis_asm_print_address, \
  (INFO).insn_info_valid = 0

  GDB_INIT_DISASSEMBLE_INFO(info, stream);

  /* If there's no disassembler, something is very wrong.  */
  if (tm_print_insn == NULL)
    abort ();

  return (*tm_print_insn) (memaddr, &info);
}
-------------------------------------------

"stream" was used until -r1.124

revision 1.124
date: 1996/07/15 23:54:22;  author: grossman;  state: Exp;  lines: +11 -8
        * defs.h printcmd.c:  Create global disassemble_info structure
        tm_print_insn_info.
        * gdbtk.c (gdb_disassemble):  Setup di.mach from
        tm_print_insn_info.mach, and set endian from TARGET_BYTE_ORDER.
        * i386-tdep.c (set_assembly_language_command):  set
        tm_print_insn_info.mach to the appropriate value for 386 or 8086
        disassembly.
        * printcmd.c (print_insn):  Move init of disassembler_info to
        _initialize_printcmd.  Set endian for disassembler here.
        * sparc-tdep.c:  Set tm_print_insn_info.mach as appropriate to
        select sparc/sparclite.
        * config/sparc/{tm-sparc.h tm-sparclite.h}:  Get rid of
        TM_PRINT_INSN.  Set TM_PRINT_INSN_MACH to
        bfd_mach_sparc/bfd_mach_sparc_sparclite.


---------

The parameter was left in rather than removing and fixing all the
routines that call it.

FWIW,

John

Re: print_insn and streams

[Andrew Cagney]

> Hi,
> 
> printcmd.c currently defines print_insn as:
> 
> static int
> print_insn (CORE_ADDR memaddr, struct ui_file *stream)
> {
>   if (TARGET_BYTE_ORDER == BIG_ENDIAN)
>     TARGET_PRINT_INSN_INFO->endian = BFD_ENDIAN_BIG;
>   else
>     TARGET_PRINT_INSN_INFO->endian = BFD_ENDIAN_LITTLE;
> 
>   if (TARGET_ARCHITECTURE != NULL)
>     TARGET_PRINT_INSN_INFO->mach = TARGET_ARCHITECTURE->mach;
>   /* else: should set .mach=0 but some disassemblers don't grok this */
> 
>   return TARGET_PRINT_INSN (memaddr, TARGET_PRINT_INSN_INFO);
> }
> 
> Is there some reason that it is ignoring the parameter stream? Can it not
> just be set in the disasm info? (TARGET_PRINT_INSN_INFO->stream = stream;)

Keith,

have you looked at gdb/mi/mi-cmd-disas.c (yes the code is nasty) which 
calls the TARGET_PRINT_INSN() with a ui-out builder.

If the main function was split in two would the second half do what 
you're after?  You can check the output by looking at the 
mi-disassemble.exp test (or the doco which might be easier to read :-).

	Andrew