From: Ronald Hecht <ronald.hecht@uni-rostock.de>
To: cgen@sourceware.org
Subject: Again: Disassembly with variable instruction size
Date: Thu, 27 Jul 2006 08:48:00 -0000 [thread overview]
Message-ID: <44C87E54.4050600@uni-rostock.de> (raw)
Hello,
the problem seems to be in <arch>-dis.c. The generated function
my_print_instruction looks wrong. It looks like this*
#undef CGEN_PRINT_INSN
#define CGEN_PRINT_INSN my_print_insn
static int
my_print_insn (CGEN_CPU_DESC cd,
bfd_vma pc,
disassemble_info *info)
{
bfd_byte buffer[CGEN_MAX_INSN_SIZE];
bfd_byte *buf = buffer;
int status;
int buflen = (pc & 3) == 0 ? 4 : 2;
int big_p = CGEN_CPU_INSN_ENDIAN (cd) == CGEN_ENDIAN_BIG;
bfd_byte *x;
/* Read the base part of the insn. */
status = (*info->read_memory_func) (pc - ((!big_p && (pc & 3) != 0) ?
2 : 0),
buf, buflen, info);
if (status != 0)
{
(*info->memory_error_func) (status, pc, info);
return -1;
}
/* 32 bit insn? */
x = (big_p ? &buf[0] : &buf[3]);
if ((pc & 3) == 0 && (*x & 0x80) != 0)
return print_insn (cd, pc, info, buf, buflen);
/* Print the first insn. */
if ((pc & 3) == 0)
{
buf += (big_p ? 0 : 2);
if (print_insn (cd, pc, info, buf, 2) == 0)
(*info->fprintf_func) (info->stream, UNKNOWN_INSN_MSG);
buf += (big_p ? 2 : -2);
}
x = (big_p ? &buf[0] : &buf[1]);
if (*x & 0x80)
{
/* Parallel. */
(*info->fprintf_func) (info->stream, " || ");
*x &= 0x7f;
}
else
(*info->fprintf_func) (info->stream, " -> ");
/* The "& 3" is to pass a consistent address.
Parallel insns arguably both begin on the word boundary.
Also, branch insns are calculated relative to the word boundary. */
if (print_insn (cd, pc & ~ (bfd_vma) 3, info, buf, 2) == 0)
(*info->fprintf_func) (info->stream, UNKNOWN_INSN_MSG);
return (pc & 3) ? 2 : 4;
}
I replaced it with the stuff from fr30-dis.c :
/* Default value for CGEN_PRINT_INSN.
The result is the size of the insn in bytes or zero for an unknown insn
or -1 if an error occured fetching bytes. */
#ifndef CGEN_PRINT_INSN
#define CGEN_PRINT_INSN default_print_insn
#endif
static int
default_print_insn (CGEN_CPU_DESC cd, bfd_vma pc, disassemble_info *info)
{
bfd_byte buf[CGEN_MAX_INSN_SIZE];
int buflen;
int status;
/* Attempt to read the base part of the insn. */
buflen = cd->base_insn_bitsize / 8;
status = (*info->read_memory_func) (pc, buf, buflen, info);
/* Try again with the minimum part, if min < base. */
if (status != 0 && (cd->min_insn_bitsize < cd->base_insn_bitsize))
{
buflen = cd->min_insn_bitsize / 8;
status = (*info->read_memory_func) (pc, buf, buflen, info);
}
if (status != 0)
{
(*info->memory_error_func) (status, pc, info);
return -1;
}
return print_insn (cd, pc, info, buf, buflen);
}
This works for me. So the bug seems to be in the generation of <arch>-dis.c
Best Regards
Ronald
*
next reply other threads:[~2006-07-27 8:48 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-07-27 8:48 Ronald Hecht [this message]
2006-07-27 16:46 ` Dave Brolley
2006-07-27 16:56 ` Dave Brolley
2006-07-27 17:02 ` Ronald Hecht
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=44C87E54.4050600@uni-rostock.de \
--to=ronald.hecht@uni-rostock.de \
--cc=cgen@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).