Date: Mon, 26 Feb 1996 18:21:15 +0500 (GMT+0500) From: "Serge A. Babkin" <babkin@hq.icb.chel.su> To: bde@zeta.org.au (Bruce Evans) Cc: hackers@freebsd.org Subject: as(1) patch & dis Message-ID: <199602261321.SAA10862@hq.icb.chel.su> In-Reply-To: <199512261054.VAA31894@godzilla.zeta.org.au> from "Bruce Evans" at Dec 26, 95 09:54:01 pm
next in thread | previous in thread | raw e-mail | index | archive | help
Hello!
I have found a bug in the assembler. It does not understand the
following construction:
_x: .long 1
_y: .long 2
mov %eax,(_y-_x)(%ebx)
^^^^^^^^
(differences as displacements)
I needed this for my experiments with kernel startup time BIOS calls. I
have fixed this bug. Commit it please if you find this patch enough good.
BTW, for fixing this bug :-) I have written a 386 disassembler. It's alpha
quality yet but already useful. I think any people who wants to
reengineer some "object-only" BSDi code will find it useful too. So
how about to commit it ?
Thanks!
The patch is:
*** 1.1 1996/01/22 12:23:07
--- config/tc-i386.c 1996/02/26 13:09:25
***************
*** 1644,1649 ****
--- 1644,1652 ----
RESTORE_END_STRING (displacement_string_end);
input_line_pointer = save_input_line_pointer;
switch (exp_seg) {
+ case SEG_DIFFERENCE:
+ i.types[this_operand] |= Disp32; /* this is an address ==> 32bit */
+ break;
case SEG_ABSENT:
/* missing expr becomes absolute 0 */
as_bad("missing or invalid displacement '%s' taken as 0",
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199602261321.SAA10862>