From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mx.sdf.org ([205.166.94.24]) by ewsd; Sat Nov 21 10:45:29 -0500 2020
Received: from sdf.org (IDENT:nicolagi@faeroes.freeshell.org [205.166.94.9])
	by mx.sdf.org (8.15.2/8.14.5) with ESMTPS id 0ALFdPa3027812
	(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256 bits) verified NO)
	for <9front@9front.org>; Sat, 21 Nov 2020 15:39:25 GMT
Received: (from nicolagi@localhost)
	by sdf.org (8.15.2/8.12.8/Submit) id 0ALFdbJC001377
	for 9front@9front.org; Sat, 21 Nov 2020 15:39:37 GMT
Message-ID: <98B0A90C39B4474EE7DEEFD6C04B56B9@sdf.org>
To: 9front@9front.org
Subject: MOVE implementation in games/mix
Date: Sat, 21 Nov 2020 15:39:21 +0000
From: nicolagi@sdf.org
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
List-ID: <9front.9front.org>
List-Help: <http://lists.9front.org>
X-Glyph: ➈
X-Bullshit: information plugin hypervisor method 

Largely unimportant but I'll share anyway...

I was toying with games/mix to do some exercises from TACP.  It seems
to me MOVE shouldn't be implemented with memcpy() but with an explicit
for loop, in order to surely comply with the description of the
behavior in case of overlapping memory regions at page 135.  It seems
to me that this is expected to be exploited in exercise 16.  I've
changed my copy of games/mix with the below; probably not idiomatic C
but seems to do the trick.  Am I totally wrong?

diff -r 551ecd2fd50b sys/src/games/mix/mix.c
--- a/sys/src/games/mix/mix.c	Thu Nov 19 23:05:26 2020 -0800
+++ b/sys/src/games/mix/mix.c	Sat Nov 21 15:36:38 2020 +0000
@@ -736,7 +736,10 @@
 	d = mval(ri[1], 0, MASK2);
 	if(d < 0 || d > 4000)
 		vmerror("Bad address MOVE %d", d);
-	memcpy(cells+d, cells+s, f*sizeof(u32int));
+	u32int *p = cells + d, *q = cells + s;
+	while(p < cells+d+f)
+		*p++ = *q++;
+	/* memcpy(cells+d, cells+s, f*sizeof(u32int)); */
 	d += f;
 	d &= MASK2;
 	ri[1] = d < 0 ? -d|SIGNB : d;