From schwab@linux-m68k.org Mon May 12 18:37:51 2014
From: Andreas Schwab <schwab@linux-m68k.org>
Message-ID: <87y4y7uf4y.fsf@igel.home>
To: Thorsten Glaser <t.glaser@tarent.de>
Cc: Manoj Srivastava <srivasta@debian.org>, debian-68k@lists.debian.org
Date: Mon, 12 May 2014 18:37:49 +0200
X-Original-Subject: Re: Guile language support in make
Subject: Fix alignment and registers for m68k

Fix bogus alignment assumptions (on some architectures, “natural”
alignment is not used, which means even an uint64_t is possibly
unaligned, or aligned to only a 2 byte (m68k) or 4 byte boundary).

Also, fix register name for m68k.
--

Thorsten Glaser <t.glaser@tarent.de> writes:

> and the guile-2.0 Build-Depends on m68k because guile does not
> work there (and nobody appears capable enough to debug it)?

--- a/libguile/gsubr.c
+++ b/libguile/gsubr.c
@@ -213,7 +213,7 @@
 */
 static const struct
 {
-  scm_t_uint64 dummy; /* ensure 8-byte alignment; perhaps there's a better way */
+  scm_t_uint64 dummy SCM_ALIGNED (sizeof (scm_t_uint64)); /* ensure 8-byte alignment; perhaps there's a better way */
   const scm_t_uint8 bytes[121 * (sizeof (struct scm_objcode) + 16
                                  + sizeof (struct scm_objcode) + 32)];
 } raw_bytecode = {
@@ -317,7 +317,7 @@ static const struct
 
 static const struct
 {
-  scm_t_uint64 dummy; /* alignment */
+  scm_t_uint64 dummy SCM_ALIGNED (sizeof (scm_t_uint64)); /* alignment */
   scm_t_cell cells[121 * 2]; /* 11*11 double cells */
 } objcode_cells = {
   0,
--- a/libguile/vm-engine.h
+++ b/libguile/vm-engine.h
@@ -74,7 +74,7 @@
 #define FP_REG asm("%r16")
 #endif
 #ifdef __mc68000__
-#define IP_REG asm("a5")
+#define IP_REG asm("a3")
 #define SP_REG asm("a4")
 #define FP_REG
 #endif