From mboxrd@z Thu Jan  1 00:00:00 1970
From: Keith Wright <kwright@gis.net>
To: crossgcc@sources.redhat.com
Cc: robert.floyd@inet.com
Subject: Re: Invalid Operands with Inline assembly
Date: Thu, 26 Oct 2000 00:12:00 -0000
Message-id: <200010260706.DAA01070@fcs9.free-comp-shop.com>
References: <39F74F0C.E2546F10@inet.com>
X-SW-Source: 2000-10/msg00109.html

> From: Robert Floyd <robert.floyd@inet.com>
> 
> I am attempting to embed inline assembly into my C code but during the
> compilation, I keep getting the error "invalid operands for opcode". 
> Here is one of my lines:
> 
> asm("mov %r14, %r15");
> 
> I don't see why this should cause an error.  mov is a valid Hitachi SH
> opcode, r14 and r15 are valid registers, the quotations are necessary as
> well as the % before each register.

Gcc uses a % to mark an argument that must be substituted
by the compiler with some C expression.  Thus you need two %%'s
to get one through the compiler and into the assembler.

Here is a chunk of embeded Motorola Coldfire code.  It's written
in CWEB, so ignore @ signs and the following character.

Warning: I don't know squat about the Hitachi SH.

@ The |trap_to_system| is the heart of the linkage between a user
program and \coldboot/.  It takes two parameters.
The first is a count of the number of arguments (actually the number
of quad words occupied by the argument list, but this should be
the same thing in most cases).  The macro will produce a trap whose
number is the same as the argument count.
The second parameter is a unique number that identifies the system call.  
To keep things consistent, new system call numbers should be created
by adding to the |enum| types in |@<declare system call numbers@>|.
There are four sets of numbers, one for each trap number.

The macro expects to be used in an environment in which the variables
|res| and |arglist| are defined.  The variable |arglist| should be
set up before invoking the macro to be a pointer to the first argument
of the system call wrapper procedure.  The variable |res| will contain
the result of the system call, which is always an integer.

That explanation, and the examples already done, should be enough
to use the macro, if you should ever need to.  To understand it,
first read [6] \S 4.31, on using in-line assembly language in \.{gcc}
and [3] \S 3.3.8 for an description of the sharp sign in \CEE/ macro
definitions.
@<macro definitions shared with \.{romcalls.c}@>=
#define trap_to_system(argc,calln)@/\
  @[asm volatile(@/\
    " move.l %1,%%d0; move.l %2,%%d1; trap #" #argc "; move.l %%d0,%0"@/\
    : "=g" (res)@/\
    : "g" (calln), "g" (arglist)@/\
    : "%d0", "%d1")@];

\item{[3]} Samuel P.~Harbison and Guy L.~Steele~Jr., {\it \CEE/ A Reference
Manual} (Fourth Edition), Prentice Hall (1995)

\item{[6]} Richard M.~Stallman, {\it Using and Porting GNU CC},
Free Software Foundation (1996)

-- 
     -- Keith Wright  <kwright@free-comp-shop.com>

Programmer in Chief, Free Computer Shop < http://www.free-comp-shop.com >
         ---  Food, Shelter, Source code.  ---

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com