This is the mail archive of the libc-hacker@sources.redhat.com mailing list for the glibc project.

Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

asm constraints on i686-linux and GCC 3.2 CVS

From: Andreas Jaeger <aj at suse dot de>
To: GNU libc hacker <libc-hacker at sources dot redhat dot com>
Cc: Richard Henderson <rth at redhat dot com>
Date: Sun, 07 Apr 2002 12:24:38 +0200
Subject: asm constraints on i686-linux and GCC 3.2 CVS


Richard,

I guess your recent changes to GCC mainline CVS introduced a number of
problems while compiling inline asms in glibc CVS on i686-linux-gnu.

I get this failure:
spinlock.c: In function `__pthread_lock':
spinlock.c:113: inconsistent operand constraints in an `asm'

The line is:
      __asm __volatile ("" : "=m" (lock->__status) : "0" (lock->__status));

and lots of warnings like:
../sysdeps/i386/i486/atomicity.h:31: warning: asm operand 1 probably doesn't match constraints
../sysdeps/i386/i486/atomicity.h:31: warning: asm operand 3 probably doesn't match constraints
../sysdeps/i386/i486/atomicity.h:40: warning: asm operand 0 probably doesn't match constraints
../sysdeps/i386/i486/atomicity.h:40: warning: asm operand 2 probably doesn't match constraints
../sysdeps/i386/i486/atomicity.h:51: warning: asm operand 1 probably doesn't match constraints
../sysdeps/i386/i486/atomicity.h:51: warning: asm operand 4 probably doesn't match constraints

where the asms are:
  __asm__ __volatile__ ("lock; xaddl %0,%1"
			: "=r" (result), "=m" (*mem) : "0" (val), "1" (*mem));
  __asm__ __volatile__ ("lock; addl %1,%0"
			: "=m" (*mem) : "ir" (val), "0" (*mem));
  __asm__ __volatile__ ("lock; cmpxchgl %3, %1; sete %0"
                        : "=q" (ret), "=m" (*p), "=a" (readval)
                        : "r" (newval), "1" (*p), "a" (oldval));

../sysdeps/i386/fpu/bits/mathinline.h:492: warning: asm operand 0 probably doesn't match constraints
../sysdeps/i386/fpu/bits/mathinline.h:503: warning: asm operand 0 probably doesn't match constraints

with these lines:
__inline_mathcodeNP (floor, __x, \
  register long double __value;						      \
  __volatile unsigned short int __cw;					      \
  __volatile unsigned short int __cwtmp;				      \
  __asm __volatile ("fnstcw %0" : "=m" (__cw));				      \
  __cwtmp = (__cw & 0xf3ff) | 0x0400; /* rounding down */		      \
  __asm __volatile ("fldcw %0" : : "m" (__cwtmp));			      \
  __asm __volatile ("frndint" : "=t" (__value) : "0" (__x));		      \
  __asm __volatile ("fldcw %0" : : "m" (__cw));				      \
  return __value)

__inline_mathcodeNP (ceil, __x, \
  register long double __value;						      \
  __volatile unsigned short int __cw;					      \
  __volatile unsigned short int __cwtmp;				      \
  __asm __volatile ("fnstcw %0" : "=m" (__cw));				      \
  __cwtmp = (__cw & 0xf3ff) | 0x0800; /* rounding up */			      \
  __asm __volatile ("fldcw %0" : : "m" (__cwtmp));			      \
  __asm __volatile ("frndint" : "=t" (__value) : "0" (__x));		      \
  __asm __volatile ("fldcw %0" : : "m" (__cw));				      \
  return __value)

What's wrong here and how can this get fixed?

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj

Follow-Ups:
- Re: asm constraints on i686-linux and GCC 3.2 CVS
  - From: Roland McGrath

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]