From owner-freebsd-current  Sat Apr  6 01:51:56 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id BAA18980
          for current-outgoing; Sat, 6 Apr 1996 01:51:56 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id BAA18975
          for <current@FreeBSD.ORG>; Sat, 6 Apr 1996 01:51:53 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id TAA20678; Sat, 6 Apr 1996 19:51:00 +1000
Date: Sat, 6 Apr 1996 19:51:00 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199604060951.TAA20678@godzilla.zeta.org.au>
To: current@FreeBSD.ORG, jkh@time.cdrom.com
Subject: Re: pgcc and kernels..
Sender: owner-current@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

>at which point the following define in machine/clock.h is flagged
>as in error:
>	
>#define I586_CYCLECTR(x) \
>        __asm __volatile(".byte 0x0f, 0x31" : "=A" (x))

>I've never been all that comfortable with gcc's asm() syntax, seeing
>as I use it pretty much never, so I'm not sure who's "right" here - us
>or gcc?

gcc is always right.  pgcc != gcc.  The "A" regclass was added on 14
June 1994 and released in 2.6.0 according ChangeLog.9.  Perhaps the
relevant parts of pgcc are older than that?    I think the Pentium
parts are mostly the old Intel work which was done long before that.

Bruce