All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Gérard Roudier" <groudier@free.fr>
To: Bernard Dautrevaux <Dautrevaux@microprocess.com>
Cc: "'dewar@gnat.com'" <dewar@gnat.com>, <bernd@gams.at>,
	<gcc@gcc.gnu.org>, <linux-kernel@vger.kernel.org>
Subject: RE: [PATCH] C undefined behavior fix
Date: Wed, 9 Jan 2002 22:59:35 +0100 (CET)	[thread overview]
Message-ID: <20020109223850.G2914-100000@gerard> (raw)
In-Reply-To: <17B78BDF120BD411B70100500422FC6309E410@IIS000>


On Wed, 9 Jan 2002, Bernard Dautrevaux wrote:

> > -----Original Message-----
> > From: dewar@gnat.com [mailto:dewar@gnat.com]
> > Sent: Wednesday, January 09, 2002 11:42 AM
> > To: bernd@gams.at; gcc@gcc.gnu.org; linux-kernel@vger.kernel.org
> > Subject: Re: [PATCH] C undefined behavior fix
> >
> >
> > <<Especially if there are cases were this optimization yields
> > a slower =
> >
> > access (or even worse indirect bugs).
> > E.g. if the referenced "volatile short" is a hardware register and the
> > access is multiplexed over a slow 8 bit bus.  There are
> > embedded systems
> > around where this is the case and the (cross-)compiler has no way to
> > know this (except it can be told by the programmer).
> > >>
> >
> > Well that of course is a situation where the compiler is
> > being deliberately
> > misinformed as to the relative costs of various machine
> > instructions, and
> > that is definitely a problem. One can even imagine hardware
> > (not such a hard
> > feat, one of our customers had such hardware) where a word
> > access works, but
> > a byte access fails due to hardware shortcuts,
>
> Tht's quite often the case with MMIO, and the only portable way to give a
> hint to the compiler that it should refrain from optimizing is "volatile";
> that's why I think the compiler should not do this optimization on volatile
> objects at all.

The C programming language and MMIO are two different things that must not
be mixed.

'volatile' is not enough a portable a concept to deal with MMIO. You also
need to tell about what is atomic and what is not atomic regarding
accesses through the involved BUS, for example. As a result, you may use
'volatile' to avoid assembly for MMIO, but you want to put the code in
architecture dependant code.

The atomicity issue also applies to memory-like access obviously, but the
compiler is expected to know about the architecture capabilities. If this
may differ for a given architecture family, then some compiler options
should be made available.

  Gérard.



  reply	other threads:[~2002-01-09 22:00 UTC|newest]

Thread overview: 203+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-01-09 10:41 [PATCH] C undefined behavior fix Bernard Dautrevaux
2002-01-09 21:59 ` Gérard Roudier [this message]
  -- strict thread matches above, loose matches on Subject: below --
2002-01-10 19:01 Bernard Dautrevaux
2002-01-10 12:18 dewar
2002-01-10 12:37 ` Erik Trulsson
2002-01-10 15:27   ` Florian Weimer
2002-01-10  9:03 Bernard Dautrevaux
2002-01-10 10:40 ` David Weinehall
2002-01-10  4:33 dewar
2002-01-10  2:16 dewar
2002-01-10  1:22 dewar
2002-01-10  1:21 dewar
2002-01-10  4:37 ` Tim Hollebeek
2002-01-09 21:54 dewar
2002-01-09 20:32 dewar
2002-01-09 21:43 ` Paul Koning
2002-01-09 20:08 dewar
2002-01-09 19:53 mike stump
2002-01-09 20:12 ` Paul Koning
2002-01-09 10:42 dewar
2002-01-09 10:40 dewar
2002-01-09  9:06 Bernard Dautrevaux
2002-01-09  9:25 ` Bernd Petrovitsch
2002-01-09  2:18 dewar
2002-01-09  2:13 dewar
2002-01-09  1:51 mike stump
2002-01-09  1:19 dewar
2002-01-09  0:51 dewar
2002-01-08 11:12 Bernard Dautrevaux
2002-01-11  9:52 ` Horst von Brand
2002-01-08  9:44 Bernard Dautrevaux
2002-01-08 11:44 ` 'jtv'
2002-01-08  0:16 mike stump
2002-01-08 11:33 ` jtv
2002-01-07 23:46 mike stump
2002-01-07 23:20 dewar
2002-01-07 23:15 dewar
2002-01-07 23:12 dewar
2002-01-07 23:11 dewar
2002-01-07 22:56 mike stump
2002-01-07 22:26 Tim McDaniel
2002-01-07 19:36 mike stump
2002-01-07 21:21 ` Theodore Tso
2002-01-08  0:19 ` Paul Mackerras
2002-01-08  1:02   ` Richard Henderson
2002-01-07 19:18 mike stump
2002-01-07 19:17 Bernard Dautrevaux
2002-01-07 18:38 mike stump
2002-01-07 13:29 Bernard Dautrevaux
2002-01-07 13:24 Bernard Dautrevaux
2002-01-07 21:49 ` jtv
2002-01-07 22:28   ` Tim Hollebeek
2002-01-07 22:16     ` jtv
2002-01-08  0:27       ` J.A. Magallon
2002-01-08 11:41         ` jtv
2002-01-08 12:36           ` Alexandre Oliva
2002-01-08 12:58             ` jtv
2002-01-09 19:47         ` Gérard Roudier
2002-01-09 20:44           ` jtv
2002-01-09 23:49           ` J.A. Magallon
2002-01-10  0:19             ` Peter Barada
2002-01-10  1:47               ` Fergus Henderson
2002-01-10 15:30                 ` Peter Barada
2002-01-10 10:13             ` Matthias Benkmann
2002-01-10 10:44           ` Bernd Petrovitsch
2002-01-06 19:32 dewar
2002-01-06 19:29 dewar
2002-01-06 19:29 dewar
2002-01-06 18:37 mike stump
2002-01-06 18:24 mike stump
2002-01-06 20:37 ` Gabriel Dos Reis
2002-01-06 21:59   ` Paul Mackerras
2002-01-06 22:19     ` Jakub Jelinek
2002-01-07  0:09       ` Tom Rini
2002-01-07  1:40         ` Alexandre Oliva
2002-01-06 18:20 mike stump
2002-01-06 16:22 dewar
2002-01-06 16:39 ` Alan Cox
2002-01-06 19:27   ` Laurent Guerby
2002-01-06 19:43     ` Alan Cox
2002-01-06 13:43 dewar
2002-01-06 13:55 ` Laurent Guerby
2002-01-06 13:41 dewar
2002-01-06 16:58 ` Gabriel Dos Reis
2002-01-06 13:16 dewar
2002-01-06 13:22 ` Gabriel Dos Reis
2002-01-06 13:07 dewar
2002-01-06 13:05 dewar
2002-01-06 13:41 ` Laurent Guerby
2002-01-06  7:56 mike stump
2002-01-06  4:26 dewar
2002-01-06  5:32 ` Paul Mackerras
2002-01-06 11:09   ` Momchil Velikov
2002-01-06  0:52 dewar
2002-01-04 22:43 dewar
2002-01-06  3:40 ` Paul Mackerras
2002-01-04 12:15 dewar
2002-01-04 12:14 dewar
2002-01-03 13:28 dewar
2002-01-04  8:38 ` Florian Weimer
2002-01-04 11:35   ` Joseph S. Myers
2002-01-05 19:25     ` jkl
2002-01-05 19:37       ` Joseph S. Myers
2002-01-05 20:06         ` jkl
2002-01-05 21:42           ` Joseph S. Myers
2002-01-06  4:09         ` Paul Mackerras
2002-01-05 20:01       ` Florian Weimer
2002-01-05 20:17         ` jkl
2002-01-05 20:51           ` Florian Weimer
2002-01-03 12:51 dewar
2002-01-03 10:35 Bernard Dautrevaux
2002-01-03 10:05 Bernard Dautrevaux
2002-01-03 10:49 ` Paul Mackerras
2002-01-03 11:59   ` Lars Brinkhoff
2002-01-03 23:53     ` Paul Mackerras
2002-01-04  9:52       ` Lars Brinkhoff
2002-01-05  6:45         ` Paul Mackerras
2002-01-06  0:20           ` Lars Brinkhoff
2002-01-09  9:28   ` Fergus Henderson
2002-01-09 10:58     ` Gabriel Dos Reis
2002-01-03  0:49 dewar
2002-01-03  0:12 dewar
2002-01-03  0:32 ` jtv
2002-01-06 19:37   ` Alexandre Oliva
2002-01-02 23:53 dewar
2002-01-02 23:59 ` Joe Buck
2002-01-03  3:12   ` Paul Mackerras
2002-01-03 15:45     ` Joe Buck
2002-01-02 17:40 Petr Vandrovec
2002-01-02 22:24 ` David Woodhouse
2002-01-02 22:39   ` Joe Buck
2002-01-02 22:59     ` David Woodhouse
2002-01-03 18:44   ` Alexandre Oliva
2002-01-02 10:02 Bernard Dautrevaux
2002-01-02 23:02 ` Paul Mackerras
2002-01-02 23:27   ` Momchil Velikov
2002-01-01 23:03 Momchil Velikov
2002-01-01 23:43 ` Tom Rini
2002-01-02  6:54   ` Momchil Velikov
2002-01-02 15:39     ` Tom Rini
2002-01-02 15:50       ` Jakub Jelinek
2002-01-02 16:45         ` Paul Koning
2002-01-02 22:56     ` Paul Mackerras
2002-01-02 23:37       ` Momchil Velikov
2002-01-03  2:51         ` Paul Mackerras
2002-01-03  6:29           ` law
2002-01-03 13:16             ` Lars Brinkhoff
2002-01-04  9:05       ` Florian Weimer
2002-01-04 22:14         ` Paul Mackerras
2002-01-02 14:56   ` Joseph S. Myers
2002-01-02 10:29 ` Florian Weimer
2002-01-02 10:41   ` Momchil Velikov
2002-01-02 13:11     ` Jakub Jelinek
2002-01-02 15:54       ` Tom Rini
2002-01-02 11:28 ` Aaron Lehmann
2002-01-02 11:40   ` Momchil Velikov
2002-01-02 15:59     ` Tom Rini
2002-01-02 11:48   ` Alan Cox
2002-01-02 19:09 ` Tom Rini
2002-01-02 20:13   ` Joe Buck
2002-01-02 20:42     ` Tom Rini
2002-01-02 21:36   ` Richard Henderson
2002-01-02 22:05     ` Tom Rini
2002-01-02 22:23       ` jtv
2002-01-02 23:12         ` Tom Rini
2002-01-02 23:45           ` jtv
2002-01-03  0:01             ` Tom Rini
2002-01-03  0:07               ` Richard Henderson
2002-01-03  0:16                 ` Tom Rini
2002-01-03  1:12                   ` Richard Henderson
2002-01-03  0:20                 ` jtv
2002-01-03  2:33                 ` Paul Mackerras
2002-01-03  8:32                   ` Richard Henderson
2002-01-03 10:10                     ` Paul Mackerras
2002-01-03 22:41                       ` Tim Hollebeek
2002-01-04  8:48                       ` Florian Weimer
2002-01-03 10:03                   ` Geert Uytterhoeven
2002-01-03  0:25               ` Alan Cox
2002-01-03  0:35                 ` David Woodhouse
2002-01-03  0:55                   ` jkl
2002-01-03  2:10                   ` Tom Rini
2002-01-03 22:39                     ` Pavel Machek
2002-01-04  9:29                       ` Paul Mackerras
2002-01-03 15:46                   ` Edgar Toernig
2002-01-03 16:48                     ` Momchil Velikov
2002-01-03 17:13                     ` jtv
2002-01-03 21:26                   ` Daniel Jacobowitz
2002-01-02 22:27       ` Richard Henderson
2002-01-02 22:35         ` Tom Rini
2002-01-02 22:44           ` Richard Henderson
2002-01-02 22:50             ` Tom Rini
2002-01-03 22:40         ` Pavel Machek
2002-01-04  8:42           ` Richard Henderson
2002-01-02 23:11   ` Paul Mackerras
2002-01-02 23:26     ` Tom Rini
2002-01-02 23:28     ` Momchil Velikov
2002-01-02 23:34       ` Tom Rini
2002-01-03  0:19         ` jtv
2002-01-03  0:29           ` Tom Rini
2002-01-03  1:03             ` jtv
2002-01-03  1:17               ` Tom Rini
2002-01-03 17:44                 ` jtv
2002-01-04  8:42   ` Florian Weimer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20020109223850.G2914-100000@gerard \
    --to=groudier@free.fr \
    --cc=Dautrevaux@microprocess.com \
    --cc=bernd@gams.at \
    --cc=dewar@gnat.com \
    --cc=gcc@gcc.gnu.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.