mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Mike Galbraith <efault@gmx.de>
To: Bill Davidsen <davidsen@tmr.com>, Andrew Morton <akpm@digeo.com>
Cc: Brice Goglin <bgoglin@ens-lyon.fr>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: copy_from_user broken on i386 since 2.5.57
Date: Fri, 24 Jan 2003 04:35:04 +0100	[thread overview]
Message-ID: <5.1.1.6.2.20030124035109.00ce15f8@pop.gmx.net> (raw)
In-Reply-To: <Pine.LNX.3.96.1030123182133.9688A-100000@gatekeeper.tmr.co m>

At 06:22 PM 1/23/2003 -0500, Bill Davidsen wrote:
>On Wed, 22 Jan 2003, Andrew Morton wrote:
>
> > Brice Goglin <bgoglin@ens-lyon.fr> wrote:
> > >
> > > Hi,
> > >
> > > Trying to compile a very very simple module for 2.5,
> > > I got an error from gcc saying that assembly code
> > > is incorrect.
> > > This problem appeared in 2.5.57 and is still here
> > > in 2.5.59.
> > > I only tried on i386.
> > >
> > > Here's a very simple example code :
> > >
> > > #define __KERNEL__
> > > #define MODULE
> > > #include "linux/module.h"
> > > #include "asm/uaccess.h"
> > >
> > > int func(void *to, const void *from) {
> > >   return __copy_from_user(to, from, 1);
> > > }
> > >
> > >
> > > Here's gcc report :
> > >
> > > mp760:~/tmp% gcc user.c -c -o user.o -Ipath_to_2.5.57/include
> > > /tmp/cceAbcRd.s: Assembler messages:
> > > /tmp/cceAbcRd.s:120: Error: `%al' not allowed with `movl'
> > > /tmp/cceAbcRd.s:124: Error: `%al' not allowed with `xorl'
> > > /tmp/cceAbcRd.s:209: Warning: using `%eax' instead of `%ax' due to 
> `l' suffix
> > > /tmp/cceAbcRd.s:213: Warning: using `%eax' instead of `%ax' due to 
> `l' suffix
> > > /tmp/cceAbcRd.s:213: Warning: using `%eax' instead of `%ax' due to 
> `l' suffix
> >
> > Add `-O2' to the compiler switches.
> >
> > No, I don't know either ;)
>
>So, does this actually fix the problem or simply optimize the suspect code
>out of existance until some later date?

Compiler bug?  With -O2, gcc-2.95.3 emits movb and xorb at the same spot.

gcc-3.2.1 doesn't have a problem with that code optimized or not... it flat 
refuses to suck up any inline assembly unless you use at least -O 
-finline-limit=1580.  (why 1580?)

         -Mike


  parent reply	other threads:[~2003-01-24  3:29 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-01-22 18:31 Brice Goglin
2003-01-22 18:49 ` Andrew Morton
2003-01-23 23:22   ` Bill Davidsen
     [not found]   ` <Pine.LNX.3.96.1030123182133.9688A-100000@gatekeeper.tmr.co m>
2003-01-24  3:35     ` Mike Galbraith [this message]
2003-01-22 18:50 ` Sam Ravnborg

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=5.1.1.6.2.20030124035109.00ce15f8@pop.gmx.net \
    --to=efault@gmx.de \
    --cc=akpm@digeo.com \
    --cc=bgoglin@ens-lyon.fr \
    --cc=davidsen@tmr.com \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®