Mailing-List: contact cygwin-developers-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-developers-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-developers/>
List-Post: <mailto:cygwin-developers AT cygwin DOT com>
List-Help: <mailto:cygwin-developers-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-developers-owner AT cygwin DOT com
Delivered-To: mailing list cygwin-developers AT cygwin DOT com
Date: Thu, 12 Dec 2002 11:48:38 -0500
From: Christopher Faylor <cgf AT redhat DOT com>
To: cygwin-developers AT cygwin DOT com
Subject: Re: quandary with pthreads
Message-ID: <20021212164838.GG10952@redhat.com>
Reply-To: cygwin-developers AT cygwin DOT com
Mail-Followup-To: cygwin-developers AT cygwin DOT com
References: <20021212163049 DOT GC10952 AT redhat DOT com> <Pine DOT WNT DOT 4 DOT 44 DOT 0212121731040 DOT 268-200000 AT algeria DOT intern DOT net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.WNT.4.44.0212121731040.268-200000@algeria.intern.net>
User-Agent: Mutt/1.5.1i

On Thu, Dec 12, 2002 at 05:38:50PM +0100, Thomas Pfaff wrote:
>
>
>On Thu, 12 Dec 2002, Christopher Faylor wrote:
>
>> On Thu, Dec 12, 2002 at 01:37:37AM -0600, Gary R. Van Sickle wrote:
>> >Ok, attached is what I'd try if I was set up to try it.  I've found that you
>> >pretty much always need the "volatile" after the __asm__, or the newer gccs want
>> >to optimize your assembly away, even if they should be able to tell that you've
>> >got side-effects.
>>
>> AFAIK, there is a volatile after the __asm__.
>>
>> How about a standard unified diff so that it is easy to see what you changed?
>
>I didn't specified it when i put the asm stuff it into a c source file.
>You might have a look at the version that i created.
>
>AFAICT the only changes between Garys inline asm version and mine is the
>addition of the missing volatile.
>Anyway, i have attached a diff.

Sigh.  So, to clarify, I had a version of things which you admitted
worked when you changed them from inline.  You indicated that you'd just
taken my versions and adapted them.  My versions used volatile, yours
didn't.

Then Gary offered to help.  I pointed him at my versions.  Instead, I
infer, he looked at your versions and noticed that volatile was missing.
And, voila, adding volatile fixed the problem.

Nope.  I'm not going to look at this.  One or both of you need to
clarify what is going on here.  Sending a diff where nearly every line
is changed (thanks to gratuitous formatting difference?) is not a help.

I certainly appreciate the effort involved in tracking down the problem.
I would appreciate a little more effort in showing what the problem
actually was.

cgf