Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Message-Id: <6.0.1.1.0.20040118192108.01f0abb0@imap.myrealbox.com>
X-Sender: tprince AT imap DOT myrealbox DOT com
Date: Sun, 18 Jan 2004 19:37:56 -0800
To: "Benson Margulies" <bim2003 AT basistech DOT com>, <cygwin AT cygwin DOT com>
From: Tim Prince <tprince AT computer DOT org>
Subject: Re: Windows Server 2003 on AMD64, One More Ping
In-Reply-To: <E8CC5F35918C084F8DBF32A5161C4003066CED@mailserver.basistec h.net>
References: <E8CC5F35918C084F8DBF32A5161C4003066CED AT mailserver DOT basistech DOT net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-IsSubscribed: yes

At 10:12 AM 1/18/2004, Benson Margulies wrote:

>TWIMC,
>
>Some time ago, I reported that fork() didn't work when running the
>current cygwin distro on the AMD64 on Windows. At the time, I debugged
>far enough to get an approximate picture of what Cygwin was doing with
>VirtualXXX calls to implement fork, and I posted some questions in the
>hopes of understanding it well enough to try to make a fix. As far as I
>could see, I didn't get a reply.
>
>To summarize, it seemed to me as if the code was making some assumptions
>about what virtual addresses ranges would be available and assigned
>under certain conditions related to fork, and that these assumptions
>were not valid on the AMD64, leading to failures.
>
>Presumably, a ground-rule of Cygwin is to program only to the documented
>Win32 API, and not to resort to the NT API substrate as illustrated in
>Nebbett.
>
>In any case, the offer is still open; if someone would be so kind as to
>offer up a summary of the design of fork(), I'd be willing to make some
>effort to diagnose and propose mods to adapt it.
>

Since this hasn't been answered by more knowledgeable people, I'll stick my 
neck out.  No, I don't believe anyone has found satisfactory support for 
fork() within the documented Win32 API.  Thus, cygwin is easily broken by 
changes which Microsoft has made in the various "64-bit" Windows 
versions.  I do believe Cygwin has ground rules of running only on released 
Windows versions for which cgf has been provided a working hardware 
platform.   If you're talking about the Physical Address Extension Windows 
with 48-bit virtual/40-bit physical- addressing for AMD64, that meets 
neither of those criteria.  Apparently, no one has been willing to provide 
any "64-bit" Windows hardware for the Cygwin project, even for the released 
ia64 Windows. If you don't have more time than I to look at the source and 
try to understand how fork() was implemented, I think you're wasting 
bandwidth.  Likewise, if you're proposing supporting a version of Windows 
which Microsoft will not permit the Cygwin project to use.  If you're 
talking about standard released 32-bit Windows running on an AMD, my 
impression is there should be no problem.


Tim Prince 


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/