Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-developers-subscribe AT sourceware DOT cygnus DOT com>
List-Archive: <http://sourceware.cygnus.com/ml/cygwin-developers/>
List-Post: <mailto:cygwin-developers AT sourceware DOT cygnus DOT com>
List-Help: <mailto:cygwin-developers-help AT sourceware DOT cygnus DOT com>, <http://sourceware.cygnus.com/ml/#faqs>
Sender: cygwin-developers-owner AT sourceware DOT cygnus DOT com
Delivered-To: mailing list cygwin-developers AT sourceware DOT cygnus DOT com
From: Chris Faylor <cgf AT cygnus DOT com>
Date: Mon, 6 Dec 1999 23:48:15 -0500
To: Mumit Khan <khan AT nanotech DOT wisc DOT edu>
Cc: cygwin-developers AT sourceware DOT cygnus DOT com
Subject: Re: (patch) enable signals in dynamically loaded Cygwin DLL
Message-ID: <19991206234815.A21564@cygnus.com>
Mail-Followup-To: Mumit Khan <khan AT nanotech DOT wisc DOT edu>,
	cygwin-developers AT sourceware DOT cygnus DOT com
References: <199912070351 DOT VAA26863 AT mercury DOT xraylith DOT wisc DOT edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0i
In-Reply-To: <199912070351.VAA26863@mercury.xraylith.wisc.edu>; from khan@nanotech.wisc.edu on Mon, Dec 06, 1999 at 09:51:31PM -0600

On Mon, Dec 06, 1999 at 09:51:31PM -0600, Mumit Khan wrote:
>Completely forgot about the synchronization used in makethread to avoid
>stack corruption, and that's why enabling signals for dynamically loaded
>Cygwin was crashing in newly created threads.
>
>This change removes the synchronization and uses dynamically allocated
>memory to the thread stub routine, which frees the memory. The overhead
>should be roughly the same, perhaps a tiny bit higher due to malloc's
>internal MT synchronization.
>
>A gotcha for future users of makethread -- if you pass a thread parameter,
>make sure it's not on the stack (which is the case currently, so it should
>be safe). 

Note that this probably affects tty initialization too.  It's not just
signals that are affected, it's any thread that is started during DLL
initialization.  Since CYGWIN=tty implies the creation of several runtime
threads, I assume that the problem is manifested there as well.

cgf