X-Recipient: archive-cygwin AT delorie DOT com
X-SWARE-Spam-Status: No, hits=-2.2 required=5.0	tests=AWL,BAYES_00,KHOP_RCVD_UNTRUST,KHOP_THREADED,RCVD_IN_DNSWL_LOW,RCVD_IN_HOSTKARMA_YE
X-Spam-Check-By: sourceware.org
Message-ID: <4FBE4DCF.1060306@sister-shadow.de>
Date: Thu, 24 May 2012 17:03:43 +0200
From: Otto Meta <otto DOT meta AT sister-shadow DOT de>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:12.0) Gecko/20120430 Thunderbird/12.0.1
MIME-Version: 1.0
To: cygwin AT cygwin DOT com
Subject: Re: 1.7.15-1: pthread_cancel and pthread_kill not working as expected
References: <20120522110238 DOT GC15843 AT calimero DOT vinschen DOT de> <4FBB93D6 DOT 90408 AT sister-shadow DOT de> <20120523085215 DOT GG9200 AT calimero DOT vinschen DOT de> <20120523160625 DOT GL9200 AT calimero DOT vinschen DOT de> <20120523171834 DOT GM9200 AT calimero DOT vinschen DOT de> <4FBD217B DOT 8000803 AT acm DOT org> <20120523175403 DOT GN9200 AT calimero DOT vinschen DOT de> <20120523201041 DOT GP9200 AT calimero DOT vinschen DOT de> <4FBE0730 DOT 6020103 AT sister-shadow DOT de> <4FBE0F0C DOT 2020206 AT sister-shadow DOT de> <20120524111946 DOT GB18736 AT calimero DOT vinschen DOT de>
In-Reply-To: <20120524111946.GB18736@calimero.vinschen.de>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
X-IsSubscribed: yes
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Id: <cygwin.cygwin.com>
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/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

On 2012-05-24 13:19, Corinna Vinschen wrote:
> You know that Cygwin is just a user space DLL, right?  There's no state
> information kept in the OS beyond the lifetime of any process using the
> Cygwin DLL.  In case of pthreads, there's no state at all shared with
> other processes.

Yes, that’s exactly why I’m confused about it.

> But even if so, if you stop *all* Cygwin processes and then start
> another one, all info from the old processes is gone and you should be
> back to normal.  If that's not the case, I would suspect a case of BLODA.

Yes, I tried that and it changed nothing. I took the chance to uninstall
some unused software and stopped all dodgy-sounding Windows services,
including Windows Defender, so that the only thing left from the BLODA
was the nVidia driver: No change.

Rebasing also didn’t help.

>> If the test code includes semaphore.h but doesn’t even use any of its
>> functions it fails right away, just like before. A reboot doesn’t help.
> Is that with the same "read" testcases you sent two days ago?  If so, I
> can't reproduce it.  I ran both tests in a loop, with and without an
> additional semaphore.h, but to no avail.  They both just work.

Yes, same tests (the ones blocking on read()), but the semaphore.h was
probably unrelated after all.

I also ran the tests continuously and discovered the following:

Running the same test several times manually from a cmd shell works a
few times, then fails. Running the async and deferred tests alternating
from cmd works, even after they failed previously.

Running one of the tests manually from bash fails most of the time.

$ while :; do ./testcase_cancel_asynchronous; done
First test run usually fails, further runs succeed. Same for the
deferred testcase.

$ sleep 1; while :; do ./testcase_cancel_asynchronous; done
All test runs succeed, including the first one. Same for the deferred
testcase.

I am now convinced that my system is toying with me.

As I actually don’t need to read from stdin with several threads and
only discovered this problem while you fixed async cancel (thanks for
that), I’m inclined to ignore this “problem” and stop wasting your
time, unless you want me to try something else to debug this.

> This is under W7 on a dual-core machine.

Same here.

Otto

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