DomainKey-Signature: a=rsa-sha1; c=nofws; d=sourceware.org; h=list-id
	:list-unsubscribe:list-subscribe:list-archive:list-post
	:list-help:sender:from:to:subject:date:message-id:references
	:in-reply-to:content-type:content-id:content-transfer-encoding
	:mime-version; q=dns; s=default; b=YvzvlMrkqBGa0vN43dMNawublAwcE
	OyHyoCI9o6xzG7iof3tiYZiIaF5mU5IpB87g6XNzlDXJ3LMwDof1ePddf6Nc1DQQ
	0GXltujozS/0TmIlBH/L8/GBL+a82k0irf1IPA7/uBkK01Hf8QNrx/gnh1Aoji6/
	xSU8sAkVysWIMc=
Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
Sender: cygwin-owner@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
From: Ken Brown <kbrown@cornell.edu>
To: "cygwin@cygwin.com" <cygwin@cygwin.com>
Subject: Re: timerfd deadlock
Date: Mon, 24 Jun 2019 20:15:48 +0000
Message-ID: <c4d32d35-a668-bca9-f783-bc5ee09ba4ff@cornell.edu>
References: <17657e77-6d7c-7367-3e6f-54e2586ed40c@cornell.edu> <20190611085221.GC3520@calimero.vinschen.de>
In-Reply-To: <20190611085221.GC3520@calimero.vinschen.de>
user-agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.7.2
received-spf: None (protection.outlook.com: cornell.edu does not designate permitted sender hosts)
Content-Type: text/plain; charset="utf-8"
Content-ID: <B0DAC2B82A5E6E47A402E6753228045B@namprd04.prod.outlook.com>
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-userprincipalname: ksb2@cornell.edu
X-IsSubscribed: yes
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from base64 to 8bit by delorie.com id x5OKGcv5010691

On 6/11/2019 4:52 AM, Corinna Vinschen wrote:
> On Jun  7 12:10, Ken Brown wrote:
>> While testing some new pipe code, I'm occasionally seeing a timerfd deadlock:
>> The main thread is stuck in timerfd_tracker::dtor at timerfd.cc:354, waiting for
>> the timerfd thread to close; and the timerfd thread is stuck at timerfd.cc:140,
>> waiting to acquire a mutex that's held by timerfd_tracker::dtor.
> 
> Hmm, I had hoped my current code does not have a deadlock situation
> anymore, but I was wrong, apparently.
> 
> Is there any chance to create an STC?

I haven't been able to create an STC.  The deadlock seems to depend on a rare 
timing fluke, and I've only seen it when running a parallel make on a clean 
emacs tree, with my new pipe code installed.

On the other hand, I think I have a simple fix, which I'll send to 
cygwin-patches.  With this patch, I ran the parallel make 15 times without 
getting the deadlock.  After reverting the patch, I got the deadlock on the 4th run.

Ken

--
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