From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 11954 invoked by alias); 5 Apr 2007 03:42:02 -0000 Received: (qmail 11945 invoked by uid 22791); 5 Apr 2007 03:42:01 -0000 X-Spam-Status: No, hits=-1.3 required=5.0 tests=AWL,SPF_HELO_PASS,SPF_PASS X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (66.187.233.31) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 05 Apr 2007 04:41:59 +0100 Received: from int-mx1.corp.redhat.com (int-mx1.corp.redhat.com [172.16.52.254]) by mx1.redhat.com (8.13.1/8.13.1) with ESMTP id l353fvIY009579 for ; Wed, 4 Apr 2007 23:41:57 -0400 Received: from pobox.corp.redhat.com (pobox.corp.redhat.com [10.11.255.20]) by int-mx1.corp.redhat.com (8.13.1/8.13.1) with ESMTP id l353fvQa022843 for ; Wed, 4 Apr 2007 23:41:57 -0400 Received: from [127.0.0.1] (sebastian-int.corp.redhat.com [172.16.52.221]) by pobox.corp.redhat.com (8.13.1/8.13.1) with ESMTP id l353fuRG010361 for ; Wed, 4 Apr 2007 23:41:56 -0400 Message-ID: <46147013.90601@redhat.com> Date: Thu, 05 Apr 2007 03:42:00 -0000 From: Andrew Cagney User-Agent: Thunderbird 1.5.0.10 (X11/20070301) MIME-Version: 1.0 To: frysk Subject: blocking waitpid based event-loop Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-IsSubscribed: yes Mailing-List: contact frysk-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: frysk-owner@sourceware.org X-SW-Source: 2007-q2/txt/msg00010.txt.bz2 Hi, The new core event-loop, that is implemented using a blocking waitpid call (and not dependent on SIGCHLD), has been fully committed The old SIGCHLD + WAITPID/NOHANG based event loop had a race where, after a SIGCHLD, it could call WAITPID/NOHANG before the kernel had made the corresponding WAITPID event ready; this leading to frysk hanging as it didn't appear to receive the event. Not that it is still disabled though. While the FC-5 test results are identical to the old event-loop, it's managing to generate some mysterious "unknown pid" messages - perhaps it is working a little too well :-) If you want to try the the new event loop, make an obvious tweak to frysk.proc.Manager.usePoll. I'll look at the possibility of adding an --event-loop={poll,wait} option. Next will be merging the ptrace-thread into the event-loop thread allowing for much of that thread's overhead to be eliminated. After that Phil and I will see about making frysk use bulk ptrace transfers. Andrew