From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <4A895028.6010205@domain.hid> Date: Mon, 17 Aug 2009 14:42:16 +0200 From: Gilles Chanteperdrix MIME-Version: 1.0 References: <4A6EECF1.9000103@domain.hid> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Xenomai-help] Pagefaults during mlocked Realtime task List-Id: Help regarding installation and common use of Xenomai List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Henri Roosen Cc: xenomai@xenomai.org Henri Roosen wrote: > We have not much progress tracing the problem so far. But in parallel > to the tracing we also updated a kernel to the latest xenomai git > status of the 2.4.8 branch, just to see whether we are not facing > problems that are solved already. > > This version however crashes and dumps its call-stack directly when > starting the latency test application. Interesting is that it crashes > at a check that is added recently in xnshadow_harden: > XENO_BUGON(NUCLEUS, !irqs_disable_hw()); > > Please find attached the .config Ok. Thanks for reporting. I am afraid there is a hole in the way Xenomai 2.4 uses the Linux preemptible context switches. I do not even understand how it is supposed to have worked at one time. It is my fault, of course, since I am always testing head with the full modification for unlocked context switches, and only 2.4 once in a while. I will test this configuration and will keep you informed as soon as I have a fix. -- Gilles