From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <4A9BE7CF.3030406@domain.hid> Date: Mon, 31 Aug 2009 17:10:07 +0200 From: Gilles Chanteperdrix MIME-Version: 1.0 References: <7a15153d0908281050l56343b26g45a8f66b87c839e9@domain.hid> <4A981B54.4020908@domain.hid> <4A981C1E.8000904@domain.hid> <7a15153d0908281209i5b133c53l4fa816993df7d0eb@domain.hid> <4A982D08.4050802@domain.hid> <7a15153d0908310753s68f36dffg817df3cb9a16db06@domain.hid> In-Reply-To: <7a15153d0908310753s68f36dffg817df3cb9a16db06@domain.hid> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Xenomai-core] [PXA + Xenomai] Latency program List-Id: Xenomai life and development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Felipe Castro Cc: xenomai@xenomai.org Felipe Castro wrote: > hi ! > > I apllied the patch that you sent me, but there are some weird! > When i run latency program the result is: > > root@domain.hid$ /usr/xenomai/bin/latency > == Sampling period: 100 us > == Test mode: periodic user-mode task > == All results in microseconds > warming up... > RTT| 00:00:02 (periodic user-mode task, 100 us period, priority 99) > RTH|-----lat min|-----lat avg|-----lat max|-overrun|----lat best|---lat > worst > RTD| 83.384| 146.769| 189.538| 1803| 83.384| > 189.538 100us is to short a period for an ARM without FCSE enabled. Actually it is also too short for an ARM with FCSE. Look at the latency: you are asking a 100us period whereas the latency is 150us on average, don't you see the problem? You can try a period of 1ms, it is more reasonable. If you want periods of 70us, you can not use an ARM. > > > And than the program just freeze, and the kernel trace: > > [ 780.136217] Xenomai: Switching display-343 to secondary mode after > exception #5 from user-space at 0x93e0 (pid 345) That is another problem which should be fixed in the next Xenomai release, which should happen real soon now. -- Gilles