| 1994 1995 1996 1997 1998 1999 2000 <2001> 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 | Index | 1994 1995 1996 1997 1998 1999 2000 <2001> 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 |
| <== Date ==> | <== Thread ==> |
|---|
| Subject: | RE: cac_recv_task |
| From: | "Jeff Hill" <[email protected]> |
| To: | "Geoff Savage" <[email protected]>, <[email protected]> |
| Date: | Tue, 10 Jul 2001 10:57:15 -0600 |
> > We have a processor that has a task running the cac_recv_task function > in vxWorks_depen.c that uses 99% of the cpu time. What is this tasks > purpose. Under what conditions is this task started? How do we avoid > this condition? > This is the thread in the CA client library that receives network messages from other hosts. I am guessing that you are experiencing a very high monitor throughput into this client or possibly you have BSP configuration resulting in a very high system tick rate. In R3.13 CA specifies the following delay when this thread blocks in the select() system call. timeout.tv_usec = (4/*ticks*/ * USEC_PER_SEC)/sysClkRateGet(); timeout.tv_sec = 0; This delay in usec is computed as (4 ticks ) (sec/tick) (usec/sec). On most vxWorks systems sysClkRateGet() returns 60 and the delay is computed to be about 66 mS. Jeff