EPICS Controls Argonne National Laboratory

Experimental Physics and
Industrial Control System

2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  Index 2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
<== Date ==> <== Thread ==>

Subject: Re: [Fwd: R3.14.8 testing: epicsMessageQueueTestHost fails on linux-x86]
From: Marty Kraimer <[email protected]>
To: [email protected]
Date: Wed, 16 Nov 2005 11:21:17 -0600
Ralph Lange wrote:

Marty Kraimer wrote:

I found the cause of the crash.

The problem was calling epicsThreadSetPriority from a thread not created via epicsThreadCreate.

I committed the change and a change to epicsThreadTest.cpp to test epicsThreadSetPriority.

When I run epicsMessageQueueTestHost on solaris it never terminates but does do everthing it should???


Is this the same issue that I reported on Nov 11th for Linux-x86? (See attachment.)


No. It was the issue Earnest raised.

For the problem shown below, I am not sure we can guarantee a fix. The test keeps creating threads until it fails. At this point it is not clear what state the underlying operating system is in.

Marty


Ralph


------------------------------------------------------------------------

Subject:
3.14.8@Linux: epicsMaxThreadsHost does not return
From:
Ralph Lange <[email protected]>
Date:
Tue, 08 Nov 2005 13:13:44 +0100
To:
EPICS Core Talk <[email protected]>

To:
EPICS Core Talk <[email protected]>


On my Linux (Debian 3.1; gcc 3.3.5) box:

epicsMaxThreadsHost does not return properly. In most cases it runs almost complete

aragon: .../base/3-14-X > bin/linux-x86/epicsMaxThreadsHost
stackSize 131072
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
number threads 1533
pthread_create error Resource temporarily unavailable
main terminating

then hangs at this point, probably in a system call (doesn't recognize control chars, has to be killed explicitly with -9). In other cases, the outputs stops /before/ the "number threads" line with the same behaviour. In the hanging state, 'ps' shows zillions of threads, which I would consider normal during that test.

I'm aware that this could be rather a Linux configuration problem than really related (or even caused) by the test from base. Nevertheless: Any hint about where I should look further?

Ralph



References:
Re: [Fwd: R3.14.8 testing: epicsMessageQueueTestHost fails on linux-x86] Marty Kraimer
Re: [Fwd: R3.14.8 testing: epicsMessageQueueTestHost fails on linux-x86] Ralph Lange

Navigate by Date:
Prev: Re: dbEvent.c line 665 Marty Kraimer
Next: Re: dbEvent.c line 665 Marty Kraimer
Index: 2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
Navigate by Thread:
Prev: Re: [Fwd: R3.14.8 testing: epicsMessageQueueTestHost fails on linux-x86] Ralph Lange
Next: Re: [Fwd: R3.14.8 testing: epicsMessageQueueTestHost fails on linux-x86] Ernest L. Williams Jr.
Index: 2002  2003  2004  <20052006  2007  2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  2019  2020  2021  2022  2023  2024 
ANJ, 02 Feb 2012 Valid HTML 4.01! · Home · News · About · Base · Modules · Extensions · Distributions · Download ·
· Search · EPICS V4 · IRMIS · Talk · Bugs · Documents · Links · Licensing ·