Re: asynDriver / epicsTimer bug

[Eric Norum <norume@aps.anl.gov>]

On May 31, 2006, at 4:27 AM, Dirk Zimoch wrote:

Hi all,

attached is a test program that can reproduce the problem. Since it uses sysClkRateGet(), it compiles only on vxWorks.

It calls pasynOctet->readRaw() once with pasynUser->timeout=1.1/sysClkRateGet() and once with pasynUser->timeout=1.0/sysClkRateGet().

In the first case, readRaw() returns asynTimeout, as expected. In the second case, readRaw() does not return.

I print the thread ID before calling readRaw(). When I do 'tt' on the thread ID, readRaw() suddenly returns asynTimeout. Very strange.

Hmm....

This sounds like a task starvation issue.  It appears that the read task is looping at a higher priority than the event timer callback task and so is preventing the callback from running.   My suspicion is that it's the vxWorks version of select() which is not rounding the timeout up to at least one tick.   If so, the problem will still be present in the latest version of ASYN. 

-- 

Eric Norum <norume@aps.anl.gov>

Advanced Photon Source                     

Argonne National Laboratory

(630) 252-4793