Experimental Physics and
Industrial Control System

"John A. Priller" <[email protected]> · Tue, 26 May 2009 13:58:19 -0400

We're considering upgrading our VxWorks and Linux IOCs from the

now-ancient R3.14.6 to R3.14.9, but in testing I'm encountering an odd

issue with access security when reading channels from either.

My test access security file (one that we've been using successfully in

R3.14.6 for years) looks like this:

	ASG(DEFAULT) {
	  RULE(1,READ)
	  RULE(1,WRITE,TRAPWRITE)
	}

My test Linux IOC ("plain") startup script is thus:
	< envPaths
	cd ${TOP}
	dbLoadDatabase("dbd/plain.dbd")
	plain_registerRecordDeviceDriver(pdbbase)
	dbLoadRecords("db/plainDb.db")
	cd ${TOP}/iocBoot/${IOC}
	asSetFilename "access.acf"
	iocInit()

And my test database has a single channel, a calc record that increments

and loops:

	record(calc, "MYCNTR0")
	{
	  field(DESC, "Counter")
	  field(SCAN, ".1 second")
	  field(CALC, "(A<B)?(A+C):D")
	  field(INPA, "MYCNTR0.VAL  NPP NMS")
	  field(INPB, "1000")
	  field(INPC, "3")
	  field(INPD, "0")
	  field(EGU,  "Counts")
	  field(ADEL, "1.00")
	}

About one time in ten if I run the command "caget MYCNTR0" (from the

same or a different Linux box) I get the following error:

	Read operation timed out: some PV data was not read.
	MYCNTR0                        *** no read access

The same sort of error sometimes crops up in EDM (1-12-3) compiled

w/R3.14.9, again not every time but some times:

	CA get control info error('MYCNTR0'): Read access denied

If I disable access security on the test IOC (comment out the

asSetFilename line) I never have a problem. caget and EDM compiled with

R3.14.6 also never have a problem. caget built for win32-x86 on a PC

also never has a problem.

This is all with plain unpatched unmodified straight-from-the-tarball

R3.14.9. I'm really hoping I'm missing something obvious! Anybody have

a clue they could lend me?

--
John A. Priller            | Phone : (517) 333-6375
MSU Cyclotron Laboratory   | Fax   : (517) 324-8126
South Shaw Lane            | Email : [email protected]
East Lansing, MI 48824     | Web   : http://www.nscl.msu.edu/~priller/

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System