Re: monitors received out of order

["Ernest L. Williams Jr." <ernesto@slac.stanford.edu>]

Tim Mooney wrote:
> On 11/2/2010 10:17 AM, Ernest L. Williams Jr. wrote:
>   
> > Tim Mooney wrote:
> >     
> > > Dear folks,
> > >
> > > I have two records, and a separate task monitoring a field that both records
> > > post, all in the same IOC.  Most of the time, my task receives monitors in the
> > > order in which they were posted.  But sometimes, the task receives a monitor
> > > from record B before it receives a previously posted monitor from record A.
> > > (I know for sure which record is posting first, because record A posts before
> > > causing record B to process.  Also, I've modified the record to set its time
> > > stamp immediately before posting, and I get another time stamp on entry to
> > > the monitor routine.  The time ordering of those stamps does not agree.)
> > > I've seen this on solaris and Linux, but not on vxWorks.
> > >       
> > Hi Tim,
> >
> > When you run an IOC on Linux the EPICS Priorities are not map properly unless 
> > you run the IOC as root.
> > All tasks are given the same OS priority.
> > Are you running the Linux IOC as root?
> >
> >     
> No.  Although the original problem was reported for a Linux system, most of my
> testing has been on solaris.  Do you know if this also applies to solaris?
>   
Sorry, for my late reply as I did not have a Solaris machine.
The same problem indeed happens on Solaris as confirmed by Stephanie 
Allison.

Here is epicsThreadShowAll on solaris 10:

flora02:iocBoot/ioctestStatsSoft>../../bin/solaris-sparc-gnu/testIocStats
epics> < st.cmd
#!../../bin/linux-x86/testIocStats
< envPaths
epicsEnvSet("ARCH","linux-x86")
epicsEnvSet("IOC","ioctestStatsSoft")
epicsEnvSet("TOP","/afs/slac.stanford.edu/u/qb/saa/spear/iocStats")
epicsEnvSet("EPICS_SITE_TOP","/afs/slac/g/spear/epics")
epicsEnvSet("EPICS_MODULES","/afs/slac/g/spear/epics/modules")
epicsEnvSet("SNCSEQ","/afs/slac/g/spear/epics/modules/seq/seq-R2-0-12-spear1")
epicsEnvSet("EPICS_BASE","/afs/slac/g/spear/epics/base/")
cd /afs/slac.stanford.edu/u/qb/saa/spear/iocStats
#
# do OS-specific startup here
#
epicsEnvSet("ENGINEER","engineer")
epicsEnvSet("LOCATION","location")
# Soft IOCs only
epicsEnvSet("STARTUP","/afs/slac.stanford.edu/u/qb/saa/spear/iocStats")
epicsEnvSet("ST_CMD","st.cmd")
## Register all support components
dbLoadDatabase("dbd/testIocStats.dbd",0,0)
testIocStats_registerRecordDeviceDriver(pdbbase)
## Load all record instances (VxWorks)
#dbLoadRecords("db/iocAdminVxWorks.db","IOC=IOCTEST")
## or load only those records for RTEMS IOCs
#dbLoadRecords("db/iocAdminRTEMS.db","IOC=IOCTEST")
## or load only those records for Soft IOCs
dbLoadRecords("db/iocAdminSoft.db","IOC=IOCTEST")
## optionally load the SCAN monitoring records
dbLoadRecords("db/iocAdminScanMon.db","IOC=IOCTEST")
dbLoadRecords("db/testIocAdminRelease.db","IOC=IOCTEST")
iocInit()
Starting iocInit
############################################################################
## EPICS R3.14.11 $R3-14-11$ $2009/08/28 18:47:36$
## EPICS Base built Aug 31 2010
############################################################################
iocRun: All initialization complete
#seq(&testSuspension)
#seq(&testCpuUse)
epics> epicsThreadShowAll
           NAME     EPICS ID   PTHREAD ID   OSIPRI  OSSPRI  STATE
         _main_        23fd8            0      0       0       OK
         errlog        2b470            2     10       0       OK
         taskwd        3e138            3     10       0       OK
     timerQueue        337a8            4     70       0       OK
          cbLow        30630            5     59       0       OK
       cbMedium        80828            6     64       0       OK
         cbHigh        808f8            7     71       0       OK
       dbCaLink        30770            8     50       0       OK
     timerQueue        30bf8            9     60       0       OK
       scanOnce        975d8           10     70       0       OK
         scan10        97f20           11     60       0       OK
          scan5        97fb8           12     61       0       OK
          scan2        98050           13     62       0       OK
          scan1        980e8           14     63       0       OK
        scan0.5        98180           15     64       0       OK
        scan0.2        98218           16     65       0       OK
        scan0.1        982b0           17     66       0       OK
        CAS-TCP        9be98           18     18       0       OK
     CAS-beacon        9bf68           19     17       0       OK
        CAS-UDP        9c038           20     16       0       OK







epicsThreadShowAll()  will clearly show this.

On the other hand Windows does map the thread priorities as a normal 
user. :)




Cheers,
Ernest


> >     
> > > I have code that misbehaves when this happens, so I started digging around
> > > and have convinced myself that I should not be relying on the time ordering of
> > > monitors received from different records (even when those records are running
> > > in the same task).  I now think I can rely on posts from a single record 
> > > arriving in
> > > time order, but not posts from different records.  I think this because 
> > > events from
> > > different record go into different queues, and there doesn't seem to be any code
> > > in the vicinity that seems worried about time ordering across event queues.
> > >
> > > Am I right about this?
> > >
> > >       
>
>