Experimental Physics and
Industrial Control System

"Kiman Ha" <[email protected]> · Sat, 13 Sep 2003 06:14:38 +0900

Dear Andrew,

Thanks for your prompt reply for my problem.

> Were there any error messages from [gnu]make when you built this application?

The base and application compile was just clearly completed. This problem was iocCore loading after record registration problem.
I tried many times recompile and excutions the base3.14.2 and base3.14.3 example IOCs.
At the Windows2K and linux OS  the base3.13.2 with base3.13.3 working well for my applications.

> You should try doing a "make rebuild" (or "gnumake rebuild") from the top 
> of your application - I suspect it didn't link properly.  If the build 
> fails or prints any warning or error messages you need to investigate and 
> fix those.  If that doesn't solve it, did you have any problems building 
> base at all?  
The base program and any application compiles clean without error message.
But the same target environment the base3.14.1.cvs version was woking well.

>Was this test application created by R3.14.3, or by a 
> different release of Base?

I'm sure, application created at the same release.

My Host environment :
-Windows2K
-Tornado2.2 (vxWorks5.5.1)
-used base : base3.13.7, base3.13.8
                    base3.14.1.cvs

Best regards,

Kiman Ha,
PLS Control group

> 
> 
> -----Original message-----
> From: "AndrewJohnson" <[email protected]>
> To: "KimanHa" <[email protected]>
> Cc: [email protected]
> Date: 09/13/2003(Sat)04:10
> Subject: Re: epics ioc3.14.3 problem at the tornado2.2
> 
> Kiman Ha wrote:
> > But base3.14.2 and base3.14.3 versions
> > test_registerRecordDeviceDriver(pdbbase) functions loding errors.
> 
> Were there any error messages from [gnu]make when you built this application?
> 
> > test_registerRecordDeviceDriver(pdbbase)
> > 
> > alignment
> > Exception current instruction address: 0x00000020
> > Machine Status Register: 0x0200b030
> > Data Access Register: 0x073b509b
> > Condition Register: 0x20002044
> > Data storage interrupt Register: 0x00005fff
> > 
> > 24aa24 vxTaskEntry    +68 : shell ()
> > 2017ec shell          +190: 201818 ()
> > 201a18 shell          +3bc: execute ()
> > 201b54 execute        +90 : 201818 ()
> > 201a18 shell          +3bc: execute ()
> > 201b9c execute        +d8 : yyparse ()
> > 21fd10 yyparse        +71c: 21e0a8 ()
> > 21e22c yystart        +96c: test_registerRecordDeviceDriver ()
> > 72b0e34 test_registerRecordDeviceDriver+12c: 20 ()
> 
> The above stack trace shows that the processor is trying to execute 
> instructions from address 0x20, which cannot be the correct address for 
> the sizeOffset routine of any record type.  This is probably an unresolved 
> reference.
> 
> > The probeem from this functions.  sizeOffset(dbEntry.precordType); 
> 
> You should try doing a "make rebuild" (or "gnumake rebuild") from the top 
> of your application - I suspect it didn't link properly.  If the build 
> fails or prints any warning or error messages you need to investigate and 
> fix those.  If that doesn't solve it, did you have any problems building 
> base at all?  Was this test application created by R3.14.3, or by a 
> different release of Base?
> 
> - Andrew
> -- 
> There are 10 types of people in the world:
> Those who understand binary, and those who don't.
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System