Re: Do I have a looping issue?
|
"Which I completely agree with! Executing a CS instruction and checking to see whether or not it causes a protection exception program interrupt is a rather piss-poor way to determine if an address is writable or not! (ISK/ISKE/IVSK immediately springs to mind)"
See the TPROT instruction on page 10-47, which was added to IEAVEVAL in MVS/XA.?
That is what IEAVEVAL?is doing. Since the TPROT instruction didnt?exist on 370, CS was the only way...
Joe
Dave Wade wrote:
> Joe,
>
> A bit of researching turned this up from 19 years ago.
>
>
>
> if you scroll down you will see:-
>
> I am now getting
>
> HCCP014I CPU0000: Protection exception CODE=0004 ILC=4
> SW=078C2000 00050528 INST=BA661000 CS 6,6,0(1)
> :00FD2070:K:06=009EC818 809ECF98 00404040 40404040 ..H.
>
> So the same as Jim reports. Volker responded
>
> This particular program interrupt does not lead to a dump. It is
> "working as designed" - the CS instruction is issued by an IBM module
> to test if a certain address is valid or not. If there is a program
> interrupt, MVS knows the address in question is not "valid" - and
> goes on with its merry business. It never bothers to write an SVC
> dump in this case
>
> So not as I suggested but working as designed.
Also notice what Binyamin Dissen wrote in response to Volker's claim:
? ?> This particular program interrupt does not lead to a dump.
? ?> It is "working as designed" - the CS instruction is issued
? ?> by an IBM module to test if a certain address is valid or not.
? ?I certainly hope not.
? ?There are much better ways to determine if an address is writable
? ?using the current psw key, especially if one is in supervisor state.
Which I completely agree with! Executing a CS instruction and checking to see whether or not it causes a protection exception program interrupt is a rather piss-poor way to determine if an address is writable or not! (ISK/ISKE/IVSK immediately springs to mind)
Is the source available to the version of MVS that's being used? Can it be fixed?
--
"Fish" (David B. Trout)
Software Development Laboratories
mail: fish@...
|
