Re: Program Interrupts (signals)
CMS runs in supervisor state only, it only uses storage keys to differentiate between nucleus (supervisor) and user storage. This is also why an application program under CMS can issue diagnose
By
Mark L. Gaubatz
·
#1089
·
|
Re: Program Interrupts (signals)
Adrian, In a virtual machine, CMS runs in supervisor state so you can alter the PSW or any storage area you like.? Of course you need to know what you are doing and which specialized instructions to
By
Bob Polmanter
·
#1088
·
|
Re: Program Interrupts (signals)
Adrian Do you know, I can¡¯t remember if you need you need to be in supervisor state. I seem to think not, just have the right storage keys set¡. ¡ while looking for info on PSW hooking I found
By
Dave Wade
·
#1087
·
|
Re: Program Interrupts (signals)
After reading a bit more. Presumably they chained the PSWs so that the injected one only did what it needed to do and "called" the official CMS new PSW for anything it did not understand? Otherwise
By
adriansutherland67
·
#1086
·
|
Re: memset help
Hi Peter ... I am testing these now and there is a bug in this one. My test harness is using unaligned big (4096) blocks and tests the contents top down. There is a mismatch on the 16th (approx) char
By
adriansutherland67
·
#1085
·
|
Re: Program Interrupts (signals)
Noted
By
adriansutherland67
·
#1084
·
|
Re: Program Interrupts (signals)
Adrian The problem with OS Macros are that at program end they clean up. So if you use system/svc202 to run an OS, and your program also uses OS macros they don¡¯t always work after the called
By
Dave Wade
·
#1083
·
|
Re: Program Interrupts (signals)
I guess it issues SPIE as its MVT code and so in its natural environment it can¡¯t snag the program exception new psw. I haven¡¯t looked at the source, but you could trace it with PER ¡ Dave Sent:
By
Dave Wade
·
#1082
·
|
Re: Program Interrupts (signals)
Joe, I was thinking for REXX, so your handler gets over written when you call another program. But for EXECIO yes you need it. Dave Sent: 21 April 2020 22:21 To: [email protected] Subject: Re:
By
Dave Wade
·
#1081
·
|
Re: Program Interrupts (signals)
Here's an example: SPIE DOITSPIE,(1,5,(7,10)) That sets an exit (DOITSPIE) for PIC 1,5 and 7-10. Joe adrian@...> wrote:
By
Joe Monk
·
#1080
·
|
Re: Program Interrupts (signals)
For completeness, the same FORTRAN program (Rexx appears to behave different with PLI and FORTRAN ): Ready; T=0.21/0.35 13:45:09 dave FI 06 TERM LOAD DAVE1 ( START Execution begins... Hi Dave IHC900I
By
[email protected]
·
#1079
·
|
Re: Program Interrupts (signals)
I am happy to try this ... but I clearly need more hints / examples! Where should I start? Can you help? A
By
adriansutherland67
·
#1078
·
|
Re: Program Interrupts (signals)
Useful thanks
By
adriansutherland67
·
#1077
·
|
Re: Program Interrupts (signals)
Is there an example assembler doing this?
By
adriansutherland67
·
#1076
·
|
Re: Program Interrupts (signals)
Not sure what you mean ... how can I use os interrupt handlers without os macros?
By
adriansutherland67
·
#1075
·
|
Re: Program Interrupts (signals)
Indeed, VM/SP does not clean up: q cmslevel VM/SP Release 5, Service Level 521 Ready; T=0.12/0.50 13:29:06 type divzero exec /* check divide by zero */ 'DMSIBM' 'q n' Ready; T=0.01/0.04 13:29:14
By
[email protected]
·
#1074
·
|
Re: Program Interrupts (signals)
Since this version of Fortran came from the OS world, it almost certainly uses SPIE to get control on a program check. Bob
By
Bob Polmanter
·
#1073
·
|
Re: Program Interrupts (signals)
The point is that I am trying to make gcclib avoid the need for a IPL if a client program happens to cause an abend ...
By
adriansutherland67
·
#1072
·
|
Re: Program Interrupts (signals)
Exactly. But I need a hint about how to catch it! How does the Fortran runtime do it?
By
adriansutherland67
·
#1071
·
|
Re: Program Interrupts (signals)
If I must I must ... I will just have to use infinite monkeys!
By
adriansutherland67
·
#1070
·
|